[Build-common-hackers] Bug#461513: cdbs: requires aclocal.m4 to be present in order to generate it

[Charles Plessy]

Le Mon, Mar 31, 2008 at 02:40:18PM +0200, Peter Eisentraut a écrit :
> Am Donnerstag, 31. Januar 2008 schrieb Charles Plessy:
> > > No, but you appear to expect that the feature *creates* the aclocal.m4
> > > file, but the variable name clearly indicates that it only *updates*.
> >
> > Ah, OK. Actually, I expected the command to do something similar as when
> > I type aclocal in command line. In that case, it indeed creates the file
> > if it is not there. Would you have something against having
> > DEB_AUTO_UPDATE_ACLOCAL doing this ?
> 
> What would be the use case for this?  If the upstream package doesn't have 
> aclocal.m4, why would a Debian package have one?

I admit that in my case it is a cornercase: it is an old package that
necessitates the autotools to be ran again with newer versions in order
to build. I do this with the DEB_AUTO_UPDATE* variables of CDBS.

http://svn.debian.org/wsvn/debian-med/trunk/packages/tacg/trunk/debian/rules?op=file&rev=0&sc=0

I placed a fake aclocal.m4 to force DEB_AUTO_UPDATE_ACLOCAL=1.10 to take
effect.

http://svn.debian.org/wsvn/debian-med/trunk/packages/tacg/trunk/aclocal.m4?op=file&rev=0&sc=0

All of this comes of course from the fact that the package seems
abandonned upstream. Actually, for other reasons I ended up thinking
that it is not suitable for Debian unless somebody revives it upstream.

Maybe you can contact Morten Kjeldgaard <mok at bioxray.au.dk>, who issued
the Ubuntu bug, to know about his use case? It may be more relevant.

https://bugs.launchpad.net/ubuntu/+source/cdbs/+bug/162015

Have a nice day,

(and thanks for CDBS, I like it a lot).

-- 
Charles Plessy
http://charles.plessy.org
Wakō, Saitama, Japan