Issue with GitHub integration

[chgad]

• safety version: GitHub integration
• Python version: 2.7.x
• Operating System:

Description

We have an status integrated for varryfying each PR going into the master branch.

I now encounter that this status is Pending for over 12 hours.

Is there currently a server issue or am I doing something wrong with the integration?

[jayfk]

There are currently no (known) global server issues. The most reasonable explanation is that one of the worker processes wasn't able to correctly set the commit status.

Is this on a private repo? If so, could you sent me the repo name and the commit sha via mail to support@pyup.io?

If not, please send me a link to the PR on GitHub and I'll have a look at it.

[chgad]

Well,
I "reencounter" this Problem, but i found smth which seeems like a bug to me.
So the pyup/saftey Integration works fine when using feature branches and Pr them into our develop branch. Now when Pring into master it stays on "waiting for satus report".

But if i do a normal commit on the develop branch (which i'd like to prevent in the future) and push this commit, the status is reported immediately.
This really looks like a Bug. @jayfk reopen this Issue ?

[jayfk]

Could you check if there are any webhooks not being sent to pyup.io and/or are throwing errors?

If you go into your repos settings, you should see a webhook pointing to pyup.io. A click on edit should show you all calls to the api (you need to scroll down a bit).

[chgad]

Strangely,

it seems that all Request were succesfull even if i encountered "perma-pending" statuses.

I'll track this for now.

[chgad]

Update:

I first thought it might be an issue of "not calling the hook when creating the PR". But i doubled checked it and yes, I'm calling the hook when a PR is opened etc. . And the Response looks fine to me.

But still, pending.

[katherine-black]

Running into something similar, if I rebase my branch and force push it, the commit status will hang in pending, but if I make a new commit and push it everything will run as expected.

To add to this, it seems like safety-ci is only looking at the first commit of multiple that are part of that webhook call, in my case the very first commit actually failed safety-ci so I'm not sure if that would cause an issue.

Let me know if I can provide any more info.

[jayfk]

Running into something similar, if I rebase my branch and force push it, the commit status will hang in pending, but if I make a new commit and push it everything will run as expected.

Thanks, that helps a lot. I'm struggling to reproduce this on a minimal test repo. Maybe that will help.

I'll keep you posted.

[jayfk]

I've just deployed some changes that hopefully fix this issue.

@chgad @ilyavw could you try reproducing the bug on one of your repos? If it still doesn't work, could you give me the exact steps that you were doing?

- git branch foo
- git checkout foo
- touch bar.txt
- git commit -a -m 'baz'
- git push origin foo
- [created PR on GitHub through UI]
- ...
- ...

[chgad]

I'm currently on vacation. I'll try it as soon as I'm back.

[katherine-black]

Unfortunately I was able to still reproduce this. Here are my steps:

- git checkout -b test-branch
- vim README.md
- git commit -am "test commit 1"
- vim README.md
- git commit -am "test commit 2"
- vim tox.ini
- git commit -am "test commit 3"
- git push
-- PR Created in github UI
-- At this point the 3 commits are pushed together and safety ci still works, as the job only runs on the last commit

- vim tox.ini
- git commit -am "test commit 3" --amend
- git push --force
-- Still working

- vim tox.ini
- git commit -am "test commit 4"
- git push
-- Added a second "buildable" commit, still working

- vim tox.ini
- git commit -am "test commit 4" --amend
- git push -f
-- At this point safety ci hangs

[jayfk]

Thanks! That should have worked, tough. (It does locally)

I believe the Server Error is unrelated. Prior to you posting this, I've received a error message from Sentry telling me it failed miserably to parse a tox.ini. That's probably where the Server Error comes from.

If you have the time, could you try that again without messing with your tox.ini?

Thanks a lot for your time helping me debugging this.

[katherine-black]

Ah, that was it! I messed up the syntax in the tox file and that's what caused the failure. I can confirm now that this issue appears to be resolved. Thank you!

[rafaelpivato]

Closing this as it is apparently fixed given the conversation above. Let me know otherwise, please.