On Fri, Feb 22, 2013 at 12:27:18PM -0800, Russ Allbery wrote:
> Steve Langasek writes:
> > I'm not sure how you've arrived at this conclusion. Have you overlooked
> > that the shlibs in the ntfs-3g package have been fixed by the maintainer
> > in unstable (as commented in bug #700677)?
> > It s
Steve Langasek writes:
> I'm not sure how you've arrived at this conclusion. Have you overlooked
> that the shlibs in the ntfs-3g package have been fixed by the maintainer
> in unstable (as commented in bug #700677)?
> It still doesn't comply with policy 8.1. But I think that's a policy
> bug
On Fri, Feb 22, 2013 at 09:44:42AM -0500, Phillip Susi wrote:
> On 2/22/2013 1:48 AM, Adam D. Barratt wrote:
> > No. Both situations are buggy and neither of them is acceptable in
> > testing.
> > [There are situations where we do knowingly introduce
> > uninstallability to testing, but those are
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/22/2013 1:48 AM, Adam D. Barratt wrote:
> No. Both situations are buggy and neither of them is acceptable in
> testing.
>
> [There are situations where we do knowingly introduce
> uninstallability to testing, but those are always short term and
On Thu, 2013-02-21 at 15:12 -0500, Phillip Susi wrote:
> On 2/21/2013 11:37 AM, Adam D. Barratt wrote:
> > Apologies if I'm missing something here, but is that a definition
> > of "fixed" that involves intentionally allowing the migration of
> > package A to make package B uninstallable in testing?
On Thu, Feb 21, 2013 at 03:12:17PM -0500, Phillip Susi wrote:
> On 2/21/2013 11:37 AM, Adam D. Barratt wrote:
> > Apologies if I'm missing something here, but is that a definition
> > of "fixed" that involves intentionally allowing the migration of
> > package A to make package B uninstallable in t
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/21/2013 11:37 AM, Adam D. Barratt wrote:
> Apologies if I'm missing something here, but is that a definition
> of "fixed" that involves intentionally allowing the migration of
> package A to make package B uninstallable in testing?
Exactly. Beca
On 21.02.2013 16:06, Phillip Susi wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/21/2013 10:32 AM, Sam Hartman wrote:
It's less of an issue for Ubuntu, but for Debian it's going to
block testing migrations of the depended package until the
depending package updates.
That could be f
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/21/2013 10:32 AM, Sam Hartman wrote:
> It's less of an issue for Ubuntu, but for Debian it's going to
> block testing migrations of the depended package until the
> depending package updates.
That could be fixed in Britney to detect this case and
> "Phillip" == Phillip Susi writes:
Phillip> Not having the .pc file and headers etc in the -dev package
Phillip> would prevent the build of anything with a decent
Phillip> pkg-config enabled build system, so that could work with a
Phillip> tweak to the policy to allow it. Th
> "Guillem" == Guillem Jover writes:
Guillem> On Tue, 2013-02-19 at 20:30:48 -0500, Sam Hartman wrote:
>> with the current packaging tools, you tend to end up producing
>> the .shlibs files in order to manage cross-package dependencies
>> within a single source package. If it
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 02/19/2013 08:30 PM, Sam Hartman wrote:
> Uh, no. If you don't install a .so, then linking with -lfoo won't
> work.
I assumed you meant to just leave the .so in the app package, not a
separate -dev package. Obviously there not being a .so in the
On Tue, 2013-02-19 at 20:30:48 -0500, Sam Hartman wrote:
> with the current packaging tools, you tend to end up producing the
> .shlibs files in order to manage cross-package dependencies within a
> single source package.
> If it were relatively easy to convince the packaging tools to handle
> dep
> "Phillip" == Phillip Susi writes:
Phillip> On 2/18/2013 1:21 PM, Sam Hartman wrote:
>> 2) don't install a .so in a -dev package.
Phillip> That might be a signal a human can understand, but the
Phillip> build system won't catch it. The goal is to make sure the
Phillip>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/18/2013 1:21 PM, Sam Hartman wrote:
> 2) don't install a .so in a -dev package.
That might be a signal a human can understand, but the build system
won't catch it. The goal is to make sure the build system doesn't
generate broken binary packages
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/18/2013 1:04 PM, Kurt Roeckx wrote:
> What could be done is that the shlibs file sets up a strict
> version dependency. It currently says: libntfs-3g 837 ntfs-3g
> udeb: libntfs-3g 837 ntfs-3g-udeb
Collin Watson suggested this as a reasonable
Hi.
I'd like to speak a bit to what tthe right answer here is rather than
what the policy currently says.
It's sometimes fairly annoying to move a library out of the default path
and to adjust the build system accordingly.
Russ did that for one of krb5's private libraries, and in retrospect, I
fee
On Sat, Feb 16, 2013 at 08:24:09PM -0500, Phillip Susi wrote:
> Package: tech-ctte
>
> I filed bug #700677 because ntfs-3g has a shared library that ubuntu's
> testdisk links to, but it does not follow the SONAME rules. It seems
> that upstream breaks ABI on every release, and the maintainer feel
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 2/18/2013 11:09 AM, Guillem Jover wrote:
> I would have thought that the correct place to discuss confusing
> sections or possible interpretations of the policy manual would be
> the policy mailing list, where improvements could get proposed if
>
On Sat, 2013-02-16 at 20:24:09 -0500, Phillip Susi wrote:
> Package: tech-ctte
>
> I filed bug #700677 because ntfs-3g has a shared library that ubuntu's
> testdisk links to, but it does not follow the SONAME rules. It seems
> that upstream breaks ABI on every release, and the maintainer feels
>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: tech-ctte
I filed bug #700677 because ntfs-3g has a shared library that ubuntu's
testdisk links to, but it does not follow the SONAME rules. It seems
that upstream breaks ABI on every release, and the maintainer feels
that the library is not
21 matches
Mail list logo
