Re: Need help with GitHub Issues migration
Oswald Buddenhagen via mc-devel <mc-devel-+hD5IHI5XseWCegYutOAJTiNl0CLU6MPYPYVAmT7z5s@public.gmane.org> Tue, 25 Feb 2025 19:49:13 +0100
| Newsgroups | gmane.comp.gnome.apps.mc.devel |
|---|---|
| Message-ID | <Z74QqQHN2LYwzAfF@ugly> |
On Tue, Feb 25, 2025 at 07:03:10PM +0100, Yury V. Zaytsev wrote: > On 25. Feb 2025, at 16:26, Oswald Buddenhagen wrote: >> is it possible to convert the PR template to a form as well? >> it's _remarkable_ how many people don't delete the instruction text ... > >Unfortunately, no, that's not possible. > drat >However, I have added an HTML comment around the instructions. I hope >that helps. > hmm, the preview renders it without the comments. but i've seen that in several projects, and the plain-text mail notifications arrive with the comments included, which is mightily annoying. so unless these were user errors on the side of the contributors, it might be better to put banners at the top and bottom that tell that the instructions are to be deleted by the contributor after being followed. well, banners in comments, maybe, for an extra layer of defense. >I wonder how you identified them? I did a SQL select for duplicate >summaries, so I missed this one. > i did the silly thing and actually scrolled through the 35 pages of tickets. i started doing that because i didn't remember at what point the break in the numbering was, but once i found some dupes i thought i'd just go on. >Did you mean that instead of copying the timeline trail, I could >reconstruct PRs from merged branches for issues that lead to merges? > yes. well, as indicated, i'm not sure what exactly would happen to the trails. the main idea would be to have a linked source branch and have the activity on it represented in the usual gh way. that seems to fundamentally match what trac does, though i never really understood the details of the latter, as i didn't contribute that way. >This is definitely less hopeless than assessing tickets with patches. >Unfortunately, the data is still far too messy to make it a largely >automated process. I also don't see much benefit in doing so... > yeah, it would just look good (a very high-level representation of coding activity), and might be nicer for some "forensics", but other than that it doesn't seem very valuable. if it was a low-hanging fruit, it might be still worthwhile. -- mc-devel mailing list mc-devel-+hD5IHI5XseWCegYutOAJTiNl0CLU6MPYPYVAmT7z5s@public.gmane.org https://lists.midnight-commander.org/mailman/listinfo/mc-devel