On Wed, Aug 17, 2016 at 12:09:27AM +0200, gregor herrmann wrote: > On Tue, 16 Aug 2016 21:13:55 +0100, Santiago Vila wrote: > > BTW: In this particular package, I decided to raise the severity after > > finding the failure myself while checking for bugs regarding > > "dpkg-buildpackage -A" > > *and* also getting unsuccessful build logs from reproducible builds as well. > > Sorry, I don't understand the connection to building with -A. Is > there a report about the package failing with it? Does it fail for > you if you build with -A?
I only build with -A, but there are a lot of packages for which a failed build with -A and a failed build without -A are essentially the same thing. For example: this package, libdevel-gdb-perl, is "Arch: all" and it has a minimal dh-style debian/rules. If "dpkg-buildpackage -A" fails, it is almost sure that it's not a bug in debian/rules or a bug in the packaging but a bug which would qualify as an "ordinary" FTBFS bug. So when I build a package with -A and it fails, I go to the reproducible builds site to see the build logs there. If it fails there too, then I know that -A has nothing to do with the failure. > So you got a failing build? [...] Yes, so far I got a failed build and three successful builds. I attach the failed build. > (And on the reproducible builds it works roughly half of the time: > https://tests.reproducible-builds.org/debian/history/libdevel-gdb-perl.html > ) That's more or less my point: For a package which is "Arch: all" like this one, a success rate of only 50% is bad enough not to be RC. Thanks.
libdevel-gdb-perl_2.02-3_amd64-20160813-1447.gz
Description: application/gzip
