Re: On autogenerated ChangeLog

[Dan Winship <danw gnome org>]

Tristan Van Berkom wrote:
> On Mon, Apr 20, 2009 at 10:58 AM, Dan Winship <danw gnome org> wrote:
> [...]
>> So, actually, what exactly IS the use case of ChangeLog if there is git
>> history on one end and NEWS on the other? Who are the people who need
>> more information than NEWS gives, but who would not want to actually
>> check out the source tree, and what information, exactly, do they need?
> 
> Generally its the tarball that is published and trusted, not the git repository.
> 
> The ChangeLog comes with the published tarball like an exported history,
> for the use of anyone who receives the tarball (the NEWS is just a quick
> resume of what happened in a release).

But that doesn't answer the question. Who are these people who read
ChangeLog, and what is it that they're doing with it, such that NEWS is
too brief, but a fully-VCS-ed source tree is unnecessary.

Eg, this subthread started when Alex suggested that we needed to put the
names of all modified files into each ChangeLog entry. It seems to me
that anyone who cares exactly which files got modified by a particular
change is going to want to see the actual diff very soon after, and so
those people are not actually part of the don't-need-a-full-checkout use
case.

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.

-- Dan