On Fri, 2007-10-05 at 10:38 +0200, Frederic Crozat wrote:
> Second, being self-contained (and therefore, ease of use for external
> maintainers / lurkers / packagers / release-team). You are only thinking
> of svn log when using viewcvs. But if you are, just like me, subscribed
> to svn-commit mailing list, it is really much easier to see directly
> what it a commit purpose (fix a typo, a bug, etc) without having to rely
> on another tool / website. Same apply when using svn log command
> directly.

I second that.  I also read svn-commits-list (filtered for Evolution
packages only) because I like to follow what the other developers are
working on.  I personally prefer to see longer commit messages because
it saves me from having to open ViewCVS for each and every commit to
find out what it was about.

Also, because we keep several ChangeLogs in Evolution and E-D-S, a
complete set of changes for a particular bug is often fractured across
several different ChangeLog files.  So I try to merge them into a single
ChangeLog entry when preparing a commit message, prepending full path
names to the filenames where necessary.  I do this for the benefit of
others reading svn-commits-list, and also to help improve "code
archaeology" [1], as Federico talked about in one of the few insightful
responses I saw in that recent thread on desktop-devel-list.

Also, thanks to Federico's post, I've been trying to write ChangeLog
entries that describe *why* instead of *what*.  See his posting for an
excellent example.

Those are just my personal habits and disciplines.  I'll follow whatever
policy the team decrees.

Matthew Barnes


Evolution-hackers mailing list

Reply via email to