Re: On autogenerated ChangeLog

[Alexander Larsson <alexl redhat com>]

On Mon, 2009-04-20 at 15:37 -0400, Germán Póo-Caamaño wrote:
> > 
> > But that's just a gut feeling and maybe it's wrong. The point is,
> > ChangeLogs were invented back when RCS-files-on-an-NFS-server was the
> > pinnacle of version control technology, and maybe what was most useful
> > then isn't what's most useful now.
> 
> Sysadmins and they take the decision if it worth an update/upgrade or
> when they should do it.
> 
> If we ask people to read ChangeLogs from git, I wonder why do we
> bother in releasing tarballs when they could them from the repository
> also? (in the sense you can get a tar.gz/tar.bz2, they are tagged,
> etc.)

This is a totally bogus comparison in so many way. First of all, not
everyone has the setup necessary to dist from the repository, secondly a
release that you can put on a ftp site, CD, in an rpm, or whatever is
clearly necessary, whereas its not at all obvious exactly how and if a
ChangeLog in the tarball is used.

For instance, if a sysadmin wants to determine wheter to update to the
new release of gtk+. Starting from zero data, how does he start? Does he
download the tarball and extract the ChangeLog? Possibly so, but it
seems vastly easier to look at e.g the release mail, the NEWS file on
the ftp site, or even the gnome cgit instance.