Front page | perl.bootstrap |
Postings from July 2000
Re: patch trail
From: Steve Fink
July 24, 2000 15:35
Re: patch trail
Message ID: 397CC4A4.1A622A48@digital-integrity.com
Joshua N Pritikin wrote:
> On Mon, Jul 24, 2000 at 01:14:15PM -0700, firstname.lastname@example.org wrote:
> > Joshua N Pritikin <email@example.com> writes:
> > > On Mon, Jul 24, 2000 at 04:33:24AM -0700, firstname.lastname@example.org wrote:
> > >> (Well, actually they have two lists, gcc and gcc-patches. But the
> > >> second is largely due to the fact that there's a requirement that
> > >> everything checked into the repository is also sent to the mailing list
> > >> as a patch. Something that I think is *wonderful* as someone watching
> > >> the project from outside, although I realize that it's a bit hard to
> > >> get to the point where you're in the habit of doing that.)
> > > Why? Perforce can send out patches *automatically*.
> > Because if the patch is at all complicated, the human also includes an
> > explanation. There's of course already a lot of this with Perl; most
> > patches do seem to go to p5p. I'm just not sure that all of them do, and
> > I've found it interesting that it's a matter of policy in gcc development.
> Still, I think automatic patch announcements are better than "gee, I
> forgot to post an explanation" three months (or years) after the fact.
(missed the beginning of the thread)
In my company, I set up CVS to automatically mail out every patch to one
of a set of -patches lists. It gives the filename, branch, and commit
message. People now use the commit log messages as a way of explaining
the patches, even if the explanation will look odd when viewed as part
of a "cvs log" output. It is working out very well (except for in one
Java tree, where one guy commits too many tiny changes for anyone to
bother keeping up with; I disabled the -patches mail there.) I've found
it to be of enormous value as long as nobody is flooded with too many
patches; if only a handful comes by every day, you end up getting
everything at least cursorily reviewed by several people.
Presumably perforce can do the same?