summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Release version 0.3.5appstream_glib_0_3_5Richard Hughes2015-03-112-1/+23
|
* trivial: Fix a potential crash spotted by clangRichard Hughes2015-03-111-0/+6
|
* trivial: Remove some dead code spotted by clangRichard Hughes2015-03-112-4/+0
|
* Add as_store_add_filter()Richard Hughes2015-03-112-0/+58
|
* trivial: Firmware files do not require screenshotsRichard Hughes2015-03-101-2/+3
|
* Support parsing .inf files into an AsAppRichard Hughes2015-03-098-113/+270
|
* trivial: Add as_app_get_release_default()Richard Hughes2015-03-092-0/+29
|
* Support Dirids in .inf filesRichard Hughes2015-03-092-0/+29
| | | | | A Dirid is a number that maps to a location in Windows, e.g. '%12%' corresponds to '%SystemRoot%\system32\drivers'
* trivial: Split off the desktop file parsing to it's own source fileRichard Hughes2015-03-094-359/+390
| | | | No logic changes other than the moved code.
* Make the .inf parser compatible with more filesRichard Hughes2015-03-089-233/+1283
|
* Relax validation checks to allow empty release sectionsRichard Hughes2015-03-062-5/+4
| | | | | | | | If an project has more than one application, and the NEWS data is encoded in the AppData file then it's entirely possible and valid that one application has no release notes at all. Resolves: https://bugzilla.gnome.org/show_bug.cgi?id=745712
* trivial: Make build more portableRichard Hughes2015-03-061-3/+3
| | | | Resolves https://github.com/hughsie/appstream-glib/issues/34
* Parse the entire description when reading AppStream releasesRichard Hughes2015-03-053-5/+17
|
* Add as_utils_parse_driver_version()Richard Hughes2015-03-054-49/+98
| | | | This allows us to use it from fwupd and GNOME Software.
* trivial: Fix a crash when subsuming an application with itselfRichard Hughes2015-03-051-0/+2
|
* trivial: Don't accept negative version componentsRichard Hughes2015-03-042-4/+9
| | | | In this case, fall back to using the timestamp.
* Sort the releases using vercmp rather than strcmpRichard Hughes2015-03-044-2/+106
|
* Do not disable packages without an architectureRichard Hughes2015-03-042-11/+41
| | | | Also add to the self tests to catch this in the future.
* trivial: Fix up two .inf parsing issuesRichard Hughes2015-03-042-1/+8
|
* trivial: Set the timestamp when subsuming if unsetRichard Hughes2015-03-041-2/+5
|
* trivial: Fix the cairo-dock blacklistRichard Hughes2015-03-041-1/+1
|
* trivial: Fix 'make check'Richard Hughes2015-03-042-6/+9
| | | | | The ctimes are stored in the .cab file, so we need to set these to something other than the git checkout date.
* Create firmware AppStream components from .cab filesRichard Hughes2015-03-0422-11/+889
| | | | | See http://blogs.gnome.org/hughsie/2015/03/03/updating-firmware-on-linux/ for more infomation on how this will be used.
* trivial: Correctly subsume AsRelease objectsRichard Hughes2015-03-041-7/+34
|
* Add support for <location> and <checksum>Richard Hughes2015-03-046-0/+211
| | | | This is needed for AppStream 0.9 support.
* trivial: Support the ID kind of 'firmware'Richard Hughes2015-03-044-0/+14
| | | | This is needed for AppStream 0.9 support.
* Sort releases in newest-to-oldest orderRichard Hughes2015-03-041-0/+13
|
* trivial: Find built test files when running make distcheckRichard Hughes2015-03-042-2/+8
|
* Add as_utils_load_inf_file()Richard Hughes2015-03-035-0/+192
| | | | This is for future use when loading firmware files.
* trivial: Fix a tiny comment header problemRichard Hughes2015-03-021-1/+1
|
* trivial: If guessing the NEVRA failed, fall back to opening the packageRichard Hughes2015-03-021-0/+7
|
* Add releases from the upstream AppData filesRichard Hughes2015-03-022-0/+28
|
* Add as_app_get_release()Richard Hughes2015-03-022-0/+28
|
* Support multiple source files for appdata-to-newsRichard Hughes2015-02-131-54/+62
| | | | | | At the moment this just duplicates the sections for each source file, but in the future we want to support merging these into one list after deduplicating the list elements.
* Merge pull request #32 from darkxst/masterRichard Hughes2015-02-111-0/+1
|\ | | | | tests: Fix DSO Linking error when built with --as-needed
| * tests: Fix DSO Linking error when built with --as-neededTim Lunn2015-02-111-0/+1
|/
* Improve the news-to-appdata section detectionRichard Hughes2015-02-101-2/+12
|
* Speed up adding extra packagesRichard Hughes2015-02-041-4/+6
| | | | Do not load the filelist if we're just checking for a common srpm name
* Do not run with a helper thread if --max-threads=1 is usedRichard Hughes2015-02-041-3/+8
|
* Fix a crash when more than one thread tries to ensure the same packageRichard Hughes2015-02-042-4/+41
| | | | | | This can happen when a package is itself an extra package of another, at the same time being a valid application itself. This is probably a packaging bug, but crashing isn't nice either.
* Allow short descriptions when validating <description> in releasesRichard Hughes2015-02-031-2/+9
|
* Output valid AppData markup when using news-to-appdataRichard Hughes2015-02-031-3/+11
|
* Fix a builder crash where a package depends on itselfRichard Hughes2015-02-021-0/+2
|
* Guess the NEVRA from the filename if possibleRichard Hughes2015-02-022-0/+65
| | | | This avoids opening every package at startup just to check the architecture.
* trivial: Add asb_package_get_version()Richard Hughes2015-02-022-0/+18
|
* Only open the package when requiredRichard Hughes2015-02-023-6/+22
| | | | This reduces the peak RSS by about 15% typically.
* Also clear the filelist and deplist of the extra packageRichard Hughes2015-02-021-0/+7
| | | | This typically halves the amount of RSS when the context is unloaded.
* Clear the dep list and the file list if we've finished with the packageRichard Hughes2015-02-023-0/+26
| | | | This is a simple way to further decrease the peak RSS value by a few hundred MB.
* Close the packages after they have been processedRichard Hughes2015-02-024-2/+63
| | | | | | | | | This means we have to re-open the files for the few packages used as extra-packages to other packages, but means we keep the RSS value to a sane value for a large number of packages. We can't go above 4GB RSS for Fedora... This also means we spend less time 'shutting down' as the packages have typically been released as part of the task, rather when the context exits.
* trivial: Do not allocate 1k of log space for each package at startupRichard Hughes2015-02-021-1/+1
|