New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Switch to using Codecov GitHub Action. #1158
Conversation
@Dead2 @iii-i Is it possible to install these packages on
https://stackoverflow.com/questions/5178416/libxml-install-error-using-pip |
5a2e2f7
to
e64e1f0
Compare
Fixed pigz workflow. |
I've installed |
e64e1f0
to
0ea8051
Compare
It looks like this command is still failing: |
Argh, of course - the build is running inside a Docker container. I've updated the image, please try again. If it works, feel free to cherry-pick iii-i@f26c347 into your PR. |
Cherry-picked. Thanks! |
@Dead2 what is the plan for buildkite? |
This does not seem to work right. Buildkite.. not decided yet, but was considering using it for automated benchmarking. |
af8d865
to
bfc4c69
Compare
I wonder why this PRs coverage does not appear in codecov.
I see nothing wrong here at all. Only possibility is the \n at the end of the file name, but it would be a bit strange if that was the error, no? Relevant part of the codecov yaml below; it replaces that part of the path with nothing, so that it sees files according to the path in the repo. So that should still be correct.
|
I noticed in the CI list that the 2 Codecov instances no longer show. |
@nmoinvaz That is a problem, but unrelated to this PR. Codecov claims it was because of the buildkite build was failing and suggested config changes. I have applied the changes, but not sure whether it will work for old PRs. |
bfc4c69
to
baa5834
Compare
Still nothing. Just to exclude some of the recent codecov problems, could you close this PR and make a new one? |
It does report properly in my branch. |
Then perhaps it is because it is a foreign repo and thus treated differently (hence our token workaround) |
Yes, it is probably due to the token not being exposed to my PR. You could create a branch in zlib-ng and cherry-pick my changes and test it there? |
Sure, but we would still need to fix the workaround, so that we get coverage data for PRs like we do today. |
It keeps saying there was an error in processing.. what does your codecov.yaml look like?
|
We really should add it as a file though. |
Do you want me to add it as part of this PR? |
Sure, if you want to. |
I can't figure out why the logs are showing |
Actually it looks like it is the |
@thomasrockhu any idea why this PR would not be showing up on the code coverage website? The reports generate properly and upload to codecov.io but fail to show on the website. They do however show for my repository, but not in the PR commit. My PR for zlib-ng repo: We are trying to move away from python uploader and to use the codecov GH action. |
df21d48
to
a72caad
Compare
Codecov Report
@@ Coverage Diff @@
## develop #1158 +/- ##
===========================================
+ Coverage 81.99% 87.08% +5.08%
===========================================
Files 97 96 -1
Lines 9080 9034 -46
Branches 1440 2234 +794
===========================================
+ Hits 7445 7867 +422
+ Misses 1068 1050 -18
+ Partials 567 117 -450
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
a72caad
to
a576d80
Compare
Thank you @thomasrockhu-codecov that seems to have fixed the issue. Not sure why though because I did check the timestamp being put in the XML and it looked right. Anyways, by turning that check off at least now there is one less point of failure. I ended up using |
@Dead2 I am seeing -5% coverage. I think these results are accurate. It could be because the benchmarks are not being run on every platform. I think after incorporating gtest (which has all the c++ tooling) we can turn on the benchmarks for more platforms. |
2755d3f
to
84bcbbb
Compare
Rebased. |
I really don't see any reason why this PR should change the coverage at all (other than the usual noise), this commit does not really make any changes to build, code or testing compared to the develop commit it is based on, so coverage should stay the same. -5% is a pretty big change. Interestingly, most of the lost coverage is related to assert() or failure modes. But I'll reluctantly approve this for merge. |
@Dead2, I think it might have something to do with |
84bcbbb
to
6a4d99f
Compare
@Dead2 looks like that gave us +5% in coverage. |
@nmoinvaz I see, looking at the docs it seems there is a mismatch with how the statistics count different parts of the code. And So, I think this new percentage is actually more useful to us than the old one was. 👍 |
This uses codecov's GitHub Actions action.
https://github.com/codecov/codecov-action
Which also uses their new uploader.