On 19-09-2011 20:21:51 +0300, Markos Chandras wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA512
> 
> On 09/19/11 20:09, Fabian Groffen wrote:
> > 
> > $EDITOR ChangeLog && repoman commit -m "added bugref"
> > 
> > (side-note: I don't see the point of your echangelog usage for this
> > case)
> > 
> > 
> Ok pretty much what I said in my previous email. In this case you edit
> the ChangeLog but the "added bugref" text will be appended to
> Changelog instead of fixing your previous commit message. In my
> opinion we have two options here

Why?  The message will not go to ChangeLog, simply because repoman would
not do that, because echangelog doesn't do that as well:

[nut:portage/mail-client/mutt] % cvs update
cvs update: Updating .
cvs update: Updating files
[nut:portage/mail-client/mutt] % echangelog "bla"
**
** NOTE: No non-trivial changed files found.  Normally echangelog
** should be run after all affected files have been added and/or
** modified.  Did you forget to cvs add?
**
** In strict mode, exiting
** If you know what you're doing there pass '--no-strict' to echangelog

> 1) Ignore commit messages that only affect ChangeLogs. This requires
> the ChangeLog entry to be kept on $VCS cause it is easier

we already do that

> OR
> 2) Remove the head -1 ChangeLog entry and insert the new one.

complex for no reason

> In both cases, the ChangeLogs need to be on $VCS

or put differently, it's the best we can do when it is decided that we
need to be able to do things that require a file (like currently already
is the case).

> The only advantage for having the post-commit server generating the
> ChangeLogs is the reduce $VCS size. I can't see any other advantages
> whatsoever

I think this is non-sense, and only complicates matters.


-- 
Fabian Groffen
Gentoo on a different level