On Thu, Dec 31, 2009, Junichi Uekawa wrote:
Note that qemubuilder is somewhat already client-server model, because
we don't quite have a good communication path between inside the qemu
session and outside the qemu session.
Inside qemu, it's a shell script using stdio (which is hooked up with
On Sun, Nov 29, 2009, Junichi Uekawa wrote:
I was thinking along the lines of running something inside the chroot
as part of build process, and passing the output back to pdebuild, and
using that to get the correct value.
However, that's a bit of work to implement, so I have so far deferred
tags 516625 + patch
stop
Hi
So I poked this further and did a more complete IPC prototype in shell
(foo.sh, attached) and integrated that into pbuilder in the attached
hackish patch.
It could be the way forward, but it touches many many things, and it's
particularly fragile stuff.
Hi,
At Wed, 30 Dec 2009 17:28:18 +0100,
Loïc Minier wrote:
[1 text/plain; iso-8859-1 (8bit)]
tags 516625 + patch
stop
Hi
So I poked this further and did a more complete IPC prototype in shell
(foo.sh, attached) and integrated that into pbuilder in the attached
hackish
I don't really want to rely on correct ARCH configuration here.
I was thinking along the lines of running something inside the chroot
as part of build process, and passing the output back to pdebuild, and
using that to get the correct value.
However, that's a bit of work to implement, so I have
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
A expensive solution would be to use cowbuilder's --execute command to
list the architecture inside the chroot
ARCHITECTURE=$(${PBUILDERROOTCMD} ${PDEBUILD_PBUILDER} --execute \
/usr/bin/dpkg-architecture $@ 2 /dev/null | \
I'm experiencing the same problem here. As others have mentioned, the
use case where this becomes a problem is when attempting to build both
amd64 and i386 packages on the same box.
For example:
DIST=sid ARCH=amd64 pdebuild --auto-debsign --buildresult /tmp/$(basename
$PWD) --debsign-k
Hi,
At Mon, 23 Feb 2009 11:26:06 +0600,
Mikhail Gusarov wrote:
[1 text/plain (quoted-printable)]
Twas brillig at 12:12:14 23.02.2009 UTC+09 when dan...@netfort.gr.jp did gyre
and gimble:
the name of changes file is generated by using $(dpkg-architecture
-qDEB_HOST_ARCH).
JU
Twas brillig at 22:27:59 23.02.2009 UTC+09 when dan...@netfort.gr.jp did gyre
and gimble:
JU I usually build the package and checks that it works before signing
JU a package (with debsign); so my workflow doesn't rely on the
JU signing feature from pdebuild.
pdebuild also failed to copy
At Mon, 23 Feb 2009 19:34:40 +0600,
Mikhail Gusarov wrote:
[1 text/plain (quoted-printable)]
Twas brillig at 22:27:59 23.02.2009 UTC+09 when dan...@netfort.gr.jp did gyre
and gimble:
JU I usually build the package and checks that it works before signing
JU a package (with debsign);
Twas brillig at 14:41:59 24.02.2009 UTC+09 when dan...@netfort.gr.jp did gyre
and gimble:
JU Is cross-building with pdebuild a common use-case?
The only relatively useful use-case I can come up with is building i386
packages on amd64 host.
JU It's rare compared to native-building with
Package: pbuilder
Version: 0.185
Severity: normal
If pdebuild is run with --debbuildopts --arch=i386, then it fails
to sign the generated changes file, because the name of changes file
is generated by using $(dpkg-architecture -qDEB_HOST_ARCH).
Seems that arch-handling should be moved to
At Mon, 23 Feb 2009 00:39:53 +0600,
Mikhail Gusarov wrote:
Package: pbuilder
Version: 0.185
Severity: normal
If pdebuild is run with --debbuildopts --arch=i386, then it fails
to sign the generated changes file, because the name of changes file
is generated by using $(dpkg-architecture
Twas brillig at 12:12:14 23.02.2009 UTC+09 when dan...@netfort.gr.jp did gyre
and gimble:
the name of changes file is generated by using $(dpkg-architecture
-qDEB_HOST_ARCH).
JU anyone care to craft a patch?
I'll try to, because it breaks not only signing the file, but also
moving the
14 matches
Mail list logo
