It's probably something in C++ backend code for Agg.  Maybe someone else 
on this list knows the last time it built and worked on Windows and can 
track down a change that may have broken it.

It will probably take some work with a debugger to figure out where it 
is crashing.  I'd be completely guessing at this point.  If you feel 
adventurous and can get a C backtrace out of the crash somehow, (does 
gdb work with mingw?), that could be useful.  Barring that, maybe one of 
the other Windows folk on this list has some ideas.

Thanks for your considerable patience ;)

Cheers,
Mike

william ratcliff wrote:
> Yes.  It crashes at the savefig line.
>
> On 8/1/07, *Michael Droettboom* <[EMAIL PROTECTED] 
> <mailto:[EMAIL PROTECTED]>> wrote:
>
>     william ratcliff wrote:
>     > Ok--I reverted to head and it fixed the undefined references
>     error and
>     > everything seems to build normally.
>     Well, that's partially good news...
>     > But, if I try to run the embedding_in_wx4.py example, it crashes
>     > python 2.5.1
>     >
>     > If I install from the .msi then things do work in python 2.5.1,
>     so it
>     > must be something with the build.
>     Just to clarify, you mean the 0.90.1 installer downloaded from
>     SourceForge?
>     >
>     > I also tried to just start python from the command window, then ran:
>     > import numpy as N
>     > x=N.array([1,2,3])
>     > from pylab import *
>     > plot(x,x)
>     > show()
>     >
>     > The plot command generates a Line2D instance,
>     > but the show() command kills python in the same way that the
>     embedding
>     > example does.  In the embedding example, I notice that the
>     CanvasFrame
>     > is generated before the crash.
>     Does it also crash if you use the Agg backend alone and save to an
>     image
>     file?
>
>     import matplotlib
>     matplotlib.use("Agg")
>     from pylab import *
>     plot([1,2,3])
>     savefig("test.png")
>
>     Cheers,
>     Mike
>     >
>     > On 8/1/07, *william ratcliff* < [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>     > <mailto:[EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>>> wrote:
>     >
>     >     Let me try this.
>     >
>     >     On 8/1/07, *Michael Droettboom* < [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>     >     <mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>> wrote:
>     >
>     >         I'm going to summarize, just to make sure I have everything
>     >         straight.
>     >
>     >         1) HEAD doesn't build (you get lots of "undefined
>     reference to
>     >         `_imp__FT_Done_FreeType'")
>     >         2) revision 3645 (yesterday before my setup.py changes),
>     >         builds, but you
>     >         got the float('inf') error (which is now fixed)
>     >         3) much earlier revisions (3608) have other issues
>     (which seem
>     >         to be
>     >         related to out-of-date wxPython code)
>     >
>     >         I have just committed a possible fix for 1).  You may
>     want to try
>     >         updating to HEAD and building that.
>     >
>     >         If that doesn't work for you, I'm worried about things
>     getting
>     >         out of
>     >         hand and annoying a lot of people with my (well intentioned)
>     >         setup.py
>     >         changes yesterday.  If that's the case, I will revert my
>     build
>     >         changes
>     >         until I have a chance to sit in front of a Windows box and
>     >         figure out
>     >         what's really going wrong.
>     >
>     >         Apologies all around,
>     >         Mike
>     >
>     >         william ratcliff wrote:
>     >         > 3608 also runs into problems of linking with the freetype
>     >         libraries.
>     >         > Some much older versions (say in the 3500s) compile, but
>     >         still crash
>     >         > python2.5.1 with errors like:
>     >         >
>     >         >
>     C:\matplotlib\trunk\matplotlib\examples>c:\python25\python.exe
>     >         > embedding_in_wx4.
>     >         > py
>     >         > embedding_in_wx4.py:24: DeprecationWarning: The wxPython
>     >         compatibility
>     >         > package i
>     >         > s no longer automatically generated or actively
>     >         maintained.  Please
>     >         > switch to th
>     >         > e wx package as soon as possible.
>     >         >   from wxPython.wx import *
>     >         >
>     >         > Thanks again for the help!
>     >         >
>     >         > William
>     >         >
>     >         > On 8/1/07, *John Hunter* < [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>     >         <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
>     <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
>     >         <mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>>
>     >         > > wrote:
>     >         >
>     >         >     On 8/1/07, william ratcliff <
>     [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
>     >         <mailto: [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>>
>     >         >     <mailto: [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>     >         <mailto:[EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>>>> wrote:
>     >         >     >     from matplotlib.mathtext import
>     math_parse_s_ft2font
>     >         >     >   File
>     >         >     >
>     "c:\python25\Lib\site-packages\matplotlib\mathtext.py",
>     >         >     > line 1182, in <mo
>     >         >     > dule>
>     >         >     >     class Vlist(List):
>     >         >     >   File
>     >         >     >
>     "c:\python25\Lib\site-packages\matplotlib\mathtext.py",
>     >         >     > line 1189, in Vli
>     >         >     > st
>     >         >     >     def vpack(self, h=0., m='additional',
>     l=float('inf')):
>     >         >     > ValueError: invalid literal for float(): inf
>     >         >
>     >         >     It's not clear to me why you should be hitting
>     mathtext
>     >         at all, since
>     >         >     the wxcursor_demo.py does not use it, as far as I can
>     >         tell.  Michael,
>     >         >     would any of your earlier work caused him to get into
>     >         mathtext
>     >         >     inapproriately?  You may need to revert a little
>     farther,
>     >         William, eg
>     >         >     to r3608 while we get this sorted out.
>     >         >
>     >         >     JDH
>     >         >
>     >         >
>     >
>     >
>     >
>
>


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Reply via email to