On Tue, Oct 25, 2011 at 11:06 AM, leif <not.rea...@online.de> wrote: > On 25 Okt., 12:04, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote: >> On 2011-10-25 09:51, Dan Drake wrote: >> > I think that would be a good idea, although if you're just running that >> > through sed, the exact patches applied would be different from the ones >> > on the trac server, which would cause some confusion. >> >> Very true, but I think that is only a minor annoyance. > > I wouldn't say it's minor, since all developers basing their work on > these patches on trac still have the problem until a ticket gets > merged. > > So checking for added trailing whitespace in patches uploaded to trac > would IMHO be a task for the patchbot.
Even better, it could be done by a trac pluging itself (at the moment when it matters most), not the patchbot, which may or may not run later. > > >> The point is that I don't like warnings. There used to be time when not >> having a ticket number in the commit message created a warning in my >> merger script. This meant I had to complain on a lot of tickets about >> the commit message. > > Me neither. Of course it would be better if the complaints were > generated automagically... > Reasonable commit messages (and btw. also patch filenames and > attachment comments) are valuable *before* a ticket gets merged as > well; the latter two are more or less irrelevant after a ticket has > been merged, since they don't appear in a release, or as soon as > you've imported a patch. (The filenames of course appear in your > Mercurial patch queues and shell history, so *are* pretty relevant > during development.) For me having ticket numbers (and comments) in a patch are *incredibly* important *after* a ticket has been merged. I would say they are way more important after than before merge. Typically, when I see some code that is suspect in Sage, I use "hg blame" to see what patch last modified that code, I look at the commit message of the patch to see the ticket number, then look at the relevant page on trac. >> Eventually, people were asking, "why doesn't your >> script automatically add the ticket number?". Since then, that's how >> things are done in the merger script. No warnings, just automatically >> (and silently) make the change. > > Well, as far as I know invalid commit messages (or more precisely ones > lacking a ticket number) never created merge conflicts. But they make future development work on Sage (especially for me) much harder. Automatically adding them on merge is a great solution. > Also, if a developer doesn't get warnings (and nobody else complains), > he/she's unlikely to change his/her practice, so they're of > educational value as well. ;-) > > > -leif > > -- > To post to this group, send an email to sage-devel@googlegroups.com > To unsubscribe from this group, send an email to > sage-devel+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org > -- William Stein Professor of Mathematics University of Washington http://wstein.org -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
