-
Notifications
You must be signed in to change notification settings - Fork 8
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
Correct way to not report coverage of individual parallel jobs #1513
Comments
hi @erik-f, I see the issue and it was my mistake to offer the workaround to remove Please let me see if I can get you another solution. In the meantime, the reason you're only getting one job per build after merging into master is that, apparently only in that context, without the I'll run some tests to see if there's a workaround and get back asap. |
It is required as per the [docs][1] and was again confirmed to be necessary in a GitHub [issue][2] created by @erik-f. [1]: https://github.com/marketplace/actions/coveralls-github-action#inputs [2]: lemurheavy/coveralls-public#1513 (comment)
@erik-f I ran a number of tests and unfortunately I can't find a workaround for you right now. I submitted some suggestions to our dev team that, given their expertise might yield such an opportunity after code changes, but, for now, I'm afraid Assuming the issue you were originally trying to workaround was avoiding check failures when these settings apply:
Have you considered setting those to values that result in them practically never firing? In this case, while you'll still see all checks, the chance of one failure blocking a merge (when total coverage doesn't change) should be insignificant. You have a wide range of values to use, but I believe these give you the widest margin: On a hunch, I made a feature change request. Can you verify you'd like to see those coverage threshold checks only apply to the overall check for total coverage? (Or at least to be able to configure as such?) |
Thank you for looking into it. We don't block PRs if coverage checks are failing, so it's more of an aesthetic problem. |
Ok, thanks for confirming. FYI, I’ve submitted two feature change requests:
|
Thanks @afinetooth for your fast feedback and helpful response! It's great to see that this will be worked on, since I assume that other projects will benefit from this as well. If it is not too much to ask, can you give us a rough estimate (ballpark time frame) when you expect either of these features to become available? |
Thanks @sloede . I'm afraid I can't say. It has to go through a PM and be evaluated, scheduled, etc. As soon as I know, I'll feed back here. Feel free to check in by email to support@coveralls.io (just reference this issue). |
@sloede @afinetooth We are currently also having the exact same problem, that some parallel coverage runs turn red because they are running a different set of files than the run from the target branch. BUT the overall coverage most often increased. Just like @sloede said, this is not stopping us from merging. But as was stated as well, it is an not well situation seeing some of the action runs red 😦 |
Thanks @mobilutz. I plus-one'd both feature requests with your comments. |
@sloede @afinetooth The workaround that we have found to this has been making
It's worth to note however that we don't consider this solution ideal because it leads to bloating of github checks: |
yup this issue is a duplicate of the incorrectly closed #1505. |
@afinetooth Any chance of giving us an update on the current status ? |
Hi @jrfnl, I don't have an update besides still in backlog, as opposed to sprint log. We're an extremely small team. I understand this is affecting a decent number of folk though and since it's breaking some of your PR workflows I will do my best to raise the priority and obtain some kind of ETA. |
Much appreciated. Standing by with bated breath... |
we still have very same issue |
@afinetooth any news about the fix? |
I've tried a workaround @lekemula suggested but no luck - I see https://coveralls.io/builds/52464915
|
We don't want to make individual parallel jobs show as failed if the total coverage didn't decrease.
As advised in #1505, we removed the
flag-name
(as opposed to the docs of the action, where it's marked as mandatory for parallel runs).We changed it in this PR, and it worked as expected in this PR:
However, after merging into master, the runs only report one job each which greatly decreases our total coverage. With the exact same workflow, we got the same terrible coverage on all runs since then.
So, what is the correct way to achieve this? Is this the correct way, even though it's against the advice in the docs? If so, why doesn't it work on master for us?
The text was updated successfully, but these errors were encountered: