Re: debtags support proposed for xcontrol
On Tue, 8 Apr 2008, Holger Levsen wrote:
Uh. So there is ./Makefile (which contains the dist target), which then
calls ./debian/rules, which then modifies ./ - this seems ugly to me too.
Well, I adopted the Makefile from Debian Edu build system which was used
to build the debian/control file (and other stuff). I don't know what you
mean with "modifies ./" and in how far this seems ugly. Could you please
be a little bit more verbose on this.
I just turned the former behaviour of the Debian Edu build system to work
for all CDDs and this is done compliant to policy (in contrast to the
former build system). The "make dist" is just a shortcut for
make -f debian/rules get-orig-source
If you hate shortcuts, just don't use it. But it does exactly what I
expect from a dist target of a Makefile and so I regard it just convinient.
In debian-edu-doc one has to call "./debian/rules get-orig-sources" directly
Feel free to ignore the make dist shortcut and you have the same.
(and this only modifies the upstream source in the package, but not the
packaging.)
But we HAVE TO generate the debian/control file. The whole sense of
debian-edu source is TO GENERATE a valid control file. This is no issue
of changing the packaging but obtaining the right dependencies for a
target distribution. There is no other chance to obtain a valid source
package otherwise.
I also share Jonas opinion that native packages should be avoided and consider
them legacy.
Well, I don't care whether the tarball is called <version>.tar.gz or
<version>-<debianversion>.orig.tar.gz. But please elaborate on the
advantages we gain by changing a running system or provide patches
how you would like to see this happen. I just do not understand how
you would like to see things working. The whole content of the source
is about packaging, but the debian/control file has to be generated
and the rules, how to do this are in the source. Please read the paper
or the code and ask here if something remains unclear. I'm keen on
hearing if something remains unclear to enhance the paper.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: