On Mar 28, 2011, at 11:21 AM, Darin Adler wrote: > On Mar 28, 2011, at 10:44 AM, Jeremy Orlow wrote: > >> If the issue is simply one of overhead, then we should allow committers to >> omit change logs when they're not necessary as well. > > Sure I am open to discussion about that. I think that some check-ins, > especially LayoutTest ones, don’t need change log entries. > > What we need is a rationale for when change log entries are needed, and when > that rationale was not met, the check-in could happen without one.
Two benefits of ChangeLog entries apply to almost any kind of change: 1) They strongly encourage you to write a reasonably good commit message. (It might be possible to achieve that without ChangeLogs, but I have never seen a project that fails to use ChangeLogs and yet has comparably detailed commit messages.) 2) They let you see the change history of a file offline, even if you're not using a version control system that stores every revision locally. I think #1 is the most important, even if it's an indirect effect. It certainly applies to test cases. It might not apply to regression test meta-files, but I think it's better to have no exceptions than a very narrow exception. Regards, Maciej _______________________________________________ webkit-dev mailing list webkit-dev@lists.webkit.org http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
