Control: tags -1 moreinfo
Hi. I was looking at this again, perhaps with a clearer head than
before.
You wrote:
> But this eventually fails with:
>
>
> dpkg-buildpackage
> -----------------
>
> Command: dpkg-buildpackage -us -uc -rfakeroot
> dpkg-buildpackage: info: source package valgrind
> dpkg-buildpackage: info: source version 1:3.15.0-2~1.gbpd4e99a
> dpkg-buildpackage: info: source distribution UNRELEASED
> dpkg-buildpackage: info: source changed by Nikolaus Rath <nikol...@rath.org>
> dpkg-source --before-build .
> dpkg-buildpackage: info: host architecture amd64
> fakeroot debian/rules clean
> dh clean --with=autoreconf
> dh_clean
> dpkg-source -b .
> dpkg-source: error: can't build with source format '3.0 (quilt)': no upstream
> tarball found at ../valgrind_3.15.0.orig.tar.{bz2,gz,lzma,xz}
> dpkg-buildpackage: error: dpkg-source -b . subprocess returned exit status 25
I don't think we should be trying to rebuild the source package at
this point. Empirically, when invoked in a directory, sbuild runs
dpkg-source directly to build the source package, even before it
prints its banner line. The source package ought not then to be
*re*built via dpkg-buildpackage.
The rune in dgit-user(7) is intended to cause the source package built
by sbuild's invocation of dpkg-source to be built as 1.0, to bypass
all of the things that can go wrong with `3.0 (quilt)`. But the
example rune doesn't also have sbuild pass those options through
dpkg-buildpackage to the 2nd invocation of dpkg-source, becauwe it's
not expecting there to *be* a 2nd invocation.
In my test suite, where I test the rune out of the manpage, I see this
in the corresponding part of the sbuild logfile:
| dpkg-buildpackage
| -----------------
|
| Command: dpkg-buildpackage -us -uc -b -rfakeroot
| dpkg-buildpackage: info: source package example
| dpkg-buildpackage: info: source version 1.1-1
| dpkg-buildpackage: info: source distribution unstable
| dpkg-buildpackage: info: source changed by Ian Jackson
<ijack...@chiark.greenend.org.uk>
| dpkg-source --before-build .
| dpkg-buildpackage: info: host architecture amd64
| fakeroot debian/rules clean
| EXAMPLE RULES TARGET clean
| ./clean-target-hook
| + test build-27ad867f-e55f-47bf-a551-48e756059a46
| dh_clean
| dh_clean: warning: Compatibility levels before 10 are deprecated (level 8 in
use)
| debian/rules build
Note the ddifference in dpkg-buildpacakge invocations:
> Command: dpkg-buildpackage -us -uc -rfakeroot
| Command: dpkg-buildpackage -us -uc -b -rfakeroot
I don't seem to be able to find in this bug a complete sbuild logfile
from a repro of the original report, but:
RTFMing sbuild suggests that the "--source" option may be implicated -
especially, via the BUILD_SOURCE sbuild config variable.
Do you have this setting in your personal configuration ?
Thanks,
Ian.
--
Ian Jackson <ijack...@chiark.greenend.org.uk> These opinions are my own.
Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk,
that is a private address which bypasses my fierce spamfilter.
