#319162: gnome-games: ftbfs [sparc] --as-needed is borken in current binutils
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
gnome-games (Debian) |
Fix Released
|
Unknown
|
|||
gnome-games (Ubuntu) |
Invalid
|
High
|
Sebastien Bacher |
Bug Description
Automatically imported from Debian bug report #319162 http://
In Debian Bug tracker #319162, Steve Langasek (vorlon) wrote : Re: Bug#319162: gnome-games: ftbfs [sparc] multiple definition of `_PROCEDURE_LINKAGE_TABLE_' | #1 |
In Debian Bug tracker #319162, Loïc Minier (lool) wrote : #319162 is a --as-needed FTBFS | #2 |
retitle 319162 #319162: gnome-games: ftbfs [sparc] --as-needed is borken in current binutils
thanks
--
Loïc Minier <email address hidden>
Come, your destiny awaits!
Debian Bug Importer (debzilla) wrote : | #3 |
Automatically imported from Debian bug report #319162 http://
In Debian Bug tracker #319162, Blars Blarson (blarson) wrote : NMU patch | #4 |
NMU patch attached.
--
Blars Blarson <email address hidden>
http://
With Microsoft, failure is not an option. It is a standard feature.
In Debian Bug tracker #319162, Blars Blarson (blarson) wrote : Fixed in NMU of gnome-games 1:2.10.1-5.1 | #5 |
tag 319162 + fixed
quit
This message was generated automatically in response to a
non-maintainer upload. The .changes file follows.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Format: 1.7
Date: Sat, 3 Sep 2005 02:45:38 +0000
Source: gnome-games
Binary: gnome-games-data gnome-games
Architecture: source sparc all
Version: 1:2.10.1-5.1
Distribution: unstable
Urgency: low
Maintainer: Josselin Mouette <email address hidden>
Changed-By: Blars Blarson <email address hidden>
Description:
gnome-games - games for the GNOME desktop
gnome-games-data - data files for the GNOME games
Closes: 319162
Changes:
gnome-games (1:2.10.1-5.1) unstable; urgency=low
.
* 0 day NUM for RC bug durring BSP
* remove -Wl,--as-needed from all Makefiles (closes: #319162)
Files:
04848eccbdd918
40f48212a790bf
f02558f8b1189f
413cea97a07099
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFDGRVZg1H
gc3tYh8Tl7p6XID
=r2RL
-----END PGP SIGNATURE-----
In Debian Bug tracker #319162, Josselin Mouette (joss) wrote : Re: Bug#319162: NMU patch | #6 |
Le vendredi 02 septembre 2005 à 20:12 -0700, Blars Blarson a écrit :
> NMU patch attached.
Oh, great. Do you know that, when a package is co-maintained by
twenty-four people and a bug isn't fixed, there may be some reasons
behind?
This pseudo-fix won't help anything. Couldn't you help fixing some
unmaintained packages causing this kind of issues, like binutils,
instead?
--
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
In Debian Bug tracker #319162, Steve Langasek (vorlon) wrote : | #7 |
On Sat, Sep 03, 2005 at 08:59:27AM +0200, Josselin Mouette wrote:
> Le vendredi 02 septembre 2005 à 20:12 -0700, Blars Blarson a écrit :
> > NMU patch attached.
> Oh, great. Do you know that, when a package is co-maintained by
> twenty-four people and a bug isn't fixed, there may be some reasons
> behind?
> This pseudo-fix won't help anything. Couldn't you help fixing some
> unmaintained packages causing this kind of issues, like binutils,
> instead?
Yeah, right. binutils is unmaintained; the issue couldn't possibly be
that you've made poor design decisions in your packages by making them
dependent on kludgy, non-default toolchain options that policy doesn't
require the toolchain to support at all...
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
<email address hidden> http://
In Debian Bug tracker #319162, Blars Blarson (blarson) wrote : | #8 |
On Sat, Sep 03, 2005 at 08:59:27AM +0200, Josselin Mouette wrote:
> Le vendredi 02 septembre 2005 ? 20:12 -0700, Blars Blarson a ?crit :
> > NMU patch attached.
>
> Oh, great. Do you know that, when a package is co-maintained by
> twenty-four people and a bug isn't fixed, there may be some reasons
> behind?
If there was a reason for this easy to fix RC bug not being fixed, you
should have sent it to the bug.
You'll need to explain why gnome-games needs -Wl,--as-needed and
shouldn't be included in etch without it before I'll consider my NMU
as incorrect. (but correct as far as I knew when I made it.)
> This pseudo-fix won't help anything. Couldn't you help fixing some
> unmaintained packages causing this kind of issues, like binutils,
> instead?
binutils doesn't have any easy to fix RC bugs that have been sitting
with no action for weeks. (or any other RC bugs)
--
Blars Blarson <email address hidden>
http://
With Microsoft, failure is not an option. It is a standard feature.
In Debian Bug tracker #319162, Josselin Mouette (joss) wrote : | #9 |
Le samedi 03 septembre 2005 à 00:31 -0700, Steve Langasek a écrit :
> Yeah, right. binutils is unmaintained; the issue couldn't possibly be
> that you've made poor design decisions in your packages by making them
> dependent on kludgy, non-default toolchain options that policy doesn't
> require the toolchain to support at all...
Policy? What does binutils working properly have to do with policy?
Policy documents existing practice, and existing practice is to use this
option. This is a regression in the toolchain for some architectures. No
more, no less. If there aren't enough skilled people to fix the
toolchain for some architectures, this isn't a good sign for the health
of that port. Today, --as-needed can't be fixed, obviously because
nobody skilled enough is willing to work on it. What are you going to do
if "default" linker options are broken tomorrow?
And after all, you're the release manager, so you'll be the one to deal
with the horrible mess of gnome-games dependencies when all indirect
dependencies are explicit. Great to see how you welcome design decisions
taken to ease your work.
--
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
In Debian Bug tracker #319162, Josselin Mouette (joss) wrote : | #10 |
Le samedi 03 septembre 2005 à 00:46 -0700, Blars Blarson a écrit :
> > Oh, great. Do you know that, when a package is co-maintained by
> > twenty-four people and a bug isn't fixed, there may be some reasons
> > behind?
>
> If there was a reason for this easy to fix RC bug not being fixed, you
> should have sent it to the bug.
You're right here. The explanation wasn't sent to all such bug reports
and this was wrong.
> You'll need to explain why gnome-games needs -Wl,--as-needed and
> shouldn't be included in etch without it before I'll consider my NMU
> as incorrect. (but correct as far as I knew when I made it.)
Your NMU is highly aggressive. When there are more than twenty
maintainers, with about half of them being very responsive, you should
at least *talk* before uploading. I don't remember seeing you on
#gnome-debian nor on any GNOME-related mailing list.
There a reason why GNOME packages are team-maintained; it is precisely
to avoid anarchic and contradictory technical decisions, and to avoid
the need for NMUs.
> > This pseudo-fix won't help anything. Couldn't you help fixing some
> > unmaintained packages causing this kind of issues, like binutils,
> > instead?
>
> binutils doesn't have any easy to fix RC bugs that have been sitting
> with no action for weeks. (or any other RC bugs)
Without arguing endlessly about binutils, I think a NMU for one of the
26 RC bugs on QA packages would have been much appreciated.
--
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
In Debian Bug tracker #319162, Steve Langasek (vorlon) wrote : | #11 |
On Sat, Sep 03, 2005 at 11:42:25AM +0200, Josselin Mouette wrote:
> Le samedi 03 septembre 2005 à 00:31 -0700, Steve Langasek a écrit :
> > Yeah, right. binutils is unmaintained; the issue couldn't possibly be
> > that you've made poor design decisions in your packages by making them
> > dependent on kludgy, non-default toolchain options that policy doesn't
> > require the toolchain to support at all...
> Policy? What does binutils working properly have to do with policy?
> Policy documents existing practice, and existing practice is to use this
> option. This is a regression in the toolchain for some architectures. No
> more, no less. If there aren't enough skilled people to fix the
> toolchain for some architectures, this isn't a good sign for the health
> of that port.
If you're going to blame build failures on the toolchain, then policy
(and release policy) is quite relevant. There are in fact quite a few
toolchain options that *are* specified in policy: -O2, -O1, -O0, -g,
-Wl,-z,-defs, -Wall... and -shared is implied, of course... if a compile
fails when using one of these options, you have grounds for demanding
that the toolchain be fixed instead of trying to work around it in your
package. If you're using other, exotic toolchain options like -O3 or
-Wl,--as-needed, I believe the burden must lie primarily with the
package maintainer, not with the toolchain maintainer.
> Today, --as-needed can't be fixed, obviously because nobody skilled
> enough is willing to work on it. What are you going to do if "default"
> linker options are broken tomorrow?
Alternate explanation: the breakage in --as-needed wasn't noticed
upstream because it's a fringe option that isn't even clearly a good
idea, so it wasn't until Ubuntu and Debian packages started building
with binutils 2.16 that anyone noticed it was broken, so the breakage is
not a reflection on the viability of the porter teams for those archs.
(AIUI, there is actually activity upstream on getting a fix for this
bug; I just don't think the binutils maintainers should drop everything
else they work on to fix --as-needed, and I don't think you should wait
on binutils before fixing these build failures in your packages.)
> And after all, you're the release manager, so you'll be the one to deal
> with the horrible mess of gnome-games dependencies when all indirect
> dependencies are explicit. Great to see how you welcome design decisions
> taken to ease your work.
Yes, library dependencies are a major concern of mine, as I wrote at
<http://
kludges don't make a good solution, as they say (paraphrased); I believe
this is a problem we need to be fixing at the root, which is libtool and
pkg-config, instead of painting over it.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
<email address hidden> http://
In Debian Bug tracker #319162, Josselin Mouette (joss) wrote : | #12 |
Le dimanche 04 septembre 2005 à 03:04 -0700, Steve Langasek a écrit :
> If you're going to blame build failures on the toolchain, then policy
> (and release policy) is quite relevant. There are in fact quite a few
> toolchain options that *are* specified in policy: -O2, -O1, -O0, -g,
> -Wl,-z,-defs, -Wall... and -shared is implied, of course... if a compile
> fails when using one of these options, you have grounds for demanding
> that the toolchain be fixed instead of trying to work around it in your
> package. If you're using other, exotic toolchain options like -O3 or
> -Wl,--as-needed, I believe the burden must lie primarily with the
> package maintainer, not with the toolchain maintainer.
Sure. However, when you try to rely on these options for a package,
things are not that easy. -O3 doesn't really bring anything and you can
drop it anytime, but -Wl,--as-needed is a feature. If a feature isn't
guaranteed to work across versions, you can't start to rely on it, so it
shouldn't even be here.
> (AIUI, there is actually activity upstream on getting a fix for this
> bug; I just don't think the binutils maintainers should drop everything
> else they work on to fix --as-needed, and I don't think you should wait
> on binutils before fixing these build failures in your packages.)
As most of GNOME 2.10 is ready to enter testing, and as I don't have the
skills to fix binutils, I don't really have a choice, and will also
upload a gnome-session without --as-needed, but this isn't a good
long-term solution. Time spent re-uploading stuff and checking complex
dependencies isn't spent fixing other bugs.
> > And after all, you're the release manager, so you'll be the one to deal
> > with the horrible mess of gnome-games dependencies when all indirect
> > dependencies are explicit. Great to see how you welcome design decisions
> > taken to ease your work.
>
> Yes, library dependencies are a major concern of mine, as I wrote at
> <http://
> kludges don't make a good solution, as they say (paraphrased); I believe
> this is a problem we need to be fixing at the root, which is libtool and
> pkg-config, instead of painting over it.
The fix for libtool is available, but requires relibtoolizing packages
at each version, so it's even more work, and it's tedious work no one is
willing to do. The GNOME team manages to handle so many packages only
because we made simple packaging operations, like new upstream versions,
a trivial operation.
I'm not aware of any solution for pkg-config. It would probably need a
large rework, e.g. separating --shared-libs and --static-libs. And even
with both of them fixed, I'm afraid we'd still have issues with random
libraries added to the linkage without the need for them, or badly
written foo-config scripts. Except in some weird cases, --as-needed
solves all these issues. It may not be the Right Thing(tm), but it makes
good packages in the end.
Regards,
--
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
In Debian Bug tracker #319162, Loïc Minier (lool) wrote : | #13 |
Hi,
On Sun, Sep 04, 2005, Steve Langasek wrote:
> Yes, library dependencies are a major concern of mine, as I wrote at
> <http://
> kludges don't make a good solution, as they say (paraphrased); I believe
> this is a problem we need to be fixing at the root, which is libtool and
> pkg-config, instead of painting over it.
I think everyone agrees that if the deps generated by usage of libtool
and pkg-config were smaller, there would be less use for --as-needed.
Still, I find there's no reason to be strongly against --as-needed for
now:
- libtool's upstream never shipped with the modifications you suggest,
but binutils did ship --as-needed, and it borke on not very common
arches; this lets me think the amount of testing each feature
received is comparable, and --as-needed isn't that experimental --
even libtool people look at it for future usage
- the required libtool's changes are shipped in Debian, and in Debian
only, and until they're shipped upstream, the amount of work it would
take to upload each source relibtoolized is insane; --as-needed is
trivial to add/remove and doesn't add work between upstream releases
- the libtool fix only affects libtool generated packages
While these are all obvious conclusions from the discussions, I'm
surprized people stand so firmly pro or against --as-needed.
binutils has a regression which has to be fixed, to my knowledge a fix
is ready (the testsuite changes are missing though), and since we need
to be able to build packages to allow some testing transitions right
now, we're removing the build-option on packages needing
uploads/builds for these transitions.
As I understand it:
very short term
- remove --as-needed each time this is required to help an ongoing
important transition
short term
- finish the binutils fix, and return to the situation where we can and
we use --as-needed because the libtool and pkg-config issues remain
medium term
- fix upstream libtool and pkg-config
Only time will show whether --as-needed is often unusable and breaks on
random arches, and hence should be dropped, or in the contrary that
libtool and pkg-config were not fixed and that --as-needed is stable
enough.
PS: Steve, I've followed you bringing the libtool issue upstream, thank
you for that, but I couldn't find any hints on your claims against
pkg-config and how it should be fixed, nor a Debian bug; I was however
pointed at the reversal of an "indirect deps" change because of
complaints. Do you have any pointer on this? Are you bringing this
upstream too?
--
Loïc Minier <email address hidden>
Fabio Massimo Di Nitto (fabbione) wrote : | #14 |
we have fixed binutils and so on.. kthxbye
Debian Bug Importer (debzilla) wrote : | #15 |
Message-ID: <email address hidden>
Date: Wed, 20 Jul 2005 02:16:12 -0700
From: Blars Blarson <email address hidden>
To: Debian Bug Tracking System <email address hidden>
Subject: gnome-games: ftbfs [sparc] multiple definition of `_PROCEDURE_
Package: gnome-games
Version: 1:2.10.1-5
Severity: serious
Justification: fails to build from source
gnome-games failed to build on a sparc buildd, duplicated on sparc pbuilder:
/bin/sh ../libtool --mode=link cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,--as-needed -o same-gnome drawing.o game.o input.o same-gnome.o ui.o ../libgames-
mkdir .libs
cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,--as-needed -o same-gnome drawing.o game.o input.o same-gnome.o ui.o -Wl,--export-
/usr/lib/
/usr/bin/ld: Disabling relaxation: it will not work with multiple definitions
collect2: ld returned 1 exit status
make[4]: *** [same-gnome] Error 1
Debian Bug Importer (debzilla) wrote : | #16 |
Message-ID: <email address hidden>
Date: Wed, 20 Jul 2005 03:19:03 -0700
From: Steve Langasek <email address hidden>
To: Blars Blarson <email address hidden>, <email address hidden>
Subject: Re: Bug#319162: gnome-games: ftbfs [sparc] multiple definition of
`_PROCEDURE_
--KsGdsel6WgEHnImy
Content-Type: text/plain; charset=us-ascii
Content-
Content-
On Wed, Jul 20, 2005 at 02:16:12AM -0700, Blars Blarson wrote:
> Package: gnome-games
> Version: 1:2.10.1-5
> Severity: serious
> Justification: fails to build from source
> gnome-games failed to build on a sparc buildd, duplicated on sparc pbuild=
er:
> /bin/sh ../libtool --mode=3Dlink cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,=
--as-needed -o same-gnome drawing.o game.o input.o same-gnome.o ui.o ../li=
bgames-
R6/lib -lglade-2.0 -lgnomeui-2 -lSM -lICE -lbonoboui-2 -lxml2 -lpthread -lz=
-lgnomecanvas-2 -lgnome-2 -lpopt -lart_lgpl_2 -lpangoft2-1.0 -lbonobo-2 -l=
gconf-2 -lgnomevfs-2 -lbonobo-activation -lORBit-2 -lgthread-2.0 -lgtk-x11-=
2.0 -lgdk-x11-2.0 -latk-1.0 -lpangoxft-1.0 -lpangox-1.0 -lpango-1.0 -lrsvg-=
2 -lgdk_pixbuf-2.0 -lm -lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0 ../lib=
games-support/
> mkdir .libs
> cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,--as-needed -o same-gnome drawing.o=
game.o input.o same-gnome.o ui.o -Wl,--export-
R6/lib /usr/lib/
b/libbonoboui-2.so /usr/lib/libxml2.so -lpthread -lz /usr/lib/
s-2.so /usr/lib/
/usr/lib/
/usr/lib/
it-2.so /usr/lib/
dk-x11-2.0.so /usr/lib/
ibpangox-1.0.so /usr/lib/
gdk_pixbuf-2.0.so -lm /usr/lib/
-ldl /usr/lib/
> /usr/lib/
c/sparc32/
> /usr/bin/ld: Disabling relaxation: it will not work with multiple definit=
ions
> collect2: ld returned 1 exit status
> make[4]: *** [same-gnome] Error 1
This bug occurs on sparc and alpha (at least), and is most likely caused by
the use of the -Wl,-O1 option. So this is a gcc/binutils regression, but
not exactly a high priority one for the toolchain folks to fix -- it may be
a good idea to disable this linker option for now.
--=20
Steve Langasek
postmodern programmer
--KsGdsel6WgEHnImy
Content-Type: application/
Content-
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFC3iUXKN6
VSuCP...
Debian Bug Importer (debzilla) wrote : | #17 |
Message-ID: <email address hidden>
Date: Mon, 8 Aug 2005 22:59:47 +0200
From: =?iso-8859-
To: <email address hidden>
Subject: #319162 is a --as-needed FTBFS
retitle 319162 #319162: gnome-games: ftbfs [sparc] --as-needed is borken =
in current binutils
thanks
--=20
Lo=EFc Minier <email address hidden>
Come, your destiny awaits!
Debian Bug Importer (debzilla) wrote : | #18 |
Message-Id: <email address hidden>
Date: Fri, 02 Sep 2005 20:32:06 -0700
From: Blars Blarson <email address hidden>
To: <email address hidden>
Cc: Blars Blarson <email address hidden>, Josselin Mouette <email address hidden>
Subject: Fixed in NMU of gnome-games 1:2.10.1-5.1
tag 319162 + fixed
quit
This message was generated automatically in response to a
non-maintainer upload. The .changes file follows.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Format: 1.7
Date: Sat, 3 Sep 2005 02:45:38 +0000
Source: gnome-games
Binary: gnome-games-data gnome-games
Architecture: source sparc all
Version: 1:2.10.1-5.1
Distribution: unstable
Urgency: low
Maintainer: Josselin Mouette <email address hidden>
Changed-By: Blars Blarson <email address hidden>
Description:
gnome-games - games for the GNOME desktop
gnome-games-data - data files for the GNOME games
Closes: 319162
Changes:
gnome-games (1:2.10.1-5.1) unstable; urgency=low
.
* 0 day NUM for RC bug durring BSP
* remove -Wl,--as-needed from all Makefiles (closes: #319162)
Files:
04848eccbdd918
40f48212a790bf
f02558f8b1189f
413cea97a07099
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFDGRVZg1H
gc3tYh8Tl7p6XID
=r2RL
-----END PGP SIGNATURE-----
Debian Bug Importer (debzilla) wrote : | #19 |
Message-Id: <email address hidden>
Date: Sat, 03 Sep 2005 08:59:27 +0200
From: Josselin Mouette <email address hidden>
To: <email address hidden>, Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--=-Degh7XXuFZv
Content-Type: text/plain; charset=ISO-8859-15
Content-
Le vendredi 02 septembre 2005 =E0 20:12 -0700, Blars Blarson a =E9crit :
> NMU patch attached.
Oh, great. Do you know that, when a package is co-maintained by
twenty-four people and a bug isn't fixed, there may be some reasons
behind?
This pseudo-fix won't help anything. Couldn't you help fixing some
unmaintained packages causing this kind of issues, like binutils,
instead?
--=20
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
--=-Degh7XXuFZv
Content-Type: application/
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQBDGUnPrSl
pfIFjLa/
=5OcN
-----END PGP SIGNATURE-----
--=-Degh7XXuFZv
Debian Bug Importer (debzilla) wrote : | #20 |
Message-ID: <email address hidden>
Date: Sat, 3 Sep 2005 00:31:53 -0700
From: Steve Langasek <email address hidden>
To: Josselin Mouette <email address hidden>, <email address hidden>
Cc: Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--X1bOJ3K7DJ5YkBrT
Content-Type: text/plain; charset=iso-8859-1
Content-
Content-
On Sat, Sep 03, 2005 at 08:59:27AM +0200, Josselin Mouette wrote:
> Le vendredi 02 septembre 2005 =E0 20:12 -0700, Blars Blarson a =E9crit :
> > NMU patch attached.
> Oh, great. Do you know that, when a package is co-maintained by
> twenty-four people and a bug isn't fixed, there may be some reasons
> behind?
> This pseudo-fix won't help anything. Couldn't you help fixing some
> unmaintained packages causing this kind of issues, like binutils,
> instead?
Yeah, right. binutils is unmaintained; the issue couldn't possibly be
that you've made poor design decisions in your packages by making them
dependent on kludgy, non-default toolchain options that policy doesn't
require the toolchain to support at all...
--=20
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
<email address hidden> http://
--X1bOJ3K7DJ5YkBrT
Content-Type: application/
Content-
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFDGVFpKN6
r2/McNCEVCn+
=haNe
-----END PGP SIGNATURE-----
--X1bOJ3K7DJ5Yk
Debian Bug Importer (debzilla) wrote : | #21 |
Message-ID: <email address hidden>
Date: Sat, 3 Sep 2005 00:46:01 -0700
From: Blars Blarson <email address hidden>
To: Josselin Mouette <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#319162: NMU patch
On Sat, Sep 03, 2005 at 08:59:27AM +0200, Josselin Mouette wrote:
> Le vendredi 02 septembre 2005 ? 20:12 -0700, Blars Blarson a ?crit :
> > NMU patch attached.
>
> Oh, great. Do you know that, when a package is co-maintained by
> twenty-four people and a bug isn't fixed, there may be some reasons
> behind?
If there was a reason for this easy to fix RC bug not being fixed, you
should have sent it to the bug.
You'll need to explain why gnome-games needs -Wl,--as-needed and
shouldn't be included in etch without it before I'll consider my NMU
as incorrect. (but correct as far as I knew when I made it.)
> This pseudo-fix won't help anything. Couldn't you help fixing some
> unmaintained packages causing this kind of issues, like binutils,
> instead?
binutils doesn't have any easy to fix RC bugs that have been sitting
with no action for weeks. (or any other RC bugs)
--
Blars Blarson <email address hidden>
http://
With Microsoft, failure is not an option. It is a standard feature.
Debian Bug Importer (debzilla) wrote : | #22 |
Message-Id: <email address hidden>
Date: Sat, 03 Sep 2005 11:42:25 +0200
From: Josselin Mouette <email address hidden>
To: <email address hidden>, Steve Langasek <email address hidden>
Cc: Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--=-CKdUH3D2Mt8
Content-Type: text/plain; charset=ISO-8859-15
Content-
Le samedi 03 septembre 2005 =E0 00:31 -0700, Steve Langasek a =E9crit :
> Yeah, right. binutils is unmaintained; the issue couldn't possibly be
> that you've made poor design decisions in your packages by making them
> dependent on kludgy, non-default toolchain options that policy doesn't
> require the toolchain to support at all...
Policy? What does binutils working properly have to do with policy?
Policy documents existing practice, and existing practice is to use this
option. This is a regression in the toolchain for some architectures. No
more, no less. If there aren't enough skilled people to fix the
toolchain for some architectures, this isn't a good sign for the health
of that port. Today, --as-needed can't be fixed, obviously because
nobody skilled enough is willing to work on it. What are you going to do
if "default" linker options are broken tomorrow?
And after all, you're the release manager, so you'll be the one to deal
with the horrible mess of gnome-games dependencies when all indirect
dependencies are explicit. Great to see how you welcome design decisions
taken to ease your work.
--=20
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
--=-CKdUH3D2Mt8
Content-Type: application/
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQBDGXABrSl
A7Pr3QUlSwYjDic
=X9JD
-----END PGP SIGNATURE-----
--=-CKdUH3D2Mt8
Debian Bug Importer (debzilla) wrote : | #23 |
Message-Id: <email address hidden>
Date: Sat, 03 Sep 2005 12:08:45 +0200
From: Josselin Mouette <email address hidden>
To: <email address hidden>, Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--=-q7OmlqF7YIz
Content-Type: text/plain; charset=ISO-8859-15
Content-
Le samedi 03 septembre 2005 =E0 00:46 -0700, Blars Blarson a =E9crit :
> > Oh, great. Do you know that, when a package is co-maintained by
> > twenty-four people and a bug isn't fixed, there may be some reasons
> > behind?
>=20
> If there was a reason for this easy to fix RC bug not being fixed, you
> should have sent it to the bug.
You're right here. The explanation wasn't sent to all such bug reports
and this was wrong.
> You'll need to explain why gnome-games needs -Wl,--as-needed and
> shouldn't be included in etch without it before I'll consider my NMU
> as incorrect. (but correct as far as I knew when I made it.)
Your NMU is highly aggressive. When there are more than twenty
maintainers, with about half of them being very responsive, you should
at least *talk* before uploading. I don't remember seeing you on
#gnome-debian nor on any GNOME-related mailing list.
There a reason why GNOME packages are team-maintained; it is precisely
to avoid anarchic and contradictory technical decisions, and to avoid
the need for NMUs.
> > This pseudo-fix won't help anything. Couldn't you help fixing some
> > unmaintained packages causing this kind of issues, like binutils,
> > instead?
>=20
> binutils doesn't have any easy to fix RC bugs that have been sitting
> with no action for weeks. (or any other RC bugs)
Without arguing endlessly about binutils, I think a NMU for one of the
26 RC bugs on QA packages would have been much appreciated.
--=20
.''`. Josselin Mouette /\./\
: :' : <email address hidden>
`. `' <email address hidden>
`- Debian GNU/Linux -- The power of freedom
--=-q7OmlqF7YIz
Content-Type: application/
Content-
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQBDGXYtrSl
EYMNGWI4ejEUqTO
=HqT2
-----END PGP SIGNATURE-----
--=-q7OmlqF7YIz
Debian Bug Importer (debzilla) wrote : | #24 |
Message-ID: <email address hidden>
Date: Sun, 4 Sep 2005 03:04:08 -0700
From: Steve Langasek <email address hidden>
To: Josselin Mouette <email address hidden>
Cc: <email address hidden>, Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--Pql/uPZNXIm1JCle
Content-Type: text/plain; charset=iso-8859-1
Content-
Content-
On Sat, Sep 03, 2005 at 11:42:25AM +0200, Josselin Mouette wrote:
> Le samedi 03 septembre 2005 =E0 00:31 -0700, Steve Langasek a =E9crit :
> > Yeah, right. binutils is unmaintained; the issue couldn't possibly be
> > that you've made poor design decisions in your packages by making them
> > dependent on kludgy, non-default toolchain options that policy doesn't
> > require the toolchain to support at all...
> Policy? What does binutils working properly have to do with policy?
> Policy documents existing practice, and existing practice is to use this
> option. This is a regression in the toolchain for some architectures. No
> more, no less. If there aren't enough skilled people to fix the
> toolchain for some architectures, this isn't a good sign for the health
> of that port.
If you're going to blame build failures on the toolchain, then policy
(and release policy) is quite relevant. There are in fact quite a few
toolchain options that *are* specified in policy: -O2, -O1, -O0, -g,
-Wl,-z,-defs, -Wall... and -shared is implied, of course... if a compile
fails when using one of these options, you have grounds for demanding
that the toolchain be fixed instead of trying to work around it in your
package. If you're using other, exotic toolchain options like -O3 or
-Wl,--as-needed, I believe the burden must lie primarily with the
package maintainer, not with the toolchain maintainer.
> Today, --as-needed can't be fixed, obviously because nobody skilled
> enough is willing to work on it. What are you going to do if "default"
> linker options are broken tomorrow?
Alternate explanation: the breakage in --as-needed wasn't noticed
upstream because it's a fringe option that isn't even clearly a good
idea, so it wasn't until Ubuntu and Debian packages started building
with binutils 2.16 that anyone noticed it was broken, so the breakage is
not a reflection on the viability of the porter teams for those archs.
(AIUI, there is actually activity upstream on getting a fix for this
bug; I just don't think the binutils maintainers should drop everything
else they work on to fix --as-needed, and I don't think you should wait
on binutils before fixing these build failures in your packages.)
> And after all, you're the release manager, so you'll be the one to deal
> with the horrible mess of gnome-games dependencies when all indirect
> dependencies are explicit. Great to see how you welcome design decisions
> taken to ease your work.
Yes, library dependencies are a major concern of mine, as I wrote at
<http://
kludges don't make a good solution, as they say (paraphrased); I believe
this is a problem we need to be fixing at the root, which is libtool and
pkg-config, instead of painting over it.
--...
Debian Bug Importer (debzilla) wrote : | #25 |
Message-Id: <email address hidden>
Date: Sun, 04 Sep 2005 14:53:22 +0200
From: Josselin Mouette <email address hidden>
To: <email address hidden>, Steve Langasek <email address hidden>
Cc: Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
--=-rCN0NqvOpFO
Content-Type: text/plain; charset=ISO-8859-15
Content-
Le dimanche 04 septembre 2005 =E0 03:04 -0700, Steve Langasek a =E9crit :
> If you're going to blame build failures on the toolchain, then policy
> (and release policy) is quite relevant. There are in fact quite a few
> toolchain options that *are* specified in policy: -O2, -O1, -O0, -g,
> -Wl,-z,-defs, -Wall... and -shared is implied, of course... if a compile
> fails when using one of these options, you have grounds for demanding
> that the toolchain be fixed instead of trying to work around it in your
> package. If you're using other, exotic toolchain options like -O3 or
> -Wl,--as-needed, I believe the burden must lie primarily with the
> package maintainer, not with the toolchain maintainer.
Sure. However, when you try to rely on these options for a package,
things are not that easy. -O3 doesn't really bring anything and you can
drop it anytime, but -Wl,--as-needed is a feature. If a feature isn't
guaranteed to work across versions, you can't start to rely on it, so it
shouldn't even be here.
> (AIUI, there is actually activity upstream on getting a fix for this
> bug; I just don't think the binutils maintainers should drop everything
> else they work on to fix --as-needed, and I don't think you should wait
> on binutils before fixing these build failures in your packages.)
As most of GNOME 2.10 is ready to enter testing, and as I don't have the
skills to fix binutils, I don't really have a choice, and will also
upload a gnome-session without --as-needed, but this isn't a good
long-term solution. Time spent re-uploading stuff and checking complex
dependencies isn't spent fixing other bugs.
> > And after all, you're the release manager, so you'll be the one to deal
> > with the horrible mess of gnome-games dependencies when all indirect
> > dependencies are explicit. Great to see how you welcome design decision=
s
> > taken to ease your work.
>=20
> Yes, library dependencies are a major concern of mine, as I wrote at
> <http://
> kludges don't make a good solution, as they say (paraphrased); I believe
> this is a problem we need to be fixing at the root, which is libtool and
> pkg-config, instead of painting over it.
The fix for libtool is available, but requires relibtoolizing packages
at each version, so it's even more work, and it's tedious work no one is
willing to do. The GNOME team manages to handle so many packages only
because we made simple packaging operations, like new upstream versions,
a trivial operation.
I'm not aware of any solution for pkg-config. It would probably need a
large rework, e.g. separating --shared-libs and --static-libs. And even
with both of them fixed, I'm afraid we'd still have issues with random
libraries added to the linkage without the need for them, or bad...
Debian Bug Importer (debzilla) wrote : | #26 |
Message-ID: <email address hidden>
Date: Sun, 4 Sep 2005 19:52:27 +0200
From: =?iso-8859-
To: Steve Langasek <email address hidden>, <email address hidden>
Cc: Josselin Mouette <email address hidden>, Blars Blarson <email address hidden>
Subject: Re: Bug#319162: NMU patch
Hi,
On Sun, Sep 04, 2005, Steve Langasek wrote:
> Yes, library dependencies are a major concern of mine, as I wrote at
> <http://
> kludges don't make a good solution, as they say (paraphrased); I believ=
e
> this is a problem we need to be fixing at the root, which is libtool an=
d
> pkg-config, instead of painting over it.
I think everyone agrees that if the deps generated by usage of libtool
and pkg-config were smaller, there would be less use for --as-needed.
Still, I find there's no reason to be strongly against --as-needed for
now:
- libtool's upstream never shipped with the modifications you suggest,
but binutils did ship --as-needed, and it borke on not very common
arches; this lets me think the amount of testing each feature
received is comparable, and --as-needed isn't that experimental --
even libtool people look at it for future usage
- the required libtool's changes are shipped in Debian, and in Debian
only, and until they're shipped upstream, the amount of work it would
take to upload each source relibtoolized is insane; --as-needed is
trivial to add/remove and doesn't add work between upstream releases
- the libtool fix only affects libtool generated packages
While these are all obvious conclusions from the discussions, I'm
surprized people stand so firmly pro or against --as-needed.
binutils has a regression which has to be fixed, to my knowledge a fix
is ready (the testsuite changes are missing though), and since we need
to be able to build packages to allow some testing transitions right
now, we're removing the build-option on packages needing
uploads/builds for these transitions.
As I understand it:
very short term
- remove --as-needed each time this is required to help an ongoing
important transition
short term
- finish the binutils fix, and return to the situation where we can and
we use --as-needed because the libtool and pkg-config issues remain
medium term
- fix upstream libtool and pkg-config
Only time will show whether --as-needed is often unusable and breaks on
random arches, and hence should be dropped, or in the contrary that
libtool and pkg-config were not fixed and that --as-needed is stable
enough.
PS: Steve, I've followed you bringing the libtool issue upstream, thank
you for that, but I couldn't find any hints on your claims against
pkg-config and how it should be fixed, nor a Debian bug; I was however
pointed at the reversal of an "indirect deps" change because of
complaints. Do you have any pointer on this? Are you bringing this
upstream too?
--=20
Lo=EFc Minier <email address hidden>
In Debian Bug tracker #319162, J.H.M. Dassen (Ray) (jdassen) wrote : Bug#319162: fixed in gnome-games 1:2.10.2-1 | #27 |
Source: gnome-games
Source-Version: 1:2.10.2-1
We believe that the bug you reported is fixed in the latest version of
gnome-games, which is due to be installed in the Debian FTP archive:
gnome-games-
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
A summary of the changes between this version and the previous one is
attached.
Thank you for reporting the bug, which will now be closed. If you
have further comments please address them to <email address hidden>,
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
J.H.M. Dassen (Ray) <email address hidden> (supplier of updated gnome-games package)
(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing <email address hidden>)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Format: 1.7
Date: Sat, 24 Sep 2005 16:09:02 +0200
Source: gnome-games
Binary: gnome-games-data gnome-games
Architecture: source i386 all
Version: 1:2.10.2-1
Distribution: unstable
Urgency: low
Maintainer: J.H.M. Dassen (Ray) <email address hidden>
Changed-By: J.H.M. Dassen (Ray) <email address hidden>
Description:
gnome-games - games for the GNOME desktop
gnome-games-data - data files for the GNOME games
Closes: 319162
Changes:
gnome-games (1:2.10.2-1) unstable; urgency=low
.
J.H.M. Dassen (Ray):
* New upstream bug fix and translations update release.
* [debian/control.in]
* Bumped Standards-Version.
* Added versioned binutils build dependency (so --as-needed works).
* [debian/rules] Revert NMU changes now --as-needed works again.
* Acknowledge NMU (Closes: #319162).
.
Loic Minier:
* Fix copyright. [debian/copyright]
Files:
ea304f5e44ab08
78282fb34f96c5
869ae311021600
578e2b456b6a9e
0214c25504d15a
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFDNWFQA+
KcQ9aKSX9w+
=OVtM
-----END PGP SIGNATURE-----
Debian Bug Importer (debzilla) wrote : | #28 |
Message-Id: <email address hidden>
Date: Sat, 24 Sep 2005 07:47:07 -0700
From: <email address hidden> (J.H.M. Dassen (Ray))
To: <email address hidden>
Subject: Bug#319162: fixed in gnome-games 1:2.10.2-1
Source: gnome-games
Source-Version: 1:2.10.2-1
We believe that the bug you reported is fixed in the latest version of
gnome-games, which is due to be installed in the Debian FTP archive:
gnome-games-
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
gnome-games_
to pool/main/
A summary of the changes between this version and the previous one is
attached.
Thank you for reporting the bug, which will now be closed. If you
have further comments please address them to <email address hidden>,
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
J.H.M. Dassen (Ray) <email address hidden> (supplier of updated gnome-games package)
(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing <email address hidden>)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Format: 1.7
Date: Sat, 24 Sep 2005 16:09:02 +0200
Source: gnome-games
Binary: gnome-games-data gnome-games
Architecture: source i386 all
Version: 1:2.10.2-1
Distribution: unstable
Urgency: low
Maintainer: J.H.M. Dassen (Ray) <email address hidden>
Changed-By: J.H.M. Dassen (Ray) <email address hidden>
Description:
gnome-games - games for the GNOME desktop
gnome-games-data - data files for the GNOME games
Closes: 319162
Changes:
gnome-games (1:2.10.2-1) unstable; urgency=low
.
J.H.M. Dassen (Ray):
* New upstream bug fix and translations update release.
* [debian/control.in]
* Bumped Standards-Version.
* Added versioned binutils build dependency (so --as-needed works).
* [debian/rules] Revert NMU changes now --as-needed works again.
* Acknowledge NMU (Closes: #319162).
.
Loic Minier:
* Fix copyright. [debian/copyright]
Files:
ea304f5e44ab08
78282fb34f96c5
869ae311021600
578e2b456b6a9e
0214c25504d15a
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFDNWFQA+
KcQ9aKSX9w+
=OVtM
-----END PGP SIGNATURE-----
On Wed, Jul 20, 2005 at 02:16:12AM -0700, Blars Blarson wrote:
> Package: gnome-games
> Version: 1:2.10.1-5
> Severity: serious
> Justification: fails to build from source
> gnome-games failed to build on a sparc buildd, duplicated on sparc pbuilder:
> /bin/sh ../libtool --mode=link cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,--as-needed -o same-gnome drawing.o game.o input.o same-gnome.o ui.o ../libgames- support/ libgames- support. la -Wl,--export- dynamic -pthread -L/usr/X11R6/lib -lglade-2.0 -lgnomeui-2 -lSM -lICE -lbonoboui-2 -lxml2 -lpthread -lz -lgnomecanvas-2 -lgnome-2 -lpopt -lart_lgpl_2 -lpangoft2-1.0 -lbonobo-2 -lgconf-2 -lgnomevfs-2 -lbonobo-activation -lORBit-2 -lgthread-2.0 -lgtk-x11-2.0 -lgdk-x11-2.0 -latk-1.0 -lpangoxft-1.0 -lpangox-1.0 -lpango-1.0 -lrsvg-2 -lgdk_pixbuf-2.0 -lm -lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0 ../libgames- support/ libgames- support. la dynamic -pthread -L/usr/X11R6/lib /usr/lib/ libglade- 2.0.so /usr/lib/ libgnomeui- 2.so -lSM -lICE /usr/lib/ libbonoboui- 2.so /usr/lib/libxml2.so -lpthread -lz /usr/lib/ libgnomecanvas- 2.so /usr/lib/ libgnome- 2.so /usr/lib/libpopt.so /usr/lib/ libart_ lgpl_2. so /usr/lib/ libpangoft2- 1.0.so /usr/lib/ libbonobo- 2.so /usr/lib/ libgconf- 2.so /usr/lib/ libgnomevfs- 2.so /usr/lib/ libbonobo- activation. so /usr/lib/ libORBit- 2.so /usr/lib/ libgthread- 2.0.so /usr/lib/ libgtk- x11-2.0. so /usr/lib/ libgdk- x11-2.0. so /usr/lib/ libatk- 1.0.so /usr/lib/ libpangoxft- 1.0.so /usr/lib/ libpangox- 1.0.so /usr/lib/ libpango- 1.0.so /usr/lib/ librsvg- 2.so /usr/lib/ libgdk_ pixbuf- 2.0.so -lm /usr/lib/ libgobject- 2.0.so /usr/lib/ libgmodule- 2.0.so -ldl /usr/lib/ libglib- 2.0.so ../libgames- support/ .libs/libgames- support. a gcc/sparc- linux-gnu/ 4.0.1/. ./../.. /../lib/ crt1.o: ../sysdeps/ sparc/sparc32/ elf/start. S:32: multiple definition of `_PROCEDURE_ LINKAGE_ TABLE_'
> mkdir .libs
> cc -g -Wall -O2 -DNO_HOWL -Wl,-O1 -Wl,--as-needed -o same-gnome drawing.o game.o input.o same-gnome.o ui.o -Wl,--export-
> /usr/lib/
> /usr/bin/ld: Disabling relaxation: it will not work with multiple definitions
> collect2: ld returned 1 exit status
> make[4]: *** [same-gnome] Error 1
This bug occurs on sparc and alpha (at least), and is most likely caused by
the use of the -Wl,-O1 option. So this is a gcc/binutils regression, but
not exactly a high priority one for the toolchain folks to fix -- it may be
a good idea to disable this linker option for now.
--
Steve Langasek
postmodern programmer