On Wed, 2007-08-22 at 13:16 -0400, Jesse Keating wrote: > On Wed, 22 Aug 2007 12:24:06 -0400 > "Colin Walters" <walters@xxxxxxxxxx> wrote: > > > Another way to ensure the changelog is consistent is to stop using > > it. Move them out of the spec for historical archives, and use the > > version control system. Not writing changes twice = win. > > Especially when most of them are on the level of "new upstream > > version" which are equivalent to comments in C like: > > /* Increment x */ > > x = x+1; > > That's great, when the SCM is easy to access and available in moments > of need. But it's not really easy to do rpm -q --changelog <foo> and > have that search whatever SCM the package came from (if it came from > one) and gather the changelogs, and order them in some way that makes > sense for the package you have currently installed. One could argue that the ChangeLog should be written once (say into the spec file) and that copy should be used as the entry in the SCM or vice-versa. jeff -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
