[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

using svn-buildpackage and pbuilder for several architectures



Dear all,

at our institute in Heidelberg, we build Debian packages for Astronomy-
related software, mostly software written in the context of the Virtual
Observatory
http://en.wikipedia.org/wiki/Virtual_observatory
http://www.ivoa.net/ .

We also use a local debarchiver repository and we aim at building our
packages for two distributions ("stable", "unstable") and two architectures
("i386", "amd64"). For this purpose, we have tried to combine
"svn-buildpackage" and "pbuilder".

In order to see how we launch the build, you may have a look at the
following lines of code from one of our shell scripts:

svnbuild-all() {
        binOnly=''
        while read dist arch; do
                ARCH=$arch DIST=$dist svn-buildpackage --svn-ignore\
                         --svn-builder="pdebuild --architecture $arch --buildresult ~/build_area --auto-debsign --debsign-k gavo@ari.uni-heidelberg.de -- $binOnly"
                binOnly="--binary-arch"
        done < ~/TARGETS
}

where ~/TARGET contains
stable i386
stable amd64 .

In this case, our "pgsphere" package is built with the option
"--binary-arch" for "amd64" and w/o this option for "i386".
The problem here is that "svn-buildpackage" creates a new Debian source
package for both "i386" and "amd64" and thus overwrites the .dsc
generated for the first when it creates the second one. As a result, we
get an error like "checksums of pgsphere_1.1.1-3_i386.changes and
pgsphere_1.1.1-3.dsc do not match" when we try to upload the i386 package
via "dput".

By using the "pbuilder" only,
sudo -E DIST=stable ARCH=i386 --build pgsphere_1.1.1-3.dsc
sudo -E DIST=stable ARCH=amd64 --build --binary-arch pgsphere_1.1.1-3.dsc
I do not run into that problem.


Is it possible to tell svn-buildpackage not to create the Debian source
package for the second (third, ...) architecture?

Best regards,
Florian



--------------------------------------------------
Florian Rothmaier
German Astrophysical Virtual Observatory
Astronomisches Rechen-Institut
University of Heidelberg
Moenchhofstrasse 12-14
69120 Heidelberg
Germany
--------------------------------------------------


Reply to: