Teodor wrote:
> On Tue, Dec 23, 2008 at 2:28 PM, Thiemo Seufer wrote:
> > Gerfried Fuchs wrote:
> >> I can confirm that the package builds fine, started a build yesterday
> >> on a mipsel box I have access to. I did build it in an uptodate unstable
> >> cow
Gerfried Fuchs wrote:
> * Teodor [2008-12-23 11:53:04 CET]:
> > On Sat, Dec 20, 2008 at 12:45 AM, John Goerzen
> > wrote:
> > > Luk Claes wrote:
> > >> unblocked
> > >
> > > Even better -- thanks!
> >
> > Luk, thanks for the unblock. Unfortunately 'bacula' cannot migrate due
> > to a build fail
Julien BLACHE wrote:
> Thiemo Seufer wrote:
>
> Hi,
>
> >> > ok, do you have a chance to test something newer preferred git-current ?
> >>
> >> If you have a kernel I can try, sure. Building one is another story,
> >> this machine
Package: mandos
Version: 1.0.2-1
Severity: serious
There was an error while trying to autobuild your package:
> Automatic build of mandos_1.0.2-1 on ball by sbuild/mips 99.99
> Build started at 20081221-2253
[...]
> ** Using build dependencies supplied by package:
> Build-Depends: debhelper (>=
Package: mrpt
Version: 0.6.4svn702-1
Severity: serious
There was an error while trying to autobuild your package:
> Automatic build of mrpt_0.6.4svn702-1 on ball by sbuild/mips 99.99
> Build started at 20081221-1619
[...]
> ** Using build dependencies supplied by package:
> Build-Depends: debhe
Package: boost1.37
Version: 1.37.0-2
Severity: serious
There was an error while trying to autobuild your package:
> Automatic build of boost1.37_1.37.0-2 on ball by sbuild/mips 99.99
> Build started at 20081216-2008
[...]
> ** Using build dependencies supplied by package:
> Build-Depends: debhe
FYI, this problem is triggered on build daemons for arm, hppa, mips,
mipsel and sparc. I was also able to reproduce it on a mips machine
with 2.5.26-rc8. However, a very similiar mips machine with a 2.6.24
kernel happily builds djvulibre.
So it looks like the kernel is at least part of the problem
Package: gir-repository
Version: 0.6.1-2
Severity: serious
gir-repository fails to build two of its modules, and then fails to
detect the problem and stop the build. This appears to be a violation
of policy 4.6.
[...]
> ltmain.sh (GNU libtool) 1.5.26 Debian 1.5.26-4 (1.1220.2.493 2008/02/01
> 1
Philipp Kern wrote:
> On Sun, Dec 07, 2008 at 04:26:32PM +0100, Thiemo Seufer wrote:
> > gluezilla needs mono, this is already reflected in the packages'
> > architecture list.
>
> I am not convinced that this should be done. Packages like gluezilla
> and libgdipl
Package: rdup
Version: 0.6.3-1
Severity: serious
There was an error while trying to autobuild your package:
> Automatic build of rdup_0.6.3-1 on mayr by sbuild/mips 99.999
> Build started at 20081208-2204
[...]
> ** Using build dependencies supplied by package:
> Build-Depends: debhelper (>= 4.
Martín Ferrari wrote:
> On Mon, Dec 8, 2008 at 09:30, Thiemo Seufer <[EMAIL PROTECTED]> wrote:
>
> >> Could you tell me if the assembler code generated by gcc looks as a
> >> reasonable replacement for the offending code you've patched?
> >
> >
Martín Ferrari wrote:
> Hi,
>
> On Sun, Dec 7, 2008 at 14:21, Thiemo Seufer <[EMAIL PROTECTED]> wrote:
>
> >> as -o tas.o tas.s
> >> tas.s: Assembler messages:
> >> tas.s:10: Error: illegal operands `ori t1,zero,12345'
> >> tas.s:11:
Package: drawterm
Version: 0.cvs+20080909-1
Severity: serious
Tags: patch
There was an error while trying to autobuild your package:
> Automatic build of drawterm_0.cvs+20080909-1 on ball by sbuild/mips 99.99
> Build started at 20081206-1246
[...]
> gcc -Wall -Wno-missing-braces -ggdb -I.. -I..
Package: buildd.debian.org
Version: -
Tags: patch
acpitail, libacpi and xbattbar-acpi work only on ACPI supporting
platforms. The respective control files encode this already, albeit
in some cases with out an lpia entry.
Index: P-a-s/Packages-arch-specific
===
Package: buildd.debian.org
Version: -
Tags: patch
whitelister needs ocamlopt, this is already reflected in the packages'
architecture list.
Index: P-a-s/Packages-arch-specific
===
--- P-a-s.orig/Packages-arch-specific 2008-12-07 1
Package: buildd.debian.org
Version: -
Tags: patch
gluezilla needs mono, this is already reflected in the packages'
architecture list.
Index: P-a-s/Packages-arch-specific
===
--- P-a-s.orig/Packages-arch-specific 2008-12-07 15:39:3
Package: buildd.debian.org
Version: -
Tags: patch
The appended patch updates three entries for sys/io.h:
- picproc doesn't need sys/io.h anymore.
- rovclock uses sys/io.h. For some reason it nevertheless built on
ia64, and the package claims to support ia64.
- rxtx uses sys/io.h. This is
Julien BLACHE wrote:
> [EMAIL PROTECTED] (Thomas Bogendoerfer) wrote:
>
> Hi,
>
> > ok, do you have a chance to test something newer preferred git-current ?
>
> If you have a kernel I can try, sure. Building one is another story,
> this machine isn't exactly fast nor quiet, as you probably know.
Uwe Kleine-König wrote:
> Package: gcc-4.3
> Version: 4.3.2-1
> Severity: wishlist
>
> Hello,
>
> I tried building a cross toolchain for powerpc following the steps on
> http://www.emdebian.org/tools/crossdev.html (i.e. installing
> binutils-$arch-linux-gnu and libc6-dev-$arch-cross, then export
Mark Purcell wrote:
> On Wednesday 29 October 2008 04:06:20 Luca Bruno wrote:
> > if the release team
> > really sees this as an issue needing an immediate fix, I won't object to
> > the choice (and if Thiemo is already there I won't be against a quick
> > NMU targeting only #500418, as I'll probab
severity 500418 serious
thanks
Mark Purcell wrote:
> severity 500418 important
> thanks
>
> On Tuesday 28 October 2008 10:57:39 Thiemo Seufer wrote:
> > severity 500418 serious
> [...]
> > Please add hppa mips mipsel to the allowed architectures in debian/control.
severity 500418 serious
thanks
A test build for hppa (on paer) succeeded now, apart from the control
file bit, and the blocking python bug is fixed as well.
Please add hppa mips mipsel to the allowed architectures in debian/control.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with
Package: boost
Version: 1.34.1-11
Severity: serious
Tags: patch
gnash FTBFS on arm, armel, hppa, mipsel due to a bug in
boost/detail/endian.hpp. The appended patch fixes it.
Thiemo
--- boost-1.34.1/boost/detail/endian.hpp.orig 2008-10-26 17:35:45.0
+
+++ boost-1.34.1/boost/detai
Luca Capello wrote:
> tags 503255 + upstream
> tags 503255 + patch
> tags 503255 + fixed-upstream
> thanks
>
> Hi David!
>
> Since Thiemo Seufer did the last SBCL upload, I now directly cc:ed him
> to be sure he reads this message.
>
> On Fri, 24 Oct 2008 18:
Package: matplotlib
Version: 0.98.1-1+lenny2
Severity: serious
Tags: patch
Matplotlib FTBFS in lenny due to a recent change in the platform
identifier for alpha, hppa, mips, mipsel and sparc. The attached
patch is an update to debian/patches/build_fix.patch which fixes
the problem.
Thiemo
Index:
Package: matplotlib
Version: 0.98.3-3
Severity: serious
Tags: patch
Matplotlib FTBFS due to a recent change in the platform identifier
for alpha, hppa, mips, mipsel and sparc. The attached patch is an
update to debian/patches/build_fix.patch which fixes the problem.
Thiemo
Index: matplotlib-0.98
Aníbal Monsalve Salazar wrote:
> On Fri, Oct 03, 2008 at 06:04:31PM +0200, Thiemo Seufer wrote:
> >IMHO a fixed version should go into lenny.
>
> 0.3.9-1 isn't in lenny, 0.3.8-1 is.
Missed that, fine with me.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTE
severity 499722 serious
thanks
IMHO a fixed version should go into lenny.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Hi Artur,
do you plan to upload a fixed version of libqglviewer for lenny?
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
severity 495100 serious
thanks
Hi Francois,
sqlitebrowser currently fails to build with sudo, this causes build
failures on the alpha and mipsel, see bug #495100. I intend to do a
NMU in the next days which includes the patch.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subj
The appended patch fixes the root cause. gcc -O2 optimizes out some
tests.
Thiemo
--- libcaca-0.99.beta15/configure.ac~ 2008-09-27 19:11:55.0 +0100
+++ libcaca-0.99.beta15/configure.ac2008-10-03 10:21:19.0 +0100
@@ -112,13 +112,13 @@ AC_TRY_COMPILE([#include ],[S
[AC_MSG
After talking with Jose I NMU'ed lcdproc to allow building on
almost all architectures again. Appended is the full diff.
Thiemo
lcdproc (0.5.2-1.2) unstable; urgency=low
* Non-maintainer upload.
* Support all architectures except s390.
* Add support for kfreebsd, thanks Cyril Brulebois.
This is the version I finally added. It explicitly points out now
that the BSD licenses are GPL-incompatible.
Thiemo
--- copyright.orig 2008-09-28 16:31:07.0 +0100
+++ copyright 2008-09-28 18:42:05.0 +0100
@@ -542,3 +542,41 @@ The `xmlrpclib' module contains the foll
tags 498477 +patch
thanks
The appended patch updates the copyright notice to cover the relevant
library licenses.
Thiemo
--- copyright.orig 2008-09-28 16:31:07.0 +0100
+++ copyright 2008-09-28 16:59:20.0 +0100
@@ -542,3 +542,40 @@ The `xmlrpclib' module contains the foll
Loïc Minier wrote:
> On Sun, Sep 28, 2008, Thiemo Seufer wrote:
> > Flumotion fails to build on mips/mipsel due to bug #499132 in python2.5.
> > A test build with a patched version of python2.5 suceeded.
>
> Err why do you report this against flumotion?
To track that it m
Package: istanbul
Version: 0.2.2-4
istanbul fails to build on mips/mipsel due to bug #499132 in python2.5.
Furthermore, mips and mipsel are currently excluded in the control file.
Please re-enable builds for mips/mipsel once #499132 is fixed.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECT
Package: flumotion
Version: 0.4.2-3.1
Severity: important
Flumotion fails to build on mips/mipsel due to bug #499132 in python2.5.
A test build with a patched version of python2.5 suceeded.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
Package: lfhex
Version: 0.4-3
Severity: serious
lfhex fails to build when qt3-dev-tools is installed because it then
picks up the QT3. The easiest way to fix this is to add a
Build-Conflicts: qt3-dev-tools
line in the control file.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with
Package: dpkg-dev
Version: 1.14.22
Tags: patch
Gcc's -mno-shared support on mips/mipsel introduces a new magic
internally used symbol named __gnu_local_gp. Since gcc-4.3 the
-mno-shared option is default, and dpkg-shlibdeps thows a warning
in many build logs. The appended blacklists the symbol.
tags 496175 +patch
severity 496175 minor
thanks
This problem happens when klibc is built on a MIPS64 kernel without
changing the personality to 32 bit (uname is mips64 instead of mips).
The official buildd setup handles this, OTOH it would be nice if we
could abandon this trick one day. The append
tags 499058 +unreproducible
thanks
I did a test build on my machine on current unstable/mips, the build
succeeded. Judging from the build logs there are infrequent segfaults
of python2.4 and python2.5 on mips/mipsel. This is probably caused by
#499132, which on MIPS results in a subtly broken dlop
Package: gparted
Version: 0.3.9-1
Severity: important
Tags: patch
gparted fails to build on arm armel mips and mipsel since a patch in the
package isn't complete. I attach an updated version. The interdiff of
both patches is:
diff -u gparted-0.3.9/src/GParted_Core.cc gparted-0.3.9/src/GParted_Cor
tags 497634 +unreproducible
thanks
I can't reproduce this build failure on current mips/unstable, it
builds fine here.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
force-merge 498216 498647 498397
tags 498216 +patch
thanks
I can confirm the patch works as expected.
Thiemo
Since gcc-4.3 mips/mipsel needs an explicit PIC flag for shared library
code.
--- a/configure
+++ b/configure
@@ -1811,7 +1811,7 @@ if enabled shared; then
# LIBOBJFLAGS may have al
peter green wrote:
>
>>
>> 106: void av_free_static(void)
>> 107: {
>> 108:while(last_static){
>> 109: av_freep(&array_static[--last_static]);
>> 110:}
>> 111:av_freep(&array_static);
>> 112: }
>>
>> A segfault in line 108 is, uhm, strange.
>>
> Agreed, either gdb is misbehaving
Reinhard Tartler wrote:
[snip]
> >> (gdb) run
> >> Starting program: /usr/bin/ffmpeg
> >> [Thread debugging using libthread_db enabled]
> >> [New Thread 0x2aad0f40 (LWP 26315)]
> >>
> >> Program received signal SIGSEGV, Segmentation fault.
> >> [Switching to Thread 0x2aad0f40 (LWP 26315)]
> >> 0x2
Package: guile-1.6
Version: 1.6.8-6
Severity: serious
Tags: patch
Guile 1.6 fails to build with gcc-4.3 on at least powerpc. The
bootstrap guile segfaults at start time. This is the same problem
as guile 1.8 had in #481378. I adapted the upstream change and
tested the resulting patch. The appended
found 495467 0.4-2
thanks
The patch isn't applied due to a bug in debian/rules, so it still
fails on ARM and MIPS. The appended change fixes it.
Thiemo
--- lfhex-0.4.orig/debian/rules 2008-09-02 00:47:27.0 +0200
+++ lfhex-0.4/debian/rules 2008-09-02 00:44:18.0 +0200
@@ -2,
Package: prelink
Version: 0.0.20071009-1
Tags: important
Please enable builds for mips and mipsel in debian/control, prelink
supports MIPS for about two years now.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Tobias Klauser wrote:
> On 2008-08-17 at 17:55:42 +0200, Thiemo Seufer <[EMAIL PROTECTED]> wrote:
> > Package: lfhex
> > Version: 0.4-1
> > Severity: important
> > Tags: patch
>
> Thanks for your bug report and for the patch. I'll create an updated
>
Package: gnudatalanguage
Version: 0.9~rc1-1
Severity: important
Tags: patch
gnudatalanguage fails to build because the -z muldefs option is set in
LDFLAGS. LDFLAGS, however, is fed to the compiler driver, so it needs
to be escaped with -Wl. (Some versions of gcc-4.3 appear to simply
ignore -z para
Package: fplll
Version: 2.1.6+20071129-1
Severity: important
Tags: patch
fplll fails to build on architectures which don't support long double.
The appended patch fixes it. Tested with a sucessful build on mips.
Thiemo
Index: fplll-2.1.6+20071129/src/dpe.h
=
Package: lfhex
Version: 0.4-1
Severity: important
Tags: patch
lfhex fails to build on mips, mipsel, arm and armel because on those
platforms g++-4.3 did already away with abs(long long) in the global
namespace. (On other platforms it is only deprecated, and will be
removed in future g++ versions.)
tags +patch
thanks
The appended update to the powerpc patch in libatomic-ops makes
the test failure go away and leads to a successful build.
Be warned that I have no idea what it does, I followed blindly
the example given in http://gcc.gnu.org/ml/gcc/2008-03/msg01094.html.
Thiemo
--- ../liba
The build still works with gcc-4.2.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Package: opentoken
Version: 3.0b-7
Severity: important
Please enable builds of opentoken for mips and mipsel. Those
architectures support gnat-4.3 now, which should be sufficient
to build opentoken.
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble
Package: libatomic-ops
Version: 1.2-4
Severity: serious
libatomic-ops currently fails to build on powerpc/unstable due to
a failing test case:
[...]
make[4]: Entering directory `/home/ths/tf/debs/libatomic-ops-1.2/tests'
cc -DHAVE_CONFIG_H -I. -I. -I../src -I./../src -g -O2 -g -Wall -O2 -DNDEB
Julien Cristau wrote:
> On Thu, Aug 14, 2008 at 22:39:31 +0200, Thiemo Seufer wrote:
>
> > --- spass-2.1.orig/debian/rules 2008-08-14 22:35:04.0 +0200
> > +++ spass-2.1/debian/rules 2008-08-14 22:09:41.0 +0200
> > @@ -20,7 +20,7 @@ endif
>
Package: spass
Version: 2.1-3
Severity: serious
Tags: patch
spass fails to build when using dpkg-buildpackage -rsudo because in
this case $(PWD) isn't set as a make variable in debian/rules. The
appended patch fixes this.
Thiemo
--- spass-2.1.orig/debian/rules 2008-08-14 22:35:04.0 +02
Package: usepackage
Version: 1.6.4-1
Severity: serious
Tags: patch
usepackage fails to build on the build daemons for mips and mipsel,
which use sudo instead of fakeroot. In this environment, $(PWD)
is not set, and the build creates a /debian directory in the
root of the host machine. The appended
Package: pasco
Version: 1.0+20040505-2
Severity: important
Tags: patch
pasco fails to build on the build daemons for alpha and mipsel, which
use sudo instead of fakeroot. This happens because the clean rule
creates a bin directory with root permissions. The appended patch
fixes it.
Thiemo
---
Package: galleta
Version: 1.0+20040505-2
Severity: important
Tags: patch
galleta fails to build on the build daemons for alpha and mipsel, which
use sudo instead of fakeroot. This shappens because the clean rule
creates a bin directory with root permissions. The appended patch fixes
it.
Thiemo
Package: libqglviewer
Version: 2.2.6-3-3
Severity: important
Tags: patch
libqglviewer fails to build on the build daemons for alpha, mips and
mipsel, which use sudo instead of fakeroot. This happens because
the clean rule invokes configure, which the creates directories
with root permissions. The
Package: sqlitebrowser
Version: 1.3-2
Severity: important
Tags: patch
sqlitebrowser fails to build on the alpha and mipsel build daemons
(which run with sudo instead of fakeroot). This happens because the
clean target changes the permissions of "Makefile" in such a setup.
The appended patch fixes
Package: gnat-gps
Version: 4.0.1-6
Severity: important
gnat-gps fails to build on at least mips/mipsel because it hardcodes
the python include path in debian/rules to python2.4 (but python-dev
provides python2.5 now).
IIRC the python 2.4 -> 2.5 transition announcement explained how to
fix this.
reassign 492740 cl-f2cl
thanks
David Bremner wrote:
> Package: cl-f2cl
> Version: 20080222-1
> Severity: important
>
> Hi;
>
> I'm not sure if this is an sbcl problem or an f2cl problem.
>
> (require 'f2cl)
>
> fails in sbcl-1.0.18.0-1
>
> the last part of the output is
>
>
> ; compiling
Christian Thalinger wrote:
> On Mon, 2008-07-07 at 12:49 +0100, Thiemo Seufer wrote:
> > I leave the Debian bug open for this one.
>
> Is there still a problem somewhere? Did I miss something?
I believe the SVN commit you pointed at is not yet in the Debian
package.
T
Christian Thalinger wrote:
> I guess you are talking about this code:
>
> http://mips.complang.tuwien.ac.at/hg/cacao/file/2effc42cfb01/src/vm/jit/mips/asmpart.S#l426
>
> This code is correct, as the fsX defines use paired registers, starting
> with a even one. Also see:
>
> http://mips.complang
Package: openmpi
Version: 1.2.7~rc2-1
Tags: patch
Severity: wishlist
I believe the appended patch is sufficient to add Linux/MIPS support
to openmpi (for both the mips and mipsel variants). That said, I wasn't
able to build and test it, due to some libtool-induced trouble:
- With just this patch,
Peter T. Breuer wrote:
> "Also sprach Debian Bug Tracking System:"
> > On Tue, Jul 01, 2008 at 11:34:03AM +0200, Peter T. Breuer wrote:
> > > Compiling kernel 2.6.24.2 on ia32, compilaton fails at final link
> > > complaining about not being able to find __udivdi3 and __umoddi3 for
> > > getnstimeo
Package: lxlauncher
Version: 0.1.6-1
Severity: serious
Lxlauncher currently fails to build on hppa, mips, mipsel and powerpc.
The appended patch is a minimal fix.
Thiemo
--- lxlauncher-0.1.6/src/inotify/inotify-syscalls.h~2008-06-25
16:06:48.0 +0100
+++ lxlauncher-0.1.6/src/inotif
Package: basket
Version: 1.0.2-4
Severity: serious
Basket fails to build in current unstable when automake 1.10 is installed:
[...]
Applying patch 16_basket_kontact_compat.diff
patching file kontact_plugin/basket.desktop
Now at patch 16_basket_kontact_compat.diff
touch debian/stamp-patched
cp -f
Package: ahven
Version: 1.2-1
Ahven fails to build on at least mips. This is probably a bug in
adabrowse 4.0.2-6, I don't know enough Ada to figure that out.
Thiemo
[...]
# build docs as well
/usr/bin/make docs
make[1]: Entering directory `/srv/ths/debian/ahven/ahven-1.2'
adabrowse -c adabrows
John Goerzen wrote:
> severity 485596 normal
> thanks
FYI, it used to build before (probably because earlier gcc versions
used less RAM). This failure prevents migration of the package to
testing.
> Thiemo Seufer wrote:
> >
> > WASH/HTML/HTMLPrelude98.hs
> > WASH/H
Package: washngo
Version: 2.12-5
Severity: serious
Washngo fails to build on 32-bit platforms with non-registrised ghc6.
This is caused by three large source (or generated source) files:
WASH/HTML/HTMLPrelude98.hs
WASH/HTML/HTMLMonad98.hs
WASH/HTML/HTMLPrelude.hs
Compiling those eventually (afte
Alexandre Oliva wrote:
> Package: linux-2.6
> Severity: normal
>
>
> In case you're not aware, the linux-libre project grew out of
> gNewSense's efforts to publish 100% Free Software extracts from the
> kernel distributed by kernel.org, later on picked up by BLAG and, more
> recently, by FSF Lati
Neil Jerram wrote:
> 2008/5/28 Thiemo Seufer <[EMAIL PROTECTED]>:
> >
> > After a closer look I believe the logic of the test is just plain wrong:
> >
> > aux (l) unsigned long l;
> > { int x; exit (l >= ((unsigned long)&x)); }
> > main ()
Neil Jerram wrote:
> 2008/5/24 Thiemo Seufer <[EMAIL PROTECTED]>:
>
> > Neil Jerram wrote:
> > >
> > > I believe those definitions came from the Boehm GC library. Do you
> > happen
> > > to know whether similar improvements have already be
Neil Jerram wrote:
[snip]
> > however, one of those patches is only a workaround.
> >
> > There are two problems:
> > - For Linux on mips/mipsel the gc definitions look decidedly strange,
> > I replaced those with more sensible variants.
>
>
> I believe those definitions came from the Boehm GC l
Package: gnat-4.3
Version: 4.3.0-5
Severity: important
Tags: patch
Gnat 4.3 doesn't support libraries on mips/mipsel, as the build logs of
ahven and libxmlada show. I believe this is an unintentional omission.
The appended patch fixes it.
Thiemo
diff -urpN gnat-4.3-4.3.0.old/debian/patches/ada
Andreas Barth wrote:
> * Matthias Klose ([EMAIL PROTECTED]) [080521 12:13]:
> > that looks ok for now, we can work on the unification of the headers
> > later.
>
> In case Thiemo agrees (or doesn't disagree) - do you plan to upload it?
> Should I NMU it? ...?
I disagree, because I was confused ab
Package: libghc6-pcre-light-dev
Version: 0.3.1-3
Severity: serious
libghc6-pcre-light-dev fails to install at least on mips due to a
misformatted installed-pkg-config file:
[...]
Selecting previously deselected package libghc6-pcre-light-dev.
(Reading database ... 260249 files and directories cur
Package: hfsprogs
Version: 332.14-3
Please enable building for the remaining Debian 32-bit architectures,
specifically: hppa mips mipsel s390.
All of those architectures (as well as powerpc and sparc) have also
variants with 64-bit extensions, but the Debian port defaults to use
the 32-bit mode,
Package: kmchart
Version: 1.2.6
Severity: serious
Kmchart fails to build on mips/mipsel, and I believe it will also fail
on any other architecture. The reason is a missing build dependency on
libqt4-opengl-dev. This used to be implicit by build-depending on
libqt4-dev, but this package was split u
Arthur Loiret wrote:
> Hi Thiemo,
>
> On Sat, May 17, 2008 at 04:13:20PM +0200, Arthur Loiret wrote:
> > With this patch /usr/include/mips64-linux-gnu/ is installed in both
> > libn32ffi-dev and lib64ffi-dev, please wait for my next patch.
>
> Here is the diff between header from mips and mips64
Package: soci
Version: 2.2.0-3
Severity: serious
Tags: patch
Soci fails to build on mips, mipsel, armel and presumably arm because
it tries to use std::snprintf. This is, however, a function in the
global namespace, and only retained as an alias in std:: for backward
compatibility purposes on othe
tags 479737 +patch
thanks
The appended patch fixes the FTBFS on mips and presumably also mipsel
and s390. I changed the Architecture: line to "any", as this seems to
be more maintainable.
Thiemo
diff -urpN lasso-2.1.98.old/debian/control lasso-2.1.98/debian/control
--- lasso-2.1.98.old/debian/
Package: xcircuit
Version: 3.6.130.dfsg-1
Severity: serious
Xcircuit fails to build with -rsudo (like the autobuilders for alpha,
mips and mipsel do). This is caused by the debian/rules clean target
which "touch"es config.status, which leads to wrong permissions for
the file.
I fail to understand
Package: tightvnc
Version: 1.3.9-1
Severity: important
Tags: patch
The appended patch adds some missing bits to the Imake configuration
for Linux/MIPS. Tested on big-endian mips, I believe mipsel will be
fine, too.
Thiemo
diff -urpN tightvnc-1.3.9.old/Xvnc/config/cf/linux.cf
tightvnc-1.3.9/Xv
Package: scribus
Version: 1.3.3.11.dfsg+svn20080214-1
Serverity: important
Please re-enable support for mips and mipsel. The earlier build failure
was caused by a bug in gcc-4.1, which was fixed upstream in gcc-4.2.
I did a test build of scribus on current mips/unstable, which worked fine.
I thin
Alexis Huxley wrote:
> Hi Arthur,
>
> > It has been closed by Arthur Loiret <[EMAIL PROTECTED]> (reply to [EMAIL
> > PROTECTED]).
>
> > /usr/bin/gcc symlink is provided by `gcc' package from `gcc-default' source
> > package, not by `gcc-4.2'.
>
> Fine, you are right, the package name might be w
Package: guile-1.8
Version: 1.8.5+1-1
Severity: serious
Tags: patch
Guile-1.8 currently fails to build on ia64, m68k, mips, mipsel and
powerpc. I created two patches which let it build again on mips,
however, one of those patches is only a workaround.
There are two problems:
- For Linux on mips/m
reopen 479529
thanks
Upstream's version of the patch is incorrect for the SIZOF_VOID_P != 8
case. For the O32 ABI, l.d/s.d expand to a instruction sequence which
handles the even/odd FPU register pairs, so the programmer can just
think of it as "load double / store double" without ever explicitly
Package: firebird2.0
Version: 2.0.3.12981.ds1-13
Severity: serious
Tags: patch
firebird fails to build on mips (and presumably will fail on arm)
because fbsvcmgr is an invalid make target. I attached updates of
port-arm.patch and port-mips.patch. For reference I appended also
the interdiff between
Package: kaffe
Version: 1.1.8-4
Severity: serious
Kaffe fails to build with:
[...]
make[4]: Leaving directory
`/srv/ths/debian/kaffe/kaffe-1.1.8/build/pthreads/tools/javatools'
make[3]: Leaving directory
`/srv/ths/debian/kaffe/kaffe-1.1.8/build/pthreads/tools/javatools'
make[3]: Entering direct
severity 479572 serious
thanks
This causes a FTBFS on mips (and presumably on all architectures
when a rebuild in a clean current unstable chroot is attempted).
Thiemo
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Frank Lichtenheld wrote:
> On Mon, May 05, 2008 at 01:44:21PM +0200, Michael Koch wrote:
> > On Mon, May 05, 2008 at 12:45:57PM +0200, Frank Lichtenheld wrote:
> > > | cc -I../../../../src -I../../../.. -I../../../../src -D__MIPS__
> > > -D__LINUX__ -ansi -pedantic -Wall -Wno-long-long
> > > -D_
version to 2007.
+ * Add mips and mipsel as supported architectures.
+
+ -- Thiemo Seufer <[EMAIL PROTECTED]> Fri, 02 May 2008 14:11:56 +0100
+
libflorist (2006-1) unstable; urgency=low
* debian/control (Maintainer): update email address, I'm now a full DD.
diff -uprN libflorist-20
Martin Michlmayr wrote:
> * Thiemo Seufer <[EMAIL PROTECTED]> [2008-03-03 15:18]:
> > > > The appended patch works around the problem. I don't tag it as patch
> > > > because its implications are not fully analyzed upstream. It works
> > > > well o
hangelog 2008-04-11 23:42:15.0 +0100
@@ -1,3 +1,10 @@
+libtexttools (3.0.0-1) unstable; urgency=medium
+
+ * Use gnat-4.3. Bump soname version to 3.0.
+ * Add mips and mipsel as supported architectures (Closes: #475572).
+
+ -- Thiemo Seufer <[EMAIL PROTECTED]> Fri, 11 Apr 2008 21:13:52
1 - 100 of 409 matches
Mail list logo
