Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 02:38
    tristanlatr commented #670
  • 01:39
    altendky review_requested #20
  • 01:39
    altendky edited #20
  • 01:15
    altendky closed #18
  • 01:15
    altendky commented #18
  • 01:15
    altendky opened #20
  • 01:12

    altendky on just_drop_them

    remove CPython 3.5 and 3.6 supp… (compare)

  • 01:09

    altendky on just_drop_them

    stop testing CPython 3.5 and 3.6 (compare)

  • 01:08
    altendky opened #19
  • 01:07

    altendky on update_github_provided_actions

    update all versions of github p… (compare)

  • 00:31
    adiroiban commented #18
  • Dec 08 18:31
    kloczek commented #670
  • Dec 08 13:14
    fizyk review_requested #453
  • Dec 08 13:12
    altendky commented #18
  • Dec 08 12:47
    codecov[bot] commented #453
  • Dec 08 12:47
    codecov[bot] commented #453
  • Dec 08 12:46
    codecov[bot] commented #453
  • Dec 08 12:46
    codecov[bot] commented #453
  • Dec 08 12:43
    codecov[bot] commented #453
  • Dec 08 12:42
    fizyk synchronize #453
Adi Roiban
@adiroiban
but on 3.11 I see that SystemEventTestsBuilder_IOCPReactorTests is reported as OK - https://github.com/twisted/twisted/actions/runs/3530577155/jobs/5922726329#step:10:1933
jean-paul.
@jean-paul.:matrix.org
[m]
I think the GitHub Actions Windows environment comes with dev tools, so it is probably working because it can build the extension itself - but most real users can't figure out how to do that.
A "binary only dependencies" configuration would be good to test.
Adi Roiban
@adiroiban
I guess we only need a re-release so that the 3.11 binaries for iocpsupport are generated
and published
to pypi
Kyle Altendorf
@altendky
https://github.com/twisted/twisted/blob/49ae3c5545dccf6f442839879e6a57ab080be649/CONTRIBUTING.md where are the extensive online docs supposed to be? (link is 404)
I don't know what happend, when I did the PR to mode the documentation I remember I left redirection pages
but it looks like I forgot/missed this page. I will fix. Thanks for the report
Adi Roiban
@adiroiban
PR ready at twisted/twisted#11777 :)
jean-paul.
@jean-paul.:matrix.org
[m]
Kyle Altendorf
@altendky
as i understand it, the core need for towncrier is to avoid the constant merge conflicts that would arise from every pr directly appending to the list of changes in the changelog file. towncrier also brings some automation and structure to the headings and lists, but that seems to me to be secondary. has anyone here instead looked at merge conflict resolution by just putting 'theirs' first and 'ours' second? i would think that would be fairly doable, though there would still be the blocker that i'm unaware of any way to inject anything like that into github etc.
jean-paul.
@jean-paul.:matrix.org
[m]
I don't think that would always work because git isn't based on a sound theory of patches.
Adi Roiban
@adiroiban
towncrier also allows you to focus on the content, without worying about RST or markdown formating, and placing the notes into different categories ... and linking back to the issues tracker
jean-paul.
@jean-paul.:matrix.org
[m]
It would work as long as the text of the conflicts doesn't accidentally look too similar.
Adi Roiban
@adiroiban
Thanks Kyle for the review.
Kyle Altendorf
@altendky
@adiroiban i remember seeing a notification for something where you were commenting on PRs from other repos not getting coverage comments. i can't remember where that was though. i was going to take a look and ask around at work. if we get around to doing anything with our coverage data we may end up going the same route so i figured there might be interest in figuring it out. do you recall where you brought that up?
the token generated by GitHub Actions for forks, doesn't have issue write permissions
even though in theory the PR author can create comments
so the issues write permissions exists , and is not about elevated permissions
Kyle Altendorf
@altendky
are you saying that there is a specific 'write to the pr being built' permission? or...
Adi Roiban
@adiroiban
the PR and Issues share the same API
the main PR converstation, is just a rebranded "issue" conversation . In GitHub
Kyle Altendorf
@altendky
right
we have a workflow that checks for conflicts and applies a label. i wonder if it works on external PRs. sadly it's a bit hard to tell since it also runs on main push triggers which would have permissions.
Adi Roiban
@adiroiban
labels need admin, not only write ... AFAIK
write is for adding a comment
setting labels, milestones, asignees .. is part of the admin scope
Kyle Altendorf
@altendky
i thought there were protected labels that were admin, otherwise just anyone with write. but, been awhile
Adi Roiban
@adiroiban
For Twisted we added the PR robot to add the "needs-review" label, as non-team member don't have access to manage the labels for a PR or an issue
Kyle Altendorf
@altendky
mm, and security is maintained by the bot not being driven by in-repo code i guess?
even if triggered by the pr trigger?
Adi Roiban
@adiroiban
yes. the bot is a separate "process" and interaction is done over webhooks and GitHub API... we are not using the GitHub Actions token
is like having a separate GitHub App for that
Kyle Altendorf
@altendky
and that's not the solution for the comments as well?
Adi Roiban
@adiroiban
keeping a separate service is complicated
I think for now, the workaround is to just only publish a commit status for forks
Kyle Altendorf
@altendky
commit status as in the checks are either red or green?
Kyle Altendorf
@altendky
is the coverage status going in the workflow run summary? that's an easy place to shove stuff that's a little more accessible and can use markdown
Adi Roiban
@adiroiban
yes. the commit status is the "checks" parts of a PR
oh hey, another place github is inconsistent about semantic newline rendering.
Adi Roiban
@adiroiban
Thanks
I wasn't aware of this feature
but I think it might work to add the diff-coverage report there
and have a commit status/check with the link to the build
Kyle Altendorf
@altendky
i haven't done much with it. i feel like there's some hazard of random ordering of outputs. or maybe it tracks each write to that file separately. or... i dunno
Adi Roiban
@adiroiban
I think each write is separate, and github will combine the output at the end