Conversation
strtgbb
force-pushed
the
pipeline-cleanup-25.6
branch
2 times, most recently
from
6443962 to
3578694
Compare
|
Why are all these sanitizers tests are run (and failed) despite marked as excluded in PR description? |
|
Upstream changed the way tests are excluded, the old system no longer works, although I did find a hack for regression. I still need to investigate how upstream is excluding tests now. |
strtgbb
force-pushed
the
pipeline-cleanup-25.6
branch
4 times, most recently
from
e00f7dd to
f0436de
Compare
strtgbb
force-pushed
the
pipeline-cleanup-25.6
branch
from
f0436de to
de3e358
Compare
…l memory issue crashing tests
strtgbb
force-pushed
the
pipeline-cleanup-25.6
branch
from
7079c5b to
e0b3a9a
Compare
arthurpassos
left a comment
arthurpassos
left a comment
There was a problem hiding this comment.
It's 2~3 years since I dealt with CICD code, my understanding is kind of limited. Sorry for the superficial review.
| python3 -m praktika run 'Dockers Build (multiplatform manifest)' --workflow "MasterCI" --ci |& tee ./ci/tmp/job.log | ||
| fi | ||
|
|
||
| fast_test: |
There was a problem hiding this comment.
Why is this being removed?
There was a problem hiding this comment.
It was added in a previous PR for debugging purposes. The issues seem to be the same as with stateless, so there is no need to keep it here.
| - name: Install awscli | ||
| shell: bash | ||
| run: | | ||
| .github/retry.sh 3 10 sudo apt-get update |
There was a problem hiding this comment.
Why remove the update?
There was a problem hiding this comment.
It triggered an auto-update, which occasionally caused the setup step to fail. I have addressed this on the runner side, so this change could be reverted.
| python3 -m praktika run 'AST fuzzer (amd_msan)' --workflow "PR" --ci |& tee ./ci/tmp/job.log | ||
| fi | ||
|
|
||
| upgrade_check_amd_debug: |
There was a problem hiding this comment.
Why is this being removed?
There was a problem hiding this comment.
It is not configured correctly for our repo. I can be re-enabled and configured in a seperate PR.
| if command -v ts &> /dev/null; then | ||
| python3 -m praktika run 'BuzzHouse (amd_debug)' --workflow "PR" --ci |& ts '[%Y-%m-%d %H:%M:%S]' | tee ./ci/tmp/job.log | ||
| else | ||
| python3 -m praktika run 'BuzzHouse (amd_debug)' --workflow "PR" --ci |& tee ./ci/tmp/job.log |
There was a problem hiding this comment.
Why remove the else? As I understand, the if statement checks if ts is available and does something else if not.
Do I get it right that if ts isn't available you don't want to do anything?
There was a problem hiding this comment.
This file is auto-generated, and the github diff is misleading.
| python3 -m praktika run 'Build (arm_release)' --workflow "ReleaseBranchCI" --ci |& tee ./ci/tmp/job.log | ||
| fi | ||
|
|
||
| build_arm_asan: |
There was a problem hiding this comment.
So, we are removing some builds with sanitizers? Why?
There was a problem hiding this comment.
We only use x86 builders and the ARM ASAN build is very difficult to cross-compile.
|
merged into |
3 participants
Changelog category (leave one):
Tasks
Exclude tests: