ci: avoid perforce/brew issues affecting macOS
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v3 Latest4 patches 5 repliesIn review
- v22 patches 5 repliesSuperseded
Patches in v3 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/4 | ci: make failure to find perforce more user friendlyci/ | +6 | 1 reply | |
| 2/4 | ci: avoid brew for installing perforceci/ | +8 | 3 replies | |
| 3/4 | ci: reintroduce prevention from perforce being quarantined in macOSci/ | +2 | 1 reply | |
| 4/4 | CI: use https, not http to download binaries from perforce.comci/ | +1 | 0 replies |
Reviews trailers per patch
None of the 4 patches has a Reviewed-by or Acked-by.
| Patch | Eric Sunshine |
|---|---|
| Series | Helped-by |
| 1/4 | Helped-by |
| 2/4 | |
| 3/4 | |
| 4/4 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH v3 0/4] ci: avoid perforce/brew issues affecting macOS
This series is an alternative to ab/ci-osx-loosen-package-requirement and includes the parts proposed on it in v2 that didn't conflict.
The first patch is just a slightly modified version from the one proposed by Junio and that adds also a message if git-lfs wasn't found just like we do now for perforce while making those messages go to stderr.
The second patch replaces the original proposal with a more complicated one that removes the use of brew as suggested by Ævar, it is also a little bigger than it really needs to be just to make sure it won't conflict with on the fly changes, and for the same reason uses some hardcoded values that would be worth fixing after.
The third patch is only needed to avoid regressions with Azure Pipelines which might be deprecated as proposed in other on the fly changes, so it might not be needed in seen, but is still needed if this progresses further first. It was made independent so it could be easy to revert if Azure support is going away just like TravisCI did.
The last patch was taken from Ævar's v2 and is the only one that will require changes in other core on the fly but the conflict resolution is trivial, and more importantly doesn't conflict when merged to seen in the current spot this branch uses.
For an example of a "successful" run with it merged on top of seen (with the previous iteration reverted) see:
https://github.com/carenas/git/runs/6140236148
Note that perforce no longer fails because someone fixed the issue with their Cask but still most of the new code is excercised correctly.
This introduces a problem identified by Ævar of making the CI runs less consistent, as they might be skipping some of the perforce integrations if installing that fails again, but the likelyhood of that failing has been reduced by removing the inherent brew flakyness, and so improving the UI so that it will be more visible when that happens is probably something that can be addressed later, once all other CI related changes are settled, or we could reintroduce the hard dependency on installing perforce for macOS (just like it is done in Linux)
Carlo Marcelo Arenas Belón (3):
ci: make failure to find perforce more user friendly
ci: avoid brew for installing perforce
ci: reintroduce prevention from perforce being quarantined in macOS
Ævar Arnfjörð Bjarmason (1):
CI: use https, not http to download binaries from perforce.com
ci/install-dependencies.sh | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-)
Signature · 1 line
2.36.0.266.g59f845bde02