summaryrefslogtreecommitdiff
path: root/.gitlab-ci.yml
Commit message (Collapse)AuthorAgeFilesLines
* ci: Use detached pipelinesBastien Nocera2023-03-151-0/+5
| | | | See https://gitlab.freedesktop.org/freedesktop/freedesktop/-/issues/438#what-it-means-for-me-a-maintainer-of-a-project-part-of-gitlabfreedesktoporg
* migrate from custom itstool to builtin msgfmt for creating translated XMLEli Schwartz2021-11-211-2/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | gettext upstream has supported this for a very long time (since 0.19.7 via commit b3c2a5a242c36fbbaa0c5b17f975d6c638598a23, released in 2015), and itstool is (mostly) a legacy of the time before gettext had proper support for these sorts of use cases. This is similar to the state of intltool, which is described at https://wiki.gnome.org/MigratingFromIntltoolToGettext During the port from autotools to meson, the legacy use of itstool was faithfully translated to meson in the only way possible: by jumping through hoops to run ninja inside ninja in order to generate the .mo files for itstool, because meson's i18n module used a flawed design and there was no "real" target to create those files, only a .PHONY run_target which other rules cannot depend on. Although meson 0.60.0 added support for real targets for the built .mo files, this changed the rules for output filenames, breaking the script. But msgfmt does not care, and anyways comes with builtin meson functions for convenient use with XML files. So let's take this opportunity to drop legacy dependencies and use the modern, builtin tooling, which fixes this bug as a side effect. Fixes #170
* Use meson dist for releasingwork/dfaure/hacking_fixDavid Faure2021-04-141-0/+6
| | | | Closes https://gitlab.freedesktop.org/xdg/shared-mime-info/-/issues/152
* xdgmime: configure as submoduleVille Skyttä2020-12-311-1/+1
| | | | | | | | | | | | | So that we can bind it to a specific revision, and help with keeping clones up to date. GitLab CI instructs to use a relative path to submodules on the same GitLab server [0], but doing so causes a meson/ninja dist failure [1]. Configure using a full https URL for now, as it appears to work with GitLab just fine. [0] https://docs.gitlab.com/ce/ci/git_submodules.html [1] https://github.com/mesonbuild/meson/issues/8144
* build: Remove autotoolsBastien Nocera2020-02-201-32/+0
|
* ci: Also build with mesonFélix Piédallu2020-02-201-1/+18
|
* ci: Clean list of dependenciesBastien Nocera2020-02-201-2/+14
| | | | So that it's easier to add, modify and remove individual items.
* build: Move data files to data/Bastien Nocera2020-02-081-1/+1
|
* build: Use gettext instead of intltoolValentin David2019-09-061-1/+1
|
* ci: Generate the spec in the CIBastien Nocera2019-08-201-1/+3
| | | | | To verify that the modified spec has a valid syntax, as done in: https://gitlab.freedesktop.org/xdg/xdg-specs/blob/master/web-export/update.py
* ci: Install findutils in the CIBastien Nocera2019-08-201-1/+1
| | | | | | Otherwise "make distcheck" will throw: /bin/sh: line 9: find: command not found and not even register that something went wrong.
* ci: Test with builddir not being a child of srcdirBastien Nocera2019-01-171-7/+9
| | | | And srcdir not being a parent of builddir.
* ci: Also test srcdir == builddirBastien Nocera2019-01-171-0/+9
|
* ci: Build with srcdir != builddirBastien Nocera2018-10-141-1/+3
|
* ci: Add CIBastien Nocera2018-10-141-0/+19