- 10 Sep, 2021 3 commits
-
-
They are only used by the gst-rs repos which aren't using this template anyway, so we don't need to have the jobs for the rest of the pipelines. Part-of: <!418>
-
Part-of: <!418>
-
Jordan Petridіs authored
since !414 landed Part-of: <!416>
-
- 09 Sep, 2021 1 commit
-
-
Jordan Petridіs authored
looks like it might not be needed anymore Part-of: <!414>
-
- 08 Sep, 2021 12 commits
-
-
Jordan Petridіs authored
With gitlab 14.2 now a job can depend on another from the same stage. Part-of: <!414>
-
Jordan Petridіs authored
Avoid conflicting with gst-build subprojects when copying over from the cache. Part-of: <!414>
-
Jordan Petridіs authored
This reverts commit 3cfedc8a. Part-of: <!414>
-
Jordan Petridіs authored
This reverts commit 3f17e16c. Part-of: <!414>
-
Jordan Petridіs authored
This reverts commit b81620c3. Part-of: <!414>
-
Jordan Petridіs authored
This reverts commit 9c21b05d
-
Jordan Petridіs authored
This reverts commit 5c536f26
-
Jordan Petridіs authored
This reverts commit 3d9ddab2
-
Build 2 images with different rust toolchains so we can test both the latest release, as well as the minimum supported compiler version and catch regressions. Part-of: <!405>
-
This makes it a bit harder to build an image locally with the script, as now the rust image requires to pass it the base image tag as a --build-arg. This way though we don't need to duplicate and keep the string of the image in sync. Another thing is that now the docker windows image build is properly passing/replacing build args and we can build non-master versions of the image too. Part-of: <!405>
-
Based on the build image for the existing windows jobs, add an image with rust toolchain and a (stable) gstreamer install for the bindings to build against. See gstreamer-rs!701 Part-of: <!405>
-
We depend on it for GStreamer builds and it was gettings pulled transiently by python3 install. Part-of: <!405>
-
- 07 Sep, 2021 1 commit
-
-
Matthew Waters authored
When building outside of cerbero, we only want to build the gst modules as necessary. Regression introduced by 03c5f1db Part-of: <!413>
-
- 17 Aug, 2021 1 commit
-
-
Nirbheek Chauhan authored
The default number of parallel jobs is two, which is too few. We can easily use four or more. Should speed up image builds and also downloading of (new) sources that aren't already cached in the image. Part-of: <!411>
-
- 29 Jul, 2021 1 commit
-
-
Matthew Waters authored
The relevant CI log dyld: Library not loaded: /Users/gst-ci/builds/gstreamer/cerbero/cerbero-build/dist/darwin_universal/x86_64/lib/liborc-0.4.0.dylib Referenced from: /Users/gst-ci/builds/xhaakon/gstreamer-sharp/cerbero-build/dist/darwin_universal/x86_64/bin/orcc Reason: image not found Part-of: <!410>
-
- 28 Jul, 2021 1 commit
-
-
Matthew Waters authored
Part-of: <!408>
-
- 30 Jun, 2021 1 commit
-
-
Tim-Philipp Müller authored
We need a newer cache for fontconfig since it renamed master branch to main now and the wrap file has depth=1 Part-of: <!407>
-
- 21 Jun, 2021 1 commit
-
-
Víctor Manuel Jáquez Leal authored
Add intel-mediasdk package to build gst-msdk plugin in gst-plugins-bad. Part-of: <!403>
-
- 07 Jun, 2021 1 commit
-
-
Jordan Petridіs authored
We need a newer cache for harfbuzz since it renamed master branch to main now and the wrap file has depth=1 gst-build!246 (comment 948717) Part-of: <!404>
-
- 04 Jun, 2021 6 commits
-
-
With gitlab 13.12 the behavior of jobs that depend of manual jobs changes, and they are considered skipped by default. But in our pipeline we want to enforce that the whole pipeline runs and succeds and the manual status is only there to reduce CI load. For more details see: https://about.gitlab.com/blog/2021/05/20/dag-manual-fix/ Part-of: <!404>
-
Jordan Petridіs authored
Part-of: <!406>
-
Jordan Petridіs authored
Part-of: <!406>
-
Part-of: <!406>
-
Part-of: <!406>
-
Part-of: <!406>
-
- 30 Apr, 2021 1 commit
-
-
Stéphane Cerveau authored
Enable auto features for gstreamer-full but disable "introspection" for now. Part-of: <!402>
-
- 30 Mar, 2021 2 commits
-
-
Stéphane Cerveau authored
Part-of: <!400>
-
Stéphane Cerveau authored
add qt5 devel package to build qt plugin Part-of: <!400>
-
- 05 Mar, 2021 2 commits
-
-
Jordan Petridіs authored
Current windows image doesn't seem to be able to spawn procceses for reason, use a different image until this is debuged ``` Running command 'c:/builds/alatiera/gst-ci/cerbero-build/build-tools/bin/meson --prefix=c:/builds/alatiera/gst-ci/cerbero-build/dist/msvc_x86_64 --libdir=lib -Ddebug=true --default-library=both -Doptimization=2 --backend=ninja --wrap-mode=nodownload -Db_vscrt=md --native-file c:/builds/alatiera/gst-ci/cerbero-build/sources/msvc_x86_64/gstreamer-1.0-1.19.0.1/_builddir/meson-native-file.txt -Dlibunwind=disabled -Ddbghelp=enabled -Dintrospection=disabled -Dexamples=disabled' failed to create process ``` Part-of: <!399>
-
Jordan Petridіs authored
Part-of: <!399>
-
- 25 Feb, 2021 3 commits
-
-
Jordan Petridіs authored
Build the images in the ci runners like we do, for the linux containers. Kudos to mesa for the powershell script that reimplements the ci-templates behavior. mesa/mesa!4304 Part-of: <!392>
-
Jordan Petridіs authored
Check the exit codes of various commands and fail appropriatly upon error. Part-of: <!392>
-
Jordan Petridіs authored
Similar to !375 Part-of: <!382>
-
- 18 Feb, 2021 1 commit
-
-
Nirbheek Chauhan authored
`test -d "${CERBERO_HOME}/dist/${ARCH}"` will cause the build to error out if the dist dir doesn't exist, which will happen if there's no cache, for example: https://gitlab.freedesktop.org/seungha.yang/gst-plugins-bad/-/jobs/7405565 What we actually want to do, is to not perform the rsync if there's no cache, since it means we want to rebuild from scratch; for instance, if the current MR uses a custom cerbero branch of the same name which won't have a prebuilt cache. Part-of: <!398>
-
- 16 Feb, 2021 1 commit
-
-
This fixes the PGP key installation issue. Part-of: <!396>
-
- 03 Feb, 2021 1 commit
-
-
Nirbheek Chauhan authored
It's purely for informative reasons. `du` will fail on the sources dir if a branch name has unicode in it due to an MSYS/MinGW bug. The long term fix is to from MSYS/MinGW to MSYS/MinGW-W64 or MSYS2/MinGW-W64. Part-of: <!395>
-