[BUG] Issue to conversion workflow starts but does not finish.

[Remi-Gau]

See the report here.

https://github.com/brainhackorg/global2021/runs/4178661874?check_suite_focus=true

@jhlegarreta pinging you but I can try to start debugging a bit later.

[jhlegarreta]

A few notes:

• I had already observed this. Cross-referencing https://github.com/brainhackorg/global2021/pull/72#issuecomment-963754572.

• As mentioned in the cross-referenced comment, when one tries the script (as provided by the refactoring in PR #72) with the following changes (I had to test on the 2020 repository since there were no projects when I made the changes):

  • Make the URL point to the 2020 repository.
  • Change the project repository regex in https://github.com/brainhackorg/global2021/blob/6106d92623711d22b6818e5d67f63b74cf15b9f2/scripts/issues_to_pages.py#L204 to Link to the project to account for the change in the section name of the template
  • Changes the filter in here https://github.com/brainhackorg/global2021/blob/6106d92623711d22b6818e5d67f63b74cf15b9f2/scripts/issues_to_pages.py#L251 to status:published, and moves the proj_url_re_match.group() block in https://github.com/brainhackorg/global2021/blob/d46594d9d7c62809b07a674d5794ef81b4dfef8e/scripts/issues_to_pages.py#L207 inside the if to avoid accessing the group member in case of Nones,

  the script does its job. I know, too many things to make it work with the 2020 repository, but it works. So I thought there was something else that made the script hang.

• I did not have a clear idea of what was going on, and assumed the changes in the mentioned PR were not introducing that bug, since I was able to run it successfully with the changes above.

Some testing shows that the script gets stuck precisely in https://github.com/brainhackorg/global2021/blob/6106d92623711d22b6818e5d67f63b74cf15b9f2/scripts/issues_to_pages.py#L204, the issue["body"] contains the issue body as expected, so the regex is having trouble to do its job. Thus,

• Given that no project was present prior to the changes in PR #72, the bug was already there before them.
• I had tested the regex when doing PR #55 (cross-referencing https://github.com/brainhackorg/global2021/pull/32#issuecomment-932805057), but maybe when testing on the issues in your fork the body of the issue changed slightly to just make the regex work.

As a side note,

• We may need to add https to the url matches in here: https://github.com/brainhackorg/global2021/blob/6106d92623711d22b6818e5d67f63b74cf15b9f2/scripts/issues_to_pages.py#L209

[Remi-Gau]

Thanks for the detailed explanation @jhlegarreta.

BTW: I did not mean to imply you had overlooked something.