Re: Ambiguous commit message in cvs-1.12

[Stefan Monnier]

>> So, which alternative do you want #a or #b?

As the PCL-CVS maintainer, #a is better, but as a CVS user, #b is better.
And #b is pretty easy to adapt to in PCL-CVS anyway.

> Definately #b.  I found the original commit output extremely
> cluttered, redundant, annoying, potentially confusing, and definitely
> unnecessary, which is why I removed it in the first place.

Looks like we agree.

> Stefan, my apologies for removing the path information by accident and
> thanks for the report.

No problem.  I'm actually surprised I only started to hear about
it recently.

> I do understand your frustration about output changing between releases,
> but you *are* referring to the feature releases.

It's annoying, but nothing to lose sleep over: fixing it generally takes
about 5min, 4 of which are taken by understanding the report from a user.

It also has the side-effect that cvs-1.12 doesn't work with the "latest"
PCL-CVS since releases of PCL-CVS are pretty rare (mostly linked to Emacs's
own releases).  But maybe this will force people to try beta-test code more
often ;-).

> If you still think this is a problem, it might be time to start
> discussing libifying a nice C API for most of the CVS functionality
> again, but I don't know anyone with the interest, necessary skills,
> and the time to do the work right now.

A C lib wouldn't be much help for PCL-CVS anyway (I wouldn't bother
porting PCL-CVS to it).

Don't take me wrong: I'm pretty happy with the feature branch, and I'm not
that bothered by message changes (as long as they're not ambiguous).  I used
a more concerned tone in my post because the same complaint was silently
ignored when sent by one of my users to the gnu.cvs.help group a couple
weeks ago.

Thank you for your work,


        Stefan