?????????? ????????? - ??????????????? - /home/agenciai/public_html/cd38d8/libosinfo.tar
???????
NEWS 0000644 00000053363 15125347732 0005266 0 ustar 00 libosinfo News ============== 1.10.0 - Feb 14, 2022 ===================== - Add API for resolving multiple tree matches - Add API for resolving multiple media matches - Add API to match between two OsinfoTree - Add API to match between two OsinfoMedia - Add API to get a complete list of firmwares - Add missing documentation of osinfo_os_add_firmware() - Add release status to osinfo-query - Add --all flag to all tools to report all matches - Fix hiding database entries - Adapt to libsoup3 which is now preferred over libsoup2 - Several CI improvements - Several translations improvements 1.9.0 - Feb 02, 2021 ==================== - Improve on the osinfo-tools' and osinfo-detect's help - Several CI improvements - Several translations improvements - Improve on debuggability of the loader code - Fix build when using clang 1.8.0 - May 29, 2020 ==================== - Several CI improvements - Several release scripts improvements - Several translations improvements - Several syntax-check improvements - Code cleanup in order to modernize the GObject usage - Add API to get whether a firmware is supported or not - Add API to get "cloud-image-username" 1.7.1 - Dec 04, 2019 ==================== - Fix usage of versioned symbols - MinGW spec file fixes and improvements - Fix pci.ids and usb.ids installation 1.7.0 - Nov 29, 2019 ==================== - Add API to expose a device driver "priority" - Improve media detection when using "volume-size" to distinguish medias - General improvements on OsinfoList APIs to avoid unsafe access - Minor POD fixes - Switch to meson build system - Add API to expose "firmware" support on OsinfoOS - Drop "env" type support from osinfo-detect - Print the architecture of the detected tree / media on osinfo-detect - Add API to create an OsinfoTree from a treeinfo file 1.6.0 - Jul 26, 2019 ==================== - Add API to get the kernel URL argument for an OsinfoOs - Improve detection of PPC ISOs - Add API to create an OsinfoMedia from location using OsinfoMediaDetectFlags - Add API to get whether a media is bootable or not - Add OsinfoTree treeinfo properties - Add API to identify an OsinfoTree - Deprecate API to guess an OsinfoOs from OsinfoTree - Add --config-file to osinfo-install-script tool - Deprecate --config usage for user & admin passwords on osinfo-install-script tool - Add API to generate an install-script and its command-line for OsinfoTree - Use libsoup for "http://" & "https://" requests for creating both OsinfoMedia & OsinfoTree - Drop GVFS dependency - Add OsinfoOs property to OsinfoTree 1.5.0 - May 09, 2019 ==================== - Add API to get an OS from OsinfoTrees & OsinfoImages - Add API to set an OS to OsinfoTrees & OsinfoImages - Add API to get the OS variants from OsinfoTrees & OsinfoImages - Add API to set the installation URL to a OsinfoInstallConfig - Add API to get all short-ids from an OsinfoProduct - Improve OsinfoLoader in order to fully support multiple short-ids values on an OsinfoProduct - Improve osinfo-detect tool's help wording - Improve osinfo-detect tool, making it aware of OsinfoTree's OS variants - Fix the loading of the architecture value of OsinfoImages - Fix usage of application id on media detection - Accept "all" as a valid fallback architecture for OsinfoTrees & OsinfoMedias - Improve treeinfo check by checking for both ".treeinfo" and "treeinfo" files - Explicitly set GVFS as a requirement in the project spec file - Move osinfo-db related tests to osinfo-db project 1.4.0 - Mar 01, 2019 ==================== - Add isodata test files for more distros - Refreshed translations from Zanata - Drop reprecated gnome-common and intltool - Fix test cases checking for medias, trees and images URIs due to libcurl issues when reusing the curl handler - Fix media identification when its identifiers have the maximum allowed size - Fix filling up the media properties when identifying it - Improvements in the tests in order to reduce code duplication - Add one test case per OS for medias, trees and images URIs tests - Add OSINFO_RELEASE_STATUS_ROLLING release status - Add API to add an installer script to a media - Add API to get all the installer scripts from a media 1.3.0 - Jan 30, 2019 ==================== - Add isodata test files for many more distros - Refreshed translations from Zanata - Refuse to load data files which do not follow the documented directory layout and file naming conventions introduced in version 0.3.0 - Avoid warnings when setting GError objects with new glib - Remove outdated FSF address from source file headers - Make several test suites collect & report all errors before exiting with failure for easy diagnosis - Avoid misc memory leaks - Add support for reporting maximum resource settings - Add support for reporting network installer resource settings - Fix data returned for osinfo_devicelinklist_get_devices - Support removal of devices during inheritance processing - Support inheritance of resources settings - Add OsinfoImage class & related APIs for reporting pre-installed disk images - Add API to determine if treeinfo metadata is provided - Add API to determine if media supports installer scripts - Add API to determine preferred install script injection methods - Add support for detecting if PPC ISO images are bootable 1.2.0 - Jun 20, 2018 ==================== - Test suite requires osinfo-db >= 20180612 - Replace use of libsoup with libcurl in test suite - Replace use of check library with GLib tester in test suite - Increase timeout for media URIs test - Drop const parameter annotations which break Rust generator - Improve error message with ISO test fails - Turn off -Wfunction-cast-type warning from gcc 8 - Remove calls to g_log_set_always_fatal in tests - Load database files in alphabetical order for predictable sorting - Fix leak of list when loading files - Fix build of tools with clang 6 - Add missing "global" annotation on some symbol versions 1.1.0 - Aug 15, 2017 ==================== - Force UTF-8 locale for new glib-mkenums - Avoid python warnings in example program - Misc test suite updates - Fix typo in error messages - Remove ISO header string padding - Disable bogus gcc warning about unsafe loop optimizations - Remove reference to fedorahosted.org - Don't hardcode /usr/bin/perl, use /usr/bin/env - Support eject-after-install parameter in OsinfoMedia - Fix misc warnings in docs - Fix error propagation when loading DB - Add usb.ids / pci.ids locations for FreeBSD - Don't include private headers in gir/vapi generation 1.0.0 - Oct 7, 2016 ==================== - Update loader to follow new layout for external database - Move all database files into separate osinfo-db package - Move osinfo-db-validate into osinfo-db-tools package 0.3.1 - Jul 01, 2016 ==================== Changes since 0.3.0: - Require glib2 >= 2.36 - Replace GSimpleAsyncResult usage with GTask - Fix VPATH based builds - Don't include autogenerated enum files in dist - Fix build with older GCC versions - Add/improve/fix data for - Debian - SLES/SLED - OpenSUSE - FreeBSD - Windows - RHEL - Ubuntu - Update README content - Fix string comparison for bootable media detection - Fix linker flags for OS-X & solaris - Fix darwin detection code - Fix multiple memory leaks 0.3.0 - Jan 08, 2016 ==================== Changes since 0.2.12: - Ditch custom udev rule since udev/blkid already expose ISO9660 properties. - New layout for database to facilitate local admin customization. Refer to docs/database-layout.txt for full details. - Many build-related fixes. - Plug many memory leaks. - Drop redundant information on PCI devices. - Fix the URI format for install scripts. - Memory corruption fixes. - Massive refactoring/cleanup of database. - Fixes to tests. - osinfo_loader_process_*() now expect a directory argument. Regular files are still supported but this support is deprecated. - Ensure Linux install scripts install qemu-guest-agent. This allows applications to do things like resyncing the clock of guest with that of host. - Add/improve/fix data on: - Alt Linux - CentOS - Debian - Fedora - FreeBSD - GNOME - KVM - Microsoft Windows 7, 8, 10 and Server 2012 - Novell Netware - OpenBSD - openSUSE - PCI and PS/2 devices - Qemu - Red Hat Enterprise Linux - Red Hat Linux - Solaris - Ubuntu - Xen - Many internal cleanups. Dependencies changed: Require libsoup >= 2.42 0.2.12 - May 27, 2015 ===================== Changes since 0.2.11: - Add API to get/set ISO9660 volume size of media. - Make use of volume size of ISO files for detection. This enables us to differentiate between DVD and netiso of Fedora, for example. - Allow install trees to be associated with specific variants of the OS. - Use direct HTTPS URLs to drivers. - Install scripts * Add API to generate script against a particular media. * Fix Fedora script to work with netiso. * Bigger /boot partition for Fedora and RHEL. * Don't specify network interface for Fedora and RHEL. - Add/improve/fix data on: - Debian 8.0 - Fedora (all versions) - GNOME 3.14 - Microsoft Windows Vista, 7 and 8.1. - OpenBSD 5.6 - openSUSE - RHEL 5.9-11, 6.6 and Atomic 7.0. - SUSE - Ubuntu 14.04 and 15.04. - Fix some leaks. - Fix and add some API documentation. - Many build fixes. 0.2.11 - Aug 19, 2014 ===================== Changes since 0.2.10: - Document hostname limitations. - Export OSINFO_*_VERSION macros. - Add desktop scripts for Debian. - Add API to query install scripts if they require an internet connection. - Some fixes to Debian JEOS script. - Add/improve/fix data on: - CentOS - Debian - FreeBSD - Microsoft Windows 7 - Microsoft Windows 8 - Microsoft Windows Server 2012 - Microsoft Vista - OpenBSD - OpenSUSE - Red Hat Enterprise Linux 7 - More ISO detection testcases. - Many performance improvements. - Fixes to documentation. - Fixes to build. 0.2.10 - Mar 20, 2014 ===================== Changes since 0.2.9: - Add API and option to osinfo-install-script utility that allows you to query the available methods to inject the installer script to the installation process. - Add JEOS installer scripts for Debian and Ubuntu. - Disable installer script for Windows 8.1 as its known not to work. - Allow XML special chars in installer script configuration values. - Fix a few build issues. - Add/improve/fix data on: - Debian - Fedora - FreeBSD - Mandrake - Mandriva - Microsoft Windows 7 - openSUSE - Solaris - Ubuntu - Qemu 0.2.9 - Dec 16, 2013 ==================== Changes since 0.2.8: - Add/improve/fix data on: - Debian - Fedora 20 - GNOME continuous 3.12 - Mageia - Microsoft Windows 7 - Microsoft Windows 8.1 - RHEL 5.8 - RHEL 6.5 - RHEL 7.0 - Support '<foo/>' syntax for booleans in database, in addition to existing '<foo>true</foo>'. - New API/XML to determine/specify: - if an OS is an actual release, a pre-release or a development snapshot. - variants of an OS. For example, professional, enterprise and ultimate editions of Windows OSs and workstation and server variants of RHEL etc. - Various fixes to build system. - More testcases. - Some other minor fixes/changes. 0.2.8 - Sep 16, 2013 ==================== Changes since 0.2.7: - Add/improve/fix data on: - ALTLinux - Fedora 19 - Fedora 20 - FreeBSD - GNOME 3.8 - GNOME 3.10 - Microsoft Windows 7 - QEMU - Ubuntu - Add testcases for media detection: - Microsoft Windows Server 2008 R2 - Fix export of public API. - Hide private API. - Fix a parallel build issue. - Fix build against latest intltool. - Drop usage of deprecated API. 0.2.7 - May 13, 2013 ==================== Changes since 0.2.6: - Add/improve/fix data on: - CentOS 6.* - Fedora 19 - GNOME 3.8 - openSUSE 12.3 - RHEL 6.4 - Ubuntu 13.04 - Add testcases for media detection: - Debian 7.0.* - Fedora 17, 18 and 19 - Ubuntu 12.10 and 13.04 - Installer script fixes/improvements: - Setup user avatar for Windows 7. - Fix against old RHEL and Fedora. - Specify installation method for Fedora (required by Fedora 19). - New API: - osinfo_platform_get_all_devices() - osinfo_install_script_generate_command_line() - build: - Take DESTDIR into account when creating symlinks. - Fix issues with installing in a tree where libosinfo was already installed. - Fix a race-condition regarding usb.ids/pci.ids setup. - Disable static libraries by default. - Some portability improments/fixes. - Fixes `make syntax-check`. - More docs and fixes to existing docs. 0.2.6 - March 18, 2013 ====================== Changes since 0.2.5: - New API to: - query signed status of device drivers. - query device driver signing requirement of installer scripts. - enable/disable installer script driver signing checks. - Use system-installed pci.ids/usb.ids files, if available. - Don't ignore vendor/device names from pci.ids/usb.ids files. - Corrections to RPM spec. 0.2.5 - March 04, 2013 ====================== Changes since 0.2.4: - Make use of post-installation drivers in Windows 7 install scripts. This implies apps can now easily setup virtio+QXL device drivers and spice-vdagent as part of Windows 7 unattended installation. - Windows 7 install script now requires product key, mainly because product key is the only way to choose product when dealing with installer media with multiple products on it. - Actually add install scripts for Windows 8. This was supposed to be merged in release 0.2.3. - Formalize architecture names in DB to align with libvirt. The main change is that i386, i486, i586, all merge to just i686, since in practice these differences haven't mattered for at least 15 years now. - Fixes to mingw RPM spec from Fedora. - Add a Windows 7 volume ID to DB. 0.2.4 - Feb 19, 2013 ==================== Changes since 0.2.3: - Fix crash in osinfo-detect against non-bootable media. - osinfo-install-script now displays names of generated files. - Add an all-in-one virtio and QXL device driver setup binary to Windows XP and 7. Same binary also installs spice-vdagent for us. - Make use of post-installation drivers in Windows XP installer scripts. - Log post-install commands of Windows XP to target disk. - Add/improve/fix data on: - QEMU/QEMU-KVM hypervisor - GNOME - openSUSE - RPM spec file changes: - Include datamaps and locale files. - Pointto libosinfo official website. - Disable udev rule on Fedora >= 19. Changes in udev > 197 and libblkid > 2.22.2 have made this rule obsolete. - Adapt to glib 2.36. - Fix some build warnings. - Some other fixes and improvements. 0.2.3 - Jan 14, 2013 ==================== Changes since 0.2.2: - Add datamaps for translating OS-neutral values to OS-specific ones, e.g some installer configuration parameters like keyboard, language and timezone etc. - New API to detect media that makes it possible to also query languages supported by the media. - Add install scripts for: - RHEL 6.x. - Microsoft Windows 8 - Fix install script for Fedora 18. - Drop support for encoding in l10n install script configuration parameters. - Fix test build issues. - Fixes and improvements to documentation. - Fix potential issues spotted by Coverity. - Fix build for translations. - osinfo-install-script tool now has options to list available configuration parameters and profiles. - Add/improve data on: - RHEL - Debian - openSUSE - Microsoft Windows 7 - Ubuntu - MacOS X - Added translations: - Ukrainian - Polish - Many other fixes and improvements. 0.2.2 - Dec 10, 2012 ==================== Changes since 0.2.1: - Loads of improvements and fixes to installer APIs and scripts. Now they are in such a good shape that next release of Boxes will make full use of them. - Add API for information on downloadable device drivers. As a starting point, information on virtio storage drivers for Windows XP and 7 is provided. - Add/improve data on: - RHEL - Windows 7 - Windows 8 - Windows XP - Enable translations for many (potential) user-visible strings. No translations submitted yet. :( - Correct default value for OsinfoMedia::installer-reboots. - Register enum types with gobject type system. - Add enum param getter/setter helpers. - OsinfoList is now instantiable and all its subclasses has been deprecated. Newer code should use OsinfoList directly. - Various other fixes and improvements. 0.2.1 - Oct 12, 2012 ==================== Changes since 0.2.0: - Fix and simplify udev rule. - Fedora: - Fix minimum RAM requirements for F16 and F17. - Add data on: - Fedora 18 - GNOME 3.6 - Ubuntu 12.10 - Fixes to doc build. - Install script: - Add get_config_param method. - Differentiate between expected/output script names. - Add more utility functions. - Add 'installer-reboots' parameter to medias. - osinfo-detect does not die of DB loading errors anymore. - More type-specific entity value getters/setters. - Fix and update RNG file. - Add 'subsystem' property/attribute to devices. 0.2.0 - Aug 31, 2012 ==================== Introducing a new set of APIs to enable application developers to easily create unattended/automated installation scripts for various operating systems. Currently only Fedora and Microsoft Windows are supported. Also provided is a commandline tool, 'osinfo-install-script' to easily create such scripts without having to write any code. Other changes since 0.1.2: - Fix build against mingw64 toolchain. - API for getting product logo URLs. - Fix architecture for a Windows 2008 media. - Remove bogus minimum storage information for Fedora. - Add varirables for system and local database paths in pkg-config. 0.1.2 - Jun 11, 2012 ==================== Changes since 0.1.1: - Fix udev rules for CDROM media. - Fix build against newer glibc-headers when building with -O0. - More complete/up2date data for: - Fedora 17 - Ubuntu - Windows - Fix data for: - Ubuntu - Correct order of elements returned by list union functions. 0.1.1 - Apr 12, 2012 ==================== Changes since 0.1.0: - Fixes to OS media detection. - Add support for more OSs: - NetBSD - Mandrake - More complete data for: - Debian - Fedora - RHEL - Ubuntu - Windows - OpenBSD - FreeBSD - Mandriva - Fix data for: - Fedora - openSUSE - Ubuntu - Windows - Add support for install tree metadata. - Add support for filtering products based on release/EOL dates. - Add support for product codenames. - Add support for 'distro'. - Also extract application ID from ISO9660 media. - Add test cases for validating ISO9660 data. - Disable obsolete media URLs. - Standardize OS family names as all lowercase. - Standardize format for OS id, short-id, distro, version. - Add RNG schema for database files. - Add a 'osinfo-db-validate' command for verifying database files. - Add a 'osinfo-query' command for searching/listing information from database. - Add a man page for the osinfo-detect command. - Add support for loading extra local database files. By default, libosinfo will expect them to be in /etc/libosinfo/db and $HOME/.local/config/libosinfo/db but new API is provided to change this behavior. - Fix generation of docs in a VPATH builder. - Fix loading of vendor and product names of USB and PCI devices. - Remove obsolete scripts for PCI/USB id conversion. - Remove redundant generic linux entries from database. - Fixes for 'make dist'. - Actually handle '--enable-vala' configure option and don't mandate vapigen if this option is not specified. - Add missing export of osinfo_entity_set_param. - Various other fixes and improvements. 0.1.0 - Feb 7, 2012 =================== This release marks the start of the API/ABI stability promise. While some APIs may become deprecated in the future, no APIs will be removed/changed in an incompatible manner. In the unlikely event that an incompatible API change becomes necessary, the library soname will be incremented to allow parallel installation of both old & new libraries. Changes since 0.0.5: - Add Utility function to retrieve OS devices by property. - Improved OS media detection. - Add missing 'const' OsinfoList function params. - Use ENV{} in udev rule. 0.0.5 - Jan 16, 2012 ==================== Changes since 0.0.4: - Add QXL to supported devices of windows and Fedora. - Distribute and install forgotten device XML files. - Add all metadata from default path, not just OS. - List supported audio devices for Windows and Fedora. - Specify 'derives-from' relationship for Fedora OSs. - Add API to retrieve all supported devices for an OS. - Fix some potential issues related to memory allocation/alignment. 0.0.4 - Dec 19, 2011 ==================== Changes since 0.0.3: - Fix silly mistakes in udev rules file. 0.0.3 - Dec 16, 2011 ==================== Changes since 0.0.2: - Add lots of more data on various different OSs out there. Now we also support some Apple OSs. - On successful OS detection, inform the application which known media of the OS the given media matched against. osinfo-detect tool and udev rule updated to make us of this. - Support applications' right to not be interested in error details. - Minor fixes/improvements to OS detection logic. - API to indicate media is live and/or installer. - Don't match non-live Fedora ISOs against live ones. - Fixes to Python demo to make it work. - Fix some potential crashes. - Fixes to RPM spec file. - Distribute RPM spec file. - Complement some getters with corresponding gobject properties. - Few other random minor fixes. WARNING: This release (slightly) breaks API with the last one. 0.0.2 - Nov 22, 2011 ==================== Changes since 0.0.1: - Improve media detection for Fedora, RHEL and Windows 7. - More complete documentation. - Fix configure script to not require 'check' when tests are disabled. - More automated documentation generation. - Add Fedora 16 metadata. - Recommend 9 GiB storage for Fedora 14-16. - Add openSUSE data. - Fix some copy&paste mistakes in OS database. - Add media information for 2011 Ubuntu releases and CentOS 6.0. - Improve README file. 0.0.1 ===== First public release. COPYING.LIB 0000644 00000063642 15125347732 0006230 0 ustar 00 GNU LESSER GENERAL PUBLIC LICENSE Version 2.1, February 1999 Copyright (C) 1991, 1999 Free Software Foundation, Inc. 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. [This is the first released version of the Lesser GPL. It also counts as the successor of the GNU Library Public License, version 2, hence the version number 2.1.] Preamble The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public Licenses are intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This license, the Lesser General Public License, applies to some specially designated software packages--typically libraries--of the Free Software Foundation and other authors who decide to use it. You can use it too, but we suggest you first think carefully about whether this license or the ordinary General Public License is the better strategy to use in any particular case, based on the explanations below. When we speak of free software, we are referring to freedom of use, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish); that you receive source code or can get it if you want it; that you can change the software and use pieces of it in new free programs; and that you are informed that you can do these things. To protect your rights, we need to make restrictions that forbid distributors to deny you these rights or to ask you to surrender these rights. These restrictions translate to certain responsibilities for you if you distribute copies of the library or if you modify it. For example, if you distribute copies of the library, whether gratis or for a fee, you must give the recipients all the rights that we gave you. You must make sure that they, too, receive or can get the source code. If you link other code with the library, you must provide complete object files to the recipients, so that they can relink them with the library after making changes to the library and recompiling it. And you must show them these terms so they know their rights. We protect your rights with a two-step method: (1) we copyright the library, and (2) we offer you this license, which gives you legal permission to copy, distribute and/or modify the library. To protect each distributor, we want to make it very clear that there is no warranty for the free library. Also, if the library is modified by someone else and passed on, the recipients should know that what they have is not the original version, so that the original author's reputation will not be affected by problems that might be introduced by others. Finally, software patents pose a constant threat to the existence of any free program. We wish to make sure that a company cannot effectively restrict the users of a free program by obtaining a restrictive license from a patent holder. Therefore, we insist that any patent license obtained for a version of the library must be consistent with the full freedom of use specified in this license. Most GNU software, including some libraries, is covered by the ordinary GNU General Public License. This license, the GNU Lesser General Public License, applies to certain designated libraries, and is quite different from the ordinary General Public License. We use this license for certain libraries in order to permit linking those libraries into non-free programs. When a program is linked with a library, whether statically or using a shared library, the combination of the two is legally speaking a combined work, a derivative of the original library. The ordinary General Public License therefore permits such linking only if the entire combination fits its criteria of freedom. The Lesser General Public License permits more lax criteria for linking other code with the library. We call this license the "Lesser" General Public License because it does Less to protect the user's freedom than the ordinary General Public License. It also provides other free software developers Less of an advantage over competing non-free programs. These disadvantages are the reason we use the ordinary General Public License for many libraries. However, the Lesser license provides advantages in certain special circumstances. For example, on rare occasions, there may be a special need to encourage the widest possible use of a certain library, so that it becomes a de-facto standard. To achieve this, non-free programs must be allowed to use the library. A more frequent case is that a free library does the same job as widely used non-free libraries. In this case, there is little to gain by limiting the free library to free software only, so we use the Lesser General Public License. In other cases, permission to use a particular library in non-free programs enables a greater number of people to use a large body of free software. For example, permission to use the GNU C Library in non-free programs enables many more people to use the whole GNU operating system, as well as its variant, the GNU/Linux operating system. Although the Lesser General Public License is Less protective of the users' freedom, it does ensure that the user of a program that is linked with the Library has the freedom and the wherewithal to run that program using a modified version of the Library. The precise terms and conditions for copying, distribution and modification follow. Pay close attention to the difference between a "work based on the library" and a "work that uses the library". The former contains code derived from the library, whereas the latter must be combined with the library in order to run. GNU LESSER GENERAL PUBLIC LICENSE TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION 0. This License Agreement applies to any software library or other program which contains a notice placed by the copyright holder or other authorized party saying it may be distributed under the terms of this Lesser General Public License (also called "this License"). Each licensee is addressed as "you". A "library" means a collection of software functions and/or data prepared so as to be conveniently linked with application programs (which use some of those functions and data) to form executables. The "Library", below, refers to any such software library or work which has been distributed under these terms. A "work based on the Library" means either the Library or any derivative work under copyright law: that is to say, a work containing the Library or a portion of it, either verbatim or with modifications and/or translated straightforwardly into another language. (Hereinafter, translation is included without limitation in the term "modification".) "Source code" for a work means the preferred form of the work for making modifications to it. For a library, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the library. Activities other than copying, distribution and modification are not covered by this License; they are outside its scope. The act of running a program using the Library is not restricted, and output from such a program is covered only if its contents constitute a work based on the Library (independent of the use of the Library in a tool for writing it). Whether that is true depends on what the Library does and what the program that uses the Library does. 1. You may copy and distribute verbatim copies of the Library's complete source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the notices that refer to this License and to the absence of any warranty; and distribute a copy of this License along with the Library. You may charge a fee for the physical act of transferring a copy, and you may at your option offer warranty protection in exchange for a fee. 2. You may modify your copy or copies of the Library or any portion of it, thus forming a work based on the Library, and copy and distribute such modifications or work under the terms of Section 1 above, provided that you also meet all of these conditions: a) The modified work must itself be a software library. b) You must cause the files modified to carry prominent notices stating that you changed the files and the date of any change. c) You must cause the whole of the work to be licensed at no charge to all third parties under the terms of this License. d) If a facility in the modified Library refers to a function or a table of data to be supplied by an application program that uses the facility, other than as an argument passed when the facility is invoked, then you must make a good faith effort to ensure that, in the event an application does not supply such function or table, the facility still operates, and performs whatever part of its purpose remains meaningful. (For example, a function in a library to compute square roots has a purpose that is entirely well-defined independent of the application. Therefore, Subsection 2d requires that any application-supplied function or table used by this function must be optional: if the application does not supply it, the square root function must still compute square roots.) These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Library, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. But when you distribute the same sections as part of a whole which is a work based on the Library, the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it. Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Library. In addition, mere aggregation of another work not based on the Library with the Library (or with a work based on the Library) on a volume of a storage or distribution medium does not bring the other work under the scope of this License. 3. You may opt to apply the terms of the ordinary GNU General Public License instead of this License to a given copy of the Library. To do this, you must alter all the notices that refer to this License, so that they refer to the ordinary GNU General Public License, version 2, instead of to this License. (If a newer version than version 2 of the ordinary GNU General Public License has appeared, then you can specify that version instead if you wish.) Do not make any other change in these notices. Once this change is made in a given copy, it is irreversible for that copy, so the ordinary GNU General Public License applies to all subsequent copies and derivative works made from that copy. This option is useful when you wish to copy part of the code of the Library into a program that is not a library. 4. You may copy and distribute the Library (or a portion or derivative of it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange. If distribution of object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place satisfies the requirement to distribute the source code, even though third parties are not compelled to copy the source along with the object code. 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a "work that uses the Library". Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. However, linking a "work that uses the Library" with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a "work that uses the library". The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. When a "work that uses the Library" uses material from a header file that is part of the Library, the object code for the work may be a derivative work of the Library even though the source code is not. Whether this is true is especially significant if the work can be linked without the Library, or if the work is itself a library. The threshold for this to be true is not precisely defined by law. If such an object file uses only numerical parameters, data structure layouts and accessors, and small macros and small inline functions (ten lines or less in length), then the use of the object file is unrestricted, regardless of whether it is legally a derivative work. (Executables containing this object code plus portions of the Library will still fall under Section 6.) Otherwise, if the work is a derivative of the Library, you may distribute the object code for the work under the terms of Section 6. Any executables containing that work also fall under Section 6, whether or not they are linked directly with the Library itself. 6. As an exception to the Sections above, you may also combine or link a "work that uses the Library" with the Library to produce a work containing portions of the Library, and distribute that work under terms of your choice, provided that the terms permit modification of the work for the customer's own use and reverse engineering for debugging such modifications. You must give prominent notice with each copy of the work that the Library is used in it and that the Library and its use are covered by this License. You must supply a copy of this License. If the work during execution displays copyright notices, you must include the copyright notice for the Library among them, as well as a reference directing the user to the copy of this License. Also, you must do one of these things: a) Accompany the work with the complete corresponding machine-readable source code for the Library including whatever changes were used in the work (which must be distributed under Sections 1 and 2 above); and, if the work is an executable linked with the Library, with the complete machine-readable "work that uses the Library", as object code and/or source code, so that the user can modify the Library and then relink to produce a modified executable containing the modified Library. (It is understood that the user who changes the contents of definitions files in the Library will not necessarily be able to recompile the application to use the modified definitions.) b) Use a suitable shared library mechanism for linking with the Library. A suitable mechanism is one that (1) uses at run time a copy of the library already present on the user's computer system, rather than copying library functions into the executable, and (2) will operate properly with a modified version of the library, if the user installs one, as long as the modified version is interface-compatible with the version that the work was made with. c) Accompany the work with a written offer, valid for at least three years, to give the same user the materials specified in Subsection 6a, above, for a charge no more than the cost of performing this distribution. d) If distribution of the work is made by offering access to copy from a designated place, offer equivalent access to copy the above specified materials from the same place. e) Verify that the user has already received a copy of these materials or that you have already sent this user a copy. For an executable, the required form of the "work that uses the Library" must include any data and utility programs needed for reproducing the executable from it. However, as a special exception, the materials to be distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable. It may happen that this requirement contradicts the license restrictions of other proprietary libraries that do not normally accompany the operating system. Such a contradiction means you cannot use both them and the Library together in an executable that you distribute. 7. You may place library facilities that are a work based on the Library side-by-side in a single library together with other library facilities not covered by this License, and distribute such a combined library, provided that the separate distribution of the work based on the Library and of the other library facilities is otherwise permitted, and provided that you do these two things: a) Accompany the combined library with a copy of the same work based on the Library, uncombined with any other library facilities. This must be distributed under the terms of the Sections above. b) Give prominent notice with the combined library of the fact that part of it is a work based on the Library, and explaining where to find the accompanying uncombined form of the same work. 8. You may not copy, modify, sublicense, link with, or distribute the Library except as expressly provided under this License. Any attempt otherwise to copy, modify, sublicense, link with, or distribute the Library is void, and will automatically terminate your rights under this License. However, parties who have received copies, or rights, from you under this License will not have their licenses terminated so long as such parties remain in full compliance. 9. You are not required to accept this License, since you have not signed it. However, nothing else grants you permission to modify or distribute the Library or its derivative works. These actions are prohibited by law if you do not accept this License. Therefore, by modifying or distributing the Library (or any work based on the Library), you indicate your acceptance of this License to do so, and all its terms and conditions for copying, distributing or modifying the Library or works based on it. 10. Each time you redistribute the Library (or any work based on the Library), the recipient automatically receives a license from the original licensor to copy, distribute, link with or modify the Library subject to these terms and conditions. You may not impose any further restrictions on the recipients' exercise of the rights granted herein. You are not responsible for enforcing compliance by third parties with this License. 11. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues), conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Library at all. For example, if a patent license would not permit royalty-free redistribution of the Library by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Library. If any portion of this section is held invalid or unenforceable under any particular circumstance, the balance of the section is intended to apply, and the section as a whole is intended to apply in other circumstances. It is not the purpose of this section to induce you to infringe any patents or other property right claims or to contest validity of any such claims; this section has the sole purpose of protecting the integrity of the free software distribution system which is implemented by public license practices. Many people have made generous contributions to the wide range of software distributed through that system in reliance on consistent application of that system; it is up to the author/donor to decide if he or she is willing to distribute software through any other system and a licensee cannot impose that choice. This section is intended to make thoroughly clear what is believed to be a consequence of the rest of this License. 12. If the distribution and/or use of the Library is restricted in certain countries either by patents or by copyrighted interfaces, the original copyright holder who places the Library under this License may add an explicit geographical distribution limitation excluding those countries, so that distribution is permitted only in or among countries not thus excluded. In such case, this License incorporates the limitation as if written in the body of this License. 13. The Free Software Foundation may publish revised and/or new versions of the Lesser General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Library specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Library does not specify a license version number, you may choose any version ever published by the Free Software Foundation. 14. If you wish to incorporate parts of the Library into other free programs whose distribution conditions are incompatible with these, write to the author to ask for permission. For software which is copyrighted by the Free Software Foundation, write to the Free Software Foundation; we sometimes make exceptions for this. Our decision will be guided by the two goals of preserving the free status of all derivatives of our free software and of promoting the sharing and reuse of software generally. NO WARRANTY 15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. 16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. END OF TERMS AND CONDITIONS How to Apply These Terms to Your New Libraries If you develop a new library, and you want it to be of the greatest possible use to the public, we recommend making it free software that everyone can redistribute and change. You can do so by permitting redistribution under these terms (or, alternatively, under the terms of the ordinary General Public License). To apply these terms, attach the following notices to the library. It is safest to attach them to the start of each source file to most effectively convey the exclusion of warranty; and each file should have at least the "copyright" line and a pointer to where the full notice is found. <one line to give the library's name and a brief idea of what it does.> Copyright (C) <year> <name of author> This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Also add information on how to contact you by electronic and paper mail. You should also get your employer (if you work as a programmer) or your school, if any, to sign a "copyright disclaimer" for the library, if necessary. Here is a sample; alter the names: Yoyodyne, Inc., hereby disclaims all copyright interest in the library `Frob' (a library for tweaking knobs) written by James Random Hacker. <signature of Ty Coon>, 1 April 1990 Ty Coon, President of Vice That's all there is to it! AUTHORS 0000644 00000006701 15125347732 0005631 0 ustar 00 libosinfo Authors ================= Current maintainers: Christophe Fergeau <cfergeau@redhat.com> Cole Robinson <crobinso@redhat.com> Daniel P. Berrange <berrange@redhat.com> Fabiano Fidêncio <fidencio@redhat.com> Guido Günther <agx@sigxcpu.org> Marc-André Lureau <marcandre.lureau@redhat.com> Pino Toscano <ptoscano@redhat.com> Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Previous maintainers: Arjun Roy <arroy@redhat.com> Patches contributed by: Alexey Shabalin <a.shabalin@gmail.com> Alexey Shabalin <shaba@altlinux.org> Andika Triwidada <andika@gmail.com> Andrea Bolognani <abologna@redhat.com> Andreas Henriksson <andreas@fatal.se> Andrew Stormont <andyjstormont@gmail.com> Anonymous <noreply@weblate.org> Antoni Bella Pérez <antonibella5@yahoo.com> Arjun Roy <arroy@redhat.com> Charles Arnold <carnold@suse.com> Christophe Fergeau <cfergeau@redhat.com> Cole Robinson <crobinso@redhat.com> Daniel P. Berrange <berrange@redhat.com> Daniel P. Berrangé <berrange@redhat.com> David Zeuthen <davidz@redhat.com> Debarshi Ray <debarshir@gnome.org> Diego Escalante Urrelo <diegoe@gnome.org> Emilio Herrera <ehespinosa57@gmail.com> Eric Blake <eblake@redhat.com> Ettore Atalan <atalanttore@googlemail.com> Fabian Affolter <mail@fabian-affolter.ch> Fabiano Fidêncio <fabiano@fidencio.org> Fabiano Fidêncio <fidencio@redhat.com> Fabio Tomat <f.t.public@gmail.com> Felipe Borges <feborges@redhat.com> Felipe Borges <felipe10borges@gmail.com> Felipe Borges <felipeborges@gnome.org> Gianluca Sforna <giallu@gmail.com> Giuseppe Scrivano <gscrivan@redhat.com> Guido Günther <agx@sigxcpu.org> Hela Basa <r45xveza@pm.me> Jan Kuparinen <copper_fin@hotmail.com> Jasper Lievisse Adriaanse <jasper@humppa.nl> Jim Meyering <meyering@redhat.com> Julien Humbert <julroy67@gmail.com> Kukuh Syafaat <kukuh@syafaat.id> Lasse Schuirmann <lasse.schuirmann@gmail.com> Laszlo Ersek <lersek@redhat.com> Lin Ma <lma@suse.com> Manuela Silva <mmsrs@sky.com> Marc-André Lureau <marcandre.lureau@gmail.com> Marc-André Lureau <marcandre.lureau@redhat.com> Marc Deslauriers <marc.deslauriers@ubuntu.com> Marcus Karlsson <mk@acc.umu.se> Martin Blanchard <tchaik@gmx.com> Matthew Booth <mbooth@redhat.com> Michal Privoznik <mprivozn@redhat.com> Mikel Olasagasti <mikel@olasagasti.info> Milo Casagrande <milo@milo.name> Natanael Copa <ncopa@alpinelinux.org> Oğuz Ersen <oguzersen@protonmail.com> Pavel Hrdina <phrdina@redhat.com> Pierre-Alain TORET <pierre-alain.toret@protonmail.com> Pino Toscano <ptoscano@redhat.com> Pino Toscano <toscano.pino@tiscali.it> Piotr Drąg <piotrdrag@gmail.com> Programmingkid <programmingkidx@gmail.com> Rafael Fonseca <r4f4rfs@gmail.com> Rafael Fontenelle <rafaelff@gnome.org> Rick Kerkhof <rick.2889@gmail.com> Ricky Tigg <ricky.tigg@gmail.com> Roman Bogorodskiy <bogorodskiy@gmail.com> Ryan Lortie <desrt@desrt.ca> Satoru Satoh <ssato@redhat.com> Seong-ho Cho <darkcircle.0426@gmail.com> simmon <simmon@nplob.com> T A Mahadevan <ta.mahadevan@gmail.com> Timm Bäder <tbaeder@redhat.com> Vera Cholasta <vbudikov@redhat.com> Věra Cholasta <vbudikov@redhat.com> Vicente Chaves de Melo <vchaves@ymail.com> Victor Toso <victortoso@redhat.com> Vincent <vincent.dao@asianux.com> Vladimir Beneš <vbenes@redhat.com> Wanlong Gao <gaowanlong@cn.fujitsu.com> Weblate <noreply@weblate.org> Yuri Chornoivan <yurchor@ukr.net> Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Zeeshan Ali <zeeshanak@gnome.org> Zeeshan Ali <zeeshan@kinvolk.io> ...send patches to get your name here... -- End README 0000644 00000002625 15125347732 0005442 0 ustar 00 libosinfo ========= libosinfo is a GObject based library API for managing information about operating systems, hypervisors and the (virtual) hardware devices they can support. It includes a database containing device metadata and provides APIs to match/identify optimal devices for deploying an operating system on a hypervisor. Via the magic of GObject Introspection, the API is available in all common programming languages with demos for javascript (GJS/Seed) and python (PyGObject). Vala bindings are also provided. libosinfo is Free Software and licenced under LGPLv2+. Information about the latest official releases can be found at: https://libosinfo.org/download/ Dependencies ============ - Required: - meson >= 0.49.0 - gobject-2.0 - gio-2.0 - libxml-2.0 - libxslt-1.0 - Optional: - gobject-introspection - Vala (build-time only) Patch submissions ================= Patch submissions are welcomed from any interested contributor. The preferred way to submit patches for libosinfo is opening a Merge Request against: https://gitlab.com/libosinfo/libosinfo General discussions can also happen at the mailing list: https://www.redhat.com/mailman/listinfo/libosinfo For further information about libosinfo please consult the project homepage https://libosinfo.org/ Testing ======= In order to test libosinfo, you should run: $ meson build $ ninja -C build check --End ChangeLog 0000644 00001620420 15125347732 0006334 0 ustar 00 2022-02-14 Victor Toso <victortoso@redhat.com> Update NEWS for 1.10.0 release 2022-02-11 Laszlo Ersek <lersek@redhat.com> loader: work around lstat()/EACCES regression in _g_local_file_info_get() In glib commit 71e7b5800a31 ("Handle MLS selinux policy better", 2010-07-08), which was made for <https://bugzilla.gnome.org/show_bug.cgi?id=623692>, an lstat() failure with error code EACCES was *masked* in function _g_local_file_info_get(). Consequently, if osinfo_loader_find_files() calls g_file_query_info() on a file that is inaccessible due to (e.g.) a missing "x" (search) permission on a leading directory, then g_file_query_info() succeeds, our "skipMissing" branch is dead, g_file_info_get_attribute_uint32() is reached, and it returns G_FILE_TYPE_UNKNOWN. As a consequence, the outer osinfo_loader_process_default_path() function can fail, even though it passes skipMissing=TRUE to osinfo_loader_process_list(). Example: > $ HOME=/root \ > OSINFO_SYSTEM_DIR=/usr/share/osinfo \ > build/tools/osinfo-query os > Error loading OS data: Can't read path /root/.config/osinfo Arguably, this situation should be handled by simply skipping the inaccessible path, as if all leading directories could be searched, and only the last pathname compontent (the filename entry) didn't exist in its direct parent directory. The glib regression was reported in 2017: https://bugzilla.gnome.org/show_bug.cgi?id=777187 and then migrated to gitlab: https://gitlab.gnome.org/GNOME/glib/-/issues/1237 but it's still not solved today. Work around the issue by honoring "skipMissing" on the G_FILE_TYPE_UNKNOWN branch. Demonstration: > $ HOME=/root \ > OSINFO_SYSTEM_DIR=/usr/share/osinfo \ > build/tools/osinfo-query os > > ** (osinfo-query:9924): WARNING **: 13:23:12.776: Can't read path /root/.config/osinfo > Short ID | Name | Version | ID > ----------------+------------------+---------+---------------------------------------- > alpinelinux3.5 | Alpine Linux 3.5 | 3.5 | http://alpinelinux.org/alpinelinux/3.5 > ... Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2051559 2022-02-04 Daniel P. Berrangé <berrange@redhat.com> tools: add an --all flag to osinfo-detect to report all matches By default it still only displays the first operating system with matching media/tree, but if --all is given, all possible matches will be displayed osinfo: add API for resolving multiple tree matches A provided tree object might match multiple entries from the database. Add an API that allows for this possibility to let applications decide how to handle it. osinfo: add API for resolving multiple media matches A provided media object might match multiple entries from the database. Add an API that allows for this possibility to let applications decide how to handle it. osinfo: allow tree matching code to report multiple entries Refactor the tree matching code so that it can be told to return all matching entries, rather than only the first. The original behaviour can be requested to short-circuit matching if the caller only requires the first match. osinfo: allow media matching code to report multiple entries Refactor the media matching code so that it can be told to return all matching entries, rather than only the first. The original behaviour can be requested to short-circuit matching if the caller only requires the first match. osinfo: honour tree architecture when matching The tree matching code currently ignores the tree architecture which is generally ok, since osinfo_tree_create_from_location will leave it set to NULL. None the less, if an architecture is provided for the unknown tree, we should only return results that match this architecture. osinfo: honour media architecture when matching The media matching code currently ignores the media architecture which is generally ok, since osinfo_media_create_from_location will leave it set to NULL. None the less, if an architecture is provided for the unknown media, we should only return results that match this architecture. guests: use a OsinfoTreeList when detecting tree matches In preparation for returning multiple tree matches, pass a OsinfoTreeList object around. osinfo: don't set the 'os' property on matched tree We just got the OsinfoTee list from the OsinfoOs object, thus we know that it already has the 'os' property set. guests: use a OsinfoMediaList when detecting media matches In preparation for returning multiple media matches, pass a OsinfoMediaList object around. osinfo: refactor tree guessing internal helper We want the return value from the tree guessing helper to indicate whether any matches were used, and have the OsinfoOs as an output parameter which can now be optional. osinfo: refactor media guessing internal helper We want the return value from the media guessing helper to indicate whether any matches were used, and have the OsinfoOs as an output parameter which can now be optional. osinfo: pull tree matching logic into public API The logic for matching an unidentified tree against a reference tree is potentially useful to applications and should be part of a public API, rather than hidden. osinfo: pull media matching logic into public API The logic for matching an unidentified media against a reference media is potentially useful to applications and should be part of a public API, rather than hidden. osinfo: add missing line break between methods 2022-01-18 Ettore Atalan <atalanttore@googlemail.com> Translated using Weblate (German) Currently translated at 81.8% (140 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/de/ Co-authored-by: Ettore Atalan <atalanttore@googlemail.com> 2022-01-18 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 73.0% (125 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2022-01-18 Daniel P. Berrangé <berrange@redhat.com> osinfo: use consistent include syntax for public headers osinfo: switch to '#pragma once' The current include files use the pattern #ifndef __OOSINFO_BLAH_H__ # define __OOSINFO_BLAH_H__ #endif /* __OOSINFO_BLAH_H__ */ Since we already mandated GCC + CLang, we can replace this with a simpler incantation #pragma once osinfo: fix bad preprocessor indentation 2021-11-29 Victor Toso <victortoso@redhat.com> loader: add check for unknown file type So we can provide a meaningful error message in case the provided path is not accessible to running process. e.g: running HOME=/root osinfo-query os we would get `Error loading OS data: Unexpected file type` and now we get `Error loading OS data: Can't read path /root/.config/osinfo` This error was first hit with v2v [0] that was leaking $USER and $HOME of root user when osinfo-query as vsdm user with `sudo -c vdsm`. The example above is a simple way to show lack of permision of osinfo-query to read the root's $HOME. [0] https://bugzilla.redhat.com/show_bug.cgi?id=1901423 Related: https://bugzilla.redhat.com/show_bug.cgi?id=1902720 2021-11-24 Fabiano Fidêncio <fabiano@fidencio.org> os: Add `_get_complete_firmware_list()` After realising the mistake of only returning **supported** firmwares on `_get_firmware_list()`, let's work that limitation around by adding a new function, `osinfo_os_get_complete_firmware_list()`, which returns the complete list of firmwares, including both supported and unsupported ones. os: _get_firmware_list() returns supported FWs Let's make sure we document that `_get_firmware_list()` returns **only** supported firmwares, rather than returning **all** the firmwares avaiable for an OS. As part of the documentation change, let's also adapt the internal functions to reflect that. os: Add missing docs for _add_firmware() osinfo_os_add_firmware() has been added as part of 1.7.0 release but, somehow, its documentation has been missing. 2021-11-12 Andrea Bolognani <abologna@redhat.com> ci: Add new targets lcitool now supports CentOS Stream 9, Debian 11 and Fedora 35. ci: Enable RPM jobs on CentOS Now that we have a recent version of meson installed from distro packages instead of PyPI, RPM builds will no longer fail. ci: Refresh generated files Notable differences: * meson is now installed from the distro repositories instead of PyPI on CentOS 8 and CentOS Stream 8. We can finally do this because the version they ship has been updated. * intltool and itstool (which are used to build osinfo-db as part of the CI pipeline) are no longer included in the container images. osinfo-db-tools moved away from the former and never adopted the latter. ci: Drop Fedora 33 It's going to be EOL in a matter of days, and lcitool already dropped support for it. 2021-10-10 Daniel P. Berrangé <berrange@redhat.com> Skip over entity reference checks if loading failed If we fail to parse one of the XML files, we abort loading the database. We then carry on to do an entity reference check which will almost certainly fail due to not having loaded many XML files. We need to skip over the entity reference check on failure to avoid spamming the user with bogus errors. Warn about unexpected files in the database To help remind users that files must end in .xml, warn about any unusually named files. There are a couple of files we should not warn about since we expect them to be present - VERSION - LICENSE The entire 'schema' sub-directory should be ignored. Common editor backups "~" and ".bak" should be ignored. Fix hiding of database entries The osinfo-db documentation says that a local admin/user can hide an entity from the system location: [quote] If the file is zero-length or points to /dev/null, then this represents a black-out override. This indicates that the ENTITY-NAME.xml file from a lower priority directory MUST NOT be loaded. [/quote] This does not, however, work at all. If a zero length file is given, it tries to parse this as XML and fails, causing the entire DB loading process to be aborted. This leaves the DB with dangling references. If a symlink to /dev/null is given, it is entirely ignored because the file is a block device and we only try to load regular files. Instead of only loading regular files, we need to load any type of file that has a .xml extension. This appears to be broken since the very first impl of the new DB loading process, despite having written the spec at the same time. 2021-10-06 Felipe Borges <felipeborges@gnome.org> tree, media: Soup.Session.send_async now takes an IO priority argument See https://libsoup.org/libsoup-3.0/SoupSession.html#soup-session-send-async tree, media: SoupMessage is a private struct in libsoup3 You can no longer directly access various structs such as SoupMessage. These are now accessed by getters and setters. See https://libsoup.org/libsoup-3.0/ch02.html tree, media: SOUP_SESSION_USER_AGENT isn't defined in libosup3 Let's use the literal "user-agent" instead. build: Require libsoup3 See https://blog.tingping.se/2021/02/23/future-of-libsoup.html This will fallback to libsoup-2.4 when libsoup3 isn't available. 2021-09-27 Daniel P. Berrangé <berrange@redhat.com> build: don't set glib version constraints for g-ir-scanner add_project_arguments() sets flags that apply to all invokations of the C compiler toolchain by meson. On the surface it sounds fine to use this for setting -DGLIB_VERSION_MIN_REQUIRED=VER -DGLIB_VERSION_MAX_ALLOWED=VER as we want all our code to be constrained by these declared glib versions to prevent us accidentally using APIS from newer glib by mistake. A subtle problem was revealed with the arrival of gobject-introspection version 1.70. The g-ir-scanner program auto-generates some glib code for handling introspection, and this generated code uses glib APIs that are newer than our declared version and this triggers compile failures tmp-introspectg6xadxkr/Libosinfo-1.0.c:251:3: error: ‘G_TYPE_FLAG_FINAL’ is deprecated: Not available before 2.70 [-Werror=deprecated-declarations] 251 | if (G_TYPE_IS_FINAL (type)) | ^~ In file included from /usr/include/glib-2.0/gobject/gobject.h:24, from /usr/include/glib-2.0/gobject/gbinding.h:29, from /usr/include/glib-2.0/glib-object.h:22, from tmp-introspectg6xadxkr/Libosinfo-1.0.c:30: /usr/include/glib-2.0/gobject/gtype.h:1050:3: note: declared here 1050 | G_TYPE_FLAG_FINAL GLIB_AVAILABLE_ENUMERATOR_IN_2_70 = (1 << 6) | ^~~~~~~~~~~~~~~~~ tmp-introspectg6xadxkr/Libosinfo-1.0.c:251:13: error: Not available before 2.70 [-Werror] 251 | if (G_TYPE_IS_FINAL (type)) | ^~~~~~~~~~~~~~~~~ This is actually harmless, because systems with an older glib will also have older g-ir-scanner and thus not be using these new APIs. We need to exclude the glib version constraints from code generated by glib tools, and thus means we have to stop using add_project_arguments() and set cflags explicitly on each target. 2021-09-06 Rafael Fontenelle <rafaelff@gnome.org> Translated using Weblate (Portuguese (Brazil)) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pt_BR/ Co-authored-by: Rafael Fontenelle <rafaelff@gnome.org> 2021-09-06 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 65.4% (112 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2021-09-03 Daniel P. Berrangé <berrange@redhat.com> ci: remove obsolete refresh script and documentation We now use lcitool's manifest feature to generate files. ci: re-generate containers/gitlab config from manifest This uses the command "lcitool manifest ci/manifest.yml" to re-generate all existing dockerfiles and gitlab CI config. ci: define a CI manifest file This is to be used with the command "lcitool manifest ci/manifest.yml" to re-generate all existing dockerfiles and gitlab CI config. In a slight change to the existing config, the manifest will always use the pre-build osinfo-db-tools package. Nothing in the libosinfo CI tasks, has a need to exercise latest git master for osinfo-db-tools. The latter is covered fine by osinfo-db CI jobs. gitlab: remove debian-9 build target This platform is no longer supported by libvirt-ci. gitlab: remove centos-7 build target This platform is no longer supported by libvirt-ci. 2021-08-25 Hela Basa <r45xveza@pm.me> Translated using Weblate (Sinhala) Currently translated at 0.5% (1 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/si/ Co-authored-by: Hela Basa <r45xveza@pm.me> 2021-08-25 Andika Triwidada <andika@gmail.com> Translated using Weblate (Indonesian) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ Co-authored-by: Andika Triwidada <andika@gmail.com> 2021-08-25 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 64.9% (111 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Translated using Weblate (Finnish) Currently translated at 33.9% (58 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2021-08-05 simmon <simmon@nplob.com> Translated using Weblate (Korean) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Co-authored-by: simmon <simmon@nplob.com> 2021-08-05 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ Co-authored-by: Oğuz Ersen <oguzersen@protonmail.com> 2021-08-05 Yuri Chornoivan <yurchor@ukr.net> Translated using Weblate (Ukrainian) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/uk/ Co-authored-by: Yuri Chornoivan <yurchor@ukr.net> 2021-08-05 Pino Toscano <toscano.pino@tiscali.it> Translated using Weblate (Italian) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/it/ Co-authored-by: Pino Toscano <toscano.pino@tiscali.it> 2021-08-05 Piotr Drąg <piotrdrag@gmail.com> Translated using Weblate (Polish) Currently translated at 100.0% (171 of 171 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pl/ Co-authored-by: Piotr Drąg <piotrdrag@gmail.com> 2021-08-05 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Co-authored-by: Weblate <noreply@weblate.org> Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ Translation: libosinfo/libosinfo 2021-08-05 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 32.3% (55 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2021-08-04 Pino Toscano <ptoscano@redhat.com> po: regenerate libosinfo.pot 2021-07-27 Daniel P. Berrangé <berrange@redhat.com> ci: refresh centos-stream container It is no longer possible to just install an extra package to run regular centos into centos stream. gitlab: use custom docker:dind image The current docker:dind container has broken default seccomp filter that results in clone3 being blocked, which in turn breaks Fedora 35 rawhide. This custom image has a workaround that causes the seccomp filter to return ENOSYS for clone3 instad of EPERM, thus triggering glibc to fallback to clone correctly. 2021-07-22 Seong-ho Cho <darkcircle.0426@gmail.com> Translated using Weblate (Korean) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Co-authored-by: Seong-ho Cho <darkcircle.0426@gmail.com> 2021-04-15 simmon <simmon@nplob.com> Translated using Weblate (Korean) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Co-authored-by: simmon <simmon@nplob.com> 2021-04-13 simmon <simmon@nplob.com> Translated using Weblate (Korean) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Co-authored-by: simmon <simmon@nplob.com> 2021-04-07 simmon <simmon@nplob.com> Translated using Weblate (Korean) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Co-authored-by: simmon <simmon@nplob.com> 2021-04-05 simmon <simmon@nplob.com> Translated using Weblate (Korean) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ko/ Added translation using Weblate (Korean) Co-authored-by: simmon <simmon@nplob.com> 2021-03-07 Hela Basa <r45xveza@pm.me> Added translation using Weblate (Sinhala) Co-authored-by: Hela Basa <r45xveza@pm.me> 2021-03-07 Ricky Tigg <ricky.tigg@gmail.com> Translated using Weblate (Finnish) Currently translated at 25.2% (43 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Translated using Weblate (Finnish) Currently translated at 23.5% (40 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Ricky Tigg <ricky.tigg@gmail.com> 2021-03-07 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 19.4% (33 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2021-03-07 Ricky Tigg <ricky.tigg@gmail.com> Translated using Weblate (Finnish) Currently translated at 18.8% (32 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Ricky Tigg <ricky.tigg@gmail.com> 2021-02-24 Fabiano Fidêncio <fidencio@redhat.com> tools: Add "Release status" info to `osinfo-query os` Release status has never been part of `osinfo-query os`, but it's been proven to be a needed information, mainly because the rolling distro may not even have their release & eol date filled up. Related: https://bugzilla.redhat.com/show_bug.cgi?id=1932258 2021-02-10 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.10.0 2021-02-02 Fabiano Fidêncio <fidencio@redhat.com> Update NEWS for 1.9.0 release 2021-01-11 Jan Kuparinen <copper_fin@hotmail.com> Translated using Weblate (Finnish) Currently translated at 15.8% (27 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fi/ Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com> 2021-01-11 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Co-authored-by: Weblate <noreply@weblate.org> Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ Translation: libosinfo/libosinfo 2021-01-08 Timm Bäder <tbaeder@redhat.com> install config: Rename a parameter 'timezone' is a global variable in time.h, so clang complains about that. build: Move bsd/mac checks to be compiler-specific They depend on the compiler in use, not the operating system 2020-12-15 Fabian Affolter <mail@fabian-affolter.ch> Translated using Weblate (German) Currently translated at 64.7% (110 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/de/ Co-authored-by: Fabian Affolter <mail@fabian-affolter.ch> 2020-12-15 Daniel P. Berrangé <berrange@redhat.com> ci: refresh containers for CentOS-8 PowerTools repo rename A recent CentOS-8 update renamed the "PowerTools" repo to "powertools" and since dnf is case sensitive wrt repo names, this broke ability to build new containers. The refresh fixes the repo name and pulls in other misc improvements to containers. 2020-12-01 Daniel P. Berrangé <berrange@redhat.com> gitlab: update refresh script for changed host naming convention gitlab: refresh containers with lcitool for fully minimized base gitlab: re-generate container images from lcitool This introduces Fedora 33 and removes some redundant packages. Note that rpm build is marked as skip on CentOS 8/Stream now, since we're using pip provide meson instead of distro meson, and thus don't have the %meson macro. 2020-11-13 Cole Robinson <crobinso@redhat.com> loader: Print filename in libxml error string This is how libvirt handles it. Would help debugging apparently random failures like these bugs: https://bugzilla.redhat.com/show_bug.cgi?id=1897356 https://bugzilla.redhat.com/show_bug.cgi?id=1651550 2020-09-15 Milo Casagrande <milo@milo.name> Translated using Weblate (Italian) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/it/ Co-authored-by: Milo Casagrande <milo@milo.name> 2020-09-14 Milo Casagrande <milo@milo.name> Translated using Weblate (Italian) Currently translated at 97.0% (165 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/it/ Co-authored-by: Milo Casagrande <milo@milo.name> 2020-09-14 Pino Toscano <toscano.pino@tiscali.it> Translated using Weblate (Italian) Currently translated at 97.0% (165 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/it/ Co-authored-by: Pino Toscano <toscano.pino@tiscali.it> 2020-09-14 Fabio Tomat <f.t.public@gmail.com> Translated using Weblate (Friulian) Currently translated at 66.4% (113 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fur/ Translated using Weblate (Spanish) Currently translated at 12.9% (22 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/es/ Added translation using Weblate (Friulian) Co-authored-by: Fabio Tomat <f.t.public@gmail.com> 2020-09-14 Alexey Shabalin <a.shabalin@gmail.com> Added translation using Weblate (Russian) Co-authored-by: Alexey Shabalin <a.shabalin@gmail.com> 2020-09-14 Mikel Olasagasti <mikel@olasagasti.info> Added translation using Weblate (Basque) Co-authored-by: Mikel Olasagasti <mikel@olasagasti.info> 2020-08-15 Emilio Herrera <ehespinosa57@gmail.com> Translated using Weblate (Spanish) Currently translated at 12.9% (22 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/es/ 2020-08-08 Rafael Fontenelle <rafaelff@gnome.org> Translated using Weblate (Portuguese (Brazil)) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pt_BR/ 2020-08-08 Kukuh Syafaat <kukuh@syafaat.id> Translated using Weblate (Indonesian) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ 2020-07-31 Julien Humbert <julroy67@gmail.com> Translated using Weblate (French) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fr/ 2020-07-29 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-07-29 Yuri Chornoivan <yurchor@ukr.net> Translated using Weblate (Ukrainian) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/uk/ 2020-07-28 Piotr Drąg <piotrdrag@gmail.com> Translated using Weblate (Polish) Currently translated at 100.0% (170 of 170 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pl/ 2020-07-28 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ 2020-07-28 Pino Toscano <ptoscano@redhat.com> po: regenerate libosinfo.pot 2020-07-28 Rafael Fontenelle <rafaelff@gnome.org> Fix misspellings 2020-07-28 Pino Toscano <ptoscano@redhat.com> tools,detect: improve command line option help Literal identifiers, such as 'media', 'tree', etc in translatable strings are a bad idea, since they will likely be translated, misguiding users reading the help of a translated version. Refactor a bit the way these bits are show to the user: - in the help of --format and --type, show a generic name of the argument value, and slightly improve the description text - add a small description text after the listing of the command line options to mention the allowed values Fixes: #42 2020-07-28 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ 2020-07-28 Rafael Fontenelle <rafaelff@gnome.org> Translated using Weblate (Portuguese (Brazil)) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pt_BR/ Translated using Weblate (Portuguese (Brazil)) Currently translated at 97.6% (165 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pt_BR/ 2020-07-28 Fabiano Fidêncio <fidencio@redhat.com> spec: Explicitly call %meson_build in the %build section It was never needed before as the %meson_install macro would perform a project rebuild. However, most recent versions of meson have changed the %meson_install macro to use `--no-rebuild`, exposing then this issue on our spec file. Fixes: #43 2020-06-22 Julien Humbert <julroy67@gmail.com> Translated using Weblate (French) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fr/ 2020-06-17 Andrea Bolognani <abologna@redhat.com> ci: Don't use RPM=skip for container builds It only makes sense for actual build jobs. ci: Check for rpmbuild availability Only try to build RPMs if the rpmbuild command is present on the system. This means we no longer have to use the RPM=skip trick on Debian-based targets, and can limit it to cases where building RPMs is actually broken for some reason. ci: Simplify rpmbuild command We don't have to worry about cleaning up after ourselves, since we're running the build inside an ephemeral container. ci: Add CentOS Stream Starting with commit d35a7dba9cc8ac3bb0647171628731c625d4e14f Author: Erik Skultety <eskultet@redhat.com> Date: Thu Jun 4 16:17:36 2020 +0200 guests: Add CentOS Stream The idea behind Stream is to get a preview of what the next RHEL-minor will look like, thus getting newer versions of packages than regular CentOS provides. It's supposed to be rolling, so currently it's based on CentOS 8, but once RHEL-9 is out, it will point to the next RHEL-9 instead. However, unlike debian-sid it doesn't have a dedicated URL resource (something like centos-stream), so we'll have to change it manually once it's flipped to point to next RHEL-9. Signed-off-by: Erik Skultety <eskultet@redhat.com> lcitool supports CentOS Stream as a build target. ci: Use $MESON_OPTS This information is part of the container image now. ci: Refresh Dockerfiles Starting with commit 25c9da0aaa94fbca674dca6dec65f6348f335e7d Author: Andrea Bolognani <abologna@redhat.com> Date: Fri May 22 17:38:10 2020 +0200 lcitool: Generate MESON_OPTS in Dockerfiles This serves the same purpose as CONFIGURE_OPTS: enable seamless cross-compilation support, in this case for Meson-based projects. Only MinGW cross-compilation targets are currently supported. Signed-off-by: Andrea Bolognani <abologna@redhat.com> lcitool will include a $MESON_OPTS environment variable, which can be used to simplify our GitLab CI build rules, in the Dockerfiles it generates. ci: Move Dockerfiles to a subdir This is consistent to how libvirt organizes its files, and ensures things will remain tidy even as we introduce new features such as, later on, Cirrus CI support. ci: Clean up refresh script a bit Remove some repetition through refactoring. 2020-06-02 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.9.0 2020-05-29 Fabiano Fidêncio <fidencio@redhat.com> Update NEWS for 1.8.0 release 2020-05-29 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ 2020-05-29 Julien Humbert <julroy67@gmail.com> Translated using Weblate (French) Currently translated at 98.8% (167 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fr/ 2020-05-28 Andrea Bolognani <abologna@redhat.com> ci: Use GitLab container registry Instead of using pre-built containers hosted on Quay, build containers as part of the GitLab CI pipeline and upload them to the GitLab container registry for later use. This does not slow down builds, because containers are only rebuilt when the corresponding Dockerfile has been modified. 2020-05-22 Manuela Silva <mmsrs@sky.com> Translated using Weblate (Portuguese (Portugal)) Currently translated at 57.9% (98 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pt_PT/ 2020-05-22 Anonymous <noreply@weblate.org> Translated using Weblate (Indonesian) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ 2020-05-22 Andika Triwidada <andika@gmail.com> Translated using Weblate (Indonesian) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ 2020-05-22 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-05-22 Yuri Chornoivan <yurchor@ukr.net> Translated using Weblate (Ukrainian) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/uk/ 2020-05-21 Piotr Drąg <piotrdrag@gmail.com> Translated using Weblate (Polish) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pl/ 2020-05-21 Anonymous <noreply@weblate.org> Translated using Weblate (Indonesian) Currently translated at 1.1% (2 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ Translated using Weblate (Polish) Currently translated at 100.0% (169 of 169 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/pl/ 2020-05-21 Manuela Silva <mmsrs@sky.com> Added translation using Weblate (Portuguese (Portugal)) 2020-05-20 Weblate <noreply@weblate.org> Update translation files Updated by "Update LINGUAS file" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ 2020-05-20 Pino Toscano <ptoscano@redhat.com> po: remove duplicated msgid in 'id' catalog Fix it manually, otherwise gettext tools cannot work with it. 2020-05-20 Weblate <noreply@weblate.org> Update translation files Updated by "Update PO files to match POT (msgmerge)" hook in Weblate. Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/ 2020-05-14 Daniel P. Berrangé <berrange@redhat.com> build-aux: use gmake for syntax-check rules Previously on FreeBSD the syntax-check rules would be a no-op because its non-GNU make wouldn't see the GNUmakefile and would report no work to do: $ MESON_SOURCE_ROOT=`pwd` build-aux/syntax-check `syntax-check' is up to date. Since we moved everything into a syntax-check.mk that we call explicitly though, we're the non-GNU make tries to execute rules that depend on GNU make features. build-aux: fix references to old maint.mk name 2020-05-13 Daniel P. Berrangé <berrange@redhat.com> build-aux: rename maint.mk to syntax-check.mk This name aligns with what the makefile actally does. build-aux: remove GNUmakefile entirely The only useful bits it still did were setting a couple of variables, which can be moved into the main maint.mk file. build-aux: merge cfg.mk into maint.mk Since we took full ownership of a cut down maint.mk there's no need to keep local additions separately in cfg.mk This fixes a bug which caused the contents of cfg.mk to be skipped since the moving of files into the build-aux dir and switch to meson. build-aux: trim obsolete maint rules Many rules in maint.mk are doing checks related to makefiles, or autoconf, or gnulib, none of which are used by libosinfo. There are also many misc rules related to publishing of releases that are never used. Trimming this out makes the resulting file more managable. 2020-05-13 Pino Toscano <ptoscano@redhat.com> po: update LINGUAS Add tr. id is not added as it has errors; hope that the update of the catalog will trigger an update in weblate, generating a correct po file. Drop Zanata leftovers libosinfo uses Weblate for translations. po: regenerate libosinfo.pot 2020-05-04 Fabiano Fidêncio <fidencio@redhat.com> README: MR is the preferred way to submit a patch Unlikely of what our README mentioned, the preferred way to submit a patch to libosinfo is opening a Merge Request against our GitLab project. 2020-05-03 Andika Triwidada <andika@gmail.com> Translated using Weblate (Indonesian) Currently translated at 0.5% (1 of 167 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/id/ 2020-05-01 Andrea Bolognani <abologna@redhat.com> ci: Use 'check-dco' container for DCO checking This container image, which has been created for this specific purporse and is already used by many projects under the libvirt umbrella, contains a nicer Python implementation of the require-dco.sh script that's guaranteed to be kept up to date. ci: Merge templates into native-build-job At this point, the only difference between the existing templates is that one of them includes the RPM build job, and since we now have a way to skip that part for specific targets, we can get rid of the duplication and merge them into a single template. ci: Introduce multilib This is a bit yucky, but it's only temporary: like the target ABI, this information is tied to the container image and as such should be baked into it. When that happens, we can simply drop this altogether, but in the meantime this bit of scripting allow us to have a single environment instead of two. ci: Merge templates into mingw-build-job The mingw32 and mingw64 templates are basically identical at this point, so let's merge them into a generic one. ci: Split off mingw-environment Follow the environment/script split, like we do everywhere else. We can have a single environment that works both for mingw32 and mingw64 builds because the only difference between the two is which ABI they target, and we can obtain that piece of information from the $ABI environment variable that's baked into the container images. ci: Move more information to the job templates In particular, we can avoid repeating the parts of the container image name that are always the same. ci: Introduce pipeline stages There's only a single stage for now. ci: Use standard template for CentOS 7 and openSUSE 15.1 The only difference between the open-coded job definition currently in use for these target platforms and the standard template is that the former don't include the RPM build job, but now that we have a way to skip that part selectively we can remove these special cases. ci: Make RPM job skippable We're sort of cheating by exiting right away when asked to skip the RPM job, but we know that it's always going to be the last job in a script and we currently have no need to skip any other job, so let's start with the simplest possible implementation and consider something more complex if and when the need arises. 2020-04-30 Andrea Bolognani <abologna@redhat.com> ci: Move builds from Ubuntu 16.04 to Ubuntu 20.04 ci: Move builds from Fedora 30 to Fedora 32 ci: Move MinGW builds from Fedora 30 to Fedora Rawhide 2020-04-30 Andika Triwidada <andika@gmail.com> Added translation using Weblate (Indonesian) 2020-04-21 Gianluca Sforna <giallu@gmail.com> Translated using Weblate (Italian) Currently translated at 97.6% (164 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/it/ 2020-04-20 Andrea Bolognani <abologna@redhat.com> prepare-release: Move release archive out of DIST_ROOT If the build has been successful, we can fish the release archive out of DIST_ROOT and expose it in the top-level source directory, just like autotools would do, for extra convenience. prepare-release: Define DIST_ROOT This is the subdirectory of BUILD_ROOT where the release archive is generated by Meson: it's already used multiple times, and we're about to introduce another one, so let's give it its own variable. prepare-release: Define BUILD_ROOT This fits nicely with INSTALL_ROOT and the recently-introduced RPMBUILD_ROOT; additionally, since $BUILD_ROOT is an absolute path, we can use it directly as _sourcedir when calling rpmbuild. prepare-release: Use a local directory for rpmbuild The current release process stores files inside the user's rpmbuild directory, usually part of $HOME, which is just bad manners. Use a local directory instead. prepare-release: Use a local directory for installation The current release process stores files inside the user's home directory, which is just bad manners. Use a local directory instead. Since we no longer have to care about the destination directory looking like an actual root filesystem, we can make the paths used when installing MinGW builds shorter. prepare-release: Use DESTDIR when installing DESTDIR and --prefix are similar but serve distinct purposes: in particular, the latter is intended for modifications that will persist after installation, eg. storing binaries under /usr/bin instead of /usr/local/bin, while the former is just a temporary change that can be easily reverted, eg. stashing files in an intermediate, unprivileged directory while building a package. Our needs during the release process are closer to what DESTDIR is designed for, so let's use that. 2020-04-17 Andrea Bolognani <abologna@redhat.com> prepare-release: Ensure rpmbuild is executable Using 'test -x' is more appropriate than 'test -r' for programs. prepare-release: Drop RESULTS This is a leftover from when the script was used to perform automated tests: in that context, singling out the output of the test suite made sense, but now we're no longer particularly interested in it. As a bonus, this allows us to drop a non-trivial chunk of fairly obscure POSIX shell compatibility code. prepare-release: Remove unnecessary cleanup We already remove all of build/ at the beginning of the build process, so there's no need to manually remove any of its subdirs here. prepare-release: Remove empty lines These are unnecessary and inconsistent with the style used in the rest of the script. prepare-release: Use correct spec file for MinGW RPMs Currently we're just building the very same spec file twice. gitignore: Minimize Things like editor backup files belong in the user's own ignore file rather than in the project's. 2020-04-16 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-04-15 Fabiano Fidêncio <fidencio@redhat.com> Add "osinfo_macros.h" to the installed headers As the macros defined there are used in the public headers, this header must be installed, otherwise we'll face errors such as: ``` /home/valabot/gnome3/include/libosinfo-1.0/osinfo/osinfo_entity.h:26:11: fatal error: osinfo/osinfo_macros.h: No such file or directory 26 | # include "osinfo/osinfo_macros.h" | ^~~~~~~~~~~~~~~~~~~~~~~~ ``` Reported-by: Rico Tzschichholz <ricotz@ubuntu.com> 2020-04-15 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-04-13 Rafael Fonseca <r4f4rfs@gmail.com> osinfo_macros: mark inline functions as UNUSED Fixes -Werror=unused-function with clang. See: https://gitlab.gnome.org/GNOME/glib/-/commit/90f93a64f26666b955fdafb56acab54e596f4cc5 osinfo_macros: auto-cleanup func for module class This will allow to manage the ownership of the class when reffing it without having to cast it to GTypeClass. See: https://gitlab.gnome.org/GNOME/glib/-/commit/ab25fa42cc6906df31352877154bb7b07f39552f 2020-04-13 Fabiano Fidêncio <fidencio@redhat.com> productfilter: Check productfilter in _clear_product_constraint{,s} Let's just add a safety check in the _clear_product_constraint() and _clear_product_constraints() functions. devicelinkfilter: Check devicelinkfilter in _get_target_filter() Let's just add a safety check in the _get_target_filter() function. filter: Check filter in _clear_constraint{,s} Let's just add a safety check in the _clear_constraint() and _clear_constraints() functions. db: Remove useless osinfo_db_finalize() declaration install_config_param: Use g_clear_object in _set_value_map() entity: Check entity in osinfo_entity_clear_param() Let's just add a safety check in _clear_param(). Group all GObjectClass override methods Use g_object_class_install_properties() Using _install_properties() helps to keep the code cleaner and also reduces the chance of having wrong props used. Use G_DEFINE_*_WITH_PRIVATE Together with this change, let's also get rid of our own _GET_PRIVATE() macros and use _get_instance_private() method instead. Use OSINFO_DECLARE_TYPE* macros macros: Add OSINFO_DECLARE_TYPE* macros OSINFO_DECLARE_TYPE is nothing else than a define to G_DECLARE_DERIVABLE_TYPE macro from GLib. OSINFO_DECLARE_TYPE_WITH_PRIVATE does exactly the same thing as OSINFO_DECLARE_TYPE plus adds a private structure to the declared type structure. OSINFO_DECLARE_TYPE_WITH_PRIVATE_AND_CLASS does exactly the same thing as OSINFO_DECLARE_TYPE_WITH_PRIVATE plus declares the class structure of the declared type. Using those new macros will help us to reduce a lot the GLib boilerplate in our codebase. 2020-04-13 Rafael Fonseca <r4f4rfs@gmail.com> Do not error out when inlining fails -Winline causes error in GCC10, e.g error: inlining failed in call to ‘OSINFO_LIST’: call is unlikely and code size would grow [-Werror=inline] That particular macro, which will be introduced in the next commit, is bas based on GLib own macros and it just so happens that if they were also compiled with -Winline, the same error would happen [1]. Instead, let just suppress those warnings. [1] https://gitlab.gnome.org/GNOME/glib/issues/2042 2020-04-13 Fabiano Fidêncio <fidencio@redhat.com> Enforce '-Wdeclaration-after-statement' Enforcing that variables are only declared in the beginning of a block makes the code easier to read. image: Add cloud-init property cloud-init has been added on 81c43558687b40d but the property has never been added. datamap: Rename _init() argument name "list" was a c&p from when OsinfoDatamap was introduced. avatar_format: Remove unused OSINFO_AVATAR_FORMAT_GET_PRIVATE() OsinfoAvatarFormat does not have a private member. tests: Remove duplicated tests There's no reason to keep testing union, filter, and intersection on types which inherit from OsinfoList. Mainly considering test-list already test those. Update Copyright year Remove Fabiano Fidêncio's Copyright statements Does not make sense to keep my own Copyright statements in 3 files of the project, which were added during 2012's Google Summer of Code. Remove Authors Authors are not often updated and do not reflect the current state of the code. media/tree: Force "wget" user-agent in SoupSession Spoiler alert, it'll sound insane ... because it is insane. Some servers treat libsoup user-agent differently than how they treat wget user-agent. With the default agent set, Soup would end up downloading the file containing the redirection information, instead of actually redirecting to one of the addresses present in that file. While when forcing "wget" as user-agent, the redirection happens as expected. This issue was faced at least with CentOS[0] and Manjaro[1] links. After a talk with Patrick Griffis, he mentioned that: "I imagine there isn't too much of a problem pretending to be wget. I would guess the assumption by these services is 'A CLI tool is downloading us and not a browser' rather than 'wget has unique behavior'.". Knowing that, let's force "wget" as user-agent on our side. [0]: http://isoredirect.centos.org/centos/8/isos/x86_64/CentOS-8.1.1911-x86_64-dvd1.iso [1]: https://osdn.net/dl/manjaro/manjaro-gnome-19.0.2-200312-linux54.iso 2020-04-11 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-04-10 Rafael Fonseca <r4f4rfs@gmail.com> Add API to track distro cloud-init usernames We already have cloud-init='true' so add a cloud-image-username='foo' to go with it. The property has been added to `OsinfoOs` instead of `OsinfoImage` because it's not expected to change in a per image basis. Instead we can define it once and inherit it, in the same way that it's done for kernel-url-argument. Fixes #39 2020-04-09 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-03-31 Fabiano Fidêncio <fidencio@redhat.com> tests: Add tests for only matching "rolling" as a fallback The reason to have this in a different commit from the code is to ease the backport of the previous patch by the distros which feel like doing that. db: Match "rolling" distros only as a fallback Let's only match "rolling" distros as a fallback. The reason we need to do so is to properly deal with distros as Manjaro, which has a mix of versioned and rolling entries. By matching the "rolling" only as a fallback, we ensure we'll have exhausted all the "versioned" possibilites in the first place. tests: Remove invalid release-status test This should be tested on osinfo-db side and leaving this test here causes failures in other non-related tests, all of them related to identify_media, which gets the release-status and ends up hitting the failures caused by this specific test. media: Fix typo in the "eject-after-install" description "installtion" should be "installation", as pointed out by Oğuz Ersen. Together with the change, libosinfo.pot has been regenerated. 2020-03-30 Andrea Bolognani <abologna@redhat.com> gitlab: Enable ccache usage The $PATH definition has always included a reference to the symlink farm that is used to enable transparent ccache usage, but since the directory itself does not exist in the container image it never did anything. Since libvirt-jenkins-ci commit 2563aebb6c5c, container images do actually contain the symlink directory, so it's finally possible to enable ccache usage. This requires including the correct directory in $PATH, as well as telling 1) ccache where to store its data and 2) GitLab that this data should be preserved across jobs. The exact cache size (50 MB) is sort of arbitrary, but it should be enough to store build artifacts for both osinfo-db-tools and libosinfo for the foreseeable future. gitlab: Create job templates These avoid having to repeat the same build steps over and over. gitlab: Use $MAKE Just like $NINJA, this is defined as part of the container's environment. gitlab: Don't define $NINJA Since libvirt-jenkins-ci commit 27cfddee8835, paths to build tools such as ninja and make are exported in the container's environment and can be used directly. 2020-03-27 Oğuz Ersen <oguzersen@protonmail.com> Translated using Weblate (Turkish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/tr/ 2020-03-26 Oğuz Ersen <oguzersen@protonmail.com> osinfo: Fix typo in the "initrd-path" description "inirtd" -> "initrd" 2020-03-25 Oğuz Ersen <oguzersen@protonmail.com> Added translation using Weblate (Turkish) 2020-03-23 Fabiano Fidêncio <fidencio@redhat.com> po: Import libosinfo.pot Weblate requires the pot file to be commited, in order to be able to add new translations to the project. This file has to be updated every time a new string is added. 2020-03-06 Pino Toscano <ptoscano@redhat.com> media: fix docs of OsinfoMediaDetectFlags Move the "Since" after the description, so the description block (including the version) is properly handled. gitignore: remove almost all old content Some of the content refers to autotools, which is not used anymore. Also, meson enforces builddir!=srcdir, so the generated build files are never directly among the sources. Hence remove almost all the existing content, as it is not relevant. Fix typo in Fabiano' surname 2020-02-28 Pino Toscano <ptoscano@redhat.com> Fix OsinfoDB typo in docs Add few missing "Since" in docs Add the version information for some APIs, so it is easier to know when they were introduced. product: fix doc typo Most probably one rogue s/OS/PRODUCT/ that changed CentOS to CentPRODUCT. firmware: fix doc of osinfo_firmware_get_firmware_type device-driver: add doc for osinfo_device_driver_get_priority 2020-02-26 Fabiano Fidêncio <fidencio@redhat.com> Add osinfo_firmware_is_supported() This function has been missed when rebasing 904b65e7a1dda and. right now, only its declaration in the header file is present. Pointed out by Pino Toscano. 2020-02-25 Andrea Bolognani <abologna@redhat.com> ci: Don't set $PKG_CONFIG_LIBDIR The pkg-config binary used for MinGW builds is cross-compilation aware, so it doesn't need us to point it in the right direction when it comes to globally-installed .pc files. 2020-02-24 Andrea Bolognani <abologna@redhat.com> ci: Perform MinGW builds in dedicated containers Instead of using the same Fedora 30 container for both native and MinGW builds, we are now building separate containers tailored for each scenario. Re-introduce MinGW builds, and have them use the new container images. The corresponding libvirt-dockerfiles commit is ea6286acbbfd. ci: Temporarily disable MinGW builds We're going to use dedicated container images for them going forward, but switching all at once would break CI so let's do it in two steps: first disable the current builds, then enable the new ones once the external requirements (container images) are in place for a smoother transition. 2020-02-16 Julien Humbert <julroy67@gmail.com> Translated using Weblate (French) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fr/ 2020-01-29 Fabiano Fidêncio <fidencio@redhat.com> loader: Avoid conflicting object ids for extension files Currently, when building the id of some of our objects, we do: - Loop over the several objects of that type; - Concatenate the id of the parent object with the loop index; By doing this, conflicts can be seen when using more than one extension file, as osinfo-db does for Windows, where pre and post installable drivers extension files are used and end up being loaded as we'd do: - Loop over pre installable drivers; - Concatenate the id of the parent object with the loop index; - Loop over post installable drivers; - Concatenate the id of the parent object with the loop index; In order to avoid such issues, let's also use the filename as part of the object id, avoiding then conflicts. It's important to note that the conflict is cause by osinfo_list_add(), which will replace the object in the list in case some other object with the very same id is already stored there. 2020-01-25 Julien Humbert <julroy67@gmail.com> Translated using Weblate (French) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.fedoraproject.org/projects/libosinfo/libosinfo/fr/ 2020-01-24 Piotr Drąg <piotrdrag@gmail.com> Translated using Weblate (Polish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.stg.fedoraproject.org/projects/libosinfo/libosinfo/pl/ 2020-01-13 Fabiano Fidêncio <fidencio@redhat.com> Silence -Wstring-plus-int The warning is triggered because we're ignoring the "treeinfo-" part of the OSINFO_TREE_PROP_TREEINFO_* string. A possible way to fix that, instead of just disabling the warning, would be doing something like: ``` const char *prop = OSINFO_TREE_PROP_TREEINFO_FAMILY; prop += strlen("treeinfo-"); ``` However, I sincerely do not think it's worth the verbosity of the solution. 2020-01-13 Piotr Drąg <piotrdrag@gmail.com> Translated using Weblate (Polish) Currently translated at 100.0% (168 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.stg.fedoraproject.org/projects/libosinfo/libosinfo/pl/ 2020-01-10 Fabiano Fidêncio <fidencio@redhat.com> ci: Build osinfo-db in all containers For some reason the job for building osinfo-db was not present in a few distros' containers. ci,centos8: Build project's RPM ci: Add openSUSE 15.1 container ci,centos8: Add DCO check 46db2e234a8ed82 introduced the DCO check. However, the check was not added to the CentOS 8 container. Partially revert "list: Update GObject boilerplate" This partially reverts commit 706b6fd0d97ed572d92c4a33882605d561ebe3c8, which introduced an ABI breakage by using G_DECLARE_DERIVABLE_TYPE when declaring an OsinfoList. 2020-01-09 Antoni Bella Pérez <antonibella5@yahoo.com> Translated using Weblate (Catalan) Currently translated at 88.1% (148 of 168 strings) Translation: libosinfo/libosinfo Translate-URL: https://translate.stg.fedoraproject.org/projects/libosinfo/libosinfo/ca/ Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2020-01-07 Fabiano Fidêncio <fidencio@redhat.com> po: Remove empty files https://gitlab.com/libosinfo/libosinfo/issues/40 2020-01-07 Daniel P. Berrangé <berrange@redhat.com> po: refresh translations from zanata 2020-01-04 Fabiano Fidêncio <fidencio@redhat.com> loader: Introduce OsinfoError enum Let's create an OsinfoError enum instead of just using 0 as the error type. This solves the following instrospection warning: ``` <unknown>:: Warning: Libosinfo: (ErrorQuarkFunction)error_quark: osinfo_error_quark: Couldn't find corresponding enumeration ``` product: Use product instead of prod as argument name Header, source, and documentation were not matching with regarding to the argument product name. Let's standardise them and use product all over the place. By doing this the following warnings will be fixed: ``` ../osinfo/osinfo_product.c:318: Warning: Libosinfo: osinfo_product_get_codename: unknown parameter 'product' in documentation comment, should be 'prod' ../osinfo/osinfo_product.c:329: Warning: Libosinfo: osinfo_product_get_release_date_string: unknown parameter 'product' in documentation comment, should be 'prod' ../osinfo/osinfo_product.c:342: Warning: Libosinfo: osinfo_product_get_eol_date_string: unknown parameter 'product' in documentation comment, should be 'prod' ../osinfo/osinfo_product.c:369: Warning: Libosinfo: osinfo_product_get_release_date: unknown parameter 'product' in documentation comment, should be 'prod' ../osinfo/osinfo_product.c:388: Warning: Libosinfo: osinfo_product_get_eol_date: unknown parameter 'product' in documentation comment, should be 'prod' ``` install_config: Don't use different variable names in .c and .h The tz variable, used in the source file, is declared in the header file as timezone. Let's just use the same name used in the header as it's more descriptive. By doing so, the following GTK-Doc warning is fixed: ``` ../osinfo/osinfo_install_config.c:231: Warning: Libosinfo: osinfo_install_config_set_l10n_timezone: unknown parameter 'tz' in documentation comment, should be 'timezone' ``` os: Move the "Since: " annotation The "Since: " annotation should be *always* added to the end of the comment block. More than that, multiplce "Since: " annotations cannot be part of a same comment block and in order to solve that, let's avoid adding a "Since: " in the end of the block, but explicitly mention when each of the values were added. By taking this approach, the following warnings can be avoided: ``` ../osinfo/osinfo_os.h:90: Warning: Libosinfo: "@OSINFO_RELEASE_STATUS_RELEASED" parameter unexpected at this location: * @OSINFO_RELEASE_STATUS_RELEASED: A released OS. ^ ../osinfo/osinfo_os.h:96: Error: Libosinfo: multiple "Since:" tags for identifier "OsinfoReleaseStatus": * Since: 1.4.0 ^ ../osinfo/osinfo_os.h:97: Warning: Libosinfo: "@OSINFO_RELEASE_STATUS_ROLLING" parameter unexpected at this location: * @OSINFO_RELEASE_STATUS_ROLLING: A rolling version of an OS and thus with no ^ ``` tree: Move the "Since: " annotation Let's move the "Since: " annotation to the end of the comment block, otherwise the following warning will be shown: ``` ../osinfo/osinfo_tree.h:41: Warning: Libosinfo: "@OSINFO_TREE_ERROR_NO_TREEINFO" parameter unexpected at this location: * @OSINFO_TREE_ERROR_NO_TREEINFO: No treeinfo found; ^ ``` install_script: Move the "Since: " annotation Let's move the "Since: " annotation to the end of the comment blocks, otherwise the following warnings will be shown: ``` ../osinfo/osinfo_install_script.h:100: Warning: Libosinfo: "@OSINFO_DEVICE_DRIVER_SIGNING_REQ_NONE" parameter unexpected at this location: * @OSINFO_DEVICE_DRIVER_SIGNING_REQ_NONE: Script do not require device drivers ^ ../osinfo/osinfo_install_script.h:123: Warning: Libosinfo: "@OSINFO_INSTALL_SCRIPT_INJECTION_METHOD_CDROM" parameter unexpected at this location: * @OSINFO_INSTALL_SCRIPT_INJECTION_METHOD_CDROM: Support injection of the ^ ../osinfo/osinfo_install_script.h:145: Warning: Libosinfo: "@OSINFO_INSTALL_SCRIPT_INSTALLATION_SOURCE_MEDIA" parameter unexpected at this location: * @OSINFO_INSTALL_SCRIPT_INSTALLATION_SOURCE_MEDIA: A media will be used as ^ ``` install_config_param: Fix annotation for _get_value_map() The function name was wrong, causing the following warning: ``` ../osinfo/osinfo_install_config_param.h:94: Warning: Libosinfo: osinfo_install_config_param_get_value_map: return value: Missing (transfer) annotation ``` install_config_param: Don't use invalid "transfer" annotation "transfer" annotation shouldn't be used with gboolean or enums, causing the following warnings: ``` ../osinfo/osinfo_install_config_param.c:252: Warning: Libosinfo: invalid "transfer" annotation for gboolean: only valid for array, struct, union, boxed, object and interface types ../osinfo/osinfo_install_config_param.c:267: Warning: Libosinfo: invalid "transfer" annotation for gboolean: only valid for array, struct, union, boxed, object and interface types ``` 2020-01-03 Fabiano Fidêncio <fidencio@redhat.com> install_config_param: Move the "Since: " annotation Let's move the "Since: " annotation to the end of the comment block, otherwise the following warning will be shown: ``` ../osinfo/osinfo_install_config_param.h:50: Warning: Libosinfo: "@OSINFO_INSTALL_CONFIG_PARAM_POLICY_NONE" parameter unexpected at this location: * @OSINFO_INSTALL_CONFIG_PARAM_POLICY_NONE: no policy defined ^ ``` loader: Remove ueseless /** PUBLIC METHODS */ comment The comment is not exactly helpful and causes the following GTK-Doc error: ``` ../osinfo/osinfo_loader.c:133: Error: Libosinfo: Skipping invalid GTK-Doc comment block: /** PUBLIC METHODS */ ^ ``` db: Remove useless /** PUBLIC METHODS */ comment The comment is not exactly helpful and causes the following GTK-Doc error: ``` ../osinfo/osinfo_db.c:183: Error: Libosinfo: Skipping invalid GTK-Doc comment block: /** PUBLIC METHODS */ ^ ``` tree: Fix typo in osinfo_tree_new() annotation trasfer -> transfer, otherwise the following warning will be shown: ``` ../osinfo/osinfo_tree.c:472: Warning: Libosinfo: unknown annotation: trasfer ``` tree: Fix osinfo_tree_new() annotation The extra line between the function name and its first parameter causes the following warning: ``` ../osinfo/osinfo_tree.c:467: Warning: Libosinfo: "@id" parameter unexpected at this location: * @id: the id of the tree to be created ^ ``` media: Don't use invalid "transfer" annotation "transfer" annotation shouldn't be used with gint or gint64, otherwise the following warnings will be shown: ``` ../osinfo/osinfo_media.c:1635: Warning: Libosinfo: invalid "transfer" annotation for gint: only valid for array, struct, union, boxed, object and interface types ../osinfo/osinfo_media.c:1768: Warning: Libosinfo: invalid "transfer" annotation for gint64: only valid for array, struct, union, boxed, object and interface types ``` media: Don't use deprecated annotations Both Type and Transfer are deprecated and should be replaced by (type TYPE) (transfer MODE), otherwise the following warnings will be shown: ``` ../osinfo/osinfo_media.c:657: Warning: Libosinfo: GObject-Introspection specific GTK-Doc tag "Type" has been deprecated, please use annotations on the identifier instead: * Type: GLib.List(utf8) ^ ../osinfo/osinfo_media.c:658: Warning: Libosinfo: GObject-Introspection specific GTK-Doc tag "Transfer" has been deprecated, please use annotations on the identifier instead: * Transfer: container ^ ``` os_variant: Fix osinfo_os_variant_new() annotation The extra line between the function name and its first parameter causes the following warning: ``` ../osinfo/osinfo_os_variant.c:129: Warning: Libosinfo: "@id" parameter unexpected at this location: * @id: the id of the os variant to be created ^ ``` avatar_format: Don't use deprecated annotations Both Type and Transfer are deprecated and should be replaced by (type TYPE) (transfer MODE), otherwise the following warnings will be shown: ``` ../osinfo/osinfo_avatar_format.c:102: Warning: Libosinfo: GObject-Introspection specific GTK-Doc tag "Type" has been deprecated, please use annotations on the identifier instead: * Type: GLib.List(utf8) ^ ../osinfo/osinfo_avatar_format.c:103: Warning: Libosinfo: GObject-Introspection specific GTK-Doc tag "Transfer" has been deprecated, please use annotations on the identifier instead: * Transfer: container ^ ``` 2019-12-18 Fabiano Fidêncio <fidencio@redhat.com> ci: Add a job to enforce DCO Signed-off-by This work is a copy-and-paste from what's been done by Daniel P. Berrangé on https://gitlab.com/bichon-project/bichon Add CONTRIBUTING.md Let's follow osinfo-db example and have a contributing guide, targetting both contributors and commiters. 2019-12-12 Fabiano Fidêncio <fidencio@redhat.com> ci: Add CentOS 8 container 2019-12-11 Andrea Bolognani <abologna@redhat.com> ci: Update name for Ubuntu images The corresponding libvirt-jenkins-ci commit is f289e64a5fd9. 2019-12-09 Roman Bogorodskiy <bogorodskiy@gmail.com> meson: respect mandir option Construct man installation dir from libosinfo_mandir instead of libosinfo_datadir. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-12-04 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.8.0 Update NEWS for 1.7.1 release build: Switch the project's version to 1.7.1 It's done as an 1.7.1 release is needed in order to fix the breakage caused by not using versioned symbols in 1.7.0 release. build: Only define {PCI,USB}_IDS in {pci,usb}-ids-path are defined Setting those regardless of {pci,usb}-ids-path being explicitly set or automatically detected is a regression compared to what we were doing when using autotools. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> build: Set {pci,usb}.ids paths In case pci.ids or usb.ids files are shipped by libosinfo, pci_ids_path and usb_ids_path must be set accordingly. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-12-03 Fabiano Fidêncio <fidencio@redhat.com> build: Use versioned symbols During 25c45ffd9a1b8f18f56974d5b9b6b2c1da8cbd0c, supported for versioned symbols were (wrongly) introduced but never used. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-12-02 Fabiano Fidêncio <fidencio@redhat.com> mingw,spec: Use %autosetup Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> mingw,spec: Drop gettext BuildRequires Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> mingw.spec: Add hwdata as BuildRequires Seems that when using `fedpkg scratch-build` we cannot rely on having network access, making the download of the usb and pci ids files to fail, causing then a build error as the files won't be found to be shipped. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> mingw.spec: Add gcc as BuildRequires meson requires both compiler and cross-compiler in order to perform a cross build. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> mingw,spec: Add intltool as BuildRequires Otherwise the translations are not generated. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> build: Install {usb,pci}.ids if not shipped In case usb.ids and pci.ids are not shipped, let's also install them as part of libosinfo installation. install_data() will automatically install the files to the standards-conforming location. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-11-29 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.8.0 build: Disable -Wcast-warn It causes issues when generating Libosinfo-scan.c on armv7hl. ``` Libosinfo-scan.c: In function ‘describe_type’: /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:216:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 216 | #define G_PARAM_SPEC_INT64(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_INT64, GParamSpecInt64)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:871:30: note: in expansion of macro ‘G_PARAM_SPEC_INT64’ 871 | GParamSpecInt64 *pspec = G_PARAM_SPEC_INT64 (spec); | ^~~~~~~~~~~~~~~~~~ /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:239:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 239 | #define G_PARAM_SPEC_UINT64(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_UINT64, GParamSpecUInt64)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:887:31: note: in expansion of macro ‘G_PARAM_SPEC_UINT64’ 887 | GParamSpecUInt64 *pspec = G_PARAM_SPEC_UINT64 (spec); | ^~~~~~~~~~~~~~~~~~~ /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:354:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 354 | #define G_PARAM_SPEC_DOUBLE(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_DOUBLE, GParamSpecDouble)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:921:31: note: in expansion of macro ‘G_PARAM_SPEC_DOUBLE’ 921 | GParamSpecDouble *pspec = G_PARAM_SPEC_DOUBLE (spec); | ^~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c: In function ‘describe_default’: /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:216:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 216 | #define G_PARAM_SPEC_INT64(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_INT64, GParamSpecInt64)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:1004:30: note: in expansion of macro ‘G_PARAM_SPEC_INT64’ 1004 | GParamSpecInt64 *pspec = G_PARAM_SPEC_INT64 (spec); | ^~~~~~~~~~~~~~~~~~ /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:239:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 239 | #define G_PARAM_SPEC_UINT64(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_UINT64, GParamSpecUInt64)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:1010:33: note: in expansion of macro ‘G_PARAM_SPEC_UINT64’ 1010 | GParamSpecUInt64 *pspec = G_PARAM_SPEC_UINT64 (spec); | ^~~~~~~~~~~~~~~~~~~ /usr/include/glib-2.0/gobject/gtype.h:2299:6: error: cast increases required alignment of target type [-Werror=cast-align] 2299 | ((ct*) g_type_check_instance_cast ((GTypeInstance*) ip, gt)) | ^ /usr/include/glib-2.0/gobject/gtype.h:484:66: note: in expansion of macro ‘_G_TYPE_CIC’ 484 | #define G_TYPE_CHECK_INSTANCE_CAST(instance, g_type, c_type) (_G_TYPE_CIC ((instance), (g_type), c_type)) | ^~~~~~~~~~~ /usr/include/glib-2.0/gobject/gparamspecs.h:354:45: note: in expansion of macro ‘G_TYPE_CHECK_INSTANCE_CAST’ 354 | #define G_PARAM_SPEC_DOUBLE(pspec) (G_TYPE_CHECK_INSTANCE_CAST ((pspec), G_TYPE_PARAM_DOUBLE, GParamSpecDouble)) | ^~~~~~~~~~~~~~~~~~~~~~~~~~ Libosinfo-scan.c:1070:31: note: in expansion of macro ‘G_PARAM_SPEC_DOUBLE’ 1070 | GParamSpecDouble *pspec = G_PARAM_SPEC_DOUBLE (spec); | ^~~~~~~~~~~~~~~~~~~ cc1: all warnings being treated as errors ``` Update NEWS for 1.7.0 release Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Refresh translations from zanata 2019-11-26 Fabiano Fidêncio <fidencio@redhat.com> os: Add "get_device_drivers_prioritized" get_device_drivers_prioritized() returns the list of highest priority device drivers for a given OS. In case management apps would like to add a fallback mechanism, the recommendation is to use get_device_drivers(), which returns all the drivers. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> device-driver: Add "priority" attribute Let's read the "priority" attribute from the "driver" element. If not set, the priority will be set to OSINFO_DEVICE_DRIVER_DEFAULT_PRIORITY (50) by default. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-11-11 Andrea Bolognani <abologna@redhat.com> ci: Remove Fedora 29 The OS is no longer supported by the libvirt-jenkins-ci project and the existing container images will be deleted shortly. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> ci: Add Fedora 31 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-11-08 Fabiano Fidêncio <fidencio@redhat.com> db: Take the media's volume size into account when sorting the medias If volume-size is not taken into account, when dealing with identical ISOs, we may end up wrongly matching an entry that has its volume-size with one which doesn't have the value, simply because the one which does not have the value may be declared first the in XML entry. Reviewed-by; Daniel P. Berrangé <berrange@redhat.com> 2019-11-06 Diego Escalante Urrelo <diegoe@gnome.org> list: Improve API safety Add checks across the API to validate objects before trying to operate on them. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Remove unnecessary casts Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Fix gtk-doc string Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Add osinfo_list_set_element_type Add the corresponding setter for osinfo_list_get_element_type() and use it internally to ensure safety. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Use osinfo_list_get_instance_private Instead of repeateadly accessing the priv structure, use the GObject defined helper to ensure safety. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Update GObject boilerplate Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> list: Use API to avoid unsafe access Use osinfo_list_get_element_type() instead of referencing members directly. This ensures that we are not trying to reference invalid members in invalid structs. Also make said function safer by using osinfo_list_get_instance_private() that checks the sanity of the OsinfoList object for us. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-10-14 Fabiano Fidêncio <fidencio@redhat.com> ci: Add Ubuntu 18 container ci: Add Ubuntu 16 container ci: Add Debian 9 container ci: Add CentOS 7 container We cannot run osinfo-db-tools-rpm job in the CentOS 7 container as it doesn't have the needed meson macros. ci: Keep the containers alphabetically sorted ci: Deal with multilib path When building a project using Meson, the project will be installed respecting the multi-library path. In order to get the right path we should do: - Debian based machines: - Use the value of `dpkg-architecture -q DEB_TARGET_MULTIARCH` - RPM based machines: - Use the value of `rpm --eval '%{_lib}'` ci: Set NINJA environment variable We'll start testing on CentOS 7, which doesn't have the "ninja" binary. Knowing this, we better have a NINJA environment variable setting accodingly. meson: Disable -Wunsafe-loop-optimizations warnings This commit is basically bringing b2574d33aa back: ''' Disable -Wunsafe-loop-optimizations warnings This warning option is not reporting broken code. Rather it is merely identifying loops which GCC would not be able to optimize fully when using -funsafe-loop-optimizations. In GCC 7 this triggers on the loader code osinfo_loader.c: In function 'osinfo_loader_process_list': osinfo_loader.c:2127:11: error: missed loop optimization, the loop counter may overflow [-Werror=unsafe-loop-optimizations] while (tmp && *tmp) { ^ The warning just needs to be turned off since the code is not broken, and we don't care about the optimization level ''' 2019-10-08 Pino Toscano <ptoscano@redhat.com> tools: Minor POD fixes - use B<> for some parameters - use C<> instead of a manually quoted identifier - make the install-script config file a verbatim block, so it is kept as-is, and not wrapped as normal text block - use L<> for linking to other man pages - switch example from Fedora 16 to Fedora 30 - other minor changes Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-09-25 Fabiano Fidêncio <fidencio@redhat.com> syms: Add osinfo_firmware_new osinfo_firmware_new() has been introduced as part of commit 904b65e7a1. However, it was not added to the libosinfo.syms file. ci: Introduce CI Let's introduce a basic CI, which basically runs an out-of-tree build with `make check`, running against the following systems: - Fedora 29, Fedora 30, Fedora Rawhide; - Debian 10, Debian Sid; meson: Disable gtk-doc for cross-builds Let's disable gtk-doc build when performing a cross-build as `./Libosinfo-scan` will be generated for the cross architecture and cannot be run from the host machine architecture. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> meson: Fix build on BSDs Unless explicitly passing "-Wno-typedef-redefinition" and "-Wno-missing-field-initializers", builds would fail not only on macOS, but also on BSDs.Knowing that, let's just use disable those warnings when not building on Linux. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> meson: Add syntax-check as part of our test-suite Let's take advantage of that we can run `make syntax-check` without the need of having a Makefile around and add it as part of our test suite. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Move files needed for `make syntax-check` to build-aux Together with the changes, let's adapt GNUmakefile and maint.mk in order to be able to run those files from build-aux. GNUmakefile: - changing both srcdir & top_srcdir, and adjusting the includes of cfg.mk and maint.mk were enough to be able to run `make syntax-check` from build-aux directory; maint.mk: - "ME" variable had to be changed to point to the right place; - Some paths expecting cfg.mk in the top_srcdir had to be changed; - tight-scope.mk target dependency has been changed from $(ME) to maint.mk; Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Support `make syntax-check` without Makefile A few dirty hacks have been done in GNUmakefile, cfg.mk, and maint.mk in order to support running `make syntax-check` without having to have a Makefile around. GNUmakefile: set srcdir & top_srcdir and remove the parts that would cause an error due to no presence of a Makefile. cfg.mk: add sc_prohibit_always-defined_macros to the list of local checks to skip as it looks for Makefile, which is generated after running ./configure. maint.mk: allow running `make syntax-check` without having a Makefile generated. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Drop autotools support Let's try to drop autotools support as meson is in place. There are a few files that were not dropped though, as: - cfg.mk - maint.mk - GNUmakefile - build-aux/bracket-spacing.pl - build-aux/gitlog-to-changelog - build-aux/useless-if-before-free - build-aux/vc-list-files From the list above, build-aux/gitlog-to-changelog is currently used by meson as part of the `dist`. All the other files, with a few changes, will be useful for meson as, with a few changes, we will be able have `ninja syntax-check` working. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> prepare-release: Adapt to use meson Let's adapt the prepare-release script to use meson build system instead of autotools. This change has been tested and is working fine on Fedora 30+. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> spec: Adapt to use meson The changes needed in the spec files are not so intrusive. One thing that has to be mentioned is the requirement for mingw-filesystem >= 107 (which is only present on Fedora 30+) in order to generate the mingw rpms for osinfo-db-tools. This huge bump in the version happened because the `%mingw_ninja_install` was only recently introduced to mingw-filesystem package. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> spec: Remove BuildRoot definition None of the currently supported distributions need that. Last one which did was EL5 (which, thankfully, is EOL now). Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> spec: Switch to %ldconfig_scriptlets This change has been part of Fedora spec file for a quite long time, but it was never replicated upstream. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Add support to meson build system Meson build system is a "simpler and easier" to understand build system that that can provide (with some work-arounds here and there) the same functionalities currently available with our current build system (autotools). Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> loader: Fix "switch-default" warning Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> readme: Add "Testing" section Let's add a short section explaning how to build & run the libosinfo tests. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Don't include config.h config.h file is only used in osinfo-db-tools for the GETTEXT_PACKAGE macro. Fortunately, when AC_CONFIG_HEADER macro is dropped from configure, -DGETTEXT_PACKAGE=... is automatically passed as a compilar argument, allowing us then to drop the file. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> maint.mk: Update to the latest gnulib Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Update exclude_file_name_regexp--sc_trailing_blank tests/isodata/* have been removed a long time ago. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Update exclude_file_name_regexp--sc_bindtextdomain Let's *not* run the check only for tests files. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Remove exclude_file_name_regexp--sc_space_tab tests/isodata has been removed a long time ago. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Remove exclude_file_name_regexp--sc_prohibit_atoi_atof Let's run sc_prohibit_atoi_atof in all files. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Remove exclude_file_name_regexp--sc_require_config_h* Those were added as a copy & paste mistake from libvirt-glib. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> cfg.mk: Remove exclude_file_name_regexp--sc_preprocessor_indentation Let's run the check in all files. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Indent includes/defines By doing this, we can stop excluding files from sc_preprocessor_indentation check. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> test-install-script: Use g_unlink() instead of unlink() Not like unlink(), g_unlink() is a portable function and using it would allow us to build this test using mingw. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-09-17 Fabiano Fidêncio <fidencio@redhat.com> test-firmware: Add tests Reviewed-by: Cole Robinson <crobinso@redhat.com> loader: Load Firmwares to OSes Let's read the <firmware> attribute from the XMLs and store this into our OS representation. Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add support to Firmware Let's add APIs for both adding a firmware to or retrieving the list of firmwares from an OS. Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Add test-firmwarelist Reviewed-by: Cole Robinson <crobinso@redhat.com> firmwarelist: Add OsinfoFirmwareList OsinfoFirmwareList has been created in order to represent a list of firmwares supported by an OS, as an OS may support various firmwares without any issue. Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Add test-firmware Reviewed-by: Cole Robinson <crobinso@redhat.com> firmware: Add OsinfoFirmware OsinfoFirmware has been created in order to represent the firmware types supported by an OS. The schema representation of this object is: <firmware arch="arch name" type="bios|efi"/> Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-09-12 Fabiano Fidêncio <fidencio@redhat.com> build: Use a minimum gettext requirement rather than an exact one In configure.ac, specify the minimum gettext version we require, rather than the exact one. This fixes a situation where the autoconf macros used for gettext will be the latest available on the system (for example, 0.20); but the copied-in Makefile.in.in will be for the exact version specified in configure.ac (in this case, 0.19). In that situation, the gettext build rules will error out at `make` time with the message: *** error: gettext infrastructure mismatch: using a Makefile.in.in from gettext version 0.19 but the autoconf macros are from gettext version 0.20 Avoid that by specifying a minimum version dependency rather than an exact one. This should not cause problems as we haven’t committed any generated or external gettext files into git, so each developer will end up regenerating the build system for their system’s version of gettext, as expected. See the subsection of https://www.gnu.org/software/gettext/manual/html_node/Version-Control-Issues.html for more information. This patch is basically the same as https://gitlab.gnome.org/GNOME/geocode-glib/commit/6ba0e38fd1987be0de8432afd93b9a5c1f487b63 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-09-02 Fabiano Fidêncio <fidencio@redhat.com> media: Check whether an InputStream is seekable SoupInputStream is not seekable, causing errors as: ``` fidencio@laerte ~/src/upstream/libosinfo $ ./tools/osinfo-detect https://cdimage.debian.org/cdimage/archive/8.10.0/powerpc/iso-dvd/debian-8.10.0-powerpc-DVD-1.iso (lt-osinfo-detect:27922): GLib-GObject-WARNING **: 18:52:29.494: invalid cast from 'SoupClientInputStream' to 'GSeekable' (lt-osinfo-detect:27922): GLib-GIO-CRITICAL **: 18:52:29.494: g_seekable_seek: assertion 'G_IS_SEEKABLE (seekable)' failed (lt-osinfo-detect:27922): GLib-GIO-CRITICAL **: 18:52:29.494: g_task_return_error: assertion 'error != NULL' failed ``` In order to avoid those, let's check whether the InputStream is seekable or not. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-08-20 Fabiano Fidêncio <fidencio@redhat.com> tools,detect: Don't leak OS when detecting trees From g_object_get() documentation: In general, a copy is made of the property contents and the caller is responsible for freeing the memory in the appropriate manner for the type, for instance by calling g_free() or g_object_unref(). Reviewed-by: Cole Robinson <crobinso@redhat.com> tools,detect: Remove 'env' output type --format env was intended for udev scripts which was dropped in commit 16cf1f797728dff9e361033937faea9f38ffbe49 Author: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Date: Fri Feb 15 19:00:37 2013 +0200 Ditch udev rule Let's remove this piece and avoid having two formats with discrepant information being displayed. Reviewed-by: Cole Robinson <crobinso@redhat.com> tools,detect: Print the arch of the detected tree/media This information is useful in order to check whether we're not only matching the right OS, but also the right architecture. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-08-16 Fabiano Fidêncio <fidencio@redhat.com> install-script: Don't expand entities when parsing XML The XML_PARSE_NOENT flag to libxml will cause it to expand all entities in the input XML document when parsing. Doing this is bad practice if the XML input file comes from an untrusted source, because it can cause the XML parser to load arbitrary files that are readable by the user running XML parsing. This is basically the same fix as 47233d0b9dc (from osinfo-db-tools) Reviewed-by: Felipe Borges feborges@redhat.com loader: Don't expand entities when parsing XML The XML_PARSE_NOENT flag to libxml will cause it to expand all entities in the input XML document when parsing. Doing this is bad practice if the XML input file comes from an untrusted source, because it can cause the XML parser to load arbitrary files that are readable by the user running XML parsing. This is basically the same fix as 47233d0b9dc (from osinfo-db-tools) Reviewed-by: Felipe Borges feborges@redhat.com 2019-08-07 Fabiano Fidêncio <fidencio@redhat.com> tree: Deal with file:// trees Let's add back the code we dropped as part of e0e06c7bc but, this time, to be executed condionatilly only when dealing with "file://", which does not require GVFS to be installed at all. https://gitlab.com/libosinfo/libosinfo/issues/35 Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Rename on_location_read() to on_soup_location_read() In order to add back the support to create a tree from a location that's a "file://...", let's, as the first step, rename on_location_read() to os_soup_location_read() as it'll make things cleaner later on. https://gitlab.com/libosinfo/libosinfo/issues/35 Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Add _create_from_treeinfo() API osinfo_tree_create_from_treeinfo() API has been added so apps can stick to whatever is their preference on downloading the treeinfo file, mainly now that ftp:// is not supported by _create_from_location(), and just pass the treeinfo file content to libosinfo in order to create an OsinfoTree. https://gitlab.com/libosinfo/libosinfo/issues/34 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-08-06 Fabiano Fidêncio <fidencio@redhat.com> examples: Remove non functional examples libosinfo examples are not working for some time. Let's remove them and, later on, we can come up with a new set of functional ones. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-08-01 Fabiano Fidêncio <fidencio@redhat.com> configure: Relax gettext requirement Since commit 3a2dd3050, we've added gettext 0.19.8 as a dependency for libosinfo. However, it broke libosinfo build on Ubuntu 16.04 as the gettext provided by them is 0.19.7. Taking a look at the "noteworthy"* changes, seems that for our usecase we can just depend 0.19.7 and keep building on Ubuntu 16.04 *: https://lists.gnu.org/archive/html/info-gnu/2016-06/msg00003.html Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-07-26 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.7.0 Update NEWS for 1.6.0 release Reviewed-by: Cole Robinson <crobinso@redhat.com> Refresh translations from zanata spec,mingw: Add mingw*-libsoup as dependency Commit ac692eaa4680 introduced libsoup dependency but forgot to add it to the mingw spec files. 2019-07-18 Fabiano Fidêncio <fidencio@redhat.com> tools,detect: Check for NULL variants Both trees & medias may return NULL in case no os-variant was found. Let's check for this and avoid a segmentation fault on calling osinfo_list_get_length() with a NULL argument. Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Add "os" property Similarly to what we have for media, let's add the "os" property to the trees. osinfo-detect code is already relying on having an "os" property and, since commit 722e7d1d4, it errors out due to the missing property: (osinfo-detect:7808): GLib-GObject-WARNING **: 09:27:58.656: g_object_get_is_valid_property: object class 'OsinfoTree' has no property named 'os Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-07-17 Fabiano Fidêncio <fidencio@redhat.com> tree: cleanup on any error & return on _create_from_location_async_helper() When calling a g_task_return_error(), we should also cleanup the CreateFromLocationAsyncData and return immediately after that. By doing so we avoid errors as: g_task_return_error: assertion 'task->ever_returned' failed. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tree: goto cleanup on any error, on on_content_read() Instead of just free'ing the CreateFromLocationAsyncData, let's just go to cleanup, which will call create_from_location_async_data_free() and return after that. By doing so we avoid errors as: g_task_return_error: assertion 'task->ever_returned' failed. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tree: Use g_clear_object in create_from_location_async_data_free() We should check whether the objects are NULL or not before calling g_object_unref(). Instead of doing so, let's use g_clear_object(), which doesn't require the check. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> spec: Drop gvfs dependency As we rely on libsoup to handle http://, https://, and ftp:// requests there's no need to depend on gvfs anymore. https://gitlab.com/libosinfo/libosinfo/issues/30 Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Use libsoup for http:// & https:// requests As osinfo_tree_create_from_location() handles non-local requests, it'd end up relying on GVFS under the hood, which would cause this API to not work when called from an app running as root. In order to avoid this situation, let's rely on libsoup for this cases. It's important to mention that glib minimum version has been bumped to 2.44 as that's the version where g_input_stream_read_all_async() has been introuced. https://gitlab.com/libosinfo/libosinfo/issues/30 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Use libsoup for http:// & https:// requests As osinfo_media_create_from_location_with_flags_async() can handle non-local cases, they'd end up relying on GVFS under the hood, which would cause those APIs to not work when called from an app running as root. In order to avoid this situation, let's rely on libsoup for these cases. https://gitlab.com/libosinfo/libosinfo/issues/30 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Use GInpuStream directly in on_location_read() Let's do this in order to always deal with a GInputStream directly and avoid dealing with a GFileInputStream or something else when we start using libsoup to deal with non-local files. Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Improve on_location_read error message readability Let's add a ": " by the end of the error message, so the prefix doesn't get concatenated with error message, increasing its readability. Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Remove not used Tree from _CreateFromLocationAsyncData struct As the OsinfoTree is not used anywhere in our code, let's just remove it. Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Fix indentation in create_from_location_async_data_free() Reviewed-by: Cole Robinson <crobinso@redhat.com> install-script: Add _generate_*_for_tree() Similarly to what we already have for media, let's add APIs for generating both the install-script and the install-script command line for trees. These new APIs take a tree as parameter and, by that, can access tree specific elements as such its URLs. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-07-16 Fabiano Fidêncio <fidencio@redhat.com> Add "Since: 1.6.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 1.5.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 1.4.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 1.3.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.13" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.12" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.11" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.10" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.9" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.8" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.7" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.6" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.3" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.2" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.1" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.2.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.1.0" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Together with this change, documentation has been added to the functions missing them. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.0.6" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.0.5" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> Add "Since: 0.0.3" to the documentation Let's add the "Since: ..." information as part of the function's documentation. It helps developers reading our docs to easily figure out when a function has been introduced. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-07-10 Fabiano Fidêncio <fidencio@redhat.com> tools,install-script: Deprecate --config {user,admin}-password Let's deprecate user-password and admin-password options of --config and also warn out whenever they're passed to osinfo-install-script. CVE-2019-13313 Libosinfo: osinfo-install-script option leaks password via command line argument. 'osinfo-install-script' is used to generate a script for automated guest installations. It accepts user and admin passwords via command line arguments, thus leaking them via process listing. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tools,install-script: Add --config-file (-f) option Let's add a new option so users can set their config from a file, instead of directly passing the values via command-line. CVE-2019-13313 Libosinfo: osinfo-install-script option leaks password via command line argument. 'osinfo-install-script' is used to generate a script for automated guest installations. It accepts user and admin passwords via command line arguments, thus leaking them via process listing. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-07-02 Fabiano Fidêncio <fidencio@redhat.com> db: deprecate osinfo_db_guess_os_from_tree() Let's deprecate osinfo_db_guess_os_from_tree() and "force" the apps to use the new/simpler API. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> tools,detect: Use osinfo_db_identify_tree() Let's switch osinfo-detect to use the new API as the old one will end up being deprecated soon. Also,, the code path for medias and trees is now way more similar with the changes. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> db: Add identify_tree() API Similarly to identify_media() API, let's add a osinfo_db_identify_tree() API, which is going to simplify a bit tree detection code on the apps level. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> tree: Fix treeinfo APIs' documentation The documention was copied & pasted from media APIs. Let's fix them ir order to actually match trees/treeinfo. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> tree: Add treeinfo-* properties Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> tree: Don't set "has-treeinfo" property's default to TRUE Doesn't make sense to set "has-treeinfo" property to TRUE by default knowing that debian and ubuntu trees do not support treeinfo. This was mistakenly added as part of commit eb53f8d6. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> tree: Remove useless properties A bunch of media properties were mistakenly added as tree properties as part of commit 0202b4db. Although this is clearly a copy & paste error, it has never been noticed so far. Reviewed-by: Cole Robinson <crobinso@redhat.com> Tested-by: Cole Robinson <crobinso@redhat.com> 2019-06-25 Fabiano Fidêncio <fidencio@redhat.com> media: Add _is_bootable() API Let's add an API in order to let the applications known whether a media created is bootable or not (according to libosinfo detection). Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Only error out if a media is not bootable when REQUIRE_BOOTABLE flag is passed It'll give more flexibility for the apps to, at least, detect a media, even if the media is not bootable. https://gitlab.com/libosinfo/libosinfo/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Rewrite create_from_location* functions Let's rewrite those functions, using the create_from_location_with_flags* functions behind. https://gitlab.com/libosinfo/libosinfo/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Add create_from_location_with_flags* APIs The newly added APIs are introduced in order to extend the current create_from_location* APIs in a way that a flag can be passed in order to decided whether a media should be created even if the media is not recognised as bootable by libosinfo. https://gitlab.com/libosinfo/libosinfo/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Add OsinfoMediaDetectFlags Let's add a new type in order to specify flags to osinfo_media_detect. This type, at least for now, contains only one specific flag that will be used to specify whether the detection of the media should or should not fail in case the media is not bootable. https://gitlab.com/libosinfo/libosinfo/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Ignore case when detecting bootable ppc ISO When detecting a PPC bootable iso, we should search by the "/ppc/bootinfo.txt" file inside the iso. However, there's no guarantee that the name of the file will be always lowercased as it depends on the extensions used to generate the ISO. The problematic ISO can be found at: http://cdimage.debian.org/cdimage/release/9.9.0/ppc64el/iso-cd/debian-9.9.0-ppc64el-netinst.iso And one curious thing to mention is that there's no TRANS.TBL file (the one used to translate, for instance, PPC -> ppc) present in the ISO. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-05-22 Fabiano Fidêncio <fidencio@redhat.com> tree: Cleanup _create_from_location_async_helper() There's no need to pass neither the URL nor the cancellable to this function as those can be taken directly from data. Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Avoid use of memory after it's freed We've been passing data->location as the @url argument of osinfo_tree_create_from_location_async_helper(), freeing it and trying to g_strdup() it as the new content of data->location. In order to avoid doing so, let's set the data->location only once, in the first caller of osinfo_tree_create_from_location_async_helper(), as its content is always going to be the same doesn't matter the treeinfo format to be used with. Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Avoid dereference of null pointer As any consumer of libosinfo API may pass NULL as the @matched argument of compare_tree(), the current code could be dereferencing a NULL pointer when calling `osinfo_tree_set_os()`. In order to avoid doing so, let's set the os to the OsinfoTree at the moment the @matched argument is set. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-05-10 Fabiano Fidêncio <fidencio@redhat.com> os: Add kernel-url-argument API Tree based installations will require a kernel-url-argument to be assed. Let's provide an API to the apps to get this info from OsinfoOS. https://gitlab.com/libosinfo/osinfo-db/issues/28 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-05-09 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.6.0 Update NEWS for 1.5.0 release Reviewed-by: Cole Robinson <crobinso@redhat.com> Refresh translations from zanata Reviewed-by: Cole Robinson <crobinso@redhat.com> docs: Add missing docs links from main docs page Reviewed-by: Cole Robinson <crobinso@redhat.com> install_script: Use the same parameter's name in .c and .h Reviewed-by: Cole Robinson <crobinso@redhat.com> install_script: Document OsinfoInstallScriptInstallationSource Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Use the same parameter's name for .c and .h Reviewed-by: Cole Robinson <crobinso@redhat.com> media: Document OSINFO_MEDIA_ERROR_NO_DIRECTORY_RECORD_EXTENT Reviewed-by: Cole Robinson <crobinso@redhat.com> install_config: Fix _get_installation_url docs Reviewed-by: Cole Robinson <crobinso@redhat.com> Revert "osinfo-detect: print os variants info for env format" This reverts commit 44cab89243220223b0ac6312237c08a514055664, as the commit was mistakenly pushed. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-05-02 Fabiano Fidêncio <fidencio@redhat.com> tests: Add tests for multiple short-ids I've decided to for testing the multiple short-ids at the os level, as it reflects better the usage of the newly added APIs. https://gitlab.com/libosinfo/libosinfo/issues/19 Reviewed-by: Cole Robinson <crobinso@redhat.com> product: Add osinfo_product_get_short_id_list() API As multiple short-ids is fully supported, let's add a new API for getting the whole list instead. https://gitlab.com/libosinfo/libosinfo/issues/19 Reviewed-by: Cole Robinson <crobinso@redhat.com> loader: Fully support multiple short-id values osinfo-db schema allows for multiple short-id, and there's already some entries that rely on that. For example, debian and ubuntu have something like: ``` <short-id>debian9</short-id> <short-id>debianstretch</short-id> ``` or ``` <short-id>ubuntu19.04</short-id> <short-id>ubuntudisco</short-id> ``` However, OsinfoLoader has always been treating "short-id" as a single property entry and, as a consequence of this, apps would always have to use the first listed id as shown below: ``` fidencio@dahmer ~/src/upstream/libosinfo $ osinfo-install-script debian9 preseed.cfg fidencio@dahmer ~/src/upstream/libosinfo $ osinfo-install-script debianstretch Error finding OS: debianstretch fidencio@dahmer ~/src/upstream/libosinfo $ osinfo-install-script ubuntu19.04 preseed.cfg fidencio@dahmer ~/src/upstream/libosinfo $ osinfo-install-script ubuntudisco Error finding OS: ubuntudisco ``` As OsinfoEntity already has support for multiple property entries, let's take advantage of this and properly treat OSINFO_PRODUCT_PROP_SHORT_ID as so. This doesn't break any compatibility with the existent code as osinfo_entity_get_param_value() will *always* return the first added entry, as it was done before. OsinfoFilter also already has support for multiple property entries and the filter works as expected. Here's the result of the very same example used above with this fix: ``` fidencio@dahmer ~/src/upstream/libosinfo $ ./tools/osinfo-install-script debian9 preseed.cfg fidencio@dahmer ~/src/upstream/libosinfo $ ./tools/osinfo-install-script debianstretch preseed.cfg fidencio@dahmer ~/src/upstream/libosinfo $ ./tools/osinfo-install-script ubuntu19.04 preseed.cfg fidencio@dahmer ~/src/upstream/libosinfo $ ./tools/osinfo-install-script ubuntudisco preseed.cfg ```` https://gitlab.com/libosinfo/libosinfo/issues/19 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-04-23 Fabiano Fidêncio <fidencio@redhat.com> osinfo-detect: Improve --help wording The "URL type" text leads to the misunderstanding that `osinfo-detect` only works for URLs and not for local medias as well. Let's change the wording in order to make it clearer. https://gitlab.com/libosinfo/libosinfo/issues/26 Reviewed-by: Cole Robinson <crobinso@redhat.com> spec: Require GVFS GVFS is required as g_file_new_from_path() and g_file_new_from_uri() rely on GVFS. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-04-12 Fabiano Fidêncio <fidencio@redhat.com> osinfo-detect: print os variants info for env format Similarly to what's been done for the plain format, let's also print os variants' information when using --format=env. osinfo-detect: Be aware of tree's os variants In the same way done for medias, make the tree detection aware of os-variants. https://gitlab.com/libosinfo/libosinfo/issues/25 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media,image,tree: Avoid critical errors when getting the os variants When getting the os variant of media, image, or tree, let's ensure their OS is not NULL* before calling osinfo_os_get_variant_list(). In case it's NULL, let's just return NULL. *: having a media, image, or tree with no OS set is a common case when the user has been creating it manually. https://gitlab.com/libosinfo/libosinfo/issues/25 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> db: Set the os to the detected tree When detecting a tree, let's also set its OS. https://gitlab.com/libosinfo/libosinfo/issues/25 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> test-image: Expand the tests to consider variants Reviewed-by: Cole Robinson <crobinso@redhat.com> image: Add get_os_variants() API osinfo-db schema allows that, but libosinfo never really had an API to support it. In order to add get_os_variants(), (get,set)_os APIs were also added. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-04-10 Fabiano Fidêncio <fidencio@redhat.com> test-image: Add test for loading an entry loader: Properly load the arch value for images osinfo_loader_image() has been considering the XML tag for the images' architecture was "architecture", while it actually is "arch". By trying to read the wrong tag, no arch was loaded, causing then any image load to fail with: CRITICAL **: 13:30:06.539: osinfo_entity_set_param: assertion 'value != NULL' failed 2019-04-03 Fabiano Fidêncio <fidencio@redhat.com> media: Fix usage of application id Christophe Fergeau pointed this issue out during the review of 6d789033e and, even though, the bogus code ended up merged (sorry, mea culpa here!). Reviewed-by: Cole Robinson <crobinso@redhat.com> test-db: Add test to cover identifying a media with "unknown" arch Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Deal with "all" media architecture Medias with "all" architecture are used as a fallback when guessing an OS from a media. Knowing that, any entry with "all" architecture should not be considered when first comparing the medias, thus those are skipped and, later on, if no match has been found, those are used. https://gitlab.com/libosinfo/libosinfo/issues/20 Reviewed-by: Cole Robinson <crobinso@redhat.com> test-db: Add test for guessing tree from OS Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Deal with "all" tree architectures Tree with "all" architecture are used as a fallback when guessing an OS from a tree. Knowing that, any entry with "all" architecture should not be considered when first comparing the trees, thus those are skipped and, later on, if no match has been found, those are used. https://gitlab.com/libosinfo/libosinfo/issues/20 Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Rename os_* to os_treeinfo_* in guess_os_from_tree() It'll make clear that we're dealing with the treeinfo attributes and not with the tree ones. https://gitlab.com/libosinfo/libosinfo/issues/20 Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Rename tree to treeinfo in guess_os_from_tree() It'll make clear that we're dealing with the treeinfo attributes and not with the tree ones. https://gitlab.com/libosinfo/libosinfo/issues/20 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-04-02 Pino Toscano <ptoscano@redhat.com> gitignore: remove old entries, simplify existing ones Remove patterns referring to old tests, and in general to files that do not exist anymore. Simplify few existing patterns into more general ones. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-03-28 Fabiano Fidêncio <fidencio@redhat.com> tests: Add test-tree.c This test, at least for now, tests: - The basic functionality of creating a OsinfoTree object; - osinfo_tree_get_os_variants() API; https://gitlab.com/libosinfo/libosinfo/issues/16 Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Add get_os_variants() API osinfo-db schema allows that, but libosinfo never really had an API to support it. In order to add get_os_variants(), (get,set)_os APIs were also added, matching the behaviour for OsinfoMedia. https://gitlab.com/libosinfo/libosinfo/issues/16 Reviewed-by: Cole Robinson <crobinso@redhat.com> tree: Also check fore "treeinfo" in addition to ".treeinfo" Akamai CDN hosted files can't start with a dot, making trees published to a CDN having their ".treeinfo" files renamed to "treeinfo". https://gitlab.com/libosinfo/libosinfo/issues/18 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-03-27 Fabiano Fidêncio <fidencio@redhat.com> Add a git-publish configuration file 2019-03-21 Fabiano Fidêncio <fidencio@redhat.com> test-os: Remove test_devices_duplication This test is now part of osinfo-db. Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Remove test_resources_uniqueness This test is now part of osinfo-db. Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Remove test_resources_minimum_recommended_maximum This test is now part of osinfo-db. Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Remove test-isodetect This test is now part of osinfo-db With this change we can get rid of the whole isodata we have. Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Remove test-*.uris Those tests are now part of osinfo-db. With this change we can get rid of the curl dependency we have. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-03-15 Fabiano Fidêncio <fidencio@redhat.com> isodata,centos: Add a new LiveGNOME isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-03-07 Cole Robinson <crobinso@redhat.com> Remove emacs 'Local variables' annotations No other virt projects have annotations like this, plus IMO .editorconfig handles this in a much nicer way and more flexible way. Ditch them Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Add toplevel .dir-locals.el This centralizes emacs config in a top level file, matching the bits sprinkled around in 'Local variable' comments Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-03-07 Fabiano Fidêncio <fidencio@redhat.com> mingw,spec: Fix the Source URL https://releases.pagure.org/libosinfo/... instead of https://releases.pagure.io/libosinfo/... Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> mingw: Use %mingw_find_Lang macro Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> spec: Use standard vala packaging pattern where vapi files are in -devel This commit brings upstream the changes done on Fedora on 42b96ea66d7b3e. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> spec: Add "BuildRequires: gcc" Reference: https://fedoraproject.org/wiki/Changes/Remove_GCC_from_BuildRoot Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> mingw,spec: Remove obsolete Group tag This commit brings upstream the changes done on Fedora on 61fea0d48407c0ee, References: https://fedoraproject.org/wiki/Changes/Remove_Group_Tag And "The Group: tag is not required in any live Fedora or EPEL release. RHEL5 did need it, but EPEL5 did not as it was supplied automatically via magic in the epel-rpm-macros package. The Packaging Guidelines have indicated that the Group: tag should not be used since March of 2017." Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> spec: Remove needless use of %defattr This commit is bringing upstream the change done on Fedora on 4a678b2f4d59, The change is related to https://docs.fedoraproject.org/en-US/packaging-guidelines/#_file_permissions Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2019-03-07 Fabiano Fidêncio <fabiano@fidencio.org> haikur1beta1: Add tests data Reviewed-by: Cole Robinson <crobinso@redhat.com> haikunightly: Update test data Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-03-07 Fabiano Fidêncio <fidencio@redhat.com> tests: Add coverage for _set_installation_url() Reviewed-by: Cole Robinson <crobinso@redhat.com> install-config: Add _set_installation_url() Although the installer URL can be taken from OsinfoTree, there are situations where OsinfoTree doesn't have a URL or even the management apps should prefer the URL set by their users instead of fallbacking to the one get from OsinfoTree. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-03-01 Fabiano Fidêncio <fidencio@redhat.com> Post release version bump to 1.5.0 Update NEWS for 1.4.0 release Refresh translations from zanata syms: Add "global" to 1.4.0 symbols Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tests: Add identify_media() tests Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> db: Add missing attributes to fill_media() Some new attributes have been added to OsinfoMedia, as: - the install-script list (during this cycle); - installer-script (1.3.0); However, none of those attributes have been filled in in the media accordingly to matched media. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-02-19 Martin Blanchard <tchaik@gmx.com> build: Prefer gettext over intltool intltool is outdated: https://wiki.gnome.org/Initiatives/GnomeGoals/GettextMigration Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> build: Prefer autoreconf over gnome-autogen gnome-common is completely deprecated: https://wiki.gnome.org/Projects/GnomeCommon/Migration Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-02-18 Fabiano Fidêncio <fidencio@redhat.com> tests: Add test-uris-common.h to the SOURCES Otherwise some breakage can be seen when building libosinfo's RPM test-uris-common.c:24:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" ^~~~~~~~~~~~~~~~~~~~ compilation terminated. make[2]: *** [Makefile:1296: test_mediauris-test-uris-common.o] Error 1 make[2]: *** Waiting for unfinished jobs.... CCLD test-entity test-mediauris.c:27:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" ^~~~~~~~~~~~~~~~~~~~ compilation terminated. make[2]: *** [Makefile:1282: test_mediauris-test-mediauris.o] Error 1 test-treeuris.c:27:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" ^~~~~~~~~~~~~~~~~~~~ compilation terminated. make[2]: *** [Makefile:1394: test_treeuris-test-treeuris.o] Error 1 test-uris-common.c:24:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" ^~~~~~~~~~~~~~~~~~~~ compilation terminated. make[2]: *** [Makefile:1408: test_treeuris-test-uris-common.o] Error 1 test-imageuris.c:27:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" ^~~~~~~~~~~~~~~~~~~~ compilation terminated. make[2]: *** [Makefile:1184: test_imageuris-test-imageuris.o] Error 1 test-uris-common.c:24:10: fatal error: test-uris-common.h: No such file or directory #include "test-uris-common.h" test-uris-common: remove empty line at EOF The empty line caused: tests/test-uris-common.c maint.mk: empty line(s) or no newline at EOF make: *** [maint.mk:926: sc_prohibit_empty_lines_at_EOF] Error 1 2019-02-18 Pino Toscano <ptoscano@redhat.com> voidlinux: add Void Linux test data gentoo: add Gentoo Linux test data archlinux: add Arch Linux test data os: add OSINFO_RELEASE_STATUS_ROLLING Add the OSINFO_RELEASE_STATUS_ROLLING value to the OsinfoReleaseStatus enum, to track the new "rolling" release-status available in the osinfo schema. 2019-02-18 Fabiano Fidêncio <fidencio@redhat.com> tests: Reduce code duplication in URI tests Let's introduce a common file that does the check of the URI for all the URI related tests. By doing this we can avoid code duplication while also avoiding to have all the tests merged together into a single one, which is not so good as limits the developer to run and wait for the whole set of tests instead of having the ability to just run the a specific one. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tests: Add ubuntu 18.04.1 and 18.04.1.0 test data Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> media: Don't assume identifiers are null terminated Identifiers as volume-id, application, publisher, and system are not null terminated and cannot be assumed as so. By assuming those are null terminated strings, libosinfo ends up not counting the last character of a MAX_* string and, consequently, not properly identifying medias that have their identifiers with the MAX_* size. One example is the ubuntu-18.04.1.0-live-server-amd64.iso media, which has as volume-id 'Ubuntu-Server 18.04.1+ LTS amd64'. As the volume-id has exactly 32 characters it's never been matched as when reading the media's volume-id it'd be read as 'Ubuntu-Server 18.04.1+ LTS amd6'. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> pureos: Add test data for pureos8 Reviewed-by: Cole Robinson <crobinso@redhat.com> 2019-02-12 Fabiano Fidêncio <fidencio@redhat.com> media: Make _supports_installer_script() aware of media installers Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> test-media: Add test for media installer script Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> media: Add installer to the media Having the installer tied to the OS is something that has been shown not optimal, unfortunately. An OS may have a few different medias and each media may support a specific type of script or, even, not support an installer script at all. With this in mind, let's have the installer scripts tied to the medias themselves. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2019-02-04 Fabiano Fidêncio <fidencio@redhat.com> tests: abort() in case curl misbehaves Although the issue with curl should be fixed by the previous commit, let's abort in case it misbehaves, avoiding then to have all the data printed on stdout. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tests: Avoid re-using the curl handler on URIs tests On CentOS, we've hit an issue that when a curl handler is re-used we end up receiving malfunctioning responses from NetBSD CDN. Dan investigated the issue found out that the problem is that libcurl in CentOS 7 at least is not ignoring the body. Sometimes it closes the keepalive connection so this doesn't matter, but other times it gives us back the body that it should be ignoring. In order to work the issue around, let's no re-use the curl handler but always open a new one for each of the URIs. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-02-04 Cole Robinson <crobinso@redhat.com> ubuntu: Add 19.04 ppc64el test data Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2019-02-04 Daniel P. Berrangé <berrange@redhat.com> tests: add one test case per OS tested The image, media & tree URIs test suites just register one test case for all the work they do. As a result when they run they just sit there printing no progress information for a long time. This changes then to register one test case per OS in the database. As a result we get progress information. An added benefit is that it is possible to now run tests for individual OS by passing the -p argument. eg test-mediauris -p /mediauris/uris/rhel3.9 2019-01-30 Daniel P. Berrangé <berrange@redhat.com> Post release version bump to 1.4.0 Update NEWS for 1.3.0 release Refresh translations from zanata 2019-01-24 Fabiano Fidêncio <fidencio@redhat.com> test-os: Fix make syntax-check While applying the last patches a few trailing whitespaces have been introduced, making `make syntax-check` fail. tools: Remove a reference to fedorahosted.org Although it's just the id used for the install script, let's remove the reference to the already sunset fedorahosted.org and just use "https://libosinfo.org/config" as the id. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-01-24 Fabiano Fidêncio <fabiano@fidencio.org> test-os: Add basic test for resources This test only ensures that libosinfo is able to properly load the resource info from an OS entry. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Make get_resources_*_without_inheritance() static With that we can get rid of osinfo_os_private.h Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> loader: Deal with network-install resources Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Deal with network-install resources network-install resources represent the resources to be used during a network installation. It may differs from the minimum resources as the network installers (not only the medias, but the trees) may download a huge amount of data to the RAM* consuming way more than the minimum or recommended RAM for the OS. *: anaconda downloads the whole stage2 of the installer into the RAM in order to perform a network installation. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> test-os: Fix typo in resources tests osinfo_os_get_recommended_resources -> osinfo_os_get_maximum_resources Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> loader: Fix typo when loading maximum resources osinfo_os_add_recommended_resources -> osinfo_os_add_maximum_resources Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Make get_maximum_resources() aware of inheritance Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Group get_*_resources() together Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Remove extra line from _get_recommended_resources() Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> os: Fix typos in resources methods methos -> method *private! -> *private*! Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2019-01-15 Fabiano Fidêncio <fidencio@redhat.com> db: improve _guess_os_from_media() checks Do not check against a distro which doesn't have *any* of the values we use to check as match_regex() would just match whatever we compare to it. tree: fix typo in the annotations: 'and' -> 'an' 2019-01-09 Fabiano Fidêncio <fidencio@redhat.com> loader: improve warning message from loader_check_id() "should" needs to be changed to "must" as since commit 00f258a6be libosinfo refuses to load non-compliant named files. README: Update download information Let's point to https://libosinfo.org/download/ instead of pointing to pagure, as it could change again in the future. 2019-01-04 Fabiano Fidêncio <fidencio@redhat.com> dragonflybsd: Add 5.4.1 test data Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> loader: refuse to load non-compliant named files There was a period of grace needed for backwards compatibility. After 3+ years, we can safely remove the code instroduced as part of 09ae3cf09 (which was released on 0.3.0). Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> README: Reword the patch submission paragraph Just mention that the libosinfo's mailing list is the *preferred* way to send patches. More info, as the one about the project accepting Merge Request, can be found in the project's website. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> README: Update download link Let's point to pagure instead of referring to the already sunset fedorahosted. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2019-01-04 Felipe Borges <felipe10borges@gmail.com> endlessos: Add Endless OS 3.4 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-20 Pino Toscano <ptoscano@redhat.com> Use g_list_free_full where easily possible Use g_list_free_full instead of g_list_foreach + g_list_free, so the lists are cleared with a single call. test-isodetect gets a void* wrapper, so there is no need to add casts everywhere. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-12-20 Fabiano Fidêncio <fidencio@redhat.com> media: Pass the source object to GTask new When creating the new GTask in osinfo_media_create_from_location_async, no source object has been passed. Consequently search_ppc_bootinfo_callback receives NULL as its source object, causing warnings when trying to unref it. Let's just pass the stream as the source object of the GTask in osinfo_media_create_from_location_async in order to avoid the issue. The issue has been introduced as part of 29f97cb7648. media: Don't try to unref a NULL entry In case g_file_read_finish() fails, the stream will always be NULL. The issue has been introduced as part of f977527c82. media: Check whether PPC ISOs are bootable PPC ISOs do not have the "El Torito" extension that describes whether the media is bootable or not. However, they have a "bootinfo.txt" file placed under "ppc" directory in order to specify the media is bootable. So, let's add a few more checks looking for "/ppc/bootinfo.txt" in case the El Torito header is not found. The whole implementation has been based on the following sources: - The ISO 9660 (ECMA-119) specification: http://www.ecma-international.org/publications/files/ECMA-ST/Ecma-119.pdf - The ISO 9660 osdev wiki page: https://wiki.osdev.org/ISO_9660 - IBM's developer article: https://www.ibm.com/developerworks/linux/library/l-detecting-bootable-ibm-power-server-iso-images/index.html And what the implementation does is: - Access the *root* directory record, which has a fixed size of 33 bytes and is located at the bytes offset 156 of the PVD; - Load the *root* directory record extent content into the memory and: - Search for the the dir/file we're looking at; - In the dir is present, we load its directory record extent content into the memory and search for the next dir/file; - In case the file is present, we return that the media is bootable; - Otherwise, we return the media is non bootable; Mind, though, that this code is *not* generic enough to be adapted in case we want to search for a different dir/file and the reason for that is related to some ISOs using a TRANS.TBL table that translates the given name of a file to the shown name of the same file. If, for some reason, we ever want to take the path of having this code more generic we'd have to: - For each directory acessed, search for TRANS.TBL; - In case it's present, load its content into a hash table and, for each dir/file, check the name of the file we're looking for with the shown name (only present in the TRANS.TBL; - Otherwise, just follow the very same code we currently have; The decision for not implementing this TRANS.TBL changes is, mainly, because it'd make the code circa twice slower than it currently is (as we'd have to search for the TRANS.TBL as the first thing ...) without any specific gain for the needs we have right now. In case there's the need, in the future, it could be implemented in the way described here. https://gitlab.com/libosinfo/libosinfo/issues/8 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Add set_non_bootable_media_error() This is a helper to set the "non bootable" error for a media. It'll become really handy in the next few patches in the series in order to avoid code duplication. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Add create_from_location_async_data() This is a helper to create the OsinfoMedia from the CreateFromLocationAsyncData. It'll become really handy in the next few patches in the series in order to avoid code duplication. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Use G_STATIC_ASSERT() Let's use G_STATIC_ASSERT() instead of creating a dummy array to check the size of the _PrimaryVolumeDescriptor struct. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Use better naming for goto labels Let's replace EXIT for cleanup and ON_ERROR for error and follow a normal naming convention for the labels. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Don't leak stream in case of errors stream has to be freed whenever an error occurr on osinfo_media_create_from_location_async() internals. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tools,install-script: Add "installation-source" parameter Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> test-install-script: Cover _(get|set)_installation_source() Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> install-script: Add _(get|set)_installation_source() As install-scripts may behave differently depending on their installation source, let's add two new methods to properly set what's the installation source being used with the script and then decide what we should do depending on the installation-source value. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> test-install-script: Cover _(get|set)_preferred_injection_method() Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> install-script: Add _(get|set)_preferred_injection_method() Those new methods are going to be used to tell the install-scripts whether the injection-method that's going to be used is. We have to do so as the command-line may be different depending on the injection-method used, for example: - fedora using cdrom, disk or floppy: ks=hd:/(vda|sda)/fedora.ks - fedora using initrd: ks=file:/fedora.ks It's important to mention that although the methods are taking GFlags, those are treated as GEnum and only one value is expected to be set, as mentioned in the documentation. Also, mind that the usage of osinfo_entity_set_param() to store the nick of the OSINFO_INSTALL_SCRIPT_INJECTION_METHOD_* is intentional as the nick is exactly what's going to be used in the install-scripts to generate the proper command-line. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-12-17 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: add FreeBSD 12.0 isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-12-14 Felipe Borges <felipeborges@gnome.org> android-x86: Add Android-x86 8.1 test data Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-12-14 Fabiano Fidêncio <fidencio@redhat.com> test-media: Add tests for OsinfoMedia attributes Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> tests: Add basic tests for OsinfoMedia Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> media: Add _supports_installer_script() Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> loader: Deal with "installer-script" media attribute Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> dragonflybsd: Add 5.4.0 test data Reviewed-by: Cole Robinson <crobinso@redhat.com> 2018-12-11 Fabiano Fidêncio <fidencio@redhat.com> tree: cleanup non-fatal errors in load_key_info() There are errors which are not fatal and just ignored in load_keyinfo. However, as those have not been cleaned up, we could see messages like: (lt-osinfo-detect:20658): GLib-WARNING **: GError set over the top of a previous GError or uninitialized memory. This indicates a bug in someone's code. You must ensure an error is NULL before it's set. The overwriting error message was: Key file does not have key “boot.iso” in group “images-x86_64” In order to avoid this, let's just call g_clear_error() after situations where an error may have been set but it can just be ignored. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> tree: cleanup load_key_info() Instead of doing the same checks for key or group not found all over the place, let's just add a new function that does that and replace the old check for the new function. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-10 Fabiano Fidêncio <fidencio@redhat.com> os: Fix _find_install_script() annotation osinfo_os_find_install_script() returns an object which ownsership is not transferred. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-10 Fabiano Fidêncio <fabiano@fidencio.org> db: improve _guess_os_from_tree() checks Do not check against a distro which doesn't have treeinfo data as match_regex() would just match whatever we compare to it. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-10 Fabiano Fidêncio <fidencio@redhat.com> test-os: Cover osinfo_tree_has_treeinfo() Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> tree: Add osinfo_tree_has_treeinfo() As not all the "trees" we stored in osinfo-db have "treeinfo" data, let's add a new method, osinfo_tree_has_treeinfo(), that can be used to check whether the tree has treeinfo or not. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-12-10 Fabiano Fidêncio <fabiano@fidencio.org> test-os: expand test_loader() to cover loading treeinfo Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> loader: properly load the treeinfo attributes treeinfo attributes haven't been loaded properly due to the change done in ab2ab35f, changing the hardcoded 9 to sizeof("treeinfo-"). The problem here is that size("treeinfo-") is 10, causing that any comparison to fail. Let's change the sizeof("treeinfo-") to strlen("treeinfo-"). Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-10 Fabiano Fidêncio <fidencio@redhat.com> test-os: Add tests for osinfo_os_find_install_script() Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> os: Do not crash when passing a NULL profile to _find_install_script() Let's ensure we're not going to crash in case a NULL profile is passed for osinfo_find_install_script(). Instead of crashing, though, we're returning an error. Also, document which are the expected values for @profile Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> os: Fix typo in _find_install_script() osinfo_find_install_script() doesn't work as expected due to a typo when getting the elements of the script list as an OsinfoList must be passed instead of an OsinfoOs. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-12-06 Vera Cholasta <vbudikov@redhat.com> alpinelinux3.8: Add test files for 3.8.1 Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-30 Fabiano Fidêncio <fidencio@redhat.com> image: Add "cloud-init" attribute to the image Some images may support cloud init customization and for those we'd like to explicitly indicate it. The "cloud-init" attribute is similar to the "live" attribute for medias, has its default value as FALSE and can be optinionally set in the XML. This commit also extends the basic image tests to also cover the cloud-images attribute. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> tests: Add test-imageuris In a quite similar way than test-mediauris and test-treeuris, let's have a test for the imageuris. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> loader: Load Images information Now that we've OsinfoImage and OsinfoImageList objects in place and OsinfoOs already provides us an API to add an OsinfoImage to an OsinfoOS, let's properly load the images' information present in osinfo-db. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> os: Introduce APIs to deal with images Let's introduce the needed APIs on os side to deal will the images. For now, there's no info being loaded as the loader part depends on one of these APIs. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> tests: Add test-imagelist test-imagelist is totally based on test-devicelist. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> Introduce OsinfoImageList object OsinfoImageList object represents has been created to easily deal with OsinfoImage objects. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> tests: Add test-image test-image tests the basic functionally of OsImage APIs. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> Introduce OsinfoImage object OsinfoImage object has been created to represent pre-installed images distributed by the OSes. We've decided to go for the simplespet approach possible that's just exposing the Image URL for an existing OS and then Apps that want to deal with the Image itself can just use libguestfs for that as they already do a quite good job on this field. https://gitlab.com/libosinfo/osinfo-db/issues/10 Reviewed-by: Pino Toscano <ptoscano@redhat.com> 2018-11-22 Christophe Fergeau <cfergeau@redhat.com> build: Use EXTRA_..._DEPENDENCIES rather than _DEPENDENCIES This allows us to use the dependencies automatically generated by automake while appending the additional one that we need. https://www.gnu.org/software/automake/manual/automake.html#index-EXTRA_005fmaude_005fDEPENDENCIES-1 « The EXTRA_*_DEPENDENCIES variable may be useful for cases where you merely want to augment the automake-generated _DEPENDENCIES variable rather than replacing it. » Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-21 Fabiano Fidêncio <fidencio@redhat.com> tests: Fix `make check` breakage for FreeBSD 11 Let's take advantage of the newly added libosinfo-impl.la that contains all the symbols that we need and link our tests against it. By doing so, we can remove the referrence to osinfo_product.c when building test-os, which has been causing test-os binary to contain 2 copies of the very same code, one linked in the test and the other from libosinfo.so. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> Tested-by: Andrea Bolognani <abologna@redhat.com> osinfo: Create an intermediate convenience library Let's create libosinfo-impl.la which is nothing else than libosinfo-1.0.la without stripping out its non-public symbols. libosinfo-impl.la can be used to link directly against our tests (as those may use private functions that we do *not* want to expose), while the "official" one (libosinfo-1.0.la) will remain the same. Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> Tested-by: Andrea Bolognani <abologna@redhat.com> 2018-11-20 Fabiano Fidêncio <fidencio@redhat.com> devicelink: Rename "removed" to "supported" After some discussion seems that a better name for the "removed" attribute would be "supported" (as it'd be more coherent with features that are about to come). While there's still no release of libosinfo or osinfo-db with the "removed" attribute, let's just rename it to "supported". Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-11-19 Fabiano Fidêncio <fidencio@redhat.com> test-devicelinklist: Remove duplicated assert Having a duplicated g_assert_true() on CentOS7 breaks `make check` as the compiler tries to make the second assert into a constant. As the second assert was a typo anyways, let's just remove it in order to fix `make check`. test-os: Add missing XML files Those XMLs should have been added as part be1f2def2 but unfortunately something went wrong while rebasing the patches. 2018-11-19 Daniel P. Berrangé <berrange@redhat.com> tests: ignore transient failures in media/tree URI tests Some times the media/tree URI tests will fail with a timeout trying to connect to the server. Treat this as non-fatal since it will usually just indicate a transient network or server failure. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> tests: display info about what failed in media/tree URI tests Make it possible to debug failures by displaying the URL, curl error code/message and http code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-19 Fabiano Fidêncio <fidencio@redhat.com> os: Fix build breakage Similarly to what was fixed by commit 6b246cd9e, let's do not use variable declarations inside "for" loop as it's only allowed in C99 mode. test-os: Add test_resources_inheritance() Now that resources are inherited between OSes that derives-from/clones some other OS, let's have a test case to ensure it's not going to break in the future. https://gitlab.com/libosinfo/osinfo-db/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Support inherited resources As resources are pretty much the same for new releases of a same OS, having them explicitly set is a source of copy-and-paste error and redundant information in our db. The approach taken to properly support inheritance of OSes is a explicitly set attribute to the resources indicating that they should be inherited. Here's an example of how the XML will look like: foo: <resources arch="all"> <minimum> <n_cpus>1</n_cpus> <storage>1</storage> </minimum> </resources> bar (which derives-from/clones foo): <resources arch="all" inherit="true"> <minimum> <n_cpus>3</n_cpus> <ram>3</ram> </minimum> </resources> When querying for bar's minimum resources, we'll get: n_cpus: 3 cpu: -1 (not set, by any of the OSes) ram: 3 storage: 1 https://gitlab.com/libosinfo/osinfo-db/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add _get_.*_resources_without_inheritance() Those two private methods have been added and those are part of the field preparation for having Resources inherited in the same way we do inherit devices. https://gitlab.com/libosinfo/osinfo-db/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> loader: Set "inherit" accordingly to the resources "inherit" is the new attribute of the resources that means whether this resources should have its (non filled) values inherited. In case the "inherit" attribute is present and as is "true", let's set it. Otherwise (the default option), there's no change from the current behaviour. https://gitlab.com/libosinfo/osinfo-db/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> resources: Add osinfo_resources_{get,set}_inherit() Those two private functions will help to keep track whether the resources should be inherited or not. https://gitlab.com/libosinfo/osinfo-db/issues/15 Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Add test_devices_inheritance_removal() now that OsinfoOs can deal with the conpect of a "removed" device, let's add some tests covering the possible situations we may face and ensure we won't regress here. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Deal with "removed" devices This commit introduces some logic to deal with "removed" devices and the way chosen to do so was to filter out the "removed" devices from the full list of devices (for both DeviceLinks and Device lists). https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add add_entity_to_list_check() The addition of this new function is basically preparing the ground for the changes that are coming in the next commits where we'll need to do the very same check in both _get_device_links_internal() and _get_devices_internal(). With that in mind, let's avoid code duplication and have a single function that can be used in both methods. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Use get_device_links_internal in _get_all callback This commit is just preparing the ground for changes that will come in the next commits where we'll start supporting "removed" devices and in order to so, we'll have to explicitly call _get_device_links_internal() forcing it to return *all* devices, even the "removed" ones, as they'll be filtered out later. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Use get_devices_internal in _get_all callback This commit is just preparing the ground for changes that will come in the next commits where we'll start supporting "removed" devices and in order to so, we'll have to explicitly call _get_devices_internal() forcing it to return *all* devices, even the "removed" ones, as they'll be filtered out later. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add osinfo_os_get_all_device_links_internal() The addition of this new function is basically preparing the ground for the changes that are coming in the next commits where osinfo_os_get_all_device_links() will have to deal with "removed" devices. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add osinfo_os_get_device_links_internal() The addition of this new function is basically preparing the ground for the changes that are coming in the next commits where osinfo_os_get_device_links() will have to deal with "removed" devices. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add osinfo_os_get_devices_internal() The addition of this new function is basically preparing the ground for the changes that are coming in the next commits where osinfo_os_get_devices() will have to deal with "removed" devices. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> os: Add osinfo_os_get_all_device_links() This function is a helper for the commits that are about come related to properly dealing with "removed" devices. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> loader: Deal with "removed" devices for an OS Let's make the loader aware of the "removed" attribute that can be set for a device linked from an OS entry. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Add test_devices_inheritance_basic() Devices are inherited from OSes that derives-from/clones some other OS. Let's have a test case to ensure it's not going to break in the future. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> test-devicelinklist: Add _get_devices() test https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Add basic tests for devicelinklist https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> devicelinklist: Fix _get_devices() osinfo_devicelinklist_get_devices() returns a list of OsinfoDevice structs and in order to do so it has to iterate over the list of OsinfoDeviceLink structs and get the device associated to each of the OsinfoDeviceLink struct. The current implementation was just trying to add the OsinfoDeviceLink struct to the list, instead of getting the devices associated to it. https://gitlab.com/libosinfo/osinfo-db/issues/13 Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Add test_devices_duplication() As devices are inherited from entries that "derives-from" or "clones" any other entries, there's no reason to keep declaring those devices in several places. By adding this test, hopefully there'll be no more devices duplication from now on. Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Add test_resources_uniqueness This test enforces that we'll never have more than one resources entry for the very same architecture in an OS entry. Reviewed-by: Cole Robinson <crobinso@redhat.com> test-os: Fix test breakage "for" loop initial declarations are only allowed in C99 mode but some old compilers (like the ones in centos7 or debian8) just don't use it by default. Reviewed-by: Andrea Bolognani <abologna@redhat.com> 2018-11-16 Fabiano Fidêncio <fidencio@redhat.com> test-os: Cover maximum resources in tests Reviewed-by: Michal Privoznik <mprivozn@redhat.com> test-os: Make test_resources_minimum_recommended generic Reviewed-by: Michal Privoznik <mprivozn@redhat.com> os: Deal with "maximum" resources In the samw way OSes have minimum and recommended resources, let's add support for "maximum" resources. https://gitlab.com/libosinfo/osinfo-db/issues/11 Reviewed-by: Michal Privoznik <mprivozn@redhat.com> popos: Add popos18.10 isodata Reviewed-by: Felipe Borges <feborges@redhat.com> test-os: move test-os-resources tests Let's have all OS related tests grouped together. Reviewed-by: Felipe Borges <feborges@redhat.com> test-os-resources: Fix typo in test_minimum_recommended_resources() Thanks to this typo the test was iterating over a single OS, always. Reviewed-by: Felipe Borges <feborges@redhat.com> rhel: Add rhel8.0 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> ubuntu: Add 19.04 prerelease isodata ubuntu18.10: Update isodata after release Reviewed-by: Didier Roche <didier.roche@canonical.com> 2018-11-15 Vicente Chaves de Melo <vchaves@ymail.com> NetBSD: add NetBSD 7.2 tests isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-09 Vera Cholasta <vbudikov@redhat.com> tests: continue after failure in mediauris & treeuris https://bugzilla.redhat.com/show_bug.cgi?id=1639108 Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-08 Fabiano Fidêncio <fidencio@redhat.com> os: Don't leak scripts list osinfo_list_get_elements() calls g_hash_table_get_values() which returns a GList that must be freed after used. For more info, please, refer to: https://developer.gnome.org/glib/unstable/glib-Hash-Tables.html#g-hash-table-get-values Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-11-01 Felipe Borges <feborges@redhat.com> rhel: Add rhel7.6 isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-11-01 Fabiano Fidêncio <fabiano@fidencio.org> tests: Update silverblue29 isodata Reviewed-by: Felipe Borges <feborges@redhat.com> tests: Update f29 isodata Reviewed-by: Felipe Borges <feborges@redhat.com> 2018-10-19 Alexey Shabalin <shaba@altlinux.org> alt: Add alt8.X, alt.p8 and alt.sisyphus isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-10-18 Věra Cholasta <vbudikov@redhat.com> test-isodetect: continue after failure https://bugzilla.redhat.com/show_bug.cgi?id=1639108 Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> test-isodetect: replace while loop with for loop Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-10-15 Fabiano Fidêncio <fidencio@redhat.com> Remove FSF address from source files Let's follow what's already done for the most part of libosinfo and recommended from http://www.gnu.org/licenses/gpl-howto.html and do not put FSF address in the source files. Instead, just link to FSF website. Based on: https://gitlab.com/libosinfo/libosinfo/commit/3343d0f9c1b0937824bee2527612047bc2e75f84 Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-10-09 Fabiano Fidêncio <fidencio@redhat.com> silverblue: Add silverblue29 (Beta) isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com> fedora: Add fedora29 (Beta) isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com> authors,maintainers: Add all gitlab's group members Let's update the list of authors/maintainers accordingly to the libosinfo's gitlab group members[0]. The list has been put in alphabetical order on purpose. [0]: https://gitlab.com/groups/libosinfo/-/group_members Reviewed-by: Cole Robinson <crobinso@redhat.com> 2018-10-08 Fabiano Fidêncio <fidencio@redhat.com> Revert "db: Force anchored patterns when matching regex" Daniel suggested that the fix should go to the osinfo-db entries that are causing the issue instead of forcing anchored patterns in libosinfo. This reverts commit e6168463f4fc659b9827b5c8694dc1c6d7d5239a. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2018-10-03 Charles Arnold <carnold@suse.com> Add SUSE CaaS Platform 3.0 isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Add SUSE CaaS Platform 2.0 isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Add SUSE CaaS Platform 1.0 isodata Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-09-24 Fabiano Fidêncio <fidencio@redhat.com> centos: Add 6.10 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-09-21 Fabiano Fidêncio <fidencio@redhat.com> ubuntu: Add isodata for 18.10 Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-09-21 Christophe Fergeau <cfergeau@redhat.com> install-script: Add propagate_libxml_error() helper This simplifies the code a bit. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> install-script: Use g_set_error_literal where appropriate Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Use OSINFO_ERROR in more places Passing a NULL error domain will trigger runtime warnings from glib. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> loader: Introduce OSINFO_ERROR Since glib commit 2.31.10~16 [1], g_set_error() will issue a warning if it's passed a NULL domain. libosinfo is doing this in a few places (try for example 'osinfo-query os winxp'). This commit adds a generic OSINFO_ERROR domain which will be used instead of NULL. It's a bit inconsistent with OSINFO_MEDIA_ERROR and OSINFO_TREE_ERROR in that they both start with "osinfo-*" while OSINFO_ERROR uses "libosinfo". However, since OsinfoLoader already makes use of that quark, changing that would arguably be an (very small) ABI break. [1] https://gitlab.gnome.org/GNOME/glib/commit/6d9f874330 Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> loader: Use OSINFO_LOADER_SET_ERROR in a few more places There are 2 calls to g_set_error(err, "%s", "constant string"); which can be replaced with OSINFO_LOADER_SET_ERROR Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> loader: Rename OSINFO_ERROR to OSINFO_LOADER_SET_ERROR This will be useful in the following commits where another OSINFO_ERROR #define will be introduced Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> 2018-09-11 Fabiano Fidêncio <fidencio@redhat.com> haiku: Add Nightly isodata https://bugzilla.redhat.com/show_bug.cgi?id=1092627 Reviewed-by: Cole Robinson <crobinso@redhat.com> haiku: Add R1/Alpha4.1 isodata https://bugzilla.redhat.com/show_bug.cgi?id=1092627 Reviewed-by: Cole Robinson <crobinso@redhat.com> haiku: Add R1/Alpha3 isodata https://bugzilla.redhat.com/show_bug.cgi?id=1092627 Reviewed-by: Cole Robinson <crobinso@redhat.com> haiku: Add R1/Alpha2 isodata https://bugzilla.redhat.com/show_bug.cgi?id=1092627 Reviewed-by: Cole Robinson <crobinso@redhat.com> haiku: Add R1/Alpha1 isodata https://bugzilla.redhat.com/show_bug.cgi?id=1092627 Reviewed-by: Cole Robinson <crobinso@redhat.com> openbsd: Add 5.3 isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> openbsd: Add 5.2 isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> openbsd: Add 5.1 isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> db: Force anchored patterns when matching regex Without forcing anchored patterns some matches may be completely wrong as "(J_)?CENN?A_X64FREV" (volume-id for Windows 10) that could be taken as a match for "HRM_CENNA_X64FREV" (volume-id of a Windows 8 ISO). Reviewed-by: Cole Robinson <crobinso@redhat.com> tests: Expand the arch's parser for isodetect Some Windows ISOs isodata, as the ones added as part of commit 857f503, uses X64 and X86 and load_iso() didn't take those into consideration. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2018-09-07 Fabiano Fidêncio <fidencio@redhat.com> centos7: Add "NetInstall" isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> centos7: Add "Minimal" isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> centos7: Add another "DVD" isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> centos7: Add another "DVD" isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> centos7: Add LiveKDE isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> centos7: Add LiveGNOME isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> popos: Add 18.04 isodata Reviewed-by: Cole Robinson <crobinso@redhat.com> win10: Add LTSB Evaluation isodata https://bugzilla.redhat.com/show_bug.cgi?id=1139394 Reviewed-by: Cole Robinson <crobinso@redhat.com> gitignore: Add test-os-resources test-os-resources has been added as part of 6a88a78143. Reviewed-by: Cole Robinson <crobinso@redhat.com> 2018-09-06 Fabiano Fidêncio <fidencio@redhat.com> rhel-atomic: Add 7.4 isodata Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> rhel-atomic: Add 7.3 isodata Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> rhel-atomic: Add missing 7.2 isodata Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> maint.mk: Adjust sc_prohibit_strcmp - There's no need to exclude STRN?EQ as those are not defined in libosinfo; - Adjust the comment and the halt message to suggest the usage of g_str_equal() instead of the non existent STRN?EQ(); https://bugzilla.redhat.com/show_bug.cgi?id=1335291 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> maint.mk: Re-enable sc_prohibit_strcmp https://bugzilla.redhat.com/show_bug.cgi?id=1335291 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> osinfo-detect: Replace strcmp() with g_str_equal() https://bugzilla.redhat.com/show_bug.cgi?id=1335291 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> loader: Replace strcmp() with g_str_equal() https://bugzilla.redhat.com/show_bug.cgi?id=1335291 Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-09-05 Daniel P. Berrangé <berrange@redhat.com> Refresh translations from zanata 2018-07-24 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> netbsd: Add NetBSD 8.0 isodata Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2018-07-20 Věra Cholasta <vbudikov@redhat.com> ubuntu: Add test files for Ubuntu18.04 Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-07-20 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> dragonflybsd : Add DragonFlyBSD 5.x isodata Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-07-18 Věra Cholasta <vbudikov@redhat.com> rhel: Add RHEL6.10 iso test files Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-07-18 Charles Arnold <carnold@suse.com> sles: Add SLE15 ISO Information Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-07-18 Věra Cholasta <vbudikov@redhat.com> Add test files for Alpinelinux 3.8 Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2018-07-03 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: add FreeBSD 11.2 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-06-20 Daniel P. Berrangé <berrange@redhat.com> Post release version bump Update NEWS for 1.2.0 release 2018-06-12 Fabiano Fidêncio <fabiano@fidencio.org> silverblue: Move isodata to its own folder Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-06-07 Fabiano Fidêncio <fabiano@fidencio.org> syms: Add missing "global" on 0.2.[7-10] versions Reported-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> syms: Fix identation of 0.2.13 symbols When adding the symbols for LIBOSINFO_0.2.13 I mistakenly added the symbols not respecting default identation used (1 tab) and ended up using 4 spaces instead. Reviewed-by: Michal Privoznik <mprivozn@redhat.com> 2018-05-31 Fabiano Fidêncio <fabiano@fidencio.org> opensuse: Add Leap 15.0 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> tests: Add test-os-resources test-os-resources has been written to avoid bug as having minimum resources greater than the recommended resources in osinfo-db. This problem was exposed by RHEL/CentOS data files, which caused some really bad UI effect in clientes using those two attributes (as GNOME Boxes). Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-05-14 Fabiano Fidêncio <fabiano@fidencio.org> fedora: Add f28 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-05-12 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> netbsd: Add NetBSD 7.1.2 isodata Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-05-12 Andrea Bolognani <abologna@redhat.com> tools: Fix compilation with Clang 6 Clang 6 doesn't like the way the last item in a GOptionEntry array is initialized: ../../tools/osinfo-install-script.c:87:12: error: missing field 'short_name' initializer [-Werror,-Wmissing-field-initializers] { NULL } ^ 1 error generated. gmake[2]: *** [Makefile:540: osinfo-install-script.o] Error 1 gmake[2]: *** Waiting for unfinished jobs.... ../../tools/osinfo-detect.c:104:12: error: missing field 'short_name' initializer [-Werror,-Wmissing-field-initializers] { NULL } ^ 1 error generated. gmake[2]: *** [Makefile:540: osinfo-detect.o] Error 1 Fix the issue by using the { 0 } initializer instead. Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-05-10 Andrea Bolognani <abologna@redhat.com> autogen: Don't hardcode --enable-gtk-doc The usual behavior is for configure to detect the availability of gtk-doc, along with other optional features, and only enable it if found. Hardcoding it this way is causing issues for MinGW builds, where pkg-config is looking in the MinGW-specific paths and can't locate the native gtk-doc.pc file, thus concluding gtk-doc must not be present and aborting the build altogether instead of simply skipping API reference generation. Both 'make distcheck' (through the DISTCHECK_CONFIGURE_FLAGS variable) and the prepare-release.sh script are already enforcing the availability of gtk-doc when preparing releases, so it's safe to use the default behavior for regular builds. 2018-05-08 Daniel P. Berrangé <berrange@redhat.com> configure: fix HAVE_CURL conditional on Win32 AM_CONDITIONAL lines must always be executed, but HAVE_CURL was inside an "if ...", which caused configure to fail when run on Win32. 2018-05-08 Andrea Bolognani <abologna@redhat.com> spec: Drop %{extra_release} It was mainly meant to be used for automatic builds through Test::AutoBuild, so it can be removed now. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> prepare-release: Drop references to Test::AutoBuild They are misleading, and no longer relevant anyway. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> maint: Rename autobuild.sh to prepare-release.sh The script was originally used by the Test::AutoBuild project to perform periodic automatic builds; however, that effort has been abandoned a long time ago, and these days libosinfo CI builds are happening on the Jenkins-based CentOS CI environment under the libvirt umbrella[1], where build recipes are maintained separately from the projects themselves. The script is still used to prepare releases, so it can't be dropped from the repository: rename it so that its purpose is more clearly communicated instead. [1] https://ci.centos.org/view/libvirt/ Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-05-04 Andrea Bolognani <abologna@redhat.com> rhel: Fix isodata for RHEL 7.5 on ppc64le There is a spurious empty line at the end of the file that causes 'make syntax-check' to fail with: [...] tests/isodata/rhel/rhel7.5/RHEL-7.5-20180322.0-Server-ppc64le-dvd1.iso.txt maint.mk: empty line(s) or no newline at EOF make: *** [maint.mk:927: sc_prohibit_empty_lines_at_EOF] Error 1 Get rid of it to solve the issue. See also https://ci.centos.org/view/libvirt/job/libosinfo-master-syntax-check/123/ Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-05-02 Felipe Borges <felipe10borges@gmail.com> rhel: Add RHEL 7.5 isodata Reviewed-by: Christophe Fergeau <cfergeau@redhat.com> 2018-04-24 Fabiano Fidêncio <fabiano@fidencio.org> tumbleweed: Add "Kubic DVD/USB Stick" isodata Reviewed-by: Charles Arnold <carnold@suse.com> tumbleweed: Add "Network Image" isodata Reviewed-by: Charles Arnold <carnold@suse.com> opensuse-42.3: Add "Network Image" isodata Reviewed-by: Charles Arnold <carnold@suse.com> 2018-04-10 Christophe Fergeau <cfergeau@redhat.com> loader: Fix leak in osinfo_loader_find_files() 'tmp' is used to iterate over the list of files, so it will be NULL at the end of the iteration, and g_list_free() will be a no-op. The initial list is stored in 'children', so free that instead. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> 2018-04-10 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> openbsd: add OpenBSD 6.3 isodata 2018-04-10 Natanael Copa <ncopa@alpinelinux.org> alpinelinux: add Alpine Linux isodata 2018-04-10 Vincent <vincent.dao@asianux.com> asianux: add Asianux isodata 2018-03-18 Daniel P. Berrangé <berrange@redhat.com> loader: process files in alphabetical order Currently when loading DB files we process them in whatever order readdir() returns them in, which is often inode order. This makes the order of loading files non-deterministic across installs, and if there are ambiguities in the data, we can in turn get different results. eg ISO images match different OS entries depending on load order. Alphabetically sorting the files doesn't remove any ambiguity that may exist, but at least gives us consistent results on every host. Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-03-18 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> fedora27: Add Atomic Workstation isodata Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-03-11 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> fedora: add Fedora 27 isodata Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-03-11 Charles Arnold <carnold@suse.com> sles: add SLE12 SP3 ISO Information Add SLE12 SP3 ISO Information Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-03-11 Felipe Borges <felipeborges@gnome.org> windows: add win2k16 test data Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> pop_os: add Pop!_OS test data Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-03-11 Christophe Fergeau <cfergeau@redhat.com> tests: Remove calls to g_log_set_always_fatal() All tests call that method because they want g_warning() to be fatal. Since the switch to GTest, this is no longer necessary as this is the default behaviour. Reviewed-by: Fabiano Fidêncio <fabiano@fidencio.org> 2018-02-23 Daniel P. Berrangé <berrange@redhat.com> m4: turn off -Wfunction-cast-type warning flag from gcc8 GCC 8 adds -Wfunction-cast-type to -Wextra, but this breaks when using G_DEFINE_TYPE which does so called "bad" function casts. 2018-02-19 Christophe Fergeau <cfergeau@redhat.com> test-isodetect: Iterate over all isodata/ subdirs When adding support for a new vendor, test-isodetect.c needs to be updated to list it, which is too easy to forget. This commit iterates over subdirs in tests/isodata/ and uses these as the vendors to test for. This means no further modification of test-isodetect.c is needed when adding a new vendor. 2018-02-16 Christophe Fergeau <cfergeau@redhat.com> NetBSD: add NetBSD 6.x and 7.x tests isodata Based on a patch from Vicente Chaves de Melo <vchaves@ymail.com> 2018-02-15 Felipe Borges <felipeborges@gnome.org> opensuse: add OpenSUSE Leap 42.3 test data 2018-02-14 Felipe Borges <felipeborges@gnome.org> freedos: add freedos 1.2 test data 2018-01-15 Christophe Fergeau <cfergeau@redhat.com> test-isodetect: Improve error message on detection failures Since the switch to the GTest framework, when the test suite fails to identify an ISO, the error message is: ERROR:test-isodetect.c:359:test_one: 'matched' should be TRUE while it used to be the much more explicit: ERROR **: ISO alpine-extended-3.7.0_rc2-x86_64.iso.txt was not matched by OS alpinelinux3.7 This commit readds the explicit message on this particular failure. 2017-12-04 Daniel P. Berrange <berrange@redhat.com> Refresh translations from zanata 2017-11-29 Pino Toscano <ptoscano@redhat.com> ubuntu: add Ubuntu 17.10 test data 2017-11-22 Zeeshan Ali <zeeshan@kinvolk.io> Drop redundant 'const' keyword from object parameters Use of 'const' parameter on object parameters is redundant, inconsistent (both internally and against other GObject libraries) and currently breaks the low-level Rust binding generator: https://github.com/gtk-rs/glib/pull/249 2017-11-06 Christophe Fergeau <cfergeau@redhat.com> tests: Add test case for Windows 10 1709 builds These are currently missing from osinfo-db https://bugzilla.redhat.com/show_bug.cgi?id=1509387 2017-11-06 Debarshi Ray <debarshir@gnome.org> Marginally simplify the code to create and run a GMainLoop Even though g_main_loop_new accepts a is_running parameter, it isn't very important since g_main_loop_run will set it to TRUE anyway. There is no requirement that it should be set before calling g_main_loop_run. The vast majority of GMainLoop users simply ignore the is_running parameter unless they are doing something out of the ordinary. 2017-11-01 Pavel Hrdina <phrdina@redhat.com> tests: get the curl response code every time (more) This failed on FreeBSD with following error: ../../tests/test-treeuris.c:47:13: error: variable 'response_code' is used uninitialized whenever 'if' condition is false [-Werror,-Wsometimes-uninitialized] tests: get the curl response code every time This failed on FreeBSD with following error: ../../tests/test-mediauris.c:47:13: error: variable 'response_code' is used uninitialized whenever 'if' condition is false [-Werror,-Wsometimes-uninitialized] Reviewed-by: Daniel P. Berrange <berrange@redhat.com> 2017-10-20 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: Add FreeBSD 10.4 isodata Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2017-10-19 Pino Toscano <ptoscano@redhat.com> tests: convert from check to the GLib testing framework GLib has shipped a testing framework for many years already, so we can make use of it, replacing the external "check". The conversion only switches framework without changing the structure of the tests, making use of the more appropriate assertion in the various places. build: make curl optional, and only for tests Look for curl only when the tests are enabled, and as optional build dependency. If curl is missing, then the two network tests are not built. Also, limit the use of curl CFLAGS/LIBS only for the tests that really need it. As side result, it is possible to drop the curl BuildRequires from the mingw spec, as it builds with --enable-tests=no. Reviewed-by: Daniel P. Berrange <berrange@redhat.com> build: fix disabling of tests Actually use the result of the configure switch --enable-tests, enabling or disabling the tests accordingly. Reviewed-by: Daniel P. Berrange <berrange@redhat.com> 2017-10-18 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> openbsd: add OpenBSD 5.9 isodata 2017-10-16 Pierre-Alain TORET <pierre-alain.toret@protonmail.com> openbsd: add OpenBSD 6.2 isodata Acked-by: Christophe Fergeau <cfergeau@redhat.com> openbsd: add OpenBSD 6.1 isodata Acked-by: Christophe Fergeau <cfergeau@redhat.com> openbsd: add OpenBSD 6.0 isodata Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2017-10-16 Pino Toscano <ptoscano@redhat.com> tests: test-mediauris: raise timeout to 5 minutes With the growing amount of distros in osinfo-db, the test takes more and more time to complete, hitting the 2 minutes timeout in case of any network hiccup (or simply when a server takes more time to reply). Hence, raise the timeout to 5 minutes, which ought to be enough now. Acked-by: Christophe Fergeau <cfergeau@redhat.com> Revert "tests: test-mediauris: skip non-http URLs" Now that test-mediauris uses libcurl, it can handle protocols different than http(s), like ftp, just fine. This reverts commit 4c9c7185475a9739af51a9fc680e5b8a97055467. Acked-by: Christophe Fergeau <cfergeau@redhat.com> Switch from libsoup to libcurl libsoup is used to check the validity of URLs in distributions in osinfo-db; OTOH it supports only HTTP(S), so this limits the checks to that protocol. To overcome this limitation, switch to libcurl: while it requires slightly more code to do the same task, it provides a bit more flexibility, and support for other protocols. No version check is performed, since the APIs used are old enough. Adapt also the README, and the packaging files. Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2017-08-15 Daniel P. Berrange <berrange@redhat.com> Post-release version bump Update NEWS for 1.1.0 release 2017-08-14 Daniel P. Berrange <berrange@redhat.com> Unset LC_ALL, as LC_CTYPE doesn't override a previous LC_ALL setting 2017-08-10 Daniel P. Berrange <berrange@redhat.com> Run glib-mkenums in en_US.UTF-8 locale The new python based glib-mkenums crashes if run in the C locale as python is (incorrectly) not 8-bit clean when processing input. 2017-08-03 Pino Toscano <ptoscano@redhat.com> rhel: add RHEL 7.4 test data 2017-08-02 Pino Toscano <ptoscano@redhat.com> tests: make the network tests run Load the default osinfo path, since the osinfo data are no more part of the libosinfo sources. The tests still fails though, but this is due to missing paths to ISOs and trees. tests: test-mediauris: show status code on failure It helps when debugging failures, to check what is actual error for an URL. tests: test-mediauris: skip non-http URLs The test uses libsoup, which supports only HTTP, and not other protocols such as FTP. Hence, allow only http and https URLs. mageia: add Mageia 6 test data 2017-08-02 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: add FreeBSD 11.1 isodata 2017-07-13 Guido Günther <agx@sigxcpu.org> test-isodetect: add Debian's arm flavours to avoid "Unknown arch" debug messages 2017-07-12 Guido Günther <agx@sigxcpu.org> test-isodetect: fix parsing of application, system and publisher id 2017-07-12 Pino Toscano <ptoscano@redhat.com> fedora: add Fedora 26 test data 2017-07-11 Guido Günther <agx@sigxcpu.org> tests: add Debian 9.0 test data 2017-05-30 Christophe Fergeau <cfergeau@redhat.com> build: Don't use private headers for gir/vapi generation Using private headers when generating the gir/vapi files mean that they are listing private methods which are not exported by the shared library, as well as types which are not present in public headers, and not meant to be used by external users. This was noticed with recent vala.git as it started erroring out on the ProductForeachFlag which is only defined in osinfo_product_private.h https://bugzilla.redhat.com/show_bug.cgi?id=1456709 2017-05-03 Pino Toscano <ptoscano@redhat.com> ubuntu: add Ubuntu 17.04 test data Acked-by: Fabiano Fidêncio <fabiano@fidencio.org> 2017-05-02 Roman Bogorodskiy <bogorodskiy@gmail.com> configure: add (usb|pci).ids locations on FreeBSD Add default locations for usb.ids and pci.ids on FreeBSD, i.e. paths used when installing this from misc/usbids and misc/pciids ports. This way users on FreeBSD don't need to explicitly set these paths using --with-(usb|pci)-ids-path flag. 2017-05-02 Daniel P. Berrange <berrange@redhat.com> Move FreeBSD 11 data into tests/ subdir Remove trailing whitespace from FreeBSD 11 data files 2017-05-02 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: add FreeBSD 11.0 isodata 2017-04-28 Daniel P. Berrange <berrange@redhat.com> Rename test/ to tests/ directory. To simplify libvirt Jenkins CI setup, rename the test directory to tests, so its name matches that used in other virt projects. 2017-04-27 Daniel P. Berrange <berrange@redhat.com> Fix error propagation in DB loader When failing to read the pci/usb ID files we failed to propagate the reported error. We also had some dead code which would never be reached due to earlier gotos. 2017-04-24 Daniel P. Berrange <berrange@redhat.com> Remove blank lines at end of file 2017-04-22 Michal Privoznik <mprivozn@redhat.com> gitignore: Ignore tags Acked-by: Christophe Fergeau <cfergeau@redhat.com> docs: Fix couple of warnings Firstly, two functions (osinfo_install_script_has_config_param_{name,list}) don't document their arguments. Secondly, gtk-doc requires enum documentation to be in one block with no empty lines in it. Then at last, private enums are marked as <private> not <skip>. Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2017-04-21 Fabiano Fidêncio <fabiano@fidencio.org> Add 'eject-after-install' parameter to OsinfoMedia If the media is an installer, thus specifies whether the media should be ejected after the installation process. Default value is true. This is mainly needed for applications like GNOME Boxes (and maybe virt-install) to be able to decide whether the media should be ejected or not when the final reboot happens during its installation process. The latter case may happen when the installer leaves some packages to be installed after rebooting the OS for the last time. The issue solved by this patch is a corner-case faced when adding the install scripts for SLES, as during its installation only one reboot is performed (so, installer-reboots attribute doesn't help us) and the media must not be ejected after the reboot in order to finish the installation. Acked-by: Daniel P. Berrange <berrange@redhat.com> 2017-04-18 Pino Toscano <ptoscano@redhat.com> centos: add CentOS 6.8 and 6.9 test data rhel: add RHEL 6.9 and 7.3 test data 2017-03-20 Guido Günther <agx@sigxcpu.org> Debian: Update Jessie test data Acked-by: Fabiano Fidêncio <fabiano@fidnecio.org> 2017-03-20 Pino Toscano <ptoscano@redhat.com> mageia: add Mageia 4 and 5 test data Acked-by: Fabiano Fidêncio <fabiano@fidencio.org> ubuntu: add Ubuntu 16.10 test data Acked-by: Fabiano Fidêncio <fabiano@fidencio.org> 2017-02-23 Roman Bogorodskiy <bogorodskiy@gmail.com> Improve Perl-related routines * Add a check for the 'perl' program to configure.ac as $(PERL) is used in various places * Do not hardcode '/usr/bin/perl' in perl scripts because on FreeBSD and maybe some other systems it's installed in a different prefix, so use '/usr/bin/env perl' instead. 2017-02-23 Daniel P. Berrange <berrange@redhat.com> Remove reference to fedorahosted.org in spec files 2017-02-22 Daniel P. Berrange <berrange@redhat.com> Disable -Wunsafe-loop-optimizations warnings This warning option is not reporting broken code. Rather it is merely identifying loops which GCC would not be able to optimize fully when using -funsafe-loop-optimizations. In GCC 7 this triggers on the loader code osinfo_loader.c: In function 'osinfo_loader_process_list': osinfo_loader.c:2127:11: error: missed loop optimization, the loop counter may overflow [-Werror=unsafe-loop-optimizations] while (tmp && *tmp) { ^ The warning just needs to be turned off since the code is not broken, and we don't care about the optimization level 2017-02-07 Charles Arnold <carnold@suse.com> suse: Add SLED-12-SP2 data suse: Add SLES-12-SP2 data opensuse: Add openSUSE Leap 42.2 data 2017-01-12 Fabiano Fidêncio <fidencio@redhat.com> fedora: Add F25 data Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-12-23 Fabiano Fidêncio <fidencio@redhat.com> media: Remove ISO header strings' padding The strings in the ISO header ("volume-id", "system-id", "publisher-id" and "application-id") have a predefined size and are padded with spaces, but nothing in on_{pvd,svd}_read() removes these padding spaces from those strings, which makes osinfo_media_get_volume_id() always return [SOME-ISO-LABEL ] rather than the expected [SOME-ISO-LABEL]. Resolves: rhbz#1408271 Reported-by: Allan Day <aday@redhat.com> Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-12-20 Christophe Fergeau <cfergeau@redhat.com> win10: Add test cases for recent Windows 10 builds Recent Windows 10 builds (build 1607) seem to have dropped the J_ at the beginning of the volume ID. These additional files make sure we've got test cases for this kind of win10 ISOs. 2016-11-06 Guido Günther <agx@sigxcpu.org> loader: Fix typo in error message test-isodetect: extend arch table Reduces the output a bit when running under G_MESSAGES_DEBUG. 2016-10-29 Guido Günther <agx@sigxcpu.org> examples: avoid PyGIWarning This fixes demo.py:3: PyGIWarning: Libosinfo was imported without specifying a version first. Use gi.require_version('Libosinfo', '1.0') before import to ensure that the right version gets loaded. from gi.repository import Libosinfo as osinfo; 2016-10-10 Daniel P. Berrange <berrange@redhat.com> Test suite requires osinfo-db to be installed 2016-10-07 Daniel P. Berrange <berrange@redhat.com> Post-release version bump Update NEWS for 1.0.0 Refresh PO files from zanata Revert "Fix mingw DLL version after bump to 1.0.0" This reverts commit dd68d8ccb0abb9cab7324da1b3a706dbf158870b. This is not required after commit 8ec2de1ff44e47d189386cb5e83d26521fc76f29 Author: Daniel P. Berrange <berrange@redhat.com> Date: Wed Aug 10 10:09:17 2016 +0100 configure: fix accidental break in soname 2016-08-11 Daniel P. Berrange <berrange@redhat.com> spec: fix path to test-suite.log - its in the test subdir 2016-08-10 Daniel P. Berrange <berrange@redhat.com> spec: preserve error exit code on test failure spec: display test results on failure autotools hides detailed test output by default, so we must cat test-suite.log on failure to actually see what went wrong. configure: fix accidental break in soname The change in version from 0.3.0 to 1.0.0 triggered a latent bug in our code which sets the library soname. This meant we created libosinfo-1.0.so.1 instead of libosinfo-1.0.so.0 Pull in the fix from libvirt which is where the original logic was copied from. Remove database files Remove the database files and add a dependency on the osinfo-db RPM loader: update to comply with new database install location The new database locations are - System location This is determined by the env variable $OSINFO_SYSTEM_DIR If not set, then defaults to /usr/share/osinfo This location is intended for use by operating system distributors to install the initial data set via a package management system like RPM or Deb - Local location This is determined by the env variable $OSINFO_LOCAL_DIR If not set, then defaults to /etc/osinfo This location is intended for use by local system administrators to install custom local data that should be available to all users on a host - User location This is determined by the env variable $OSINFO_USER_DIR If not set, then defaults to $XDG_CONFIG_HOME/osinfo If that is not set, then defaults to $HOME/.config/osinfo This location is intended for use by unprivileged users wishing to install local data for use by their applications Adapt to use those, but include temporary support for looking at the legacy local & user directory locations, and the OSINFO_DATA_DIR env variable for back-compat. 2016-07-18 Daniel P. Berrange <berrange@redhat.com> Remove osinfo-db-validate tool Remove the osinfo-db-validate tool which is now available via osinfo-db-tools. Add a RPM dep on osinfo-db-tools to pull it in for upgradability. Assume availability of gobject introspection during RPM build The gobject introspection support is available for all currently targetted platforms, so remove RPM conditionals. Fix mingw DLL version after bump to 1.0.0 2016-07-08 Daniel P. Berrange <berrange@redhat.com> fedora: record when virtio devices were made available We currently only record virtio-net, virtio-blk data against Fedora. virtio-9p, virtio-balloon and virtio-console were also all available in Fedora 9 kernels, with virtio-rng added in Fedora 10 and virtio-scsi in Fedora 17. The virtio-1.0 devices all appeared in Fedora 23 guest kernels Add missing virtio PCI devices We were missing virtio-rng and virtio-scsi, as well as all the virtio 1.0 device variants, which have different PCI IDS The pci.ids file lists the following known devices 1000 Virtio network device 1001 Virtio block device 1002 Virtio memory balloon 1003 Virtio console 1004 Virtio SCSI 1005 Virtio RNG 1009 Virtio filesystem 1041 Virtio network device 1042 Virtio block device 1043 Virtio console 1044 Virtio RNG 1045 Virtio memory balloon 1048 Virtio SCSI 1049 Virtio filesystem 1050 Virtio GPU 1052 Virtio input Of which, only 1001, 1001, 1002 and 1009 were recorded in libosinfo. They would have appeared via pci.ids file loading, but they would have lacked friendly names. 2016-07-05 Marc-André Lureau <marcandre.lureau@redhat.com> fedora: Add F24 info & testcases Fedora 24 with virtio-gpu support. 2016-07-04 Marc-André Lureau <marcandre.lureau@gmail.com> device: add virtio-gpu virtio-gpu is a non-legacy device, so device_id is always 0x10 + 0x1040. 2016-07-01 Daniel P. Berrange <berrange@redhat.com> Post-release version bump to 1.0.0 Bump major version to reflect that that we're splitting up the codebase in next release Update NEWS for 0.3.1 release 2016-07-01 Christophe Fergeau <cfergeau@redhat.com> test: Remove empty line at end of XML test file This fixes a make syntax-check failure. install-script: Return glib-allocated string from ::apply_xslt() Strings created by libxml2 should be freed with xmlFree while glib ones should be freed with g_free, so it's cleaner to allocate the string returned from ::apply_xslt() with the glib allocator as it's meant to be g_free'ed later on. osinfo: Add GDestroyNotify callback to g_task_return_pointer calls The GDestroyNotify callback passed to g_task_return_pointer() is used to free the associated pointer in case g_task_propagate_pointer() is not called (error cases, _finish() never called, ...). When the GTask owns the pointer it's passed through g_task_return_pointer() (for example the _finish() function is (transfer full) and only calls g_task_propagate_pointer()), we need to set a GDestroyNotify when calling g_task_return_pointer() or we may have leaks in corner cases. tools: Fix OsinfoOs leak in osinfo-install-script The OsinfoOs instance returned by osinfo_media_get_os() must be unref'ed when no longer needed. osinfo-install-script has code doing: if (media == NULL) { os = find_os(); } else { os = osinfo_media_get_os(); } find_os() does not return a ref'ed OsinfoOs to the caller, while osinfo_media_get_os() does. In order to make it possible to release the ref returned by osinfo_media_get_os(), this commit changes find_os() to always ref the OsinfoOs instance it returns. Moreover, one of the codepaths in find_os() looks like it was potentially returning an invalid pointer: os = OSINFO_OS(osinfo_list_get_nth(OSINFO_LIST(filteredList), 0)); g_object_unref(filteredList); The OsinfoOs instance was probably kept alive by references held outside of the filtered list, but it's safer to keep a reference to ourselves, which this commit is doing anyway in order to fix find_os() memory handling. install-script: Fix OsinfoOs leaks The OsinfoOs instance returned by osinfo_media_get_os must be unref'ed when no longer needed. This fixes: ==29859== 3,926 (96 direct, 3,830 indirect) bytes in 1 blocks are definitely lost in loss record 2,286 of 2,295 ==29859== at 0x65E19D0: g_type_create_instance (gtype.c:1854 ==29859== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==29859== by 0x65C55ED: g_object_new_valist (gobject.c:2040) ==29859== by 0x65C58A0: g_object_new (gobject.c:1624) ==29859== by 0x5339F69: osinfo_os_new (osinfo_os.c:201) ==29859== by 0x5347FFD: osinfo_loader_get_os (osinfo_loader.c:464) ==29859== by 0x534B439: osinfo_loader_os (osinfo_loader.c:1410) ==29859== by 0x534BFB1: osinfo_loader_root (osinfo_loader.c:1599) ==29859== by 0x534C399: osinfo_loader_process_xml (osinfo_loader.c:1680) ==29859== by 0x534CFEE: osinfo_loader_process_file_reg_xml (osinfo_loader.c:1877) ==29859== by 0x534DA1F: osinfo_loader_process_list (osinfo_loader.c:2160) ==29859== by 0x534DCEA: osinfo_loader_process_path (osinfo_loader.c:2232) ==29859== by 0x402310: test_script_file (test-install-script.c:146) ==29859== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==29859== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==29859== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==29859== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==29859== by 0x51178EB: srunner_run (check_run.c:782) ==29859== by 0x40303D: main (test-install-script.c:349) ==29859==- ==29859== 3,926 (96 direct, 3,830 indirect) bytes in 1 blocks are definitely lost in loss record 2,287 of 2,295 ==29859== at 0x65E19D0: g_type_create_instance (gtype.c:1854) ==29859== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==29859== by 0x65C55ED: g_object_new_valist (gobject.c:2040) ==29859== by 0x65C58A0: g_object_new (gobject.c:1624) ==29859== by 0x5339F69: osinfo_os_new (osinfo_os.c:201) ==29859== by 0x5347FFD: osinfo_loader_get_os (osinfo_loader.c:464) ==29859== by 0x534B439: osinfo_loader_os (osinfo_loader.c:1410) ==29859== by 0x534BFB1: osinfo_loader_root (osinfo_loader.c:1599) ==29859== by 0x534C399: osinfo_loader_process_xml (osinfo_loader.c:1680) ==29859== by 0x534CFEE: osinfo_loader_process_file_reg_xml (osinfo_loader.c:1877) ==29859== by 0x534DA1F: osinfo_loader_process_list (osinfo_loader.c:2160) ==29859== by 0x534DCEA: osinfo_loader_process_path (osinfo_loader.c:2232) ==29859== by 0x402626: test_script_data (test-install-script.c:199) ==29859== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==29859== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==29859== by 0x51178EB: srunner_run_tcase (check_run.c:373) install-script: Fix leak in _template_loaded() The file content obtained through g_file_load_contents_finish() was never freed. This fixes: ==29859== 16,384 bytes in 1 blocks are definitely lost in loss record 2,294 of 2,295 ==29859== at 0x4C2DC57: realloc (vg_replace_malloc.c:785) ==29859== by 0x684FF1F: g_realloc (gmem.c:159) ==29859== by 0x681DE68: g_array_maybe_expand (garray.c:779) ==29859== by 0x681E783: g_array_set_size (garray.c:555) ==29859== by 0x681F328: g_byte_array_set_size (garray.c:1752) ==29859== by 0x626F143: load_contents_read_callback (gfile.c:6885) ==29859== by 0x6289129: async_ready_callback_wrapper (ginputstream.c:532) ==29859== by 0x62B0342: g_task_return_now (gtask.c:1107) ==29859== by 0x62B0378: complete_in_idle_cb (gtask.c:1121) ==29859== by 0x684A702: g_main_dispatch (gmain.c:3154) ==29859== by 0x684A702: g_main_context_dispatch (gmain.c:3769) ==29859== by 0x684AAAF: g_main_context_iterate.isra.29 (gmain.c:3840) ==29859== by 0x684ADD1: g_main_loop_run (gmain.c:4034) ==29859== by 0x402442: test_script_file (test-install-script.c:165) ==29859== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==29859== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==29859== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==29859== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==29859== by 0x51178EB: srunner_run (check_run.c:782) ==29859== by 0x40303D: main (test-install-script.c:349) install-script: Fix leak in error case in generate_entity_xml() The key list returned by osinfo_entity_get_param_keys() was freed in the nominal case, but not in the error case. install-script: Fix GFile leaks This fixes: ==12193== 89 (32 direct, 57 indirect) bytes in 1 blocks are definitely lost in loss record 1,826 of 2,310 ==12193== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==12193== by 0x684FE58: g_malloc (gmem.c:94) ==12193== by 0x6866D42: g_slice_alloc (gslice.c:1025) ==12193== by 0x686736D: g_slice_alloc0 (gslice.c:1051) ==12193== by 0x65E1821: g_type_create_instance (gtype.c:1857) ==12193== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==12193== by 0x65C50CC: g_object_newv (gobject.c:1928) ==12193== by 0x65C58B3: g_object_new (gobject.c:1621) ==12193== by 0x6320EF9: _g_local_file_new (glocalfile.c:301) ==12193== by 0x98512F7: g_daemon_vfs_get_file_for_path (gdaemonvfs.c:383) ==12193== by 0x98513A1: g_daemon_vfs_get_file_for_uri (gdaemonvfs.c:407) ==12193== by 0x5334B09: osinfo_install_script_generate_async_common (osinfo_install_script.c:959) ==12193== by 0x5335170: osinfo_install_script_generate_for_media_async (osinfo_install_script.c:1216) ==12193== by 0x402B8A: test_script_file (test-install-script.c:157) ==12193== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12193== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==12193== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==12193== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==12193== by 0x51178EB: srunner_run (check_run.c:782) ==12193== by 0x401F82: main (test-install-script.c:345) ==12193== 89 (32 direct, 57 indirect) bytes in 1 blocks are definitely lost in loss record 1,827 of 2,310 ==12193== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==12193== by 0x684FE58: g_malloc (gmem.c:94) ==12193== by 0x6866D42: g_slice_alloc (gslice.c:1025) ==12193== by 0x686736D: g_slice_alloc0 (gslice.c:1051) ==12193== by 0x65E1821: g_type_create_instance (gtype.c:1857) ==12193== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==12193== by 0x65C50CC: g_object_newv (gobject.c:1928) ==12193== by 0x65C58B3: g_object_new (gobject.c:1621) ==12193== by 0x6320EF9: _g_local_file_new (glocalfile.c:301) ==12193== by 0x98512F7: g_daemon_vfs_get_file_for_path (gdaemonvfs.c:383) ==12193== by 0x98513A1: g_daemon_vfs_get_file_for_uri (gdaemonvfs.c:407) ==12193== by 0x402899: test_script_data (test-install-script.c:191) ==12193== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12193== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==12193== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==12193== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==12193== by 0x51178EB: srunner_run (check_run.c:782) ==12193== by 0x401F82: main (test-install-script.c:345) = test-install-script: Fix various small leaks The value returned by osinfo_install_script_generate_for_media_finish() was not freed in several codepaths, and the mainloop was not destroyed when no longer needed. Data read with g_file_load_contents was also not freed when no longer used. This fixes: ==12193== 528 bytes in 1 blocks are definitely lost in loss record 2,279 of 2,310 ==12193== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==12193== by 0x5F6D459: xmlStrndup (xmlstring.c:45) ==12193== by 0x556EF95: xsltSaveResultToString (xsltutils.c:1780) ==12193== by 0x53345C2: osinfo_install_script_apply_xslt (osinfo_install_script.c:824) ==12193== by 0x53346EC: osinfo_install_script_apply_template (osinfo_install_script.c:853) ==12193== by 0x5334A7A: osinfo_install_script_generate_async_common (osinfo_install_script.c:942) ==12193== by 0x5335170: osinfo_install_script_generate_for_media_async (osinfo_install_script.c:1216) ==12193== by 0x40298A: test_script_data (test-install-script.c:212) ==12193== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12193== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==12193== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==12193== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==12193== by 0x51178EB: srunner_run (check_run.c:782) ==12193== by 0x401F82: main (test-install-script.c:345) ==12193== 16 bytes in 1 blocks are definitely lost in loss record 772 of 2,310 ==12193== at 0x4C2DA60: calloc (vg_replace_malloc.c:711) ==12193== by 0x684FEB0: g_malloc0 (gmem.c:124) ==12193== by 0x684ABF1: g_main_loop_new (gmain.c:3929) ==12193== by 0x402373: test_script_datamap (test-install-script.c:284) ==12193== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12193== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==12193== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==12193== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==12193== by 0x51178EB: srunner_run (check_run.c:782) ==12193== by 0x401F82: main (test-install-script.c:345) ==29859== 16,384 bytes in 1 blocks are definitely lost in loss record 2,295 of 2,295 ==29859== at 0x4C2BAEE: malloc (vg_replace_malloc.c:298) ==29859== by 0x4C2DCA1: realloc (vg_replace_malloc.c:785) ==29859== by 0x684FF1F: g_realloc (gmem.c:159) ==29859== by 0x681DE68: g_array_maybe_expand (garray.c:779) ==29859== by 0x681E783: g_array_set_size (garray.c:555) ==29859== by 0x681F328: g_byte_array_set_size (garray.c:1752) ==29859== by 0x62775AE: g_file_load_contents (gfile.c:6748) ==29859== by 0x4025E9: test_script_data (test-install-script.c:195) ==29859== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==29859== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==29859== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==29859== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==29859== by 0x51178EB: srunner_run (check_run.c:782) ==29859== by 0x40303D: main (test-install-script.c:349) productfilter: Free 'supportDate' in ::finalize() This fixes: ==4878== 8 bytes in 1 blocks are definitely lost in loss record 8 of 319 ==4878== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==4878== by 0x6849E58: g_malloc (gmem.c:94) ==4878== by 0x682A114: g_date_new_dmy (gdate.c:289) ==4878== by 0x5335C6D: osinfo_productfilter_add_support_date_constraint (osinfo_productfilter.c:210) ==4878== by 0x401E8B: test_supportdate (test-product.c:171) ==4878== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==4878== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==4878== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==4878== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==4878== by 0x51178EB: srunner_run (check_run.c:782) ==4878== by 0x40165E: main (test-product.c:221) productfilter: Fix GDate leak osinfo_productfilter_matches_default() was not freeing the GDate instances returned by osinfo_product_get_{release,eol}_date(). This fixes: ==14496== 16 bytes in 2 blocks are definitely lost in loss record 141 of 382 ==14496== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==14496== by 0x684FE58: g_malloc (gmem.c:94) ==14496== by 0x6830114: g_date_new_dmy (gdate.c:289) ==14496== by 0x5337012: date_from_string (osinfo_product.c:345) ==14496== by 0x533706D: osinfo_product_get_release_date (osinfo_product.c:354) ==14496== by 0x5337F6F: osinfo_productfilter_matches_default (osinfo_productfilter.c:279) ==14496== by 0x532CC8D: osinfo_filter_matches (osinfo_filter.c:288) ==14496== by 0x532D51B: osinfo_list_add_filtered (osinfo_list.c:272) ==14496== by 0x532DD52: osinfo_list_new_filtered (osinfo_list.c:443) ==14496== by 0x401B35: test_supportdate (test-product.c:140) ==14496== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==14496== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==14496== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==14496== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==14496== by 0x51178EB: srunner_run (check_run.c:782) ==14496== by 0x40166E: main (test-product.c:219) ==14496==- ==14496== 16 bytes in 2 blocks are definitely lost in loss record 142 of 382 ==14496== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==14496== by 0x684FE58: g_malloc (gmem.c:94) ==14496== by 0x6830114: g_date_new_dmy (gdate.c:289) ==14496== by 0x5337012: date_from_string (osinfo_product.c:345) ==14496== by 0x53370C8: osinfo_product_get_eol_date (osinfo_product.c:364) ==14496== by 0x5337F92: osinfo_productfilter_matches_default (osinfo_productfilter.c:280) ==14496== by 0x532CC8D: osinfo_filter_matches (osinfo_filter.c:288) ==14496== by 0x532D51B: osinfo_list_add_filtered (osinfo_list.c:272) ==14496== by 0x532DD52: osinfo_list_new_filtered (osinfo_list.c:443) ==14496== by 0x401B35: test_supportdate (test-product.c:140) ==14496== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==14496== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==14496== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==14496== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==14496== by 0x51178EB: srunner_run (check_run.c:782) ==14496== by 0x40166E: main (test-product.c:219) test-product: Fix two small leaks This fixes: ==14496== 456 (48 direct, 408 indirect) bytes in 1 blocks are definitely lost in loss record 373 of 382 ==14496== at 0x65E1964: g_type_create_instance (gtype.c:1854) ==14496== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==14496== by 0x65C50CC: g_object_newv (gobject.c:1928) ==14496== by 0x65C58B3: g_object_new (gobject.c:1621) ==14496== by 0x53375FD: osinfo_productfilter_new (osinfo_productfilter.c:93) ==14496== by 0x401982: test_supportdate (test-product.c:121) ==14496== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==14496== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==14496== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==14496== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==14496== by 0x51178EB: srunner_run (check_run.c:782) ==14496== by 0x40166E: main (test-product.c:219) ==14496==- ==14496== 1,964 (48 direct, 1,916 indirect) bytes in 1 blocks are definitely lost in loss record 379 of 382 ==14496== at 0x65E1964: g_type_create_instance (gtype.c:1854) ==14496== by 0x65C36AA: g_object_new_internal (gobject.c:1781) ==14496== by 0x65C55ED: g_object_new_valist (gobject.c:2040) ==14496== by 0x65C58A0: g_object_new (gobject.c:1624) ==14496== by 0x5338257: osinfo_productlist_new (osinfo_productlist.c:81) ==14496== by 0x401946: test_supportdate (test-product.c:116) ==14496== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==14496== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==14496== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==14496== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==14496== by 0x51178EB: srunner_run (check_run.c:782) ==14496== by 0x40166E: main (test-product.c:219) test: Fix osinfo_list_get_elements() leak The returned list must be freed after use. This fixes: ==13583== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==13583== by 0x684FE58: g_malloc (gmem.c:94) ==13583== by 0x6866D42: g_slice_alloc (gslice.c:1025) ==13583== by 0x68462B5: g_list_prepend (glist.c:311) ==13583== by 0x683A06B: g_hash_table_get_values (ghash.c:1814) ==13583== by 0x532D2CC: osinfo_list_get_elements (osinfo_list.c:212) ==13583== by 0x401961: test_uris (test-mediauris.c:87) ==13583== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==13583== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==13583== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==13583== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==13583== by 0x51178EB: srunner_run (check_run.c:782) ==13583== by 0x4016A4: main (test-mediauris.c:146) entity: Fix osinfo_entity_get_param_keys() leak The content of the GList returned by osinfo_entity_get_param_key() is not owned by the caller, but by OsinfoEntity. However, the "id" field added to the list was owned by the caller, and could not be freed properly. This commit adds a const char * to the GList rather than adding a g_strdup'ed string in order to fix this leak. ==12129== 3 bytes in 1 blocks are definitely lost in loss record 7 of 269 ==12129== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==12129== by 0x684FE58: g_malloc (gmem.c:94) ==12129== by 0x686889E: g_strdup (gstrfuncs.c:363) ==12129== by 0x532BB81: osinfo_entity_get_param_keys (osinfo_entity.c:300) ==12129== by 0x401AFF: test_multi_props_clear (test-entity.c:266) ==12129== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12129== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==12129== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==12129== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==12129== by 0x51178EB: srunner_run (check_run.c:782) ==12129== by 0x401369: main (test-entity.c:344) test-entity: Fix osinfo_entity_get_param_keys() leak The list returned by osinfo_entity_get_param_keys() must be freed. ==20812== 24 bytes in 1 blocks are definitely lost in loss record 109 of 262 ==20812== at 0x4C2BBAD: malloc (vg_replace_malloc.c:299) ==20812== by 0x684FE58: g_malloc (gmem.c:94) ==20812== by 0x6866D42: g_slice_alloc (gslice.c:1025) ==20812== by 0x6846A13: g_list_append (glist.c:261) ==20812== by 0x532BB88: osinfo_entity_get_param_keys (osinfo_entity.c:300) ==20812== by 0x40160F: test_empty_props (test-entity.c:71) ==20812== by 0x5117535: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==20812== by 0x51178EB: srunner_iterate_tcase_tfuns (check_run.c:231) ==20812== by 0x51178EB: srunner_run_tcase (check_run.c:373) ==20812== by 0x51178EB: srunner_iterate_suites (check_run.c:195) ==20812== by 0x51178EB: srunner_run (check_run.c:782) ==20812== by 0x40314A: main (test-entity.c:344) 2016-07-01 Fabiano Fidêncio <fidencio@redhat.com> osinfo-query: return failure when no matches are found According to `man osinfo-query`: "The exit status will be 0 if matching entries were found, or 1 if not matches were found". So, let's adapt the program in order to match the manual. Also, this patch was written keeping the current behavior about printing the Short ID, Name, Version, ID table. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1351718 Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-06-03 Christophe Fergeau <cfergeau@redhat.com> test: Make warnings fatal Since standard output is hidden by default while we run make check, it's very easy to miss test cases emitting new warnings. This commit uses g_log_set_fatal_mask() to ensure that the test cases will abort if a warning is emitted while they run. This matches the behaviour of the GTest framework. test: Silence expected warning in test-os os: Fix invalid URLs Some URLs have become obsolete since the last time test-treeuris/test-mediauris were run, this fixes these 2 tests. test: Adjust test database directory layout Since the database reorganization, the files must follow a directory layout matching their id, otherwise a warning is emitted. ubuntu: Fix initrd/kernel paths The initrd extensions are either .lz or .gz depending on the version (server/desktop), and the base directory is either casper or install. Some of the powerpc/parisc entries could be improved as the ISOs contain multiple kernels and initrds (eg 32/64 bit versions). ubuntu: Mark the sparc ISOs as sparc64 These ISOs only contain a 64 bit kernel. ubuntu: Add Ubuntu 16.04 data 2016-06-02 Christophe Fergeau <cfergeau@redhat.com> opensuse: Fix id of x11-keyboard datamap The correct id is http://x.org/x11-keyboard and not http://libosinfo.fedorahosted.org/x11-keyboard opensuse: Rename autoinst files according to their id This causes warnings when running ./test/test-isodetect if they are not named as expected: ** (process:28952): WARNING **: Entity http://opensuse.org/opensuse/autoyast/desktop should be in file install-script/opensuse.org/opensuse-autoyast-desktop.xml not install-script/opensuse.org/autoinst-desktop.xml ubuntu: Add Ubuntu 15.10 test cases Detection was already present, but not the corresponding test files. rhel: Add RHEL 6.8 data rhel: Add RHEL Atomic 7.2 data 2016-05-23 Fabiano Fidêncio <fidencio@redhat.com> Revert "debian: Allow empty passwords" The workaround that has been used so far doesn't work. For the user account, the password is indeed removed in the end of the installation, but then login in from GDM is impossible. Although it works from a VT (depending on the installed version), this is not something obvious that every user would try. So, requiring the user password seems the best to do for now and when a proper solution is found, the user password can be set to optional again. For the root account, the password is not removed in the of the installation and ends up being set as "dummyPa55w0rd", something that the user would never guess, unless they have access to the libosinfo code. So, requiring the admin password seems the best to do for now and when a proper solution is found the admin password can be set to optional again. This reverts commit 087a8f8a23895e2d792b0f6d26a9d0c01a6c7e87. Conflicts: data/install-script/debian.org/debian-preseed-desktop.xml.in Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1336108 https://bugzilla.redhat.com/show_bug.cgi?id=1336109 Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> 2016-05-18 Fabiano Fidêncio <fidencio@redhat.com> debian,installer: Set user avatar Set user avatar in the desktop installation's profile based on what is done in Fedora and openSUSE scripts. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1336112 Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> debian,installers: Enable ftp mirror Currently, after completing the installation, the users are not able to install any software in the VM as the only entry in /etc/apt/sources.list is the installation media itself. Also, desktop installation's profile is broken as qemu-guest-agent is not part of the installation media and we have no mirror set up for downloading the package from. Let's fix these issues enabling the ftp mirror and removing the DVD from the /etc/apt/sources.list. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1336110 https://bugzilla.redhat.com/show_bug.cgi?id=1336111 Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> 2016-05-17 Fabiano Fidêncio <fidencio@redhat.com> debian: Add Debian Testing info Is worth mentioning that Debian Testing is a rolling release done every week. That's the reason the upgrades/derives-from fields are omitted in this patch and that the devices must be added here as well. Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-05-13 Programmingkid <programmingkidx@gmail.com> configure.ac: Fix darwin detection code The $host variable would be set to "x86_64-apple-darwin10.8.0" when using Mac OS X. So to detect it, we need to add another asterisk to the case. Acked-by: Fabiano Fidêncio <fidencio@redhat.com> 2016-05-13 Fabiano Fidêncio <fidencio@redhat.com> win,installer: Fix "Windwos" typo Acked-by: Christophe Fergeau <cfergeau@redhat.com> win,installer: Fix misplaced parenthesis The commit 17809d54 inserted a wrong parenthesis in the test done for skipping the ProductKey during the WindowsPE phase of the script. In this patch, the bogus expression "count(media/variant != 0)" is replaced by "count(media/variant) != 0". Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-05-11 Vladimir Beneš <vbenes@redhat.com> debian,installer: Add ssh-server for installation Installing ssh-server is needed for testing purposes. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1251668 Acked-by: Fabiano Fidêncio <fidencio@redhat.com> 2016-05-11 Andrew Stormont <andyjstormont@gmail.com> build: Fix "--no-undefined" linker's flag for solaris Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1302339 Acked-by: Fabiano Fidêncio <fidencio@redhat.com> 2016-05-11 Fabiano Fidêncio <fidencio@redhat.com> build: Fix "--no-undefined" linker's flag for OSX Apple's linker doesn't have support to "--no-undefined" and we must use something equivalent, like: "-undefined,error". The patch was test by Programmingkid and his positive feedback can be seen here: https://www.redhat.com/archives/libosinfo/2016-May/msg00164.html Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1305016 Acked-by: Christophe Fergeau <cfergeau@redhat.com> 2016-05-11 Christophe Fergeau <cfergeau@redhat.com> media: Fix misplaced parenthesis in strncmp use Due to a misplaced parenthesis, bootable media detection was doing strncmp(.., .., sizeof(..) != 0) instead of strncmp(.., .., sizeof(..)) != 0 This means we were comparing 1 character from BOOTABLE_TAG rather than the full string. This was found when trying to compile libosinfo with CC=clang 2016-05-10 Fabiano Fidêncio <fidencio@redhat.com> win10: Enable unattended installation Unattended installation works properly, apart from the user avatar not being set for Windows 8 (or newer)[0]. [0]: https://bugzilla.redhat.com/show_bug.cgi?id=1328637 Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win,installer: Improve "setting product key" logic Only a few variants of Windows 8 and Windows 8.1, like Enterprise, Enterprise Debug and Professional, have to have the ProductKey skipped during the WindowsPE phase of the installation script. This patch re-writes the "skipping ProductKey" logic in a cleaner way, making it work also for Windows 10 (and all its variants). Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Generalize "No Variant" volume ID expression Let's generalize the expression for x86 volume IDs in order to cover more volume IDs. The expression is based on the one used for the x86_64 volume IDs. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Add two x86_64 volume IDs Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Generalize "Enterprise LTSB" volume ID expression Let's generalize the x86 volume ID expression in order to cover more volume IDs. The expression is based on the one used for the x86_64 volume IDs. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Add two Enterprise LTSB (x86_64) volume IDs LTSB stands for Long Term Servicing Branch. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Generalize "Enterprise" volume ID expression Let's generalize the expression for x86 volume IDs in order to cover more volume IDs. The expression is based on the one used for the x86_64 volume IDs. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Add two Enterprise x86_64 volume IDs Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Generalize "Education" volume ID expression Let's generalize the x86 volume ID expression in order to cover more volume IDs, based on the ones already covered in the x86_64 volume ID expression. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win10: Add an Education x86_64 volume ID Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> 2016-05-09 Fabiano Fidêncio <fidencio@redhat.com> win8,installer: Tune based on variants As Windows 8.1, Windows 8 (most likely) has to have the ProductKey skipped during the WindowsPE phase of the script for the enterprise, enterprise-debug and professional variants. "Most likely" is said because I haven't been able to test the Windows 8 Professional variant as it needs a VLK and MSDN tells me that only administrators can access keys in the Volume Licensing Service Center (VLSC). Also, I haven't been able to start an installation using none of the debug ISOs, as the system reboots just a minute (or so) after booting up. It seems to be a problem with the ISOs themselves, as I've downloaded a few ISOs (different locales) and faced the same issue. win8: Generalize "No Variant" vol ID expression Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8: Add professional variant Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8: Add debug variant Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8: Add enterprise-debug variant Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8: Add enterprise variant Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Enable express installation Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Enable express installation Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse,installer: Add DESKTOP script The script has been tested and works as expected for the most recent releases of OpenSUSE (Leap 4.21 and Tumbleweed). A big thanks to Lasse Schuirmann, who had worked on a similar script, from which I got the chroot-scripts part, during his GSoC project. Co-author: Lasse Schuirmann <lasse.schuirmann@gmail.com> Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse,installer: Add JEOS script The script has been tested and works as expected for the most recent releases of OpenSUSE (Leap 4.21 and Tumbleweed). Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Bump recommended storage size Although the values are correct, it is impossible to have a normal (graphical) usage of a system with the storage's recommended size. So, instead of using 5GB as recommended storage size, let's increase this value to 20GB. Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Add USB Tablet device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Add virtio-net device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Add virtio-block device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Add QXL device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-tumbleweed: Add AC97 device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Add USB Tablet device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Add virtio-net device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Add virtio-block device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Add QXL device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Add AC97 device Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> opensuse-42.1: Fix volume-id match Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> 2016-05-04 Fabiano Fidêncio <fabiano@fidencio.org> win8.1: Add installer-reboots value Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8.1: Re-enable unattended installation Now the unattended installation works properly, apart from the user avatar not being set for Windows 8 (or newer)[0]. Due to the Windows policy on hostname and therefore apps might need to be patched for the installer scripts to work[1]. [0]: https://bugzilla.redhat.com/show_bug.cgi?id=1328637 [1]: https://bugzilla.gnome.org/show_bug.cgi?id=765175 This reverts commit 49ae46cda4fbd2e95dc794cb64aec27117351c85. Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> 2016-05-04 Fabiano Fidêncio <fidencio@redhat.com> installer: Use l10n-language as UI language Since commit 57b459e09 we set 'EN-US' as UI language instead of 'EN-GB'. However, this hack is only necessary for Windows 7 and can cause issues when installing newer versions of the Windows, as the installer, somehow, gets confused by having different values set to UI language, System Locale and User Locale and ends up prompting a window for letting the user to set these values. By limiting the hack to Windows 7, this patch fixes the installation of 'EN-GB' ISOs of Windows 8 and 8.1 (the ones tested for now). Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8.1,installer: Force Windows activation I've noticed that when installing Windows 8.1 the activation is not done by the end of the unattended installation and, actually, the serial number shown in the Activate Windows program is different than the one used to perform the installation. Also, when tring to activate the software it fails with the "Windows can't be activate. Try again later" message and no useful details about the error. So, in order to avoid this issue, let's force the activation in the FirstLogon phase of the installation script by directly editing the Slmgr.vbs options for Volume Activation. For more info about the Slmgr.vbs Options for Volume Activation, please, take a look on: https://technet.microsoft.com/en-us/library/dn502540.aspx Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> 2016-05-04 Fabiano Fidêncio <fabiano@fidencio.org> win8.1,installer: Tune based on variants So, I ended up doing tests with all ISOs that I have access from MSDN and here is the conclusion: Windows 8.1 Problematic: en_windows_8_1_enterprise_debug_checked_x64_dvd_2756358.iso en_windows_8_1_enterprise_n_debug_checked_x64_dvd_2942525.iso en_windows_8_1_enterprise_n_x64_dvd_2971845.iso en_windows_8_1_enterprise_x64_dvd_2971902.iso en_windows_8_1_pro_n_vl_x64_dvd_2971969.iso Non-problematic: en_windows_8_1_n_debug_checked_x64_dvd_2707894.iso en_windows_8_1_debug_checked_x64_dvd_2707208.iso en_windows_8_1_n_x64_dvd_2707896.iso Windows 8.1 with Updates: Problematic: en_windows_8.1_enterprise_n_with_update_x64_dvd_6050225.iso en_windows_8.1_enterprise_with_update_x64_dvd_6054382.iso en_windows_8.1_pro_n_vl_with_update_x64_dvd_6050969.iso en_windows_8.1_pro_vl_with_update_x64_dvd_6050880.iso Non-problematic: en_windows_8.1_n_with_update_x64_dvd_6051677.iso en_windows_8.1_with_update_x64_dvd_6051480.iso Based on this info, the best way to go is: - Set the ProductKey during the WindowsPE phase of the setup only in the following conditions: - any Windows media that is not Windows 8.1 (as it was before) - Windows 8.1 media has no variant - Windows 8.1 media contains on its variant debug but doesn't contain enterprise on it These rules will cover medias with no variant and with "debug" variant that are not enterprise. All the rest (Enterprise, Enterprise Debug and Professional) will have the ProductKey skipped during the WindowsPE phase of the setup. Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8.1: Clean up vol-id entries - IR[M135]_CEN?NA_X*FRE is already covered by Enterprise Volume ids. - IR[M3]_CCSN?A_X*FRE is now covered by No Variant Volume ids. - IR[M15]_CPRN?A_X*FREV is now covered by Professional Volume Ids. Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8.1: Adjust the vol-id for enterprise x86 Use the same pattern of Volume ids used for enterprise x86_64. Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8.1: Add a few more professional volume ids Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8.1: Add two more "no variant" volume ids Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8.1: Add one more enterprise iso data Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8: Add installer-reboots value Acked-by: Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win8: Add comment about the user avatar issue User avatar is not set as expected after performing a unattended installation of Windows 8 (or newer). Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win8: Add missing cmd script for installer Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> win7: Remove two duplicated volume id entries The entries removed are: - GSP1RMCPRVOL (Windows 7 Professional x86) - GSP1RMCPRXVOL (Windows 7 Professional x86_64) Acked-by: Christophe Fergeau <cfergeau@redhat.com> Acked-by: Zeeshan Ali (Khattak) <zeeshanak gnome org> 2016-04-20 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: add information about 10.2 and 10.3 2016-04-19 Christophe Fergeau <cfergeau@redhat.com> opensuse: Remove invalid upgrades/derives-from links tumbleweed has upgrades and derives-from nodes using non-existent IDs. This causes runtime warnings when running test-isodata. 2016-04-19 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: use proper version value for 10.1 10.0 -> 10.1 README: adjust wording a little. README: update homepage 2016-03-24 Charles Arnold <carnold@suse.com> data: add missing openSUSE distro information Add missing openSUSE distro information. Added os data and isodata for the following distros: openSUSE Leap 42.1 openSUSE Tumbleweed data: add missing SLES/SLED distro information Add missing SLES/SLED distro information. Added os data and isodata for the following distros: SUSE Linux Enterprise Server 12 SP1 SUSE Linux Enterprise Desktop 12 SP1 Added isodata for the following distros: SUSE Linux Enterprise Server 11 SP4 SUSE Linux Enterprise Desktop 11 SP4 2016-03-03 Roman Bogorodskiy <bogorodskiy@gmail.com> Fix build with older GCC Build with older gcc fails with: CC libosinfo_1_0_la-osinfo_avatar_format.lo In file included from ../osinfo/osinfo_os.h:30, from ../osinfo/osinfo.h:55, from osinfo_avatar_format.c:28: ../osinfo/osinfo_media.h:68: error: redefinition of typedef 'OsinfoMedia' ../osinfo/osinfo_install_script.h:43: error: previous declaration of 'OsinfoMedia' was here To avoid redefinition, use forward-declaration in a similar way like it's done for OsinfoOs in osinfo/osinfo_os.h. 2016-02-05 Guido Günther <agx@sigxcpu.org> Update CD URLs to latest Debian Jessie point release 2016-01-21 Fabiano Fidêncio <fidencio@redhat.com> build: Fix .vapi generation After a267020e272a generation of the .vapi file broke. It no longer has path_pattern property while get_path_pattern() is still there. Adding nodist_libosinfo_1_0_include_HEADERS to the list of the files passed to the .gir generation solves the problem. 2016-01-13 Daniel P. Berrange <berrange@redhat.com> don't include autogenerated enum files in dist Previous commit 77cf2730eb6e5503a9bd40e6baf7cc0b393a9dc7 reverts 4e488678d4af0e54da7400851d554bc4c4497c76 (fixes to the VPATH build) because it was said to break the make distcheck. After examining the problem, it seems this commit was not in fact the problem - it merely exposed the existing bug elsewhere. The real problem is that we were including the autogenerated enum files in the dist. So when doing builds from git those files were in $builddir, but when doing builds from dist those files were in $srcdir. Except that 'make distclean' would delete the enum files, which would again cause them to appear in $builddir. So the real fix is stop including the enum files in the dist build: fix VPATH based builds The VPATH build fails because the enum generator is told to look for two generated files in the $srcdir instead of builddir. We should not in fact loom the generated files at all when building enums, so they can be removed. The g-ir-scanner is also missing a -I$(build_dir) to let it find the generated files. Finally the gtk-doc tools were not being told to look in the $(build_dir) autobuild.sh: enable gtk doc build when testing autobuild.sh: fix flag to turn on werror In commit 97091136bd037aa44da13873ac910320b9ad7e6b we changed to use gnulib's compiler warnings code, but forgot to change the autobuild.sh flag to match. It should be --enable-werror autobuild.sh: run parallel make for faster builds Set MAKEFLAGS to use -j option matching number of available CPUs. autobuild.sh: set a default value for AUTOBUILD_INSTALL_ROOT 2016-01-11 Fabiano Fidêncio <fidencio@redhat.com> build: Don't use USE_GNOME2_MACROS The following warning was shown when running autogen.sh: ***Warning*** USE_GNOME2_MACROS is deprecated, you may remove it from autogen.sh build: Don't set PKG_NAME The following warning was shown when running autogen.sh: ***Warning*** PKG_NAME is deprecated, you may remove it from autogen.sh Use SOUP_VERSION_MAX_ALLOWED In order to avoid using a too new libsoup API. Use GLIB_VERSION_MAX_ALLOWED In order to avoid using a too new GLib API. Remove checks/code for GLib older than 2.35.1 As now we depend on GLib 2.36, let's remove all checks/code that depend on an older version than the one we require. tree: Add an indentation level for the if block Use GTask instead of GSimpleAsyncResult Instead of using GSimpleAsyncResult, use the new GTask API, which is much more straightforward. For using the new GTask API, let's bump GIO (part of GLib) dependency version to 2.36. what is safe based on major distro support: - Debian Jessie: glib-2.42 - RHEL-7.1: glib-2.40 - SLES12: glib-2.38 - Ubuntu LTS 14.04: glib-2.40 2016-01-08 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.3.0 Revert "build: fix VPATH based builds" This reverts commit 4e488678d4af0e54da7400851d554bc4c4497c76. It breaks `make distcheck` here so reverting it for release 0.3.0: -------------------- if test "/home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/../../docs/reference" != "/home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/docs/reference" ; then \ files=`echo Libosinfo-docs.xml Libosinfo-sections.txt Libosinfo-overrides.txt Libosinfo.types`; \ if test "x$files" != "x" ; then \ for file in $files ; do \ destdir=`dirname /home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/docs/reference/$file`; \ test -d "$destdir" || mkdir -p "$destdir"; \ test -f /home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/../../docs/reference/$file && \ cp -pf /home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/../../docs/reference/$file /home/zeenix/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/docs/reference/$file || true; \ done; \ fi; \ fi make[4]: *** No rule to make target '../../osinfo/*.c', needed by 'scan-build.stamp'. Stop. make[4]: *** Waiting for unfinished jobs.... touch setup-build.stamp make[4]: Leaving directory '/extra-data/checkout/gnome/libosinfo/libosinfo-0.3.0/_build/sub/docs/reference' ------------------- 2015-12-16 Daniel P. Berrange <berrange@redhat.com> Switch to 0.3.0 for next release Since we have had a major change in the way we process the libosinfo database files, this would be a good time to switch to 0.3.0 for release numbers 2015-12-14 Fabiano Fidêncio <fidencio@redhat.com> centos: Add CentOS 6.7 data About the URL field I've opted to use a server who cleary is related to CentOS and that was recommended by CentOS people from #centos-devel at freenode. Those URLs will point the user to select from another URL (from USA or Europe) where the ISOs can actually be downloaded. centos: Add CentOS 6.6 data About the URL field I've opted to use a server who cleary is related to CentOS and that was recommended by CentOS people from #centos-devel at freenode. Those URLs will point the user to select from another URL (from USA or Europe) where the ISOs can actually be downloaded. 2015-11-20 Christophe Fergeau <cfergeau@redhat.com> rhel: Add RHEL 6.7 data rhel: Add RHEL7 codename As documented on https://access.redhat.com/articles/3078?tour=7#RHEL7 rhel: Add RHEL7 release dates rhel: Remove i686 medias for RHEL7 RHEL7 never had i686 ISOs released. rhel: Add RHEL 7.2 data altlinux: Fix ALT Linux 7.0 detection Publisher ID should be all upper-case. altlinux: Fix URLs to ALTLinux 7.0 The ones currently used return a 404 error. ubuntu: Fix URLs to ubuntu 15.10 'vivid' was mistakenly used in the URL path leading to a 404 error. 15.10 is wily, vivid was 15.04. ubuntu: Fix URLs to old ubuntu versions Once they are no longer supported, they move to old-releases.ubuntu.com. fedora: Fix URLs to old fedora versions Once a fedora version becomes unsupported, its ISOs are moved from download.fedoraproject.org to archive.fedoraproject.org debian: Update URLs to latest minor releases During the lifetime of a major release, there are regular minor releases of updated isos. Availability of ISOs for older minor releases seem not to be guaranteed, leading to failures in test-mediauris (with LIBOSINFO_NETWORK_TESTS set in the environment). This commit updates debian 6/7/8 URLs to the latest minor update. debian: Fix Debian 6.0.10 detection The current regex for Debian only expects a single digit as the micro version number, but Debian 6 goes up to 10. configure: bump libsoup version to 2.42 By bumping libsoup version, we are fixing a few deprecated warnings in our tests (use SoupSession directly rather than SoupSessionAsync) and as a side effect we don't need libsoup-gnome anymore. Based on a patch from Fabiano Fidêncio <fidencio@redhat.com> 2015-11-19 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora,script: Use more generic URL The URL we have been pointing Anaconda to, seems to only work for released versions and not for Alpha and Beta. Let's use a URL that works for both released and development snapshots. rhel,script: Install qemu-guest-agent for desktop profile. Agent package has been available since RHEL6 and our installer script is only listed by RHEL 6 and greater so we can ask for it to be installed unconditionally. fedora,script: Install qemu-guest-agent For desktop profile. 2015-11-13 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> debian,script: Install qemu-guest-agent For desktop profile. fedora: Add F23 info & testcases 2015-11-09 Daniel P. Berrange <berrange@redhat.com> schema: allow multiple <name> elements under OS variant The OS <variant> tag must allow multiple <name> elements because there can be translations present. 2015-11-06 Christophe Fergeau <cfergeau@redhat.com> db: Fix valgrind issues with language detection match_languages() returns a list containing a single string, but given the way language_code_from_raw() works, this string may actually point to already freed memory by the time match_languages() ends. Subsequent uses of this list will cause accesses to free'd memory. This commit changes match_languages() so that it uses the string itself (before it's freed) rather than letting the caller do this. This fixes: ==28623== Invalid read of size 1 ==28623== at 0x4C2BC22: strlen (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28623== by 0x6A6B1F2: g_strdup (gstrfuncs.c:362) ==28623== by 0x553FCA2: osinfo_entity_add_param (osinfo_entity.c:253) ==28623== by 0x5554BD8: osinfo_media_set_languages (osinfo_media.c:1256) ==28623== by 0x555A78A: fill_media (osinfo_db.c:641) ==28623== by 0x555ABED: osinfo_db_identify_media (osinfo_db.c:709) ==28623== by 0x4030AA: test_one (test-isodetect.c:346) ==28623== by 0x40339D: test_windows (test-isodetect.c:395) ==28623== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==28623== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==28623== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==28623== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==28623== by 0x532BB7C: srunner_run (check_run.c:782) ==28623== by 0x403A1B: main (test-isodetect.c:500) ==28623== Address 0x89914b0 is 0 bytes inside a block of size 4 free'd ==28623== at 0x4C29D6A: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28623== by 0x6A525ED: g_free (gmem.c:189) ==28623== by 0x55586A6: match_languages (osinfo_db.c:117) ==28623== by 0x555A76C: fill_media (osinfo_db.c:639) ==28623== by 0x555ABED: osinfo_db_identify_media (osinfo_db.c:709) ==28623== by 0x4030AA: test_one (test-isodetect.c:346) ==28623== by 0x40339D: test_windows (test-isodetect.c:395) ==28623== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==28623== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==28623== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==28623== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==28623== by 0x532BB7C: srunner_run (check_run.c:782) ==28623== by 0x403A1B: main (test-isodetect.c:500) ==28623== Block was alloc'd at ==28623== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==28623== by 0x6A524D8: g_malloc (gmem.c:94) ==28623== by 0x6A6B277: g_strndup (gstrfuncs.c:425) ==28623== by 0x6A5EC90: g_match_info_fetch (gregex.c:1005) ==28623== by 0x55583DF: get_raw_lang (osinfo_db.c:57) ==28623== by 0x5558672: match_languages (osinfo_db.c:113) ==28623== by 0x555A76C: fill_media (osinfo_db.c:639) ==28623== by 0x555ABED: osinfo_db_identify_media (osinfo_db.c:709) ==28623== by 0x4030AA: test_one (test-isodetect.c:346) ==28623== by 0x40339D: test_windows (test-isodetect.c:395) ==28623== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==28623== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==28623== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==28623== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==28623== by 0x532BB7C: srunner_run (check_run.c:782) ==28623== by 0x403A1B: main (test-isodetect.c:500) 2015-11-05 Christophe Fergeau <cfergeau@redhat.com> tests: Add testcases for ID collision in OsinfoList This will detect issues fixed by the commit fixing osinfo_list_add. list: Fix replacement in osinfo_list_add OsinfoList maintains a mapping between a const char *entity_id and an OsinfoEntity object. It does so by using both a GPtrArray and a GHashTable. When an entity is added to an OsinfoList, a pointer from the OsinfoEntity is used as the key (through osinfo_entity_get_id()) and the OsinfoEntity is g_object_ref'ed and used as a value. The OsinfoEntity is also appended to the array mentioned previously. The use of g_hash_table_new_full() ensures that the OsinfoEntity will be unref'ed when it's removed from the hash table. However, when trying to add an entity with the same id as one already present in the OsinfoList, several issues occur: - g_hash_table_insert() is used, which means the preexisting entity with the same id is going to get removed, and thus g_object_unref'ed. However, the _old_ key will still be used by the hash table, which means that if the OsinfoList was holding the last reference to the old OsinfoEntity, the key stored in the hash table is going to point to freed memory - similarly, we make no attempt at removing the preexisting entity from the GPtrArray, so this will also contain stale data This commit makes sure we remove entities with the same id from the GPtrArray when this situation occurs. It uses g_hash_table_replace() rather than g_hash_table_insert() which replaces both the key and the value if the hash table already contains the same key. This fixes some valgrind issues or crashes that occurred when running make check in a libosinfo tree with 2 databases, one in the old layout and one using the new layout. test-isodetect: Fix GDataInputStream leak in load_iso() This fixes: ==12890== 1,024,600 (15,040 direct, 1,009,560 indirect) bytes in 235 blocks are definitely lost in loss r ==12890== at 0x67E2C88: g_type_create_instance (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C4B7A: ??? (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C6A14: g_object_new_valist (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C6D80: g_object_new (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x40244F: load_iso (test-isodetect.c:101) ==12890== by 0x4029B1: load_distro (test-isodetect.c:205) ==12890== by 0x402B28: load_distros (test-isodetect.c:252) ==12890== by 0x402C51: load_isos (test-isodetect.c:283) ==12890== by 0x403058: test_one (test-isodetect.c:331) ==12890== by 0x4033B8: test_windows (test-isodetect.c:387) ==12890== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12890== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==12890== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==12890== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==12890== by 0x532BB7C: srunner_run (check_run.c:782) ==12890== by 0x4037B9: main (test-isodetect.c:496) test-isodetect: Fix leak when loading ISO data The string returned by g_data_input_stream_read_line() must be freed when no longer needed. This fixes: ==12890== 199,410 bytes in 6,710 blocks are definitely lost in loss record 2,166 of 2,169 ==12890== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==12890== by 0x6A514D8: g_malloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x646F9AE: g_data_input_stream_read_line (in /usr/lib64/libgio-2.0.so.0.4600.1) ==12890== by 0x402806: load_iso (test-isodetect.c:135) ==12890== by 0x4029B1: load_distro (test-isodetect.c:205) ==12890== by 0x402B28: load_distros (test-isodetect.c:252) ==12890== by 0x402C51: load_isos (test-isodetect.c:283) ==12890== by 0x403058: test_one (test-isodetect.c:331) ==12890== by 0x4033B8: test_windows (test-isodetect.c:387) ==12890== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12890== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==12890== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==12890== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==12890== by 0x532BB7C: srunner_run (check_run.c:782) ==12890== by 0x4037B9: main (test-isodetect.c:496) test-isodetect: Fix GFileInfo leak This fixes: ==12890== 10,438 (520 direct, 9,918 indirect) bytes in 13 blocks are definitely lost in loss record 2,151 ==12890== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==12890== by 0x6A514D8: g_malloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A68622: g_slice_alloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A68CBD: g_slice_alloc0 (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x67E2B63: g_type_create_instance (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C4B7A: ??? (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C6460: g_object_newv (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x67C6D93: g_object_new (in /usr/lib64/libgobject-2.0.so.0.4600.1) ==12890== by 0x652836D: ??? (in /usr/lib64/libgio-2.0.so.0.4600.1) ==12890== by 0x652587B: ??? (in /usr/lib64/libgio-2.0.so.0.4600.1) ==12890== by 0x6481253: g_file_enumerator_next_file (in /usr/lib64/libgio-2.0.so.0.4600.1) ==12890== by 0x402B7A: load_distros (test-isodetect.c:246) ==12890== by 0x402C51: load_isos (test-isodetect.c:283) ==12890== by 0x403058: test_one (test-isodetect.c:331) ==12890== by 0x4033B8: test_windows (test-isodetect.c:387) ==12890== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12890== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==12890== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==12890== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==12890== by 0x532BB7C: srunner_run (check_run.c:782) ==12890== by 0x4037B9: main (test-isodetect.c:496) test-isodetect: Fix leak of 'langs' GList This fixes: ==12890== 1,032 bytes in 43 blocks are definitely lost in loss record 2,129 of 2,169 ==12890== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==12890== by 0x6A514D8: g_malloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A68622: g_slice_alloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A47EA9: g_list_copy_deep (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x55400A8: osinfo_entity_get_param_value_list (osinfo_entity.c:424) ==12890== by 0x555476E: osinfo_media_get_languages (osinfo_media.c:1238) ==12890== by 0x402CCA: test_langs (test-isodetect.c:300) ==12890== by 0x4031E8: test_one (test-isodetect.c:350) ==12890== by 0x4033B8: test_windows (test-isodetect.c:387) ==12890== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12890== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==12890== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==12890== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==12890== by 0x532BB7C: srunner_run (check_run.c:782) ==12890== by 0x4037B9: main (test-isodetect.c:496) db: Fix leak of 'variants' GList The GList returned by osinfo_entity_get_param_value_list() must be freed when no longer needed. This fixes: ==12890== 2,112 bytes in 88 blocks are definitely lost in loss record 2,139 of 2,169 ==12890== at 0x4C28C50: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==12890== by 0x6A514D8: g_malloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A68622: g_slice_alloc (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x6A47EA9: g_list_copy_deep (in /usr/lib64/libglib-2.0.so.0.4600.1) ==12890== by 0x55400A8: osinfo_entity_get_param_value_list (osinfo_entity.c:424) ==12890== by 0x555A58F: fill_media (osinfo_db.c:654) ==12890== by 0x555A8C4: osinfo_db_identify_media (osinfo_db.c:709) ==12890== by 0x4030C5: test_one (test-isodetect.c:338) ==12890== by 0x4033B8: test_windows (test-isodetect.c:387) ==12890== by 0x532B78A: tcase_run_tfun_nofork.isra.9 (check_run.c:390) ==12890== by 0x532BB7C: srunner_iterate_tcase_tfuns (check_run.c:231) ==12890== by 0x532BB7C: srunner_run_tcase (check_run.c:373) ==12890== by 0x532BB7C: srunner_iterate_suites (check_run.c:195) ==12890== by 0x532BB7C: srunner_run (check_run.c:782) ==12890== by 0x4037B9: main (test-isodetect.c:496) = l10n: Update translations from zanata 2015-10-22 Daniel P. Berrange <berrange@redhat.com> loader: re-add support for loading files explicitly Some existing applications pass an XML file to the osinfo_loader_process_path method, so we must continue to support this, even though we want apps to switch to using a directory of XML files, one per entity. This requires us to re-add the 'skipMissing' flag for use when loading files. The other change is that we need to disable the code that checks relative paths when loading a single file, instead of files inside a directory. loader: remove leading / from relative path Fix the warning ** (gnome-boxes:5499): WARNING **: Entity http://opensuse.org/opensuse/11.4 should be in file /os/opensuse.org/opensuse-11.4.xml not os/opensuse.org/opensuse-11.4.xml by removing the leading '/', since g_file_get_relative_path does not use a leading '/' 2015-10-20 Daniel P. Berrange <berrange@redhat.com> loader: don't return NULL from a void method loader: implement new semantics for entity overriding This implements the new scheme for allowing entities to be overridden. If an XML file relative path matches one in an earlier datbase directory, the original file is not loaded at all. Any files in a ENTITY-NAME.d will now augment earlier data rather than replace it. eg consider these files 1. /usr/share/libosinfo/db/os/fedoraproject.org/fedora-22.xml 2. /etc/libosinfo/db/os/fedoraproject.org/fedora-22.xml 3. $HOME/.config/libosinfo/db/os/fedoraproject.org/fedora-22.d/extra.xml File #1 will not get loaded, because file #2 is detected as a full override. File #2 will get loaded to define the main Fedora 22 OS, and then file #3 will get loaded to augment the main data. Most of the PCI and USB device info in the local DB is just augmenting the pci.ids and usb.ids master files, so these need to move to .d directories now loader: skip leading whitespace in USB/PCI ids strings When parsing the usb.ids & pci.ids files, we must skip any leading whitespace in the vendor/device/subsystem strings. loader: sanity check entity filenames when loading When we have loaded the id attribute of the entity, check it against the entity filename to ensure its XML document was held in a file with the modern naming. This merely prints out a warning for now to avoid breaking existing users, but in the future, files with non-compliant names will not be loaded at all. loader: rework handling of pci.ids and usb.ids Currently we have the option to download pci.ids/usb.ids or symlink to the distro provided version given to configure. Change configure so that it searches for the files in a variety of expected locations. This ensures we always use the external files, in any modern Linux distro, without the user needing to give an arg. We look for them in /usr/share/hwdata/{pci.ids,usb.ids} (RHEL/Fedora) /usr/share/misc/{pci.ids,usb.ids} (Ubuntu/Debian/Gentoo) /usr/share/{pci.ids,usb.ids} (SLES/OpenSuse) Instead of loading the pci.ids/usb.ids files as part of the main database file enumeration process, explicitly load them from their expected location. This avoids the need to symlink the distro provided files into the database dir. loader: rewrite loader to separate file enumeration from loading The current loader code enumerates files, processing them as it finds them. This changes it into a two stage process, first enumerating all files and then loading all files. This will facilitate later patches which will change the way files are enumerated. loader: re-arrange code for loading list of directories Re-arrange the loading code so that there is a single helper method that takes an array of GFile * to load. This will facilitate future code for handling overrides between directories. loader: pull out code for building database paths Create standalone helper methods for constructing a GFile object for the system/local/user database paths. data: remove "virtio" and "ide" as buses The virtio devices don't use a virtio bus, rather than virtio protocol is layered over another bus like pci. IDE is not a bus type we need to track compatibility for. tests: switch to use XML from builddir only All XML files are given a .in suffix, even if they don't contain translations, so we can assume all .xml files are now in the build dir, and nothing in the source dir needs loading. Update ignore files for new database layout data: remove pluralization from data directories eg instead of platforms/qemu.org/qemu-1.2.0.xml we just have platform/qemu.org/qemu-1.2.0. The top level directory name now matches the entity name exactly data: split devices into one file per device data: split windows-* into one file per install script data: split ubuntu into one file per install script data: split rhel into one file per install script data: split fedora into one file per install script data: split debian into one file per install script data: split x11-keyboard into one file per datamap data: split windows-lang into one file per datamap data: split xen into one file per platform data: split rhel-xen into one file per platform data: split qemu into one file per platform data: split kvm into one file per platform data: split windows into one file per OS data: split ubuntu into one file per OS data: split suse into one file per OS data: split solaris into one file per OS data: split rhl into one file per OS data: split rhel into one file per OS data: split opensuse into one file per OS data: split openbsd into one file per OS data: split netware into one file per OS data: split netbsd into one file per OS data: split mandriva into one file per OS data: split mandrake into one file per OS data: split mageia into one file per OS data: split macos into one file per OS data: split gnome into one file per OS data: split freebsd into one file per OS data: split fedora into one file per OS data: split dos into one file per OS data: split debian into one file per OS data: split centos into one file per OS data: split altlinux into one file per OS data: rename 'hypervisors' to 'platforms' The database directory name matches the object name, except in the case of the platform object, where the directory is named "hypevisors". Change the dir name to "platforms" to be consistent. data: standard install script URI format Fix the URI format for install scripts to take the scheme http://$domain-name/$distro/$script/$profile where "$script" is the type of install script (ie kickstart, preseed, unattend, etc) for the distro in question. data: change domain name for x11 keyboard map The X11 keyboard mapping is associated with the X.org site, not the libosinfo site. data: change PS2 device URIs to refer to ibm.com The PS2 computer spec originated with IBM, so the correct URI to refer to is ibm.com not qemu.org data: fix URI scheme use for ib700 device The ib700 device currently uses a URI of "isa://ib700". Change the URI to refer to the product vendor (http://ibasetechnologies.net) and use a path of "isa/ib700" for the device. data: change URIs for netware OS The common URI path format is distro/version, so make netware URIs comply. data: fix typos in altlinux, debian and rhl datasets The versions for altlinux 5/6/7 are inconsistent, sometimes using 1 digit, sometimes 2 digits and sometimes 3 digits. Change them all to use 2 digits as was practice for earlier versions. Also remove codenames from the product name. Debian buzz was 1.1, not 1.0 - there was never an official 1.0 release. The version of RHL 6.2 was wrong data: change GNOME URIs to include 'gnome' prefix Change GNOME URIs to include 'gnome' as the first part of the URI path, since this will be used to identify filenames later. data: change QEMU/KVM ID URIs to include product string Include "qemu" and "qemu-kvm" as the first part of the path component of the URIs, since this will be used to form a filename later. data: fix RHEL Xen platform versions & adjust IDs Currently the RHEL Xen platform is returning the same version number for multiple entities. Fix this by suffixing the Xen version with the RHEL version. Change the short-id and URI IDs so they are of the form 'xen-rhel-$VERSION' eg http://redhat.com/xen-rhel/3.1.0.5.6 xen-rhel-3.1.0.5.6 data: sanitize Xen device ID URIs and add missing devs The Xen devices have a formal device ID value, which is used in Xen bus. We should record that in the <product-id> field and add a corresponding <product> string. Change the device ID URIs so that they are of the format http://xen.org/xen/$PRODUCT-ID Finally add various devices that were missing. data: remove duplication in Xen hypervisor devices Every Xen hypervisor platform has duplicated the list of supported devices. The RHEL Xen platforms also missed out the <upgrades> relationships. Remove all the device duplication and rely on inheritance instead. data: change PCI and USB device ID URIs Currently PCI and USB device URIs look like http://pciids.sourceforge.net/v2.2/pci.ids/8086/2668 http://www.linux-usb.org/usb.ids/80ee/0021 We want to associate filenames with URIs in a well-defined format. These existing URIs, particular the PCI one, are too long in the path component. It is also desirable to use the industry standards body as the domain name, rather than the domain of the specific project we happen to import data from right now. Thus the URIs are changed to http://pcisig.com/pci/1af4/1000 http://usb.org/usb/80ee/0021 data: remove most information about PCI devices We already load PCI/USB information from the pci.ids/usb.ids files, so should not duplicate it again in XML files. The only info we need to retain are the device classes. docs: add document describing database organization Add a document that sets out the desired end goal for the database organization that the following patches will move towards. docs: describe entity URI ID scheme Add a doc describing the best practice guidelines around the format of URIs for various entities. 2015-10-16 Christophe Fergeau <cfergeau@redhat.com> test-isodetect: Don't set size when block size is unknown When parsing an isoinfo file, we don't try to set the volume size if the 'vol_size' was not set. This commit does the same when 'blk_size' is not set as volume size is set to vol_size * blk_size. test-isodetect: Ignore empty isoinfo fields While parsing isoinfo files, test-isodata is going to add empty strings ("") to the libosinfo database for the fields it can parse. This commit ignores them instead. test-isodetect: Fix invalid read when parsing isoinfo files While parsing the isoinfo files, we check if the current line starts with some prefix, but then we skip one byte more when we pass the string to osinfo_entity_set_param(). This accounts for a space which comes right after the prefix. It can happen that the line being parsed ends right after the prefix with no additional space, in which case we'd be accessing invalid memory when we try to skip the space. This commit adds the additional space to the prefix check, and uses strlen() rather than a hardcoded len when skipping the prefix later on. This fixes ==10921== Conditional jump or move depends on uninitialised value(s) ==10921== at 0x4C2BC29: strlen (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==10921== by 0x6A6B222: g_strdup (gstrfuncs.c:362) ==10921== by 0x553F4F2: osinfo_entity_set_param (osinfo_entity.c:190) ==10921== by 0x402716: load_iso (test-isodetect.c:141) ==10921== by 0x4029C7: load_distro (test-isodetect.c:203) ==10921== by 0x402B3E: load_distros (test-isodetect.c:250) ==10921== by 0x402C67: load_isos (test-isodetect.c:281) ==10921== by 0x40306E: test_one (test-isodetect.c:329) ==10921== by 0x4032E1: test_rhel (test-isodetect.c:367) ==10921== by 0x532B78A: ??? (in /usr/lib64/libcheck.so.0.0.0) ==10921== by 0x532BB7C: srunner_run (in /usr/lib64/libcheck.so.0.0.0) ==10921== by 0x403A4C: main (test-isodetect.c:490) 2015-10-12 Jasper Lievisse Adriaanse <jasper@humppa.nl> data: add OpenBSD 5.7 and 5.8 https://bugzilla.redhat.com/show_bug.cgi?id=1270735 2015-10-07 Daniel P. Berrange <berrange@redhat.com> data: add missing Xen 3.1.0 release The RHEL Xen platforms reference a Xen 3.1.0 release which was missing entirely. The 3.1.0 was not officially released as a tar.gz but several enterprise distros shipped it based off the GIT tag. Similarly 3.1.1 & 3.1.2 were not officially released. These are not added since no distro is thought to have used them. data: fix bad upgrades reference for KVM platform The KVM 1.2.0 platform is referencing a non-existant platform via the upgrades relationship. data: fix reference to windows install script for vista Previous commit commit 12e13186165f9f6ecb5e1eb92fea450502bb81c2 Author: Fabiano Fidêncio <fabiano@fidencio.org> Date: Thu Aug 16 15:54:11 2012 -0300 Add desktop profile for Windows Changed the URI for the existing install scripts to have a 'desktop' or 'jeos' suffix. It added both new install scripts to existing OS entries, but forgot to update the Windows vista entry, so it was left pointing to a non-existant script. loader: identify referenced but not defined entities When loading the database we often have to instantiate entities to represent relationships (upgrades, derives, etc) before the entity is actually defined. This makes the code liable to bugs as we might instantiate entities which are never defined. This extends the loader so that it tracks all entities that are created via references, and once loading is complete it prints out a warning if any referenced entities were not defined fully. This identifies a number of mistakes in our current database that the following patches will fix. loader: fix memory leaks when loading data The 'lang' attribute needs to be free'd with xmlFree. The install script config param needs to be unref'd since when it is added to the install script an extra ref is taken. The install script avatar needs to be unref'd since when it is added to the install script an extra ref is taken. 2015-09-25 Daniel P. Berrange <berrange@redhat.com> build: fix path to locale files in mingw RPM The mingw RPM spec listed the locale files in the libdir instead of datadir test: load XML files from builddir too If the builddir is separate from the sourcedir we must load files from the former too, otherwise we miss all the XML files which are translated. build: fix ChangeLog generation in VPATH build When running 'make dist' from a VPATH, the ChangeLog file is not generated because it doesn't find the .git directory build: fix VPATH based builds The VPATH build fails because the enum generator is told to look for two generated files in the $srcdir instead of builddir. We should not in fact loom the generated files at all when building enums, so they can be removed. The g-ir-scanner is also missing a -I$(build_dir) to let it find the generated files. Finally the gtk-doc tools were not being told to look in the $(build_dir) 2015-09-24 Christophe Fergeau <cfergeau@redhat.com> rhel: Add RHELSA 7.1 volume ID rhel: Narrow down RHELSA 7.0 volume ID regexp It's currently matching RHELSA.*aarch64 regardless of the version number present in the volume ID. This means it will likely match all RHELSA 7.x releases in addition to RHELSA 7.0 win10: Add one volume ID Data was provided by Vaclav Ehrlich win8.1: Add 2 more volume IDs The regex matching x86-64 enterprise variants is not as strict as it could: it's going to match IR5_CENNA_X64FREV or IR1_CENA_X64FREV for example for which we don't have test cases. However, I'm confident this will match en_windows_8.1_enterprise_n_with_update_x64_dvd.iso and en_windows_8.1_enterprise_x64_dvd.iso if such ISOs exist, and that it's not going to give us false positives. Since it might allow us to identify unknown ISOs without false positives, I went for it. Data was provided Vaclav Ehrlich 2015-09-23 Marc Deslauriers <marc.deslauriers@ubuntu.com> ubuntu: Update Ubuntu information - Fix typos - Use official release names instead of codenames - Update release and eol dates - Update repo URLs for eol releases 2015-09-17 Christophe Fergeau <cfergeau@redhat.com> fedora: Add Fedora 20 for PowerPC volume ID Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 centos: Add CentOS 6.3 Live volume ID Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 freebsd: Add FreeBSD 10.1 volume ID Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 win: Add Windows 8.1 volume IDs Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 win: Add more Windows 8 test cases These ISOs are already covered by libosinfo volume IDs. Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 win: Add Windows 7 volume IDs Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 win: Add Windows Server 2012 R2 volume IDs Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 win: Add Windows Server 2012 volume IDs Data provided by John Snow in https://bugzilla.redhat.com/show_bug.cgi?id=1212599 2015-09-11 Christophe Fergeau <cfergeau@redhat.com> win: Fix wrong comment added in 7b454eff 'debug' variant is introduced by a /* Starter */ comment. 2015-09-10 Christophe Fergeau <cfergeau@redhat.com> Add Windows 10 Education volume-id 2015-09-02 Daniel P. Berrange <berrange@redhat.com> Fix sun/oracle confusion in solaris metadata There are multiple mistakes in the solaris metadata - Solaris 11 was released after Oracle acquisition, so vendor should be Oracle - The product name of Solaris >= 11 includes "Oracle" and not "Sun", so should be "Oracle Solaris 11" - The original product names for earlier versions of Solaris did not include "Sun", so just "Solaris 10" - The domain name in the URL for OpenSolaris should be sun.com not oracle.com - The domain name in the URL for Solaris 11 should be oracle.com Resolves https://bugzilla.redhat.com/show_bug.cgi?id=1257151 2015-08-26 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Ditch udev rule New udev/blkid expose more ISO9660 properties that apps can grab and set on install media for detection to work. As a result our udev rule becomes redundant. This improves system bootup time conciderably if there is an ISO9660 media inserted during bootup as our udev rule needed to parse the XML etc and is therefore pretty slow (keeping in mind that a few seconds are a long time in the land of modern udev/systemd). 2015-08-17 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Bump version 2015-07-23 Victor Toso <victortoso@redhat.com> windows: Add a win7 volume ID Iso downloaded directly from MSDN with the name: Windows 7 Home Premium with Service Pack 1 (x64) - DVD (English) Related: https://bugzilla.gnome.org/show_bug.cgi?id=752635 2015-07-17 Christophe Fergeau <cfergeau@redhat.com> fedora: Fix fedora-unknown regexp It's currently matching fedora 21 and fedora 22, causing make distcheck failures. 2015-07-16 T A Mahadevan <ta.mahadevan@gmail.com> Added support for qemu upto version 2.3.0 2015-07-09 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> rhel: Add data about RHEL 7.1 2015-06-04 Lin Ma <lma@suse.com> suse: Add opensuse 13.2 info 2015-05-27 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.12 NEWS: Fix year for last 2 releases fedora: Add F22 info This time around, we also recommend RAM and storage to some nice sane values. Test data for each media is included. fedora,script: More generic test for netinstaller Instead of checking if variant is 'netinst', let's check if it contains 'netinst' in it. In Fedora 22, networkinstaller media are provided for both Server and Workstation variant so we'll need two different variants for both but the same installer script could be used with both. 2015-05-26 Christophe Fergeau <cfergeau@redhat.com> ubuntu: Add Ubuntu 15.04 info ubuntu: Add Ubuntu 14.04 info debian: Add Debian 8.0 info 2015-05-26 Lin Ma <lma@suse.com> suse: Add sled11 sp1, sp2, sp3 and sp4 Add entries for SUSE Linux Enterprise Desktop 11 SP1, SP2, SP3 and SP4. I can't find out the release date information about these SLED distros, will update them when I figure them out. Change in V1: Removed the fix for sles 11 32bit to another patch. suse: Fix the volume-id information for sles 11.x 32bit 2015-05-19 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Enable installer scripts for F21 Now that these scripts work for at least one of the variants (which was always the case for all Fedora versions we supported the installer scripts for anyway), let's enable them. fedora,script: Fix netinst auto-install The fedora install script was written for and tested against installer DVDs so it does not work for netiso. This patch fixes the issue by removing the install method commands and adding remote repos if the media is detected to be a netiso. https://bugzilla.redhat.com/show_bug.cgi?id=1178787 fedora: Separate variant for F21 netiso tools,install-script: Allow arg to be a media file Now that we have API to generate scripts for specific media and since it is the recommended path to generating scripts, let's allow users of this tool to provide an installer or live media as argument. tests: Pass media to script generator Now that we have an API to pass media to script generation methods, let's test that API instead since its not only the recommended API but also this way we get to test both passing of OS and media information to script template. One of the test cases still uses the existing API on purpose to keep the case of 'no media available' covered as well. Recommend osinfo_install_script_generate_*for_media() Recommend the use of the new script generation API that takes an identified media object as argument rather than OS, in the docs. Add osinfo_install_script_generate_command_line_for_media() Add variant of osinfo_install_script_generate_command_line() that takes (an identified) media instead of an OS. Currently no commandline template needs access to media but lets add this to be future safe and be consistent with other osinfo_install_script_generation*() API. Add osinfo_install_script_generate_output_for_media*() Add variant of osinfo_install_script_generate_for_media*() that generates the script into a file. Add osinfo_install_script_generate_for_media*() Add API that allows apps to pass the media matched against the OS they are generating the script for. Scripts can use this information to make decisions about how installation should be setup. For example, we are going to use this in the Fedora installer script to decide whether to install from ISO or from network, based on whether or not media is a network installer. install-script: Add some missing docs 2015-05-19 Lin Ma <lma@suse.com> suse: Add sles11 sp1, sp2, sp3 and sp4 Add entries for SUSE Linux Enterprise Server 11 SP1, SP2, SP3 and SP4. 2015-05-19 Christophe Fergeau <cfergeau@redhat.com> vista: Re-add volume id removed by mistake In my recent commit adding 2 new Windows Vista IDs, I mistakenly removed a preexisting one, breaking make distcheck in the process. 2015-05-18 Christophe Fergeau <cfergeau@redhat.com> win: Add ISO data for 2 more Vista releases 2015-05-12 Christophe Fergeau <cfergeau@redhat.com> test-isodetect: Remove space before parens This fixes make syntax-check which is currently failing with: GEN sc_bracket_spacing_check test/test-isodetect.c:153: blk_size = (gint64) atoll (line + 23); test/test-isodetect.c:155: vol_size = atoll (line + 16); 2015-05-07 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> install-script: Minor fix to a doc comment build: Ignore 'strings exceeding C90 limit' warning We don't really need to care about C90 compliance (it's pretty old standard by now) to have to limit our strings to 509 characters. tools,install-script: Add missing '\n' to error msg 2015-05-06 Cole Robinson <crobinso@redhat.com> rhel: Add AArch64 ISO RHELSA will be its own release for a short while, but long term it's expected to just be another RHEL architecture. So piggyback on the rhel7.0 entry for now. If needed we can add a toplevel rhelsa entry later. schema: Sync arch names with libvirt Next patch will add aarch64, and we will may have ppc64le variants in the future too 2015-04-29 Charles Arnold <carnold@suse.com> suse: Add sles12 and sled12 Add entries for SUSE Linux Enterprise Server and Desktop 12. suse: Change vendor from Novell to SUSE SUSE has been run as an independent entity since about 2011 and while all things Novell have been absorbed into the parent company (now Micro Focus International), Novell no longer exists as a company. suse: Rename Suse to SUSE Rename Suse to SUSE because this is the official spelling. 2015-04-20 Christophe Fergeau <cfergeau@redhat.com> tests: Rename ISO test data from .iso to .iso.txt A few files containing iso-info -d -i output have a .iso extension instead of a .txt extension. test-isodetect only looks at .txt files, so they need to be renamed. make check still passes after these changes. 2015-04-10 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> windows: Add a win8.1 volume ID Add volume ID of a Windows 8.1 enterprise ISO and test data for it. 2015-03-17 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> test: Add all F21 ISO testcases fedora: Add F21 server netiso media entries fedora: Add volume size info for F21 ISOs tests,fedora: Add testcases for all F20 ISOs fedora: Separate media entry for each ISO Different ISOs have different volume sizes so to be able to add volume size info to all medias, we have to create separate media entry for each ISO. Moreover, apart from the fact that this makes one particular XML file a lot bigger than we'd like it to be, this change makes a lot of sense anyway. This will also make addition of sequence numbers to multi-disc installer medias, much easier later. test: Replace F19 & F21 Alpha ISO data w/ released Both F19 and F21 has been released. Not only its better to keep test data against released ISOs rather than Alpha ISOs, the volume-size is obviously different between released and Alpha ISO so if we want to add volume-size for these ISOs to our database, we need this change to avoid breaking the tests. db: Use volume-size info for detection If volume-size is known of a media, ensure that it's the same as media being compared with, for detection. This will enable us to differentiate ISOs with exactly same volume-id, system-id and application-id. It's possible that we encounter two different ISOs that have the same volume-size, not just the other IDs but chances of that happening are extremely small. Even if that happens, this patch doesn't make things worse for that scenario. test-isodetect: Load volume size info Our data on various ISOs provide volume size information. Lets load that information so that its part of our tests. schema: Allow volume-size attribute for ISO9660 media Allow medias to specify the size of ISO9660 volumes. media: Extract volume size from ISO9660 device/image When creating a media object from an ISO9660 device/image, extract volume size too along with other other information. media: API to get/set ISO9660 volume size 2015-03-15 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Correct indentation in .syms file Bump version 2015-03-04 Rick Kerkhof <rick.2889@gmail.com> Add a Windows 7 CD ID 2015-03-04 Christophe Fergeau <cfergeau@redhat.com> os: Free minimum/recommended resources in _finalize() These OsinfoResourcesList objects will be leaked otherwise. 2015-02-27 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> gnome: Add entry for 3.14 2015-02-23 Daniel P. Berrange <berrange@redhat.com> Switch to using zanata for translations Push new pot with cd po make libosinfo.pot zanata push Pull new translations with cd po zanata pull 2015-02-11 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> script,rhel: Specify boot partition size for < 7 Seems even RHEL 6.6 Anaconda can't handle the --recommended option of part command so lets not use that unless its RHEL >= 7. 2015-02-04 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> rhel,script: Don't specify interface Its not guaranteed that we'll have 'eth0' available or that it will be named 'eth0' by kernel. Since kickstart allows us not to specify the name of the interface, lets omit that and let kickstart choose the best device for us. This change has only been tested against RHEL 7.0. Would probably be best to test against 6.x too. 2015-01-26 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora,script: Don't specify interface Its not guaranteed that we'll have 'eth0' available or that it will be named 'eth0' by kernel. Since kickstart allows us not to specify the name of the interface, lets omit that and let kickstart choose the best device for us. 2015-01-06 Daniel P. Berrange <berrange@redhat.com> Allow tree/media elements in any order Put all child elements for <tree> and <media> inside an <interleave> to allow them to appear in any order 2015-01-05 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Add/update media & tree info for F21 Fedora now has variants and simply updating the ISO URLs is not a good way forward. This patch adds information about media/tree of each variant seperately. Unfortunately, there is no installer media for Workstation variant and even though Server netiso is actually generic enough to let user install Workstation as well, both netiso and full installer ISO have the exact same volume-id so we can't simply assign netiso a (fake) generic variant and be done with it. :( For more details: https://bugzilla.redhat.com/show_bug.cgi?id=1178787 schema: Allow variants for install trees Allow install trees to be associated with specific variants of the OS. 2014-12-15 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Mark F21 as released F21 was release on Dec 3, 2014. Add release date for F21 and update release status. 2014-12-01 Lasse Schuirmann <lasse.schuirmann@gmail.com> openSUSE: Add information on factory images openSUSE: Add information on future versions 2014-11-14 Giuseppe Scrivano <gscrivan@redhat.com> oses: Add RHEL Atomic 7.0 2014-11-06 Jasper Lievisse Adriaanse <jasper@humppa.nl> Add OpenBSD 5.6 2014-11-06 Satoru Satoh <ssato@redhat.com> Add rhel-5.9..11 and rhel-6.6 info 2014-10-14 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora,rhel: Bigger boot partition I had a RHEL VM where I got into trouble when yum wasn't able to upgrade system as my boot partition was too small to fit new kernel image. Kickstart format allows us to just say that we want the size to be whatever is the recommended value so lets make use of that. I tested against RHEL 7.0 and /boot was allocated 512MB. 2014-10-13 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win: Use direct HTTPS URLs to drivers HTTP URLs we have in DB redirect to the actual HTTPS URLs. Its better to point to the actual URL anyway (for efficiency for one) but currently this also breaks driver download in Boxes 3.14.x as it can't handle indirection at the moment. https://bugzilla.gnome.org/show_bug.cgi?id=738168 2014-10-11 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> windows-unattend: Remove bogus architecture attribute This was supposed to be removed when we moved this script template from Boxes to libosinfo. I think it hasn't caused any issues so far because we declare this attribute again soon after and probably Windows ignores the first value and takes the last one it parses. 2014-09-25 Giuseppe Scrivano <gscrivan@redhat.com> oses: Fix upgrades and derives-from for Ubuntu 14.04 2014-09-24 Cole Robinson <crobinso@redhat.com> Add symlink to allow using OSINFO_DATA_DIR=$srcdir Creates a symlink 'db' pointing to 'data'. Now If I want to test commits to ensure osinfo-detect works, I do: OSINFO_DATA_DIR=`pwd` ./tools/osinfo-detect oses: Add Fedora21 This just adds a plain Fedora21 entry. F21 is going to have different versions like 'Workstation' and 'Server' but we don't seem to distinguish for other distros so I just stuck with the pattern here. This currently doesn't work for detecting Workstation media, since their volume IDs are truncated. Bug tracking that: https://bugzilla.redhat.com/show_bug.cgi?id=1145264 2014-09-23 Cole Robinson <crobinso@redhat.com> oses: Add Ubuntu 14.04 LTS 2014-09-22 Cole Robinson <crobinso@redhat.com> gitignore: Ignore build-aux/compile Which appears on f21 m4: Update manywarnings from gnulib Fixes these noisy errors on Fedora 21: gcc: warning: switch '-Wmudflap' is no longer supported 2014-09-17 Fabiano Fidêncio <fidencio@redhat.com> Add a missing win8.1 i686 entry Just replicating an x86_64 entry to the i686 volume-ids Add more win8.1 data to DB And replicating the x86_64 entry to the i686 volume-ids 2014-08-19 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.11 2014-08-19 Christophe Fergeau <cfergeau@redhat.com> Add win2k12r2 info Add win2k12 info win8.1: Add missing volume id win7: Add missing volume id Add openSUSE 13.1 Live detection It's relying on the iso application id data to differentiate between 32 and 64 bit versions. Different live ISOs will have a different application ID, so this could be used to differentiate between KDE and GNOME ISOs for example. build-sys: Generate glib enum files in $(builddir) Since b4c5a1b5c which introduced the generation of an osinfo_version.h file when configure is run, the glib enum files will be regenerated during make distcheck instead of using the one from the tarball. osinfo/Makefile.am currently attempts to generate these files in $(srcdir), which causes make distcheck to fail as it makes $(srcdir) read-only. This patch makes sure these files are generated in $(builddir) and that they are cleaned up afterwards. 2014-08-16 Lasse Schuirmann <lasse.schuirmann@gmail.com> debian: Add needs-internet flag to JEOS script Maybe it is possible to avoid the need of an internet connection in this script but in the current version debian will notify the user with a "bad archive mirror" message if the connection is not present. This question seems not to be meant to be preseeded and I don't know a way to do it. debian: Allow empty passwords This uses the same workaround as the Fedora express installation script. If an empty password is given, the password is set to a dummy password to prevent the installer from erroring out and deleted later. debian: Better resources for wheezy We currently specify recommended RAM for wheezy to be 512MiB and storage to be 5.4GB. Given that the user might not only want to install a bare system but also do something with the system 20GB of disk space and 1GiB of RAM are a reasonable amount. Furthermore the debian desktop express installation script will fail with less than about 15GB or so. 2014-08-12 Giuseppe Scrivano <gscrivan@redhat.com> syntax-check: fix warning osinfo/osinfo_install_script.c:1333:can not maint.mk: undesirable word sequence 2014-08-11 Lasse Schuirmann <lasse.schuirmann@gmail.com> debian: Add desktop installation script This script needs an internet connection because it seems not possible to install the desktop that is present on the ISO. Choosing the 'desktop' package always installs GNOME which is not always shipped with the ISO hence is fetched from the internet. More info: https://wiki.debian.org/tasksel#Desktop_environment install-script: Add needs-internet param Add API to query install scripts whether they can work properly without an internet connection or not. debian: Fix JEOS installation script grub-installer does not accept the boot device unless given after a certain command. See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=666974 for more info. 2014-08-06 Giuseppe Scrivano <gscrivan@redhat.com> rhel7: set resources for ppc64 At least 2 GiB of memory are needed by the Anaconda installer. https://access.redhat.com/articles/rhel-limits cfg.mk: enable sc_prohibit_atoi_atof and skip it for tests Avoid the usage of atoi and atof which are believed to be not safe, as they don't do any error reporting. cfg.mk: enable and fix sc_space_tab It checks that tab indentation is not used together with white spaces indentation. cfg.mk: enable more syntax-check tests maint.mk: update from gnulib 2014-08-06 Fabiano Fidêncio <fidencio@redhat.com> win7: Add one more volume ID to DB 2014-08-01 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> install-script: Default to DISK injection method If script does not declare an injection method, assume DISK as every script must at least support one injection method and there is no need to require scripts to always explicitly declare supported injection methods. 2014-07-29 Roman Bogorodskiy <bogorodskiy@gmail.com> freebsd: drop ISO info Apparently, FreeBSD ISOs have the same metadata for every version, so it's not possible to distinguish releases. Drop this information from OS data to avoid ambiguity in version detection. 2014-07-28 Roman Bogorodskiy <bogorodskiy@gmail.com> Add FreeBSD 9.3 EOL info obtained from http://www.freebsd.org/security/security.html. 2014-07-28 Giuseppe Scrivano <gscrivan@redhat.com> Update RHEL7 status and add ISOs info Add CentOS 7.0 Add CentOS 6.5 2014-07-11 Lasse Schuirmann <lasse.schuirmann@gmail.com> debian: Assume defaults for l10n & disk This introduces default values so that no questions are asked if localization or target disk are not given or 'C'. 2014-06-05 Jasper Lievisse Adriaanse <jasper@humppa.nl> OpenBSD supports virtio net/disk Fix typo for FreeBSD There's a small typo in an XML comment. 2014-06-05 Giuseppe Scrivano <gscrivan@redhat.com> osinfo_loader: avoid multiple calls to OSINFO_ENTITY osinfo_loader: do not use xpath to read localized strings osinfo_loader: replace some xpath queries with direct data access osinfo_loader: introduce a compiled xpaths cache entity: use g_hash_table_replace instead of remove+insert 2014-06-04 Giuseppe Scrivano <gscrivan@redhat.com> enforce bracket spacing with a syntax rule The build-aux/bracket-spacing.pl script was copied from libvirt. osinfo: fix spacing to satisfy sc_bracket_spacing_check tools: fix spacing to satisfy sc_bracket_spacing_check test: fix spacing to satisfy sc_bracket_spacing_check osinfo_loader: don't wrap return value with parentheses 2014-05-22 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> docs: Add missing docs links from main docs page 2014-05-21 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> install-config: Document hostname limitations Since Microsoft Windows has several limitations on allowed characters and number of characters in ComputerName (which is what hostname is translated to), lets inform the app developers about these limitations. http://technet.microsoft.com/en-us/library/ff715676.aspx 2014-05-16 Giuseppe Scrivano <gscrivan@redhat.com> openbsd: add missing 5.* releases 2014-05-14 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win7,installer: 'EN-GB' not supported as UI language Apparently 'EN-GB' is not supported as UI language but only as locale by win7 installer. Lets override 'EN-GB' to 'EN-US' for UI language so unattended installation doesn't break on EN-GB locale. http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx 2014-05-12 Jasper Lievisse Adriaanse <jasper@humppa.nl> Update OpenBSD information and add test data 2014-04-24 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win7: Put pro n sp1 regex under correct architecture In commit 39f22d2 (win7: Generalize volume-id for professional), I generalized the regex of windows 7 professional to include N SP1 but I ended up adding the regex for both i386 and x86_64 architectures under i386. This patch fixes that. 2014-04-23 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win7: Generalize volume-id for professional Generalize volume-id to include 'N SP1' variant. 2014-04-16 Cole Robinson <crobinso@redhat.com> Export OSINFO_*_VERSION values There isn't any to get the version number when using libosinfo from python. Export the VERSION macros, similar to what gtk does. 2014-04-14 Cole Robinson <crobinso@redhat.com> freebsd: 10.0 is released, and supports virtio dates from: http://www.freebsd.org/security/ virtio support from: http://www.freebsd.org/cgi/man.cgi?query=virtio&sektion=4 And the fact that the gnome 3.12 freebsd image requires it. 2014-04-12 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> rhel: Add testcase for a 6.5 ISO 2014-03-26 Cole Robinson <crobinso@redhat.com> windows: Add some win7 and vista volume IDs These are from DVD media I snagged off MSDN years ago. Unfortunately I renamed them a long time ago and I don't know the original file names, hence the generic 'unknown' in the test data names. Update .gitignore It was missing mageia.xml. Ignore all xml files from data/oses as they will be generated from .xml.in files in order to integrate translations. Based on a patch from Cole Robinson. 2014-03-20 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.10 2014-03-20 Giuseppe Scrivano <gscrivan@redhat.com> osinfo-install-script: read config values as strings Prefer xmlNewDocRawNode to xmlNewDocNode when the content is a raw string. Solves this problem: $ osinfo-install-script rhel6.5 -c "admin-password=a&b" error : unterminated entity reference b 2014-03-14 Christophe Fergeau <cfergeau@redhat.com> README: Fix typo This fixes a typo in the 'Patch submission' header 2014-03-06 Giuseppe Scrivano <gscrivan@redhat.com> installer,rhel: Add injection method info installer,fedora: Add injection method info installer,windows: Add injection method info 2014-03-06 Fabiano Fidêncio <fabiano@fidencio.org> ubuntu: Add jeos install-script (version >= 12.04) script based on: https://github.com/clalancette/oz/tree/master/oz/auto Co-Author: Giuseppe Scrivano <gscrivan@redhat.com> 2014-03-06 Giuseppe Scrivano <gscrivan@redhat.com> debian: add installer script osinfo-install-script: support --list-injection-methods installer: Add get_injection_methods() The API allows applications to query the available methods to inject the script to the installation process. 2014-02-28 Giuseppe Scrivano <gscrivan@redhat.com> README: fix mailing list URL 2014-02-21 Fabiano Fidêncio <fidencio@redhat.com> win8.1: Disable unattended installation When win8.1 entry was added it was mostly a c&p from a previous version and the installer part was mistakenly also copied. Since win8.1 unattended installation has never been tested and some people have reported it not to work, lets disable it for now and re-enable when we assure it works. 2014-01-26 Christophe Fergeau <cfergeau@redhat.com> Add hypervisor data for newer QEMUs 2014-01-24 Giuseppe Scrivano <gscrivan@redhat.com> oses: add missing "distro" element for Debian 7 2014-01-17 Ryan Lortie <desrt@desrt.ca> Switch to using the upstream m4 macros for GObject-introspection integration. https://fedorahosted.org/libosinfo/ticket/11 2014-01-07 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Uncomment installer URLs Now that Fedora 20 is released and these URLs are valid, time to enable this data. win7: Add data/test for an enterprise N ISO 2014-01-07 Ryan Lortie <desrt@desrt.ca> build: add libraries via _LIBADD, not _LDFLAGS Make sure we use the _LIBADD variable for adding our dependencies. Adding libraries via _LDFLAGS means that we have our libraries appearing before the user's LDFLAGS environment variable, which is bad, because the LDFLAGS from pkg-config could contain -L for system paths and jhbuild sets LDFLAGS to -L${prefix}/lib so that we can find the libraries in the user's jhbuild install path. If the system path flags come before the user's directory then we'll get the wrong version. Also: automake documentation says it's wrong and would actually error out on this if it wasn't hidden behind a variable substitution. https://fedorahosted.org/libosinfo/ticket/12 2013-12-17 Giuseppe Scrivano <gscrivan@redhat.com> ubuntu: add Saucy Salamander ubuntu: mark 13.04 as released solaris: add Solaris 11 mandriva: add Mandriva Business Server 1.0 mandrake: remove wrong codename for 10.2 freebsd: add version 9.2 and 10.0 freebsd: add missing eol-date Taken from: http://www.freebsd.org/security/ 2013-12-17 Fabiano Fidêncio <fidencio@redhat.com> opensuse: Add 13.1 information 2013-12-16 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.9 2013-12-16 Fabiano Fidêncio <fidencio@redhat.com> debian: Update URL for 6 and 7 debian: fix ISO's recogntion for debian7 Debian 7 is using 7.\d.\d format of ISOs instead of the usual 7.0.\d 2013-12-15 Giuseppe Scrivano <gscrivan@redhat.com> test: distribute test-os.xml It fixes "make check" within a distributed tarball build: specify the full path to libosinfo.syms It fixes "make check" in a VPATH build 2013-12-11 Daniel P. Berrange <berrange@redhat.com> Remove trailing blank line in test file Add scripts for validating the libosinfo.syms file Import two test scripts from libvirt code which validate that all symbols in libosinfo.syms exist in the ELF binary, and also validate the alphabetical sorting. These are hooked up to run with 'make check'. Sort symbols in libosinfo.syms alphabetically We had mostly tried to sort symbols alphabetically but had screwed up in a large number of places. Remove non-existant osinfo_device_driver_set_signed sym The libosinfo.syms file listed osinfo_device_driver_set_signed which does not exist in the C code. This causes build errors on Win32 platforms which don't tolerate non-existant symbols 2013-12-06 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win7,xp: Version driver location This is so that apps can differentiate between different versions of drivers. 2013-12-05 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> rhel: Add comment for future 6.x entries 2013-12-04 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> install-script,rhel: Adjust for RHEL 7 rhel: Add RHEL 7.0 info install-script,rhel: Make it work for 'Server' variant Apparently there is neither office suite nor 'internet-applications' package group on (at least the first) 'Server' variant ISO and they are not exactly needed by default on workstation/client either. User can easily install them later if needed. install-script,rhel: Specify installation method/source This is the same change as we made to fedora install script in commit d04f022cc30d03f451e27284b6311309cf1dee9f. While installation method/source has always been mandatory, Anaconda has ignored us not specifying it so far. Our luck runs out in RHEL 7 though, where unattended installation breaks if this is not explicitly specified. rhel: Add info on 6.5 2013-11-29 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> win7: List all variants List all variants and map appr. medias to them. tools: osinfo-detect prints variant If a particular variant is matched against, osinfo-detect should print the name of that variant rather than the OS itself. media: Add variant info API/XML Add: * ability to associate media to one or more variants of the OS in the 'media' XML node. For example: <os id="http://microsoft.com/win/7"> <short-id>win7</short-id> <name>Microsoft Windows 7</name> .. <variant id="starter"> <name>Microsoft Windows 7 Starter</name> </variant> <variant id="home-basic"> <name>Microsoft Windows 7 Home Basic</name> </variant> <variant id="home-premium"> <name>Microsoft Windows 7 Home Premium</name> </variant> .. <media installer-reboots="2" arch="x86_64"> <variant id="home-premium"/> <iso> .. </iso> </media> </os> * API to query variants of the OS, media is associated with. os: Add variant info API/XML Add: * ability to specificy various variants of an OS under the 'os' XML node. For example: <os id="http://microsoft.com/win/7"> <short-id>win7</short-id> <name>Microsoft Windows 7</name> .. <variant id="starter"> <name>Microsoft Windows 7 Starter</name> </variant> <variant id="home-basic"> <name>Microsoft Windows 7 Home Basic</name> </variant> <variant id="home-premium"> <name>Microsoft Windows 7 Home Premium</name> </variant> .. </os> * API to add and query variants of an OS. Introducing OsinfoOsVariantList Just a OsinfoList subclass that is meant to contain Variant objects. Introducing OsinfoOsVariant This is a new entity class that will represent variants of an OS. For example professional, enterprise and ultimate editions of Windows OSs and workstation and server variants of RHEL etc. 2013-11-26 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Add generic entry for future releases This is to ensure that we do recognise medias of future (pre-)releases of Fedora and therefore give apps some clue about what they are dealing with and some rough estimate of required resources. gnome: Add info for gnome-continuous 3.12 gnome-continuous is continuous integration system so images produced by it track the git master of all modules and now that GNOME 3.10 is out and many projects have branched for 3.10 maintainance, these images are already 3.12 (3.11 at the moment but thats splitting hair I guess). 2013-11-26 Christophe Fergeau <cfergeau@redhat.com> Add testcase for loading OS XML There was no test testing that OS XML data is correctly loaded. This test is also an opportunity to test the various ways of representing a boolean in XML. 2013-11-26 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Add release status info to OS entries Applications can use this to determine if an OS is an actual released OS, a pre-release or a development snapshot. Co-author: Christophe Fergeau <cfergeau@redhat.com> 2013-11-13 Christophe Fergeau <cfergeau@redhat.com> Fix warnings in OsinfoLoader boolean support I always build with -O0, which sometimes makes me some warnings. The newly introduced boolean support in OsinfoLoader triggers some warnings when building with the default flags: osinfo_loader.c: In function 'osinfo_loader_boolean': osinfo_loader.c:200:255: error: return makes integer from pointer without a cast [-Werror] g_return_val_if_fail(ctxt != NULL, NULL); ^ osinfo_loader.c:201:257: error: return makes integer from pointer without a cast [-Werror] g_return_val_if_fail(xpath != NULL, NULL); ^ osinfo_loader.c: In function 'osinfo_loader_entity.isra.8': osinfo_loader.c:331:48: error: 'value_bool' may be used uninitialized in this function [-Werror=maybe-uninitialized] osinfo_entity_set_param_boolean(entity, keys[i].name, value_bool); Stop using <foo>true</foo> in libosinfo database libosinfo can now parse <foo/> as an alternative to <foo>true</foo> to represent a boolean. As the former is the preferred representation, switch to using this in the data libosinfo ships. Teach osinfo_loader_entity() how to parse booleans We want to support both <foo>true</foo> and <foo/> to specify booleans. Now that osinfo_loader_entity additional keys have type information, we can use a dedicated boolean parser for boolean keys. Add type info to osinfo_loader_entity() 'key' argument When using 'string(./foo)' as an XPath query, if foo does not exist, an empty string ("") will be returned. Thus osinfo_loader_string() ignores empty strings during parsing. This happens during regular parsing as we try to lookup string(./foo[lang(XX-xx)]) for localized nodes, and ignore its result if the XPath query returns "" However, we want to represent booleans as <foo/>. Boolean values are currently parsed using osinfo_loader_string(). Parsing <foo/> as a string would return an empty string. This means such nodes would be ignored as the current parser is not able to make the difference between an existing node with empty content and a non-existing node. By associating type information to the additional keys we want osinfo_loader_entity() to parse, we'll be able to indicate which values are booleans, and to parse them accordingly. 2013-11-06 Christophe Fergeau <cfergeau@redhat.com> Allow to pass a NULL 'keys' argument to osinfo_loader_entity() Some callers don't want to pass in any additional keys, and would have to resort to build a dummy empty key list to pass to osinfo_loader_entity(). It's better to allow for that 'keys' argument to be NULL and have osinfo_loader_entity() deal with it. 2013-10-16 Fabiano Fidêncio <fabiano@fidencio.org> win7: Add another win7 ISO data configure: Fix mailing list address 2013-10-15 Christophe Fergeau <cfergeau@redhat.com> Add mageia ISO data 2013-10-11 Giuseppe Scrivano <gscrivan@redhat.com> Fix RHEL 5.8 data 2013-10-07 Fabiano Fidêncio <fabiano@fidencio.org> win8.1: Add Windows 8.1 support This commit adds the volume IDs and test cases data for the Windows 8.1 ISOs I have access in MSDN 2013-10-03 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Post release version bump 2013-09-27 Christophe Fergeau <cfergeau@redhat.com> Use latest warnings.m4 from gnulib The one we were using does not work properly with clang, causing the build process to try to use -f/-W options that are not supported by clang. 2013-09-25 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Enable installer for Fedora 20 2013-09-16 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.8 Update .gitignore windows: Use latest (0.65) version of spice-guest-tools 2013-09-13 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> gnome: Add info about gnome-continuous 3.10 images Unlike the previous os entries for 'gnome', that were targetted towards Fedora-based live ISOs, this one is targetted for qcow2 (and hopefully in future ISOs) images produced by gnome-continuous: https://wiki.gnome.org/GnomeContinuous http://build.gnome.org/ostree/buildmaster/images/z/current/ Would be really nice for libosinfo to be able to detect OS from images too but that might not be feasible so at least for now apps would have to do their own detection (perhaps through guestfs). Once OS matched to a libosinfo's OS definition, apps can at least use libosinfo to figure minimum/recommended resources, which devices are supported and a nice default name etc to be able to create a decent/working VM given an installed disk image. The resources set were based on Colin Walters advice: <zeenix> walters: hi. Would you be able to approximate min. system requirements for 3.10 images? <zeenix> walters: i was thinking of just copying the ones from fedora 19 (at least for now) <walters> zeenix: offhand i'd say dual core and 1G of memory <drago01> is llvmpipe really happy with just two cores? <ebassi> "happy" is probably too strong <ebassi> but it works reasonably well <ebassi> you probably don't want to be on battery power, though <ebassi> or, at least, you won't be for long if you do <drago01> ok 2013-09-12 Fabiano Fidêncio <fidencio@redhat.com> Fix "propfile" typo in osinfo-install-script.c 2013-09-11 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> gnome: Sync 3.8 min. requirements with F19 Also remove comment about copying from F17/18 as 3.8 inheriting from F19 is indication enough of the relationship. fedora: Add data about F20 Its mostly just (edited) copy&paste of F19 data. fedora: More correct/appr. system reqs for F19 These values are based on F19 docs: http://fedorapeople.org/groups/docs/release-notes/en-US/sect-Release_Notes-Welcome_to_Fedora_.html#hardware_overview Unfortunately the docs do not specify any recommended resources so this patch also removes the made-up recommended resources info. Apps will have to figure those on their own, perhaps based on mininum resources. 2013-09-10 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> gnome: Link 3.8 to Fedora 19 GNOME 3.8 is more alike & related to Fedora 19 than it is to GNOME 3.6, so we better specify that it inherits from that. fedora: Fix version of F19 media trees fedora: Uncomment media info about F19 Uncomment some info that was supposed to be uncommented after release of F19 but got forgotten. Also correct URLs of live ISOs. 2013-09-09 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> build: Fix build against latest intltool The use of IT_PROG_INTLTOOL and AM_GNU_GETTEXT together is not supported anymore apparently. Without this change you'll get this error at the end of configure script run: config.status: error: po/Makefile.in.in was not created by intltoolize. 2013-09-06 Eric Blake <eblake@redhat.com> build: avoid missing file on parallel build Fabiano Fidêncio reported a case where make failed due to one generated file not existing in time for the generation of the other: GEN osinfo_enum_types.c Can't open ./osinfo_enum_types.h: No such file or directory at /usr/bin/glib-mkenums line 296, <> line 3630. * osinfo/Makefile.am (osinfo_enum_types.c): Add dependence on .h file. 2013-08-28 Christophe Fergeau <cfergeau@redhat.com> Don't install private headers Export more missing symbols All these symbols are present in the public headers but not listed in libosinfo.syms. They were found using abi-compliance-checker by comparing one build where the use of the .syms file is disabled (all non-static functions get exported), and one regular build. Move osinfo_install_script_add_config_param to private header When this method was added, we forgot to add it to the .sym file. It's meant to work with the osinfo_install_script_*_config_param() API, but these methods are not really consistent with the rest of libosinfo API, they belong in an OsinfoInstallConfigParamList class which was added in 789adb2a. We could decide to export this method, but since noone should have been able to use it, it's better to not export it now, and have one less deprecated method in our ABI. 2013-08-27 Christophe Fergeau <cfergeau@redhat.com> Export more missing OsinfoProduct symbols They were listed in public header files, but not exported in libosinfo.syms. 2013-08-27 Giuseppe Scrivano <gscrivan@redhat.com> Export the symbol osinfo_product_get_codename 2013-08-13 Daniel P. Berrange <berrange@redhat.com> Remove deprecated SOUP_TYPE_PROXY_RESOLVER_GNOME 2013-08-13 Alexey Shabalin <a.shabalin@gmail.com> Add ALTLinux support 2013-07-29 Marcus Karlsson <mk@acc.umu.se> freebsd: Add OS info for FreeBSD 9.1 freebsd: Add OS info for FreeBSD 8.4 freebsd: Add OS info for FreeBSD 8.3 2013-06-06 Christophe Fergeau <cfergeau@redhat.com> Add hypervisor data for newer QEMUs 2013-06-03 Christophe Fergeau <cfergeau@redhat.com> Fix ubuntu 11.10 detection The ISO system ID is not set to LINUX on these ISOs but is empty Adjust ubuntu URLs Some older releases have been moved from http://releases.ubuntu.com to http://old-releases.ubuntu.com Add test case for win2k8r2 ISO Add volume id for Windows 7 Enterprise N SP1 2013-05-13 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.7 2013-05-09 Fabiano Fidêncio <fidencio@redhat.com> debian,test: Add OS info/testcase for Debian 7.0.* 2013-04-27 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> fedora: Specify installation method/source While installation method/source has always been mandatory, Anaconda has ignored us not specifying it so far. Our luck runs out in Fedora 19 though, where unattended installation breaks if this is not explicitly specified. I'm not sure if hard coding the method is a good idea though. Ideas? 2013-04-19 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> gnome: Add data for GNOME 3.8 Its mostly just adjusted copy&paste of 3.6 data. The actual live media is not yet available for 3.8 itself but for 3.7.91/2 they are available and people have been using them already so better add some info already so those ISOs are recognized at least. Before anyone asks, the 3.7.91/2 ISOs use the same volume ID as it will be on 3.8 ISO. 2013-04-18 Christophe Fergeau <cfergeau@redhat.com> ubuntu: Add 13.04 ISO data It's unreleased yet, but should make testing in VMs easier. ubuntu: Add 12.10 test data fedora: Add Fedora 19 Alpha detection This adds recognition of F19 Alpha ISO to libosinfo database. As VMs are often used to test pre-releases of OSes, having libosinfo detection for such ISOs early in the distro development process is a nice thing to have. 2013-04-14 Fabiano Fidêncio <fidencio@redhat.com> rhel,fedora: Add kickstart commandline installer: API to generate commandline info Each OS has a specific method to handle unattended installations. To start an unattended installation one needs to pass a command line to the kernel. This API enables apps to query such command line, if any. 2013-04-12 Fabiano Fidêncio <fidencio@redhat.com> rhel,installer: Change RHEL's install script expected filename Although Kickstart doesn't care about the script filename, once the path to the file is correct, let's use the correct distribution name as expected filename. Fix "FASLSE" typo in osinfo_install_script_get_can_pre_install_drivers() API doc 2013-04-12 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> (Very late) post-release version bump 2013-04-09 Fabiano Fidêncio <fidencio@redhat.com> Update .gitignore test: Add testcase for Fedora 18 test: Add testcase for Fedora 17 centos,test: Add OS info/testcase for CentOS 6.4 centos,test: Add OS info/testcase for CentOS 6.3 centos,test: Add OS info/testcase for CentOS 6.2 centos,test: Add OS info/testcase for CentOS 6.1 centos,test: Add OS info/testcase for CentOS 6.0 rhel,test: Add OS info/testcase for RHEL 6.4 fedora,rhel,installer: Don't use "--activate" This option doesn't work with RHEL < 6.1 and the documentation (http://fedoraproject.org/wiki/Anaconda/Kickstart#network) says 'Device of the first network command is activated if network is required, e.g. in case of network installation or using vnc.' 2013-04-09 Daniel P. Berrange <berrange@redhat.com> Auto-generate AUTHORS file from GIT logs Rather than trying to manually keep track of authors, just auto-generate the list from GIT logs 2013-04-09 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> installer,win7: Setup user avatar for desktop profile While I failed to find a way to setup avatar for the user we create, I did find this method to setup avatar for admin. Since we setup our user as admin, this in the end means the same as setting up user's avatar. 2013-04-08 Daniel P. Berrange <berrange@redhat.com> Disable static libraries by default Every source file is currently built twice by libtool, once for the shared library and once for the static library. Static libs are not commonly packaged by distros and slow down compilation time by almost 100% compared to a shared-only build time. Time for non-parallel make shared only: 22 secs shared + static: 40 secs Those few people who really want them, can pass --enable-static to configure Disabling them by default requires use of LT_INIT. We don't need to support older libtool, so drop use of AM_PROG_LIBTOOL entirely. Also add the 'win32-dll' flag, since that should have been there already 2013-04-08 Christophe Fergeau <cfergeau@redhat.com> build: Make sure $(pkgdatadir)/db exists When creating symlinks to usb.ids/pci.ids, we assume the destination directory already exists. However, this assumption could very well break during parallel builds if we try to create the symlinks before anything create the db/ directory. This commit creates it explicitly as part of the symlink rule to avoid any potential issue. Issue pointed out by Samuli Suominen <ssuominen@gentoo.org> build: Use $(MKDIR_P) rather than mkdir -p Add osinfo_platform_get_all_devices() documentation Use 'unsigned int' rather than enum type for flag param Setting an enum variable to a value that is not part of the enum (as will happen when passing a bitwise-or flag param to osinfo_product_foreach_related) is undefined behaviour in C. Add osinfo_product_foreach_related() documentation It's not exported, but API doc never hurts. Fix relations that osinfo_platform_get_all_devices follows We want to consider both platforms that are upgraded from (qemu 1.2.0 -> qemu 1.2.1) and derived from (qemu 1.2.0 -> qemu-kvm 1.2.0). Should have been part of commit da20e3a 2013-04-03 Christophe Fergeau <cfergeau@redhat.com> Implement osinfo_platform_get_all_devices() Implement osinfo_os_get_all_devices() with osinfo_product_foreach_related() Add osinfo_product_foreach_related() This method iterates over all products that are related to a given product. This will be useful to implement osinfo_platform_get_all_devices(). 2013-03-29 Christophe Fergeau <cfergeau@redhat.com> Replace tab with space Fixes make syntax-check Add Andreas to AUTHORS Use g_getenv() instead of getenv() It's likely to be more portable Fix a few string leaks 2013-03-27 Christophe Fergeau <cfergeau@redhat.com> build: Try to remove symlink before trying to create it Symlink creation fails if a file/symlink with the same name already exists. This causes issues when installing in a tree where libosinfo was already installed. build: More portable symlink creation Follow the recommendations from http://www.nondot.org/sabre/Mirrored/autoconf-2.12/autoconf_4.html#IDX111 regarding symlink creation. 2013-03-27 Andreas Henriksson <andreas@fatal.se> build: Take DESTDIR into account when creating symlinks 2013-03-25 Christophe Fergeau <cfergeau@redhat.com> Fix 'retreives' typo in osinfo_os_get_all_devices API doc 2013-03-22 Christophe Fergeau <cfergeau@redhat.com> Add openSUSE data ISO information by courtesy of Frédéric Crozat <fcrozat@suse.com> 2013-03-18 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.6 2013-03-15 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> installer: API to query device driver signing requirement Some OS vendors recommend or require device drivers to be signed by them before these device drivers could be installed on their OS. This recommendation/requirement then also applies to the installer scripts of that particular OS. This API enables apps to query such requirement/recommendation, if any. install-config,winxp,win7: API to enable/disable driver signing While I thought that I had solved the problem of Windows requiring signed device drivers and QXL driver being unsigned, I could't be more wrong: * The registry key magic I used for disabling driver signature checks on XP seems to be far from reliable. I tested it many many times but on a weird broken version of XP home edition that I can't seem to have access to anymore. I now tested against both home and professional editions both with and without this registry key magic and I observed the same result in both cases: Drivers do get installed but they remain unused by the OS after installation. The only reliable way of effectively disabling signture checks during installation is through the 'DriverSigningPolicy' option in .sif file, which means disabling signature checks permanently. * On Windows 7, disabling integrity checks and test signing after drivers' installation disables the already installed drivers too if they are not signed. * The reason I thought QXL was functional at first was that automatic resolution setting was working. Turns out that unlike on Linux, on windows automatic resolution setting only requires spice-vdagent where as QXL is only required for arbitrary resolutions. So to make QXL working out of the box, I'm afraid we don't have any choice but to disable driver signature checks permanently. Since signature checks is a security measure from vendors, we need to leave it to applications to decide whether they want to do this or not. API to query signed status of device drivers Some OS vendors recommend or require device drivers to be signed by them before these device drivers could be installed on their OS. An API to query signed status of the device driver will be useful for apps to be able to make a decision whether they want to use the driver or not. Later patches add API for querying signature requirements from a script and to possibly disable these checks. 2013-03-14 Christophe Fergeau <cfergeau@redhat.com> Don't ignore vendor/device name in pci/usb id parser These values were parsed but ignored using ignore_value(), leading to very incomplete device information when using pci.ids/usb.ids. As generally this data comes from <device> nodes in libosinfo database, which takes precedence over the pci.ids/usb.ids data, this bug is not visible without removing the additional <device> data from the osinfo database. 2013-03-12 Christophe Fergeau <cfergeau@redhat.com> Remove FSF address from source file headers The FSF moved a while ago which made the address we use in every source file header invalid. Follow the recommendation from http://www.gnu.org/licenses/gpl-howto.html and don't put any address in these headers, just a link to the fsf website. 2013-03-06 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Post-release version bump 2013-03-05 Christophe Fergeau <cfergeau@redhat.com> spec: Don't require udev on newer distros 96c66b8 disabled building of the udev rule on newer Fedoras, however the .spec still has a Requires: udev, which is not relevant when the udev rule is not built. Allow to use system pci.ids/usb.ids files Most distros already ship copies of pci.ids/usb.ids. This commit allows to make use of these rather than the ones shipped with libosinfo. This is achieved through the use of --with-usb-ids-path and --with-pci-ids-path configure flags. 2013-03-05 Cole Robinson <crobinso@redhat.com> spec: Add explicit dep on pod2man pod2man used to be in the default Fedora build root, but that's changed in current rawhide. The containing package is perl-podlators, but since that package doesn't exist on F18 and earlier, it's easiest just to depend on the needed binary. 2013-03-04 Zeeshan Ali (Khattak) <zeeshanak@gnome.org> Release 0.2.5
| ver. 1.6 |
Github
|
.
| PHP 8.2.30 | ??????????? ?????????: 0.03 |
proxy
|
phpinfo
|
???????????