[Pkg-shadow-devel] build system and debian/rules discussion

[Nicolas François]

Let's feed the troll a little bit more.

On Sat, Oct 15, 2005 at 08:07:54AM +0200, bubulle at debian.org wrote:
> > While the subject originally brought up had to do with dpatch, I've in
> > fact heard some similar rants on things like cdbs, though until now I
> > hadn't really seen the whole picture on the level of difficulty that
> > using a variety of different build and patch helpers throughout the
> > archives imposes on QA, bug reporters looking over source, etc.
> 
> 
> I absolutely, completely and definitely disagree with Joey on that
> topic, espcecially when it comes at Debian patches to upstream source
> handling.

There is a least one real problem that Joey points: there is no policy for
the name of the rule that patches the source (hence the patch, patched,
please_apply_all_patches, ...).

This is an issues for some automatic packages checker (the i18n spider,
debian-audit come in mind), and for human beings.

> Remember the nightmare we had with shadow's diff .gz file when we came
> up and started working on the package (the diff.gz was bigger than the
> original source code).

I don't think his point his that patch helpers (dpatch, quilt) are only
crap. I think he would like all this to be handled internaly by
maintainers. I don't think he would complain if quilt was used by the
maintainers, to maintain a set of patches, and to generate the
package.diff.gz (which is the only way to automatically patch the source
at the package extraction time).

IMO, the second point of his blog is that advanced source control systems
are much more efficient to maintain a set of patch. I also tend to agree
with that.

The problems with these source control systems are that they require much
more time to get used to, and an infrastructure (do I have to buy a domain
to host my arch/darcs branch, can I host it at home? (i.e. it will be down
when I sleep)).

I hope I didn't distorded Joey's opinion,

PS: What do you think of German's apostrophs;)?
-- 
Nekral