Changelogs are *not* a lot of work: they *save* a lot of work! Just my personal opinion, of course, but this is how *I* work: 1. make changes 2. diff with current CVS > patch 3. read patch and write the changelog for the patch 4. attach patch to the changelog and save it in my personal patch archive 5. check in changes and use changelog as CVS log message
This way: 1. I have a changelog 2. I get to see my changes again, and fix bugs before I check them in (you'd be surprised how many bugs you'll kill in their infancy like this) I've taken up this way of working after spending much too much time hunting down bugs that should never have been checked in in the first place. It works! I now have time for coffee again! I also run the testsuite of the project each time before I check in major changes or before I set a release tag, to make sure there are no regressions; and I make a new test for every bug I find to make sure it won't be revived. It may seem like extra work, but it really isn't! rlc On Wed, 18 Jun 2003 [EMAIL PROTECTED] wrote: > We have a software development lab with about 10 people submitting > changes to CVS sometimes almost daily. With so many little changes, our CVS log > entries are not very verbose. > > If I did an acceptable ChangeLog, then that would mean I would > need an entry in right format for *each* of the many daily contributions. > > I want to follow correct GNU style but this ChangeLog file just seems like > too much work. Doesn't everyone agree? > > If you have any advice or ways to streamline the process I would very > much like to hear it. > > Chris > > P.S. By the way, I am aware that one can peel off the CVS log entries > to make ChangeLogs. The problem with this is that I believe that implies > I would have to have all CVS log entries in proper ChangeLog format. > and I could not just be sloppy with the zillion CVS contributions I do. > >
