Matplotlib's built-in mathematical expression parser produces garbage
output for math text:
http://img3.imageshack.us/img3/8194/matplotlibmathtext.png
I'm using the distribution packages from Fedora 11. I guess there's an
issue with the fonts on my system but I can't figure out how to proceed
with pinpointing the issue.
Regards,
Can you add
"verbose.level : debug-annoying"
to your ~/.matplotlib/matplotlibrc file and send us the output produced?
FWIW, I just tested the matplotlib package on FC11 and it seems to work fine on my system, so it's possibly a configuration issue.
You might want to try removing for font cache file in:
~/.matplotlib/fontList.cache
Mike
Kamran Riaz Khan wrote:
Matplotlib's built-in mathematical expression parser produces garbage
output for math text:http://img3.imageshack.us/img3/8194/matplotlibmathtext.png
I'm using the distribution packages from Fedora 11. I guess there's an
issue with the fonts on my system but I can't figure out how to proceed
with pinpointing the issue.Regards,
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA
Can you add
"verbose.level : debug-annoying"
to your ~/.matplotlib/matplotlibrc file and send us the output produced?
FWIW, I just tested the matplotlib package on FC11 and it seems to work
fine on my system, so it's possibly a configuration issue.You might want to try removing for font cache file in:
~/.matplotlib/fontList.cache
Thanks for the quick reply. I deleted the fontList.cache file and added
verbose.level : debug-annoying to my matplobrc file. The next run of
pyplot example gave me:
The plot didn't appear but I guess that's the default behavior of
verbose level option. I removed the debug-annoying and reran the example
which in turn gave me the same image output as before.
On 09/11/2009 07:28 PM, Michael Droettboom wrote:
Kamran Riaz Khan wrote:
Matplotlib's built-in mathematical expression parser produces garbage
output for math text:http://img3.imageshack.us/img3/8194/matplotlibmathtext.png
I'm using the distribution packages from Fedora 11. I guess there's an
issue with the fonts on my system but I can't figure out how to proceed
with pinpointing the issue.
--
Kamran Riaz Khan.
The log doesn't show any actual plotting. Did you run the same example with debug-annoying turned on? Turning debugging on shouldn't change any behavior -- only output more debugging information to the console.
Cheers,
Mike
Kamran Riaz Khan wrote:
On 09/11/2009 07:28 PM, Michael Droettboom wrote:
Can you add
"verbose.level : debug-annoying"
to your ~/.matplotlib/matplotlibrc file and send us the output produced?
FWIW, I just tested the matplotlib package on FC11 and it seems to work
fine on my system, so it's possibly a configuration issue.You might want to try removing for font cache file in:
~/.matplotlib/fontList.cache
Thanks for the quick reply. I deleted the fontList.cache file and added
verbose.level : debug-annoying to my matplobrc file. The next run of
pyplot example gave me:The plot didn't appear but I guess that's the default behavior of
verbose level option. I removed the debug-annoying and reran the example
which in turn gave me the same image output as before.Kamran Riaz Khan wrote:
Matplotlib's built-in mathematical expression parser produces garbage
output for math text:http://img3.imageshack.us/img3/8194/matplotlibmathtext.png
I'm using the distribution packages from Fedora 11. I guess there's an
issue with the fonts on my system but I can't figure out how to proceed
with pinpointing the issue.
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA
There seems to be something really, really wrong here. As running the
_same_ example with debug-annoying turned on just terminates the
execution at:
backend Agg version v2.2
backend_agg.new_figure_manager
On 09/11/2009 10:51 PM, Michael Droettboom wrote:
The log doesn't show any actual plotting. Did you run the same example
with debug-annoying turned on? Turning debugging on shouldn't change
any behavior -- only output more debugging information to the console.
--
Regards,
Kamran Riaz Khan.
That *is* strange. I'm at a bit of a loss. Is it segfaulting? Can you get a traceback from gdb?
Since I'm having no trouble with the FC11 packages, that suggests it's something related to a difference in configuration. Are you doing anything unusual with fontconfig? Are you 64-bit or 32-bit (I'm on 32-bit, for what it's worth)? Are you running the freetype or freetype-freeworld package?
What does:
fc-list "Computer Modern"
display?
Sorry for all the questions, but I'm grasping for something here 
Cheers,
Mike
Kamran Riaz Khan wrote:
On 09/11/2009 10:51 PM, Michael Droettboom wrote:
The log doesn't show any actual plotting. Did you run the same example
with debug-annoying turned on? Turning debugging on shouldn't change
any behavior -- only output more debugging information to the console.
There seems to be something really, really wrong here. As running the
_same_ example with debug-annoying turned on just terminates the
execution at:backend Agg version v2.2
backend_agg.new_figure_manager
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA
That *is* strange. I'm at a bit of a loss. Is it segfaulting? Can you
get a traceback from gdb?
Since I'm having no trouble with the FC11 packages, that suggests it's
something related to a difference in configuration. Are you doing
anything unusual with fontconfig? Are you 64-bit or 32-bit (I'm on
32-bit, for what it's worth)? Are you running the freetype or
freetype-freeworld package?
What does:
fc-list "Computer Modern"
display?
Nothing 
, I think this might be the culprit. However, I have
cm-unicode's ttf fonts installed:
What fonts does matplotlib exactly need here?
Sorry for all the questions, but I'm grasping for something here
I appreciate you putting in your time and effort for this seemingly
bizarre issue 
.
On 09/14/2009 06:33 PM, Michael Droettboom wrote:
Cheers,
Mike
--
Regards,
Kamran Riaz Khan.
