Bug#821924: libvigraimpex: multi_array_chunked_hdf5.hxx corrupted on mipsel

[YunQiang Su]

On Wed, Apr 20, 2016 at 10:10 PM, YunQiang Su <wzss...@gmail.com> wrote:
> On Wed, Apr 20, 2016 at 9:28 PM, Jakub Wilk <jw...@debian.org> wrote:
>> Package: libvigraimpex-dev
>> Version: 1.10.0+git20160211.167be93+dfsg-2
>> User: debian-m...@lists.debian.org
>> Usertags: mipsel
>>
>> $ sed -ne 264p
>> libvigraimpex-1.10.0+git20160211.167be93+dfsg/include/vigra/multi_array_chunked_hdf5.hxx
>> | cat -A
>>// H5Pset_chunk_cache (dapl, rdcc_nslots, rdcc_nbytes, rdcc_w0);$
>>
>> But in the mips binary package, this line is corrupted:
>>
>> $ sed -ne 264p /usr/include/*/multi_array_chunked_hdf5.hxx | cat -A
>>   ^@// H5Ps^@t_chunk cache (eapl, rdcc_nslots, rdcc_nbytes,
>> rdcc_w0);$
>>
>
> quite strange, only mipsel has this problem, mips64el/mips(eb) doesn't
> have this problem.
> Let me have a check.

I rebuild this package on Octean CPU, and didn't get this problem.
Another loongson's bug?
Let me have a check again

>
>>
>> I have no idea how could this happen. It smells like a hardware problem. (I
>> CCed debian-mips@l.d.o and mipsel@buildd.d.o.)
>>
>> --
>> Jakub Wilk
>>
>
>
>
> --
> YunQiang Su



-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#821924: libvigraimpex: multi_array_chunked_hdf5.hxx corrupted on mipsel

[YunQiang Su]

On Wed, Apr 20, 2016 at 9:28 PM, Jakub Wilk <jw...@debian.org> wrote:
> Package: libvigraimpex-dev
> Version: 1.10.0+git20160211.167be93+dfsg-2
> User: debian-m...@lists.debian.org
> Usertags: mipsel
>
> $ sed -ne 264p
> libvigraimpex-1.10.0+git20160211.167be93+dfsg/include/vigra/multi_array_chunked_hdf5.hxx
> | cat -A
>// H5Pset_chunk_cache (dapl, rdcc_nslots, rdcc_nbytes, rdcc_w0);$
>
> But in the mips binary package, this line is corrupted:
>
> $ sed -ne 264p /usr/include/*/multi_array_chunked_hdf5.hxx | cat -A
>   ^@// H5Ps^@t_chunk cache (eapl, rdcc_nslots, rdcc_nbytes,
> rdcc_w0);$
>

quite strange, only mipsel has this problem, mips64el/mips(eb) doesn't
have this problem.
Let me have a check.

>
> I have no idea how could this happen. It smells like a hardware problem. (I
> CCed debian-mips@l.d.o and mipsel@buildd.d.o.)
>
> --
> Jakub Wilk
>



-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#805893: Please confirm the patch.

[YunQiang Su]

I tested this patch. It works well for mips64el.

On Fri, Nov 27, 2015 at 11:40 AM, YunQiang Su <wzss...@gmail.com> wrote:
> On Wed, Nov 25, 2015 at 6:23 PM, Ghislain Vaillant <ghisv...@gmail.com> wrote:
>> This is the relevant portion of the proposed patch:
>>
>>>  ifeq ($(DEB_HOST_ARCH),ppc64el)
>>>  DEB_CFLAGS_MAINT_APPEND += -mno-altivec
>>> +DEB_CXXFLAGS_MAINT_APPEND += -mno-altivec
>>> +endif
>>> +# Build without Altivec to prevent FTBFS on ppc64el.
>>> +ifneq (,$(filter mips64%el mips64%el,$(DEB_HOST_ARCH)))
>>> +DEB_CFLAGS_MAINT_APPEND += -mxgot
>>> +DEB_CXXFLAGS_MAINT_APPEND += -mxgot
>>>  endif
>>
>> But the explanation is:
>>
>>> filter mips%64 mips%64el,$(DEB_HOST_ARCH)
>>> is due to we may has some other mips64 architectures in future,
>>> for example: mips64r6{el}.
>>
>> The filter string differs between the patch and explanation:
>
> The patch is correct.
>
> Sorry for it.
>
>>
>> filter mips64%el mips64%el
>>
>> and
>>
>> filter mips%64 mips%64el.
>>
>> One is twice the same (typo?), the other has the percent sign placed
>> differently.
>>
>> Please confirm the correct filter command, otherwise I cannot apply this
>> patch.
>>
>> Thanks,
>> Ghis
>
>
>
> --
> YunQiang Su



-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#805893: Please confirm the patch.

[YunQiang Su]

On Wed, Nov 25, 2015 at 6:23 PM, Ghislain Vaillant <ghisv...@gmail.com> wrote:
> This is the relevant portion of the proposed patch:
>
>>  ifeq ($(DEB_HOST_ARCH),ppc64el)
>>  DEB_CFLAGS_MAINT_APPEND += -mno-altivec
>> +DEB_CXXFLAGS_MAINT_APPEND += -mno-altivec
>> +endif
>> +# Build without Altivec to prevent FTBFS on ppc64el.
>> +ifneq (,$(filter mips64%el mips64%el,$(DEB_HOST_ARCH)))
>> +DEB_CFLAGS_MAINT_APPEND += -mxgot
>> +DEB_CXXFLAGS_MAINT_APPEND += -mxgot
>>  endif
>
> But the explanation is:
>
>> filter mips%64 mips%64el,$(DEB_HOST_ARCH)
>> is due to we may has some other mips64 architectures in future,
>> for example: mips64r6{el}.
>
> The filter string differs between the patch and explanation:

The patch is correct.

Sorry for it.

>
> filter mips64%el mips64%el
>
> and
>
> filter mips%64 mips%64el.
>
> One is twice the same (typo?), the other has the percent sign placed
> differently.
>
> Please confirm the correct filter command, otherwise I cannot apply this
> patch.
>
> Thanks,
> Ghis



-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#790284: ssm: FTBFS with glibc 2.21 and gcc-5

[YunQiang Su]

On Sat, 27 Jun 2015 14:02:19 -0700 Daniel Schepler  wrote:
> Source: ssm
> Version: 1.3-2
> Severity: normal
>
> From my pbuilder build log, using a setup preferring glibc and gcc-defaults
> from experimental:

I NMUed this package to unstable with the attached patch.

>
> ...
>dh_fixperms -O--dbg-package=libssm1-dbg
>dh_strip -O--dbg-package=libssm1-dbg
>dh_makeshlibs -O--dbg-package=libssm1-dbg
> dpkg-gensymbols: warning: some symbols or patterns disappeared in the symbols 
> file: see diff output below
> dpkg-gensymbols: warning: debian/libssm1/DEBIAN/symbols doesn't match 
> completely debian/libssm1.symbols
> --- debian/libssm1.symbols (libssm1_1.3-2+bpb6_amd64)
> +++ dpkg-gensymbolsJTEyob   2015-06-14 08:50:56.737157172 +
> @@ -433,7 +433,7 @@
>   _ZTSN3ssm6XAlignE@Base 1.3
>   _ZTSN3ssm8SortDistE@Base 1.3
>   _ZTSN3ssm9MultAlignE@Base 1.3
> - _ZTV7CStream@Base 1.3
> +#MISSING: 1.3-2+bpb6# _ZTV7CStream@Base 1.3
>   _ZTVN3ssm10GraphMatchE@Base 1.3
>   _ZTVN3ssm10XAlignTextE@Base 1.3
>   _ZTVN3ssm12SortMappingsE@Base 1.3
> dh_makeshlibs: failing due to earlier errors
> debian/rules:9: recipe for target 'binary' failed
> make: *** [binary] Error 2
> dpkg-buildpackage: error: fakeroot debian/rules binary gave error exit status 
> 2
> --
> Daniel Schepler
>
>


ssm.debdiff
Description: Binary data
-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#804768: eigen3: cause opencv ftbfs when LDBL_MANT_DIG greater than 112

[YunQiang Su]

Package: eigen3
Version: 3.3~alpha-1

When build opencv on mips64el and sparc64, we met

[ 28%] Building CXX object
modules/contrib/CMakeFiles/opencv_contrib.dir/src/rgbdodometry.cpp.o
cd /«BUILDDIR»/opencv-2.4.9.1+dfsg/obj-mips64el-linux-gnuabi64/modules/contrib
&& /usr/bin/c++   -DCVAPI_EXPORTS -g -O2 -fstack-protector-strong
-Wformat -Werror=format-security -D_FORTIFY_SOURCE=2-fsigned-char
-W -Wall -Werror=return-type -Werror=address -Werror=sequence-point
-Wformat -Werror=format-security -Wmissing-declarations -Wundef
-Winit-self -Wpointer-arith -Wshadow -Wsign-promo -Wno-narrowing
-Wno-delete-non-virtual-dtor -fdiagnostics-show-option -pthread
-fomit-frame-pointer -ffunction-sections -g -O2
-fstack-protector-strong -Wformat -Werror=format-security  -DNDEBUG
-fPIC -isystem /«BUILDDIR»/opencv-2.4.9.1+dfsg/obj-mips64el-linux-gnuabi64
-isystem /usr/include/eigen3 -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/contrib/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/contrib/src -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/obj-mips64el-linux-gnuabi64/modules/contrib
-isystem /«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/core/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/flann/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/imgproc/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/highgui/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/features2d/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/calib3d/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/ml/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/objdetect/include -isystem
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/video/include -include
"/«BUILDDIR»/opencv-2.4.9.1+dfsg/obj-mips64el-linux-gnuabi64/modules/contrib/precomp.hpp"
-Winvalid-pch  -o CMakeFiles/opencv_contrib.dir/src/rgbdodometry.cpp.o
-c /«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/contrib/src/rgbdodometry.cpp
In file included from
/usr/include/eigen3/unsupported/Eigen/MatrixFunctions:56:0,
 from
/«BUILDDIR»/opencv-2.4.9.1+dfsg/modules/contrib/src/rgbdodometry.cpp:65:
/usr/include/eigen3/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h:
In function 'void Eigen::internal::matrix_exp_compute(const
MatrixType&, ResultType&)':
/usr/include/eigen3/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h:351:33:
error: 'StdStemFunctions' was not declared in this scope
 result = arg.matrixFunction(StdStemFunctions::exp);
 ^
/usr/include/eigen3/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h:351:63:
error: expected primary-expression before '>' token
 result = arg.matrixFunction(StdStemFunctions::exp);

This is caused by
unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h, in it:

   void matrix_exp_compute(const MatrixType& arg, ResultType )
{
#if LDBL_MANT_DIG > 112 // rarely happens
  typedef typename traits::Scalar Scalar;
  typedef typename NumTraits::Real RealScalar;
  typedef typename std::complex ComplexScalar;
  if (sizeof(RealScalar) > 14) {
result = arg.matrixFunction(StdStemFunctions::exp);
return;
  }
#endif

-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#754753: healpix-cxx: FTBFS on armel: testsuite timeout

[YunQiang Su]

On Mon, 14 Jul 2014 02:11:42 +0200 Cyril Brulebois k...@debian.org wrote:
 Source: healpix-cxx
 Version: 3.11.2-7
 Severity: serious
 Justification: FTBFS

 Hi,

 your package no longer builds on armel:
 | libtool: link: g++ -g -O2 -fstack-protector --param=ssp-buffer-size=4 
 -Wformat -Werror=format-security -fopenmp -fno-tree-vectorize -fno-math-errno 
 -funsafe-math-optimizations -fno-trapping-math -fno-rounding-math 
 -fno-signaling-nans -fcx-limited-range -fomit-frame-pointer -Wl,-z -Wl,relro 
 -o hpxtest Healpix_cxx/hpxtest.o  ./.libs/libhealpix_cxx.a -lcfitsio 
 -lpthread -fopenmp
 | make[1]: Leaving directory '/«PKGBUILDDIR»'
 |dh_auto_test -a
 | make[1]: Entering directory '/«PKGBUILDDIR»'
 | make  check-TESTS
 | make[2]: Entering directory '/«PKGBUILDDIR»'
 | make[3]: Entering directory '/«PKGBUILDDIR»'
 | make[3]: *** Deleting file 'hpxtest.log'
 | Terminated
 | make[2]: *** [check-TESTS] Terminated
 | make[3]: *** [hpxtest.log] Terminated
 | make[1]: *** [check-am] Terminated
 | Makefile:1557: recipe for target 'check-TESTS' failed
 | Makefile:1578: recipe for target 'hpxtest.log' failed
 | Makefile:1763: recipe for target 'check-am' failed
 | make: *** wait: No child processes.  Stop.
 | make: *** Waiting for unfinished jobs
 | make: *** wait: No child processes.  Stop.
 | Build killed with signal TERM after 150 minutes of inactivity

It is not a real FTBFS indeed.
The test 'hpxtest' cost to much time and armel has a quite poor performance, so
the test is killed by sbuild.

So if we output some information when building, it will not be killed by sbuild.

in `test-drivers':

$@ $log_file 21

--

$@ 21 | tee $log_file

will fix this problem.



 Full build log:
   
 https://buildd.debian.org/status/fetch.php?pkg=healpix-cxxarch=armelver=3.11.2-7stamp=1405237595

 Mraw,
 KiBi.



-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#754753: healpix-cxx: FTBFS on armel: testsuite timeout

[YunQiang Su]

I NMUed this package with the attached patch to 5-delay queue.

If any objection, contact me or cut it.

On Mon, Mar 30, 2015 at 5:25 PM, YunQiang Su wzss...@gmail.com wrote:
 On Mon, 14 Jul 2014 02:11:42 +0200 Cyril Brulebois k...@debian.org wrote:
 Source: healpix-cxx
 Version: 3.11.2-7
 Severity: serious
 Justification: FTBFS

 Hi,

 your package no longer builds on armel:
 | libtool: link: g++ -g -O2 -fstack-protector --param=ssp-buffer-size=4 
 -Wformat -Werror=format-security -fopenmp -fno-tree-vectorize 
 -fno-math-errno -funsafe-math-optimizations -fno-trapping-math 
 -fno-rounding-math -fno-signaling-nans -fcx-limited-range 
 -fomit-frame-pointer -Wl,-z -Wl,relro -o hpxtest Healpix_cxx/hpxtest.o  
 ./.libs/libhealpix_cxx.a -lcfitsio -lpthread -fopenmp
 | make[1]: Leaving directory '/«PKGBUILDDIR»'
 |dh_auto_test -a
 | make[1]: Entering directory '/«PKGBUILDDIR»'
 | make  check-TESTS
 | make[2]: Entering directory '/«PKGBUILDDIR»'
 | make[3]: Entering directory '/«PKGBUILDDIR»'
 | make[3]: *** Deleting file 'hpxtest.log'
 | Terminated
 | make[2]: *** [check-TESTS] Terminated
 | make[3]: *** [hpxtest.log] Terminated
 | make[1]: *** [check-am] Terminated
 | Makefile:1557: recipe for target 'check-TESTS' failed
 | Makefile:1578: recipe for target 'hpxtest.log' failed
 | Makefile:1763: recipe for target 'check-am' failed
 | make: *** wait: No child processes.  Stop.
 | make: *** Waiting for unfinished jobs
 | make: *** wait: No child processes.  Stop.
 | Build killed with signal TERM after 150 minutes of inactivity

 It is not a real FTBFS indeed.
 The test 'hpxtest' cost to much time and armel has a quite poor performance, 
 so
 the test is killed by sbuild.

 So if we output some information when building, it will not be killed by 
 sbuild.

 in `test-drivers':

 $@ $log_file 21

 --

 $@ 21 | tee $log_file

 will fix this problem.



 Full build log:
   
 https://buildd.debian.org/status/fetch.php?pkg=healpix-cxxarch=armelver=3.11.2-7stamp=1405237595

 Mraw,
 KiBi.





-- 
YunQiang Su


healpix-cxx.debdiff
Description: Binary data
-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#779409: herwig++: Illegal instruction on mips64el when Herwig++ run LHC.run -N500 -d1 -D

[YunQiang Su]

::go(long, long, bool)
() from /usr/lib/mips64el-linux-gnuabi64/libThePEG.so.15
#29 0x000120004aac in HerwigRun (runname=..., seed=optimized
out, tag=..., N=500, tics=optimized out, resume=optimized out,
keepid=optimized out) at Herwig++.cc:228
#30 0x000120002d7c in main (argc=optimized out,
argv=0xffb418) at Herwig++.cc:125

-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#762775: pynfft: FTBFS: dh: unable to load addon sphinxdoc

[YunQiang Su]

On Wed, 24 Sep 2014 22:56:11 -0400 Aaron M. Ucko u...@debian.org wrote:
 Source: pynfft
 Version: 1.3.2-1
 Severity: serious
 Justification: fails to build from source

 Builds of pynfft in minimal environments geared for building its
 architecture-dependent binary packages (e.g., on the autobuilders)
 have been failing:

   dh clean --with python2,python3,sphinxdoc --buildsystem=pybuild
   dh: unable to load addon sphinxdoc: Can't locate 
 Debian/Debhelper/Sequence/sphinxdoc.pm in @INC (you may need to install the 
 Debian::Debhelper::Sequence::sphinxdoc module) (@INC contains: /etc/perl 
 /usr/local/lib/«arch»/perl/5.20.0 /usr/local/share/perl/5.20.0 
 /usr/lib/«arch»/perl5/5.20 /usr/share/perl5 /usr/lib/«arch»/perl/5.20 
 /usr/share/perl/5.20 /usr/local/lib/site_perl .) at (eval 13) line 2.
   BEGIN failed--compilation aborted at (eval 13) line 2.

   make: *** [clean] Error 2
   debian/rules:10: recipe for target 'clean' failed

 Could you please either conditionalize the usage of --with sphinxdoc
 appropriately or move python-sphinx into the main Build-Depends field?
 In the latter case, please bump the version to (= 1.2.2+dfsg-2~) to
 avoid running into errors if there is no actual documentation to
 install.  (See #745690.)

I will NMU it with the attached patch.


 Thanks!




pynfft.debdiff
Description: Binary data
-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#766134: coinutils: lapack and blas support is not really enabled

[YunQiang Su]

Package: src:coinutils
Version: 2.9.15-1
Severity: serious
Tag: patched

liblapack and libblas are in the building depends list while in build log,
there are lines:

checking whether -lblas has BLAS... no
checking for COIN-OR package Blas... skipped check via pkg-config,
redirect to fallback
checking for COIN-OR package Blas (fallback)... no, dependency
coinblas not available
checking whether -llapack has LAPACK... no
checking for COIN-OR package Lapack... skipped check via pkg-config,
redirect to fallback
checking for COIN-OR package Lapack (fallback)... no, dependency
coinlapack not available

If add gfortran to build-deps, it works well like this[1]

checking whether -lblas has BLAS... yes: -lblas
checking whether LAPACK is already available with BLAS library... no
checking whether -llapack has LAPACK... yes: -llapack


When blas and lapack enabled, coinor-libcoinutils-dev should depends on
liblapack-dev and libblas-dev. if not so, coinor-osi will ftbfs, as
the pkgconfig file
contains such informations.


[1]. 
http://mips64el.debian.net/debian/buildlog/c/coinutils_2.9.15-1/coinutils_2.9.15-1_mips64el-20140905-1939.build


diff -Nru coinutils-2.9.15/debian/control coinutils-2.9.15/debian/control
--- coinutils-2.9.15/debian/control 2014-09-04 19:47:26.0 +0800
+++ coinutils-2.9.15/debian/control 2034-12-15 13:37:17.0 +0800
@@ -4,7 +4,7 @@
 Maintainer: Debian Science Team
debian-science-maintainers@lists.alioth.debian.org
 Uploaders: Miles Lubin miles.lu...@gmail.com
 Build-Depends: debhelper (= 9), doxygen, graphviz, liblapack-dev,
- libbz2-dev, zlib1g-dev, autotools-dev
+ libbz2-dev, zlib1g-dev, autotools-dev, gfortran
 Standards-Version: 3.9.5
 Vcs-Git: git://anonscm.debian.org/debian-science/packages/coinutils.git
 Vcs-Browser: 
http://anonscm.debian.org/gitweb/?p=debian-science/packages/coinutils.git
@@ -31,7 +31,7 @@
 Package: coinor-libcoinutils-dev
 Section: libdevel
 Architecture: any
-Depends: ${shlibs:Depends}, ${misc:Depends}, coinor-libcoinutils3 (=
${binary:Version})
+Depends: ${shlibs:Depends}, ${misc:Depends}, coinor-libcoinutils3 (=
${binary:Version}), liblapack-dev, libblas-dev
 Provides: libcoinutils-dev
 Conflicts: libcoinutils-dev
 Replaces: libcoinutils-dev

-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#763007: fftw3: add mips64 and mips64el into long double support list

[YunQiang Su]

On Sat, 27 Sep 2014 10:02:09 +0800 YunQiang Su wzss...@gmail.com wrote:
 Package: fftw3
 Version: 3.3.4-1

 On mips64 and mips64el, double is 64bit and long double is 128bit.
 So please enable them for long double support.

I NMUed this package with delay-5, with arm64 and ppc64el to the long
double support list.


 --
 YunQiang Su




fftw3.debdiff
Description: Binary data
-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#746998: pyfftw: FTBFS on architectures without long double

[YunQiang Su]

On Thu, 22 May 2014 08:00:51 +0100 Ghislain Vaillant ghisv...@gmail.com wrote:
 Hi Aurelien,

 Thanks for reporting this issue.

 I'd have to check with upstream the best approach for dealing with this.
 I am not sure whether you can  enable / disable support for specific
 precisions in pyfftw. I think the author assumed that all these
 precisions were available.

 I could restrict the package build to selected architectures that
 provide all precisions for now, whilst upstream finds a solution to
 this. How does that sound ? In this case, I'd need to know which
 architectures are currently failing besides mips64.

This bug is about mipsel/mips aka 32bit mips, not about mips64(el).
You are confused due to he uses a 64bit kernel with 32bit userland.

I noticed that in fftw3, long double support is enabled on i386,
while the long double on i386 is 12Byte, quite strange,
as other ones are all 16Byte.


 Please let me know what you think,

 Cheers,

 Ghislain



-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#763007: fftw3: add mips64 and mips64el into long double support list

[YunQiang Su]

Package: fftw3
Version: 3.3.4-1

On mips64 and mips64el, double is 64bit and long double is 128bit.
So please enable them for long double support.

-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#755578: python-vtk: ImportError: invalid mode parameter on mips64el

[YunQiang Su]

Package: python-vtk
Version: 5.8.0-17.3

On mips64el platform

syq@thor python
Python 2.7.7 (default, Jun  6 2014, 13:33:55)
[GCC 4.9.0] on linux2
Type help, copyright, credits or license for more information.
 import vtk
Traceback (most recent call last):
  File stdin, line 1, in module
  File /usr/lib/pymodules/python2.7/vtk/__init__.py, line 41, in module
from vtkCommonPython import *
ImportError: invalid mode parameter
 from vtkCommonPython import *
Traceback (most recent call last):
  File stdin, line 1, in module
ImportError: No module named vtkCommonPython




on AMD64 platform

syq@wrath python
Python 2.7.7 (default, Jun  3 2014, 16:16:56)
[GCC 4.8.3] on linux2
Type help, copyright, credits or license for more information.
 import vtk

syq@wrath python
Python 2.7.7 (default, Jun  3 2014, 16:16:56)
[GCC 4.8.3] on linux2
Type help, copyright, credits or license for more information.
 from vtkCommonPython import *
Traceback (most recent call last):
  File stdin, line 1, in module
ImportError: No module named vtkCommonPython



-- 
YunQiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#748319: gmp: use ABI=64 on mips64(el) and fix symbol file

[Yunqiang Su]

Package: gmp
Version: 6.0.0+dfsg-4

This patch pass ABI=64 to configure and fix symbols on mips64(el).

-- 
Yunqiang Su


gmp-mips64.patch
Description: Binary data
-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#746277: arpack failed to build on mips64el due to missing in symbol files

[Yunqiang Su]

Package: arpack
Version: 3.1.5-2

debian/libparpack2.symbols
there are some symbols marked as

!mips !mipsel !mips64 !mips64el !s390 !s390x !powerpcspe !m68k !arm64
!hppa !ppc64 !sh4 !sparc64 !x32

Please add mips64 and mips64el into them.


-- 
Yunqiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Bug#743497: cpl-plugin-xsh: run test failed on mips64el

[Yunqiang Su]

Package: cpl-plugin-xsh
Version: 2.3.0+dfsg-3

buildd@lm6100:/tmp/cpl-plugin-xsh/cpl-plugin-xsh-2.3.0+dfsg/xsh/tests$
./test_xsh_correl_gaussians
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : STEP_MIN=-21 maxpos=0
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : a=1.74405e-321 b=nan c=nan
fraction=nan
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : maxpos=0 width_i=100
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : fractionary delta=nan
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : function1d corerel: shift -21
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: :   mk4 short_len=57
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: :   correl max[0]=-1
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: :   nsteps 43
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: :   fraction=nan
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : moses_shift=nan
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : shift=99
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : correl max=0
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : Shift: expected: 5.15
computed: -21 accuracy:5.07767
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : max=5.43037e-312 maxpos=98
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : maxpos my determination: 198
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : len_corr=199
[ INFO  ] Test-XSH_CORRELATE_GAUSSIANS: : a=0 b=-3 c=7.90505e-321 fraction=-0.5
[ ERROR ] Test-XSH_CORRELATE_GAUSSIANS: main: An error occurred,
dumping error trace:
[ ERROR ] Test-XSH_CORRELATE_GAUSSIANS: main:
[ ERROR ] Test-XSH_CORRELATE_GAUSSIANS: main: Access beyond boundaries
[ ERROR ] Test-XSH_CORRELATE_GAUSSIANS: main:  in [1]cpl_vector_get()
at cpl_vector.c:829
[ ERROR ] Test-XSH_CORRELATE_GAUSSIANS: main:


-- 
Yunqiang Su

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers