[PATCH v3] Re: avoid useless if-before-free tests

[Jim Meyering]

Michael Matz wrote:
> On Mon, 18 Apr 2011, Janne Blomqvist wrote:
>
>> (Why we, in the age of non-sucky version control, persist in keeping
>> manual changelog files is beyond me..)
>
> The one single reason why I'm very happy about our ChangeLog file policy
> is that I can grep it easily.  Listing the changelog of the whole tree via
> svn (or anything else) up to some years ago is very slow and generates
> quite some load on the server.  Sure there are various work-arounds for
> that, but I'm happy to not have to resolve to such work-arounds.

I doubt anyone would be happy to get rid of ChangeLog files altogether.
In coreutils and other projects, we have a compromise.
There is no VC'd ChangeLog file, yet every release tarball includes
the expected (and up to date) ChangeLog file.  How?

It generates that ChangeLog file from the git commit log.
Currently it does it only when creating a distribution tarball
but it'd be easy to add a target to recreate all (or a subset of)
generated ChangeLog files

Doing something similar for gcc would be trivial
if you'd be ok to switch to a single top-level ChangeLog file,
but I doubt folks would go for that.
However, it wouldn't be much more work to automatically
generated ChangeLog entries for all 44 of the existing files.

... in case anyone ever wants to go that route.