matplotlib-devel Mailing List for matplotlib

Chris <ceball@...> writes:

> 
> Hi,
> 
> I posted this same message to matplotlib-users a couple of days ago, but now I
> realize that this list is the more appropriate place.
> 
> I'm having trouble building matplotlib 0.99.1.1 (transcript below). 
> 
> I'm using copies of Python (2.5.1) and Tcl/Tk (8.5.5) that I have
> built myself, and that are apparently working fine. I can use this 
> exact procedure to build 0.91.4 without any problems.
> 
> Any suggestions would be greatly appreciated - thanks!
> 
> Chris
> 
> $ cd matplotlib-0.99.1.1
> $ env PREFIX=/a/b/ LD_LIBRARY_PATH=/a/b/lib  /a/b/bin/python setup.py build

I'm not sure what changed in matplotlib between 0.91.4 and 0.99.1.1, but the 
procedure below seems to work for me:

$ tar -xzf matplotlib-0.99.1.1.tar.gz
$ cd matplotlib-0.99.1.1/
$ rm setup.cfg
$ /a/b/bin/python setup.py build_ext -L$/a/b/lib/
$ env PREFIX=/a/b/ LD_LIBRARY_PATH=/a/b/lib /a/b/bin/python setup.py build

Chris

Michael Oevermann <mic...@tu...> writes:

> is there a way to use patterns instead of colors for pie and bar
> charts and for filling in general? It would be very useful to have
> this feature for printing (and publishing) in black and white or
> greyscale.

Yes: http://matplotlib.sourceforge.net/examples/pylab_examples/hatch_demo.html

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Hi,

is there a way to use patterns instead of colors for pie and bar charts 
and for
filling in general? It would be very useful to have this feature for 
printing
(and publishing) in black and white or greyscale.

Best regards

Michael

Hi,

I posted this same message to matplotlib-users a couple of days ago, but now I
realize that this list is the more appropriate place.

I'm having trouble building matplotlib 0.99.1.1 (transcript below). 

I'm using copies of Python (2.5.1) and Tcl/Tk (8.5.5) that I have
built myself, and that are apparently working fine. I can use this 
exact procedure to build 0.91.4 without any problems.

Any suggestions would be greatly appreciated - thanks!

Chris

$ cd matplotlib-0.99.1.1
$ env PREFIX=/a/b/ LD_LIBRARY_PATH=/a/b/lib  /a/b/bin/python setup.py build
============================================================================
BUILDING MATPLOTLIB
            matplotlib: 0.99.1.1
                python: 2.5.1 (r251:54863, Feb  5 2009, 13:11:08)  [GCC
                        4.1.2 20071124 (Red Hat 4.1.2-42)]
              platform: linux2

REQUIRED DEPENDENCIES
                 numpy: 1.2.1
             freetype2: 9.10.3

OPTIONAL BACKEND DEPENDENCIES
                libpng: 1.2.10
               Tkinter: Tkinter: 50704, Tk: 8.5, Tcl: 8.5
                  Gtk+: no
                        * Building for Gtk+ requires pygtk; you must be able
                        * to "import gtk" in your build/install environment
       Mac OS X native: no
                    Qt: no
                   Qt4: no
                 Cairo: no

OPTIONAL DATE/TIMEZONE DEPENDENCIES
              datetime: present, version unknown
              dateutil: matplotlib will provide
                  pytz: matplotlib will provide
adding pytz

OPTIONAL USETEX DEPENDENCIES
                dvipng: 1.5
           ghostscript: 8.15.2
                 latex: 3.141592
               pdftops: 3.00

[Edit setup.cfg to suppress the above messages]
============================================================================
pymods ['pylab']
packages ['matplotlib', 'matplotlib.backends', 'matplotlib.projections',
'mpl_toolkits', 'mpl_toolkits.mplot3d', 'mpl_too\
lkits.axes_grid', 'matplotlib.sphinxext', 'matplotlib.numerix',
'matplotlib.numerix.mlab', 'matplotlib.numerix.ma', 'matp\
lotlib.numerix.linear_algebra', 'matplotlib.numerix.random_array',
'matplotlib.numerix.fft', 'matplotlib.delaunay', 'pytz\
', 'dateutil', 'dateutil/zoneinfo']
running build
running build_py
copying lib/matplotlib/mpl-data/matplotlibrc ->
build/lib.linux-i686-2.5/matplotlib/mpl-data
copying lib/matplotlib/mpl-data/matplotlib.conf ->
build/lib.linux-i686-2.5/matplotlib/mpl-data
running build_ext
building 'matplotlib.backends._tkagg' extension
g++ -pthread -shared build/temp.linux-i686-2.5/src/agg_py_transforms.o
build/temp.linux-i686-2.5/src/_tkagg.o build/temp.\
linux-i686-2.5/CXX/cxx_extensions.o build/temp.linux-i686-2.5/CXX/cxxsupport.o
build/temp.linux-i686-2.5/CXX/IndirectPyth\
onInterface.o build/temp.linux-i686-2.5/CXX/cxxextensions.o -L/usr/lib
-L/usr/lib -L/usr/local/lib -L/usr/lib -L/usr/lib6\
4 -L/usr/local/lib -L/usr/lib -L/usr/lib64 -ltk8.5 -ltcl8.5 -lstdc++ -lm
-lfreetype -lz -lstdc++ -lm -o build/lib.linux-i\
686-2.5/matplotlib/backends/_tkagg.so
/usr/bin/ld: cannot find -ltk8.5
collect2: ld returned 1 exit status
error: command 'g++' failed with exit status 1
make: *** [matplotlib] Error 1

$ ls /a/b/lib/*tk*
lib/libtk8.5.so  lib/libtkstub8.5.a  lib/tkConfig.sh
...

On Tue, Sep 29, 2009 at 3:07 PM, Barry Scott <ba...@ba...> wrote:
>
> On 28 Sep 2009, at 17:27, William Stein wrote:
>
>> On Mon, Sep 28, 2009 at 9:14 AM, John Hunter <jd...@gm...> wrote:
>>>
>>> On Mon, Sep 28, 2009 at 10:48 AM, William Stein <ws...@gm...> wrote:
>>>>
>>>> If you could just try running your matplotlib build on
>>>> bsd.math.washington.edu and reporting about whether or not it works,
>>>> that would be very helpful, since it will indicate whether the problem
>>>> is somehow in Sage or in pyCXX or Matplotlib itself or something else.
>>>
>>> I was able to build into a mostly clean environment (rebuild all deps
>>> zlip, freetype, png) using my OSX make file in, but I had to update
>>> the MACOSX_DEPLOYMENT_TARGET  to 10.6 (I've committed this to svn
>>> HEAD).
>>>
>>>
>>> PREFIX=~/devtest make -f make.osx fetch deps mpl_build mpl_install
>>>
>>> But even simple tests are failing with::
>>>
>>> jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
>>> PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
>>> 'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
>>> import *; plot([1,2,3]); savefig("test")'
>>> Traceback (most recent call last):
>>> File "<string>", line 1, in <module>
>>> File
>>> "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/pyplot.py",
>>> line 7, in <module>
>>>  from matplotlib.figure import Figure, figaspect
>>> File
>>> "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/figure.py",
>>> line 16, in <module>
>>>  import artist
>>> File
>>> "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/artist.py",
>>> line 6, in <module>
>>>  from transforms import Bbox, IdentityTransform, TransformedBbox,
>>> TransformedPath
>>> File
>>> "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/transforms.py",
>>> line 34, in <module>
>>>  from matplotlib._path import affine_transform
>>> ImportError:
>>> /Users/jdh2358/devtest/lib/python2.6/site-packages/matplotlib/_path.so:
>>> no appropriate 64-bit architecture (see "man python" for running in
>>> 32-bit mode)
>
> This error has nothing to do the PyCXX isn't it a build problem not crating
> a fat binary with -arch x86_64 for _path.so. Use lipo -info to see what is
> inside _path.so.
>
>>>
>>> I'm attaching my build output in case anyone sees anything that might
>>> be triggering this 32bit/64bit problem (see attached for full output).
>>> I did not rebuild numpy and this may be the problem since the failure
>>> is in the _path module.  I'll give that a try next
>>
>> I think it's definitely an issue with PyCXX.   I installed the latest
>> version of PyCXX and tried their test example program, and got the
>> dreaded "Abort trap":
>
>
>>
>> bash-3.2$ python
>> Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)
>> [GCC 4.2.1 (Apple Inc. build 5646)] on darwin
>> Type "help", "copyright", "credits" or "license" for more information.
>>>>>
>>>>> import CXX.example
>>
>> CXX is installed.
>> The support files you need are in the PYTHON/etc/CXX directory.
>> The include files are in the distutils include path already.
>> Just refer to them as "CXX/CXX_Objects.h", etc.
>>
>> range object created 0x1003140b0
>> range object destroyed 0x1003140b0
>>>>>
>>>>> CXX.example.test()
>>
>> Example Test starting
>> Trying to convert a NULL to an Py::Int
>> Abort trap
>
>> bash-3.2$ pwd
>>
>> /Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/matplotlib/pycxx-6.1.1/Demo/Python2
>>
>> ----
>>
>> Note that installing PyCXX and their test program was not easy.  I
>> watched their "python setup.py install" fail repeatedly, then I would
>> copy missing files where they should go, and iterate (about 6 times
>
> I never use the setup.py - I keep trying to delete it from the kits but have
> users want
> it and contribute fixes.
>
> I do not recommend installing PyCXX into a system for a couple of reasons.
>
> On one system you may well end up using more then one version of PyCXX.
> For example one older version of maintenance and a newer version for new
> development.
>
> YOu may have more then one project using PyCXX that need compiler options
> that conflict. If you install PyCXX into the system the object code will be
> wrong
> so some projects.
>
> Your experience is convincing me to delete setup.py to avoid this in future.
>
> However the tests do all work on 10.6. You will need a new make file to use
> Apple's python 2.6.1. The PyCXX 6.1.1 kit does not have one to do this.
>
>
>
>
> Save it to the top of the PyCXX files, where all the other make files are.
>
> Then you can build and run all tests with:
>
> $ make -f example_mac_apple_py26.mak clean test
>
> All the tests will pass without error.
>
> Attached is the log of the build and test on Mac OS X 10.6.1 I ran just now
> to confirm this.
>
>
>
>
>
> I suspect that the PyCXX that you installed with setup.py is not correct.
> I suggest you delete it from out of the system and point directly to the
> headers and C and C++ files you need out of a exploded tarball.
>
> This is what I do for PySVN for example.
>
>>
>>
>> I've cc'd the PyCXX maintainer Barry Scott.  Barry, do you have access
>> to an OS X 10.6 install, since it seems PyCXX is broken there?  If
>> not, I can give you an account on a machine at University of
>> Washington.   PyCXX is the single component of Sage
>> (http://sagemath.org) that is holding us back from finishing the OS X
>> 10.6 port.
>>
>> -- William
>
> Let me know how else I can help.

Thanks for all your ideas.   Inspired by your remarks I tried again
and the first thing I attempted was to take matplotlib's build lines
for building ft2font (which uses PyCXX) and just paste them into my
console instead of letting them be built under setup.py.  Amazingly
enough doing that completely fixes all the problems.  That suggests
there is something amis with the environment that Python's distutils
and matplotlib create when building ft2font + pyCXX.

My solution for Sage for now will be to rerun those lines explicitly
when building Sage on OS X 10.6, and hope for John Hunter to be able
to replicate the problem and come up with something better in the next
few weeks (or I'll come back to it later).

William

On 28 Sep 2009, at 17:27, William Stein wrote:

> On Mon, Sep 28, 2009 at 9:14 AM, John Hunter <jd...@gm...>  
> wrote:
>> On Mon, Sep 28, 2009 at 10:48 AM, William Stein <ws...@gm...>  
>> wrote:
>>> If you could just try running your matplotlib build on
>>> bsd.math.washington.edu and reporting about whether or not it works,
>>> that would be very helpful, since it will indicate whether the  
>>> problem
>>> is somehow in Sage or in pyCXX or Matplotlib itself or something  
>>> else.
>>
>> I was able to build into a mostly clean environment (rebuild all deps
>> zlip, freetype, png) using my OSX make file in, but I had to update
>> the MACOSX_DEPLOYMENT_TARGET  to 10.6 (I've committed this to svn
>> HEAD).
>>
>>
>> PREFIX=~/devtest make -f make.osx fetch deps mpl_build mpl_install
>>
>> But even simple tests are failing with::
>>
>> jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
>> PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
>> 'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
>> import *; plot([1,2,3]); savefig("test")'
>> Traceback (most recent call last):
>> File "<string>", line 1, in <module>
>> File "/Users/jdh2358/devtest//lib/python2.6/site-packages/ 
>> matplotlib/pyplot.py",
>> line 7, in <module>
>>   from matplotlib.figure import Figure, figaspect
>> File "/Users/jdh2358/devtest//lib/python2.6/site-packages/ 
>> matplotlib/figure.py",
>> line 16, in <module>
>>   import artist
>> File "/Users/jdh2358/devtest//lib/python2.6/site-packages/ 
>> matplotlib/artist.py",
>> line 6, in <module>
>>   from transforms import Bbox, IdentityTransform, TransformedBbox,
>> TransformedPath
>> File "/Users/jdh2358/devtest//lib/python2.6/site-packages/ 
>> matplotlib/transforms.py",
>> line 34, in <module>
>>   from matplotlib._path import affine_transform
>> ImportError: /Users/jdh2358/devtest/lib/python2.6/site-packages/ 
>> matplotlib/_path.so:
>> no appropriate 64-bit architecture (see "man python" for running in
>> 32-bit mode)

This error has nothing to do the PyCXX isn't it a build problem not  
crating
a fat binary with -arch x86_64 for _path.so. Use lipo -info to see  
what is
inside _path.so.

>>
>> I'm attaching my build output in case anyone sees anything that might
>> be triggering this 32bit/64bit problem (see attached for full  
>> output).
>> I did not rebuild numpy and this may be the problem since the failure
>> is in the _path module.  I'll give that a try next
>
> I think it's definitely an issue with PyCXX.   I installed the latest
> version of PyCXX and tried their test example program, and got the
> dreaded "Abort trap":


>
> bash-3.2$ python
> Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)
> [GCC 4.2.1 (Apple Inc. build 5646)] on darwin
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import CXX.example
> CXX is installed.
> The support files you need are in the PYTHON/etc/CXX directory.
> The include files are in the distutils include path already.
> Just refer to them as "CXX/CXX_Objects.h", etc.
>
> range object created 0x1003140b0
> range object destroyed 0x1003140b0
>>>> CXX.example.test()
> Example Test starting
> Trying to convert a NULL to an Py::Int
> Abort trap

> bash-3.2$ pwd
> /Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/ 
> matplotlib/pycxx-6.1.1/Demo/Python2
>
> ----
>
> Note that installing PyCXX and their test program was not easy.  I
> watched their "python setup.py install" fail repeatedly, then I would
> copy missing files where they should go, and iterate (about 6 times

I never use the setup.py - I keep trying to delete it from the kits  
but have users want
it and contribute fixes.

I do not recommend installing PyCXX into a system for a couple of  
reasons.

On one system you may well end up using more then one version of PyCXX.
For example one older version of maintenance and a newer version for  
new development.

YOu may have more then one project using PyCXX that need compiler  
options
that conflict. If you install PyCXX into the system the object code  
will be wrong
so some projects.

Your experience is convincing me to delete setup.py to avoid this in  
future.

However the tests do all work on 10.6. You will need a new make file  
to use
Apple's python 2.6.1. The PyCXX 6.1.1 kit does not have one to do this.

John Hunter <jdh2358@...> writes:

> 
> On Wed, Sep 23, 2009 at 8:47 AM, Lev Givon <lev@...> wrote:
> > contain a setup.cfg file:
> >
> > $ tar zft matplotlib-0.99.1.tar.gz |grep setup.cfg
> > matplotlib-0.99.1/setup.cfg
> > matplotlib-0.99.1/setup.cfg.template
> 
> It seems to depend on which mirror you get the file from.  From Voxel,
> I see setup.cfg but from "German Research Network (Berlin, Germany) "
> I do not see it.  We may just need time for the mirrors to update.  I
> probably should have used a different file name...

John Hunter <jdh2358@...> writes:

> 
> On Wed, Sep 23, 2009 at 8:47 AM, Lev Givon <lev@...> wrote:
> > contain a setup.cfg file:
> >
> > $ tar zft matplotlib-0.99.1.tar.gz |grep setup.cfg
> > matplotlib-0.99.1/setup.cfg
> > matplotlib-0.99.1/setup.cfg.template
> 
> It seems to depend on which mirror you get the file from.  From Voxel,
> I see setup.cfg but from "German Research Network (Berlin, Germany) "
> I do not see it.  We may just need time for the mirrors to update.  I
> probably should have used a different file name...

I also find setup.cfg in 0.99.1.1.tar.gz, and had to delete it to begin 
building:

$ wget 
http://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-
0.99.1/matplotlib-0.99.1.1.tar.gz?use_mirror=kent
...
20:44:41 (2.11 MB/s) - `matplotlib-0.99.1.1.tar.gz' saved [11905737/11905737]

$ tar xf matplotlib-0.99.1.1.tar.gz 
$ cd matplotlib-0.99.1.1
$ ls setup.cfg
setup.cfg


Chris

Hello Eric,

The functions that I've created make it possible to generate a discrete 
(piecewise-constant) or continuous (piecewise-linear) colormap and 
register it at a single shot.  These functions also accept a list of 
thresholds if the user wants to specify non-default thresholds.  It 
seems as though the best alternatives require multiple steps and a 
considerable amount of fiddling.  Furthermore, my functions are 
thoroughly documented, while the alternatives are not.  I've found that 
it often takes less time to write something myself than to figure out 
how to use an undocumented function (which may not even do what I 
need).  If you agree that my functions provide a better interface, then 
I hope that you and John Hunter can get them incorporated into matplotlib.

Yours,

Phillip

Eric Firing wrote:
> Dr. Phillip M. Feldman wrote:
>> After experimenting with colormaps for a while, I was able to make both
>> discrete (piecewise-constant) and continuous (piecewise-linear) 
>> colormaps
>> work. Although colormaps can be created directly using
>> LinearSegmentedColormap from the matplotlib.colors package, this is a
>> tedious and error-prone process. So, I compiled a set of three interface
>> functions. (I wrote two of these myself, and got one from the SciPy
>> website). The two functions that I wrote permit one to define a discrete
>> (piecewise-constant) and continuous (piecewise-linear) colormap 
>> directly via
>> a sequence of colors and a set of thresholds specified as lists. Each 
>> color
>> may be specified either via an RGB tuple or via an English color name 
>> known
>> to webcolors.name_to_rgb. I'm going to submit all of this to the 
>> matplotlib
>> developers forums in the hopes of getting it incorporated into 
>> matplotlib.
>
> Something I should have mentioned earlier: for the discrete case, one 
> good option is to use a ListedColormap to make a map with only a few 
> colors, and then use a BoundaryNorm to handle the mapping from data 
> values to colors.  See 
> http://matplotlib.sourceforge.net/examples/pylab_examples/image_masked.html?highlight=image_masked 
>
> for a BoundaryNorm example.  Unfortunately, we don't have an example 
> of using it with a ListedColormap, but it should be straigtforward; 
> the key point is to set the BoundaryNorm ncolors kwarg to match the 
> actual number of colors in the colormap.
>
> Eric
>

After experimenting with colormaps for a while, I was able to make both
discrete (piecewise-constant) and continuous (piecewise-linear) colormaps
work. Although colormaps can be created directly using
LinearSegmentedColormap from the matplotlib.colors package, this is a
tedious and error-prone process. So, I compiled a set of three interface
functions. (I wrote two of these myself, and got one from the SciPy
website). The two functions that I wrote permit one to define a discrete
(piecewise-constant) and continuous (piecewise-linear) colormap directly via
a sequence of colors and a set of thresholds specified as lists. Each color
may be specified either via an RGB tuple or via an English color name known
to webcolors.name_to_rgb. I'm submitting this in the hopes that this
functionality will be incorporated into matplotlib.

There's one thing that I've not been able to resolve: When attempting to
display all defined colormaps using colormaps_test.py (attached), colormaps
that I've defined do not show up. So, it would appear that registering a
colormap makes it possible to access that colormap only if the name is known
to the program that wants to use it.
http://www.nabble.com/file/p25665294/make_cmap.py make_cmap.py 
http://www.nabble.com/file/p25665294/colormaps_test.py colormaps_test.py 
-- 
View this message in context: http://www.nabble.com/new-interface-functions-for-LinearSegmentedColormap-tp25665294p25665294.html
Sent from the matplotlib - devel mailing list archive at Nabble.com.

Jouni K. Seppänen <jk...@ik...> writes:

> I committed something based on this, and a new rc parameter
> savefig.extension that sets the filename extension when you call savefig
> with a bare filename. The pdf tests seem to be working, at least for me,
> but I am sure that the code can be improved.

The buildbot was getting errors, since the build environments don't have
gs. I changed the tests so that this isn't an error. It might be better
to make it a known fail, but is it possible for the image comparison
decorator to turn one test function into several cases? I.e., the png
case could be pass/fail, and the pdf case a known fail if there is no
Ghostscript.

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

On Mon, Sep 28, 2009 at 9:35 AM, John Hunter <jd...@gm...> wrote:
> On Mon, Sep 28, 2009 at 11:14 AM, John Hunter <jd...@gm...> wrote:
>
>> But even simple tests are failing with::
>>
>> jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
>> PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
>> 'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
>> import *; plot([1,2,3]); savefig("test")'
>> Traceback (most recent call last):
>>  File "<string>", line 1, in <module>
>>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/pyplot.py",
>> line 7, in <module>
>>    from matplotlib.figure import Figure, figaspect
>>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/figure.py",
>> line 16, in <module>
>>    import artist
>>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/artist.py",
>> line 6, in <module>
>>    from transforms import Bbox, IdentityTransform, TransformedBbox,
>> TransformedPath
>>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/transforms.py",
>> line 34, in <module>
>>    from matplotlib._path import affine_transform
>> ImportError: /Users/jdh2358/devtest/lib/python2.6/site-packages/matplotlib/_path.so:
>> no appropriate 64-bit architecture (see "man python" for running in
>> 32-bit mode)
>>
>> I'm attaching my build output in case anyone sees anything that might
>> be triggering this 32bit/64bit problem (see attached for full output).
>>  I did not rebuild numpy and this may be the problem since the failure
>> is in the _path module.  I'll give that a try next
>
>
> Same issue with numpy HEAD

One thing to keep in mind is that the default for GCC on OS X 10.6 is
to build 64-bit binaries.   With OS X 10.5 the default for GCC was to
make 32-bit binaries.  (To get 64-bit you used to have to do "-m64",
but that is now the default.)   So you really have to start from
scratch.

 -- William

On Mon, Sep 28, 2009 at 11:14 AM, John Hunter <jd...@gm...> wrote:

> But even simple tests are failing with::
>
> jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
> PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
> 'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
> import *; plot([1,2,3]); savefig("test")'
> Traceback (most recent call last):
>  File "<string>", line 1, in <module>
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/pyplot.py",
> line 7, in <module>
>    from matplotlib.figure import Figure, figaspect
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/figure.py",
> line 16, in <module>
>    import artist
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/artist.py",
> line 6, in <module>
>    from transforms import Bbox, IdentityTransform, TransformedBbox,
> TransformedPath
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/transforms.py",
> line 34, in <module>
>    from matplotlib._path import affine_transform
> ImportError: /Users/jdh2358/devtest/lib/python2.6/site-packages/matplotlib/_path.so:
> no appropriate 64-bit architecture (see "man python" for running in
> 32-bit mode)
>
> I'm attaching my build output in case anyone sees anything that might
> be triggering this 32bit/64bit problem (see attached for full output).
>  I did not rebuild numpy and this may be the problem since the failure
> is in the _path module.  I'll give that a try next


Same issue with numpy HEAD

On Mon, Sep 28, 2009 at 9:14 AM, John Hunter <jd...@gm...> wrote:
> On Mon, Sep 28, 2009 at 10:48 AM, William Stein <ws...@gm...> wrote:
>> If you could just try running your matplotlib build on
>> bsd.math.washington.edu and reporting about whether or not it works,
>> that would be very helpful, since it will indicate whether the problem
>> is somehow in Sage or in pyCXX or Matplotlib itself or something else.
>
> I was able to build into a mostly clean environment (rebuild all deps
> zlip, freetype, png) using my OSX make file in, but I had to update
> the MACOSX_DEPLOYMENT_TARGET  to 10.6 (I've committed this to svn
> HEAD).
>
>
>  PREFIX=~/devtest make -f make.osx fetch deps mpl_build mpl_install
>
> But even simple tests are failing with::
>
> jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
> PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
> 'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
> import *; plot([1,2,3]); savefig("test")'
> Traceback (most recent call last):
>  File "<string>", line 1, in <module>
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/pyplot.py",
> line 7, in <module>
>    from matplotlib.figure import Figure, figaspect
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/figure.py",
> line 16, in <module>
>    import artist
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/artist.py",
> line 6, in <module>
>    from transforms import Bbox, IdentityTransform, TransformedBbox,
> TransformedPath
>  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/transforms.py",
> line 34, in <module>
>    from matplotlib._path import affine_transform
> ImportError: /Users/jdh2358/devtest/lib/python2.6/site-packages/matplotlib/_path.so:
> no appropriate 64-bit architecture (see "man python" for running in
> 32-bit mode)
>
> I'm attaching my build output in case anyone sees anything that might
> be triggering this 32bit/64bit problem (see attached for full output).
>  I did not rebuild numpy and this may be the problem since the failure
> is in the _path module.  I'll give that a try next

I think it's definitely an issue with PyCXX.   I installed the latest
version of PyCXX and tried their test example program, and got the
dreaded "Abort trap":

bash-3.2$ python
Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)
[GCC 4.2.1 (Apple Inc. build 5646)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import CXX.example
CXX is installed.
The support files you need are in the PYTHON/etc/CXX directory.
The include files are in the distutils include path already.
Just refer to them as "CXX/CXX_Objects.h", etc.

range object created 0x1003140b0
range object destroyed 0x1003140b0
>>> CXX.example.test()
Example Test starting
Trying to convert a NULL to an Py::Int
Abort trap
bash-3.2$ pwd
/Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/matplotlib/pycxx-6.1.1/Demo/Python2

----

Note that installing PyCXX and their test program was not easy.  I
watched their "python setup.py install" fail repeatedly, then I would
copy missing files where they should go, and iterate (about 6 times


I've cc'd the PyCXX maintainer Barry Scott.  Barry, do you have access
to an OS X 10.6 install, since it seems PyCXX is broken there?  If
not, I can give you an account on a machine at University of
Washington.   PyCXX is the single component of Sage
(http://sagemath.org) that is holding us back from finishing the OS X
10.6 port.

 -- William

On Mon, Sep 28, 2009 at 10:48 AM, William Stein <ws...@gm...> wrote:
> If you could just try running your matplotlib build on
> bsd.math.washington.edu and reporting about whether or not it works,
> that would be very helpful, since it will indicate whether the problem
> is somehow in Sage or in pyCXX or Matplotlib itself or something else.

I was able to build into a mostly clean environment (rebuild all deps
zlip, freetype, png) using my OSX make file in, but I had to update
the MACOSX_DEPLOYMENT_TARGET  to 10.6 (I've committed this to svn
HEAD).


  PREFIX=~/devtest make -f make.osx fetch deps mpl_build mpl_install

But even simple tests are failing with::

jdh2358@bsd:~> LD_LIBRARY_PATH=~/devtest/lib/
PYTHONPATH=~/devtest/lib/python2.6/site-packages/ /usr/bin/python -c
'import matplotlib; matplotlib.use("Agg"); from matplotlib.pyplot
import *; plot([1,2,3]); savefig("test")'
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/pyplot.py",
line 7, in <module>
    from matplotlib.figure import Figure, figaspect
  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/figure.py",
line 16, in <module>
    import artist
  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/artist.py",
line 6, in <module>
    from transforms import Bbox, IdentityTransform, TransformedBbox,
TransformedPath
  File "/Users/jdh2358/devtest//lib/python2.6/site-packages/matplotlib/transforms.py",
line 34, in <module>
    from matplotlib._path import affine_transform
ImportError: /Users/jdh2358/devtest/lib/python2.6/site-packages/matplotlib/_path.so:
no appropriate 64-bit architecture (see "man python" for running in
32-bit mode)

I'm attaching my build output in case anyone sees anything that might
be triggering this 32bit/64bit problem (see attached for full output).
 I did not rebuild numpy and this may be the problem since the failure
is in the _path module.  I'll give that a try next

jdh2358@bsd:mpl> head -30 build.out
export PKG_CONFIG_PATH="/Users/jdh2358/devtest//lib/pkgconfig" &&\
        export MACOSX_DEPLOYMENT_TARGET=10.6 &&\
        export CFLAGS="-arch i386 -arch ppc
-I/Users/jdh2358/devtest//include
-I/Users/jdh2358/devtest//include/freetype2 -isysroot
/Developer/SDKs/MacOSX10.6.sdk" &&\
        export LDFLAGS="-arch i386 -arch ppc
-L/Users/jdh2358/devtest//lib
-syslibroot,/Developer/SDKs/MacOSX10.6.sdk" &&\
        python2.6 setup.py build
============================================================================
BUILDING MATPLOTLIB
            matplotlib: 1.0.svn
                python: 2.6.1 (r261:67515, Jul  7 2009, 23:51:51)  [GCC
                        4.2.1 (Apple Inc. build 5646)]
              platform: darwin

REQUIRED DEPENDENCIES
                 numpy: 1.2.1
             freetype2: found, but unknown version (no pkg-config)
                        * WARNING: Could not find 'freetype2' headers in any
                        * of '.', './freetype2'.

OPTIONAL BACKEND DEPENDENCIES
                libpng: found, but unknown version (no pkg-config)
                        * Could not find 'libpng' headers in any of '.'
               Tkinter: Tkinter: 67083, Tk: 8.5, Tcl: 8.5
              wxPython: no
                        * wxPython not found
                  Gtk+: no
                        * Building for Gtk+ requires pygtk; you must be able
                        * to "import gtk" in your build/install environment
       Mac OS X native: yes
                    Qt: no
                   Qt4: no

On Mon, Sep 28, 2009 at 8:55 AM, William Stein <ws...@gm...> wrote:
> On Mon, Sep 28, 2009 at 8:47 AM, John Hunter <jd...@gm...> wrote:
>> ---------- Forwarded message ----------
>> From: Tony S Yu <ts...@gm...>
>> Date: Mon, Sep 28, 2009 at 8:41 AM
>> Subject: Re: [matplotlib-devel] Fwd: OS X 10.6 port
>> To: John Hunter <jd...@gm...>
>> Cc: matplotlib development list <mat...@li...>
>>
>>
>>
>> On Sep 28, 2009, at 2:14 AM, John Hunter wrote:
>>
>>> in case anyone has some suggestions, I'm forwarding this from the sage list
>>>
>>>
>>> ---------- Forwarded message ----------
>>> From: William Stein <ws...@gm...>
>>> Date: Sun, Sep 27, 2009 at 10:51 PM
>>> Subject: OS X 10.6 port
>>> To: sage-devel <sag...@go...>, John Hunter <jd...@gm...>
>>>
>>>
>>> Hi,
>>>
>>> I spent several hours yesterday trying to get matplotlib for Sage to
>>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>>> another test machine (bsd.math) the workaround from my laptop doesn't
>>> work.  So at this point Sage still does not support OS X 10.6.
>>>
>>> The problem is in matplotlib's C++ wrapper for freetype2.  It's a
>>> Python extension module implemented directly in C++, and it simply
>>> doesn't work correctly at all.  For example, whenever it tries to
>>> raise a Python exception, it just terminates Python with
>>>
>>> "11713 Abort trap              sage-ipython "$@" -i"
>
> The stack trace looks like this:
>
> #4  0x000000010cd907e2 in ~PythonExtension [inlined] () at
> /Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/matplotlib/CXX/Extensions.hxx:778
> #5  0x000000010cd907e2 in FT2Font::FT2Font (this=0x10baf2b10,
> facefile=@0x10bab6da0) at src/ft2font.cpp:778
> #6  0x000000010cd90e98 in ft2font_module::new_ft2font (this=<value
> temporarily unavailable, due to optimizations>, args=<value
> temporarily unavailable, due to optimizations>) at src/ft2font.cpp:778
> #7  0x000000010cd94912 in
> Py::ExtensionModule<ft2font_module>::invoke_method_varargs
> (this=0x10ce05f10, name=@0x7fff5fbf7030, args=@0x7fff5fbf6fe0) at
> Extensions.hxx:422
> #8  0x000000010cda3190 in method_varargs_call_handler
> (_self_and_name_tuple=<value temporarily unavailable, due to
> optimizations>, _args=<value temporarily unavailable, due to
> optimizations>) at CXX/cxx_extensions.cxx:1403
> ...
>
>
>> I thought I'd chime in since I recently upgraded to OSX 10.6. I don't
>> think I ever got this error, so I probably didn't have the same issue,
>> but just in case...
>>
>> I reinstalled freetype2, but I'm not sure if I needed to. I also made
>> a change to setupext.py so that the setup script could find freetype2,
>> but this doesn't seem to be your problem because your errors don't
>> occur during the build.
>>
>> Did you happen to build on top of an old install? I had to clean out
>> the old compiled extensions before everything would build properly. I
>> had to clean out the build directory (obviously), but there are also
>> *.so files in lib/matplotlib and lib/matplotlib/backends.
>>
>> I should note that I'm built on top of the system python (2.6.1) and I
>> did not use make.osx.
>>
>
> I did a 100% clean build from source of Sage, which means building
> Python and essentially all dependencies (including freetype) in a
> single self-contained directory.    I've replicated this on maybe 5 or
> 6 different OS X 10.6 machines.  It is a total show stopper for
> releasing the next version of Sage, and is in fact the only problem
> left (of many) in porting Sage to OS X 10.6
>
> I used the latest version of XCode, which comes with OS X 10.6:
>  gcc version 4.2.1 (Apple Inc. build 5646)
>
> The version of Python  is:
>        Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)
>
> The Abort trap appears when building the font cache the moment an
> exception is raised in C++ code.  By hacking around that exception
> getting raised, the Abort trap happens when some exception somewhere
> else happens in Matplotlib's C++ code.  Etc.
>
> I'm going to try looking into pyCXX now, since that seems the best hope.
>

I rebuild matplotlib using the newest version of pyCXX (which has seen
4 new releases since the version in matplotlib), but the problem
remains the same.  (I've also tried using the latest version of
libfreetype.)

bash-3.2$ python
Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)
[GCC 4.2.1 (Apple Inc. build 5646)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import pylab
Abort trap

-----

(gdb) bt
#0  0x00007fff82adf096 in __kill ()
#1  0x00007fff82b800c2 in abort ()
#2  0x00000001015d0d51 in uw_init_context_1 ()
#3  0x00000001015d14cc in _Unwind_Resume ()
#4  0x000000010182f7f3 in ~PythonExtension [inlined] () at
/Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/matplotlib/CXX/Python2/ExtensionOldType.hxx:778
#5  0x000000010182f7f3 in FT2Font::FT2Font (this=0x101171620,
facefile=@0x7fff5fbf74c0) at src/ft2font.cpp:778
#6  0x000000010182fee8 in ft2font_module::new_ft2font (this=<value
temporarily unavailable, due to optimizations>, args=<value
temporarily unavailable, due to optimizations>) at src/ft2font.cpp:778
#7  0x0000000101830a80 in
Py::ExtensionModule<ft2font_module>::invoke_method_varargs
(this=<value temporarily unavailable, due to optimizations>,
method_def=<value temporarily unavailable, due to optimizations>,
args=<value temporarily unavailable, due to optimizations>) at
ExtensionModule.hxx:184
#8  0x000000010183f8d0 in method_varargs_call_handler
(_self_and_name_tuple=<value temporarily unavailable, due to
optimizations>, _args=<value temporarily unavailable, due to
optimizations>) at CXX/cxx_extensions.cxx:1714
#9  0x00000001000b3b25 in PyEval_EvalFrameEx ()
#10 0x00000001000b4ce0 in PyEval_EvalCodeEx ()
#11 0x00000001000b2afd in PyEval_EvalFrameEx ()
#12 0x00000001000b4ce0 in PyEval_EvalCodeEx ()

On Mon, Sep 28, 2009 at 8:47 AM, John Hunter <jd...@gm...> wrote:
> ---------- Forwarded message ----------
> From: Tony S Yu <ts...@gm...>
> Date: Mon, Sep 28, 2009 at 8:41 AM
> Subject: Re: [matplotlib-devel] Fwd: OS X 10.6 port
> To: John Hunter <jd...@gm...>
> Cc: matplotlib development list <mat...@li...>
>
>
>
> On Sep 28, 2009, at 2:14 AM, John Hunter wrote:
>
>> in case anyone has some suggestions, I'm forwarding this from the sage list
>>
>>
>> ---------- Forwarded message ----------
>> From: William Stein <ws...@gm...>
>> Date: Sun, Sep 27, 2009 at 10:51 PM
>> Subject: OS X 10.6 port
>> To: sage-devel <sag...@go...>, John Hunter <jd...@gm...>
>>
>>
>> Hi,
>>
>> I spent several hours yesterday trying to get matplotlib for Sage to
>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>> another test machine (bsd.math) the workaround from my laptop doesn't
>> work.  So at this point Sage still does not support OS X 10.6.
>>
>> The problem is in matplotlib's C++ wrapper for freetype2.  It's a
>> Python extension module implemented directly in C++, and it simply
>> doesn't work correctly at all.  For example, whenever it tries to
>> raise a Python exception, it just terminates Python with
>>
>> "11713 Abort trap              sage-ipython "$@" -i"

The stack trace looks like this:

#4  0x000000010cd907e2 in ~PythonExtension [inlined] () at
/Users/was/build/sage-4.1.2.alpha4/spkg/build/matplotlib/trunk/matplotlib/CXX/Extensions.hxx:778
#5  0x000000010cd907e2 in FT2Font::FT2Font (this=0x10baf2b10,
facefile=@0x10bab6da0) at src/ft2font.cpp:778
#6  0x000000010cd90e98 in ft2font_module::new_ft2font (this=<value
temporarily unavailable, due to optimizations>, args=<value
temporarily unavailable, due to optimizations>) at src/ft2font.cpp:778
#7  0x000000010cd94912 in
Py::ExtensionModule<ft2font_module>::invoke_method_varargs
(this=0x10ce05f10, name=@0x7fff5fbf7030, args=@0x7fff5fbf6fe0) at
Extensions.hxx:422
#8  0x000000010cda3190 in method_varargs_call_handler
(_self_and_name_tuple=<value temporarily unavailable, due to
optimizations>, _args=<value temporarily unavailable, due to
optimizations>) at CXX/cxx_extensions.cxx:1403
...


> I thought I'd chime in since I recently upgraded to OSX 10.6. I don't
> think I ever got this error, so I probably didn't have the same issue,
> but just in case...
>
> I reinstalled freetype2, but I'm not sure if I needed to. I also made
> a change to setupext.py so that the setup script could find freetype2,
> but this doesn't seem to be your problem because your errors don't
> occur during the build.
>
> Did you happen to build on top of an old install? I had to clean out
> the old compiled extensions before everything would build properly. I
> had to clean out the build directory (obviously), but there are also
> *.so files in lib/matplotlib and lib/matplotlib/backends.
>
> I should note that I'm built on top of the system python (2.6.1) and I
> did not use make.osx.
>

I did a 100% clean build from source of Sage, which means building
Python and essentially all dependencies (including freetype) in a
single self-contained directory.    I've replicated this on maybe 5 or
6 different OS X 10.6 machines.  It is a total show stopper for
releasing the next version of Sage, and is in fact the only problem
left (of many) in porting Sage to OS X 10.6

I used the latest version of XCode, which comes with OS X 10.6:
  gcc version 4.2.1 (Apple Inc. build 5646)

The version of Python  is:
        Python 2.6.2 (r262:71600, Sep 27 2009, 13:41:53)

The Abort trap appears when building the font cache the moment an
exception is raised in C++ code.  By hacking around that exception
getting raised, the Abort trap happens when some exception somewhere
else happens in Matplotlib's C++ code.  Etc.

I'm going to try looking into pyCXX now, since that seems the best hope.

William

On Mon, Sep 28, 2009 at 4:54 AM, John Hunter <jd...@gm...> wrote:
> On Mon, Sep 28, 2009 at 4:47 AM, Dr. David Kirkby
> <dav...@on...> wrote:
>> William Stein wrote:
>>>
>>> Hi,
>>>
>>> I spent several hours yesterday trying to get matplotlib for Sage to
>>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>>> another test machine (bsd.math) the workaround from my laptop doesn't
>>> work.  So at this point Sage still does not support OS X 10.6.
>>
>>
>> Just to add (which I've also CC'ed to John), mathplotlib (at least the
>> version in Sage), is not respecting CXX properly. I specified CC as the Sun
>> C compiler, and CXX as the Sun C++ compiler. The code then tries to build a
>> C++ file (src/ft2font.cpp) with the Sun C compiler, which fails.
>>
>> /opt/xxxsunstudio12.1/bin/cc -DNDEBUG -O -xcode=pic32
>> -DPY_ARRAYAUNIQUE_SYMBOL=MPL_ARRAY_API
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/lib/python2.6/site-packages/numpy/core/include
>> -I/usr/sfw/include -I/usr/sfw/include/freetype2 -I/usr/local/include -I.
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/python2.6
>> -c src/ft2font.cpp -o build/temp.solaris-2.10-sun4u-2.6/src/ft2font.o
>> cc: No valid input files specified, no output generated
>> error: command '/opt/xxxsunstudio12.1/bin/cc' failed with exit status 1
>>
>> I've not looked at this in detail, so it could be a Sage-specific issue.
>
> I forwarded the original message on to matplotlib-devel.  It would be
> helpful if when replying to this message or any in this thread if we
> CC the matplotlib-devel mailing list.  I'm the list moderator so I can
> approve these messages and you won't need to subscribe.  I don't have
> any immediate answers for you now, but hopefully we can come up with
> some.

If you could just try running your matplotlib build on
bsd.math.washington.edu and reporting about whether or not it works,
that would be very helpful, since it will indicate whether the problem
is somehow in Sage or in pyCXX or Matplotlib itself or something else.

 -- William

Michael Droettboom wrote:
> This is a very long-standing bug in distutils:
> 
>   http://bugs.python.org/issue1222585
> 
> While distutils distinguishes between C and C++ for *linking*, it does 
> not do so for *compiling*.  This doesn't matter for gcc and msvc (which 
> are arguably the two most popular compilers for Python), but probably 
> does for most commercial Unix compilers.  I have no idea why this patch 
> has never been applied in all these years -- but please jump on that bug 
> report so we can see about getting it fixed.
> 
> As for us at STScI, we just build matplotlib with gcc on Solaris and it 
> works fine.
> 
> Mike


I've registered on the python bug tracker, and submitted a request that 
this bug is solved. The patch on there is against a 3-year old version 
of python, so I'd rather not try to integrate it myself, especially as 
the patch is non-trivial.

I expected that trying to build Sage with Sun Studio would uncover some 
bugs in other software, but even I am amazed at just how many bugs this 
uncovers.


Dave





> 
> John Hunter wrote:
>> On Mon, Sep 28, 2009 at 4:47 AM, Dr. David Kirkby
>> <dav...@on...> wrote:
>>  
>>> William Stein wrote:
>>>    
>>>> Hi,
>>>>
>>>> I spent several hours yesterday trying to get matplotlib for Sage to
>>>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>>>> another test machine (bsd.math) the workaround from my laptop doesn't
>>>> work.  So at this point Sage still does not support OS X 10.6.
>>>>       
>>> Just to add (which I've also CC'ed to John), mathplotlib (at least the
>>> version in Sage), is not respecting CXX properly. I specified CC as 
>>> the Sun
>>> C compiler, and CXX as the Sun C++ compiler. The code then tries to 
>>> build a
>>> C++ file (src/ft2font.cpp) with the Sun C compiler, which fails.
>>>
>>> /opt/xxxsunstudio12.1/bin/cc -DNDEBUG -O -xcode=pic32
>>> -DPY_ARRAYAUNIQUE_SYMBOL=MPL_ARRAY_API
>>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/lib/python2.6/site-packages/numpy/core/include 
>>>
>>> -I/usr/sfw/include -I/usr/sfw/include/freetype2 -I/usr/local/include -I.
>>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/
>>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/python2.6 
>>>
>>> -c src/ft2font.cpp -o build/temp.solaris-2.10-sun4u-2.6/src/ft2font.o
>>> cc: No valid input files specified, no output generated
>>> error: command '/opt/xxxsunstudio12.1/bin/cc' failed with exit status 1
>>>
>>> I've not looked at this in detail, so it could be a Sage-specific issue.
>>>     
>>
>> I forwarded the original message on to matplotlib-devel.  It would be
>> helpful if when replying to this message or any in this thread if we
>> CC the matplotlib-devel mailing list.  I'm the list moderator so I can
>> approve these messages and you won't need to subscribe.  I don't have
>> any immediate answers for you now, but hopefully we can come up with
>> some.
>>
>> JDH
>>
>> ------------------------------------------------------------------------------ 
>>
>> Come build with us! The BlackBerry&reg; Developer Conference in SF, CA
>> is the only developer event you need to attend this year. Jumpstart your
>> developing skills, take BlackBerry mobile applications to market and 
>> stay ahead of the curve. Join us from November 9&#45;12, 2009. 
>> Register now&#33;
>> http://p.sf.net/sfu/devconf
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>   
> 

On Sep 28, 2009, at 2:14 AM, John Hunter wrote:

> in case anyone has some suggestions, I'm forwarding this from the  
> sage list
>
>
> ---------- Forwarded message ----------
> From: William Stein <ws...@gm...>
> Date: Sun, Sep 27, 2009 at 10:51 PM
> Subject: OS X 10.6 port
> To: sage-devel <sag...@go...>, John Hunter <jd...@gm... 
> >
>
>
> Hi,
>
> I spent several hours yesterday trying to get matplotlib for Sage to
> work on OS X 10.6.  On my laptop everything works perfectly, but on
> another test machine (bsd.math) the workaround from my laptop doesn't
> work.  So at this point Sage still does not support OS X 10.6.
>
> The problem is in matplotlib's C++ wrapper for freetype2.  It's a
> Python extension module implemented directly in C++, and it simply
> doesn't work correctly at all.  For example, whenever it tries to
> raise a Python exception, it just terminates Python with
>
> "11713 Abort trap              sage-ipython "$@" -i"

I thought I'd chime in since I recently upgraded to OSX 10.6. I don't  
think I ever got this error, so I probably didn't have the same issue,  
but just in case...

I reinstalled freetype2, but I'm not sure if I needed to. I also made  
a change to setupext.py so that the setup script could find freetype2,  
but this doesn't seem to be your problem because your errors don't  
occur during the build.

Did you happen to build on top of an old install? I had to clean out  
the old compiled extensions before everything would build properly. I  
had to clean out the build directory (obviously), but there are also  
*.so files in lib/matplotlib and lib/matplotlib/backends.

I should note that I'm built on top of the system python (2.6.1) and I  
did not use make.osx.

-Tony

This is a very long-standing bug in distutils:

   http://bugs.python.org/issue1222585

While distutils distinguishes between C and C++ for *linking*, it does 
not do so for *compiling*.  This doesn't matter for gcc and msvc (which 
are arguably the two most popular compilers for Python), but probably 
does for most commercial Unix compilers.  I have no idea why this patch 
has never been applied in all these years -- but please jump on that bug 
report so we can see about getting it fixed.

As for us at STScI, we just build matplotlib with gcc on Solaris and it 
works fine.

Mike

John Hunter wrote:
> On Mon, Sep 28, 2009 at 4:47 AM, Dr. David Kirkby
> <dav...@on...> wrote:
>   
>> William Stein wrote:
>>     
>>> Hi,
>>>
>>> I spent several hours yesterday trying to get matplotlib for Sage to
>>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>>> another test machine (bsd.math) the workaround from my laptop doesn't
>>> work.  So at this point Sage still does not support OS X 10.6.
>>>       
>> Just to add (which I've also CC'ed to John), mathplotlib (at least the
>> version in Sage), is not respecting CXX properly. I specified CC as the Sun
>> C compiler, and CXX as the Sun C++ compiler. The code then tries to build a
>> C++ file (src/ft2font.cpp) with the Sun C compiler, which fails.
>>
>> /opt/xxxsunstudio12.1/bin/cc -DNDEBUG -O -xcode=pic32
>> -DPY_ARRAYAUNIQUE_SYMBOL=MPL_ARRAY_API
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/lib/python2.6/site-packages/numpy/core/include
>> -I/usr/sfw/include -I/usr/sfw/include/freetype2 -I/usr/local/include -I.
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/python2.6
>> -c src/ft2font.cpp -o build/temp.solaris-2.10-sun4u-2.6/src/ft2font.o
>> cc: No valid input files specified, no output generated
>> error: command '/opt/xxxsunstudio12.1/bin/cc' failed with exit status 1
>>
>> I've not looked at this in detail, so it could be a Sage-specific issue.
>>     
>
> I forwarded the original message on to matplotlib-devel.  It would be
> helpful if when replying to this message or any in this thread if we
> CC the matplotlib-devel mailing list.  I'm the list moderator so I can
> approve these messages and you won't need to subscribe.  I don't have
> any immediate answers for you now, but hopefully we can come up with
> some.
>
> JDH
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry&reg; Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay 
> ahead of the curve. Join us from November 9&#45;12, 2009. Register now&#33;
> http://p.sf.net/sfu/devconf
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>   

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

I forgot a detail: you have to use a script to edit the commandline 
arguments to pass to gcc:

#!/bin/sh
y=""
for x in $*
do
       case "$x"
       in
       -xcode*)
               echo skip $x
               ;;
       *)
               y="$y $x"
               ;;
       esac
done
gcc $y

and then do

CC=gcc_hack python setup.py build

(where gcc_hack is the path to the above script)

Mike

John Hunter wrote:
> On Mon, Sep 28, 2009 at 4:47 AM, Dr. David Kirkby
> <dav...@on...> wrote:
>   
>> William Stein wrote:
>>     
>>> Hi,
>>>
>>> I spent several hours yesterday trying to get matplotlib for Sage to
>>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>>> another test machine (bsd.math) the workaround from my laptop doesn't
>>> work.  So at this point Sage still does not support OS X 10.6.
>>>       
>> Just to add (which I've also CC'ed to John), mathplotlib (at least the
>> version in Sage), is not respecting CXX properly. I specified CC as the Sun
>> C compiler, and CXX as the Sun C++ compiler. The code then tries to build a
>> C++ file (src/ft2font.cpp) with the Sun C compiler, which fails.
>>
>> /opt/xxxsunstudio12.1/bin/cc -DNDEBUG -O -xcode=pic32
>> -DPY_ARRAYAUNIQUE_SYMBOL=MPL_ARRAY_API
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/lib/python2.6/site-packages/numpy/core/include
>> -I/usr/sfw/include -I/usr/sfw/include/freetype2 -I/usr/local/include -I.
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/
>> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/python2.6
>> -c src/ft2font.cpp -o build/temp.solaris-2.10-sun4u-2.6/src/ft2font.o
>> cc: No valid input files specified, no output generated
>> error: command '/opt/xxxsunstudio12.1/bin/cc' failed with exit status 1
>>
>> I've not looked at this in detail, so it could be a Sage-specific issue.
>>     
>
> I forwarded the original message on to matplotlib-devel.  It would be
> helpful if when replying to this message or any in this thread if we
> CC the matplotlib-devel mailing list.  I'm the list moderator so I can
> approve these messages and you won't need to subscribe.  I don't have
> any immediate answers for you now, but hopefully we can come up with
> some.
>
> JDH
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry&reg; Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay 
> ahead of the curve. Join us from November 9&#45;12, 2009. Register now&#33;
> http://p.sf.net/sfu/devconf
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>   

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

John Hunter wrote:
> I spent several hours yesterday trying to get matplotlib for Sage to
> work on OS X 10.6.  On my laptop everything works perfectly, but on
> another test machine (bsd.math) the workaround from my laptop doesn't
> work.  So at this point Sage still does not support OS X 10.6.
>
> The problem is in matplotlib's C++ wrapper for freetype2.  It's a
> Python extension module implemented directly in C++, and it simply
> doesn't work correctly at all.  For example, whenever it tries to
> raise a Python exception, it just terminates Python with
>
> "11713 Abort trap              sage-ipython "$@" -i"
>   
Can you run this inside of gdb and get a traceback?
> I'm cc'ing John Hunter -- author of matplotlib, in case he has any
> advice to share.  In particular, he has an account on
> bsd.math.washington.edu and build tests matplotlib there, so I'm
> curious if he has any issues with that.  Note that Sage builds
> freetype from source, which could be relevant.
>
> I will try updating freetype in Sage (our version is slightly out of
> date), and also checking if there is some sort of conflict with a
> system-wide freetype.  My worry is that Matplotlib has some weird
> custom C++ code for writing extensions that maybe uses signals or
> something to raise exceptions, and it is just broken when used with OS
> X 10.6.  Hopefully this is not that the case, and something special to
> Sage is happening.
>   
CXX is the name of the C++ wrapper mechanism we use and is reasonably 
well-maintained.  I can't find out if it has been tested on 10.6 yet, 
but it's certainly been tested on the gcc version that comes with it.  
There are no signals involved, only standard C++ exceptions which are 
delegated to Python exceptions.  Is it possible there is any C++ code 
linked to this that was generated with another compiler version?

Cheers,
Mike

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

On Mon, Sep 28, 2009 at 4:47 AM, Dr. David Kirkby
<dav...@on...> wrote:
> William Stein wrote:
>>
>> Hi,
>>
>> I spent several hours yesterday trying to get matplotlib for Sage to
>> work on OS X 10.6.  On my laptop everything works perfectly, but on
>> another test machine (bsd.math) the workaround from my laptop doesn't
>> work.  So at this point Sage still does not support OS X 10.6.
>
>
> Just to add (which I've also CC'ed to John), mathplotlib (at least the
> version in Sage), is not respecting CXX properly. I specified CC as the Sun
> C compiler, and CXX as the Sun C++ compiler. The code then tries to build a
> C++ file (src/ft2font.cpp) with the Sun C compiler, which fails.
>
> /opt/xxxsunstudio12.1/bin/cc -DNDEBUG -O -xcode=pic32
> -DPY_ARRAYAUNIQUE_SYMBOL=MPL_ARRAY_API
> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/lib/python2.6/site-packages/numpy/core/include
> -I/usr/sfw/include -I/usr/sfw/include/freetype2 -I/usr/local/include -I.
> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/
> -I/export/home/drkirkby/sage/gcc32/sage-4.1.2.alpha2/local/include/python2.6
> -c src/ft2font.cpp -o build/temp.solaris-2.10-sun4u-2.6/src/ft2font.o
> cc: No valid input files specified, no output generated
> error: command '/opt/xxxsunstudio12.1/bin/cc' failed with exit status 1
>
> I've not looked at this in detail, so it could be a Sage-specific issue.

I forwarded the original message on to matplotlib-devel.  It would be
helpful if when replying to this message or any in this thread if we
CC the matplotlib-devel mailing list.  I'm the list moderator so I can
approve these messages and you won't need to subscribe.  I don't have
any immediate answers for you now, but hopefully we can come up with
some.

JDH

in case anyone has some suggestions, I'm forwarding this from the sage list


---------- Forwarded message ----------
From: William Stein <ws...@gm...>
Date: Sun, Sep 27, 2009 at 10:51 PM
Subject: OS X 10.6 port
To: sage-devel <sag...@go...>, John Hunter <jd...@gm...>


Hi,

I spent several hours yesterday trying to get matplotlib for Sage to
work on OS X 10.6.  On my laptop everything works perfectly, but on
another test machine (bsd.math) the workaround from my laptop doesn't
work.  So at this point Sage still does not support OS X 10.6.

The problem is in matplotlib's C++ wrapper for freetype2.  It's a
Python extension module implemented directly in C++, and it simply
doesn't work correctly at all.  For example, whenever it tries to
raise a Python exception, it just terminates Python with

"11713 Abort trap              sage-ipython "$@" -i"

I'm cc'ing John Hunter -- author of matplotlib, in case he has any
advice to share.  In particular, he has an account on
bsd.math.washington.edu and build tests matplotlib there, so I'm
curious if he has any issues with that.  Note that Sage builds
freetype from source, which could be relevant.

I will try updating freetype in Sage (our version is slightly out of
date), and also checking if there is some sort of conflict with a
system-wide freetype.  My worry is that Matplotlib has some weird
custom C++ code for writing extensions that maybe uses signals or
something to raise exceptions, and it is just broken when used with OS
X 10.6.  Hopefully this is not that the case, and something special to
Sage is happening.

William

--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org

Stéfan van der Walt wrote:
> Hi all,
> 
> When saving plots (using plt.figsave and matplotlib from SVN)
> consisting of subplots, the layout is a bit messed up.  As an example,
> have a look at
> 
> http://mentat.za.net/refer/segmentation.pdf

Wow, that's really ugly!  I suspect that for anyone to provide any help 
on this, you will need to supply a minimal script to show what you are 
doing.  I take it you are starting with something that looks reasonable 
on the screen, but comes out quite different when you save it?

Eric

> 
> I have tried adjusting the figure size, the font size, the w- and
> hspace, etc.  Is this a known problem, or am I doing something silly?
> 
> Thanks for your time,
> Stéfan
>