matplotlib-users Mailing List for matplotlib

Hi! I have some code importing MPL and wxmpl; presently, I have version 
0.90.1 of the former installed and 1.2.8 of the latter. I hadn't run 
this code in a while; when I last did, in the late spring sometime, it 
worked fine. Now, when I do (from the command line), a call in it to 
<class-derived-from-wxmpl.PlotPanel>.Axes.clear() results in an error 
seq. ending in AttributeError: VectorLineCollection instance has no 
attribute 'get_xdata'. One "catch": the "old" version still exists as a 
py2app-ed stand-alone; I tried it and sure enough, it still works fine. 
By searching its Mac App file tree, I was able to determine that this 
old py2app-ed version is using version 0.90.0 of MPL; unfortunately, it 
(appears to be) using a compiled version of wxmpl, so I don't know how 
to determine what version of that its using. Did the update of MPL from 
0.90.0 to 0.90.1 change anything that might result in this error; has 
there been a change in wxmpl which might result in this error? (Between 
then and now, I switched to a new Mac and installed everything from 
scratch, so it's very likely that my present version of wxmpl is 
different from the one I had installed when I created the py2app.) Any 
other ideas? (I'm at a total loss.) Thanks!
DG
-- 
ERD/ORR/NOS/NOAA <http://response.restoration.noaa.gov/emergencyresponse/>

On 2007年10月08日, Michael Droettboom apparently wrote:
> If there are any particular things that are particularly 
> bad, please send an example plot that exercises it. 
I have noticed that scatter plots take a very long time to 
render. (Use say four subplots each with three hundred points.)
Cheers,
Alan Isaac

Alan G Isaac wrote:
> On 2007年10月08日, Michael Droettboom apparently wrote:
>> http://sourceforge.net/tracker/index.php?func=detail&aid=1738494&group_id=80706&atid=560720 
>> This patch applies cleanly against 0.90.1 
> 
> Yes, that fixes it.
> 
> Perhaps I should mention that the PDF rendering is very slow.
Yes, PDF is probably one of the slowest backends in matplotlib, mainly 
due to the inherent complexity of the file format, with all of its 
cross-references etc. If there are any particular things that are 
particularly bad, please send an example plot that exercises it.
Cheers,
Mike
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

On 2007年10月08日, Michael Droettboom apparently wrote:
> http://sourceforge.net/tracker/index.php?func=detail&aid=1738494&group_id=80706&atid=560720 
> This patch applies cleanly against 0.90.1 
Yes, that fixes it.
Perhaps I should mention that the PDF rendering is very slow.
Thanks!
Alan Isaac

Hello to all,
I have found a peculiar behavior of matplotlib function errorbar.
See the code below and execute it all in once.
What I get is that first image is ok.
Second one loses all errorbar vertical line.
I noticed that while i was generating some graphs with errorbar in a for cycle 
and i tried to plot also some dots over each one. 
import pylab as p
p.figure(1)
p.errorbar([1,2,3,4],[1,2,3,4],[4,2,5,3])
p.hold(True)
# here plot what you wants
p.hold(False)
p.savefig('test_1.png')
p.close()
p.figure(1)
p.errorbar([1,2,3,4],[1,2,3,4],[4,2,5,3])
p.hold(True)
# here plot what you wants
p.hold(False)
p.savefig('test_2.png')
p.close()
-- 
--
Emanuele Passera

This is a problem using matplotlib 0.90.1 with Python 2.5.
See here:
http://sourceforge.net/tracker/index.php?func=detail&aid=1738494&group_id=80706&atid=560720
This patch applies cleanly against 0.90.1
Cheers,
Mike
Alan G Isaac wrote:
> I'm having troubles saving figures as PDF.
> Matplotlib version 0.90.1
> Simple example below.
> 
> Cheers,
> Alan Isaac
> 
> 
> Python 2.5.1 (r251:54863, Apr 18 2007, 08:51:08) [MSC v.1310 
> 32 bit (Intel)] on win32
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import pylab
>>>> fig = pylab.figure()
>>>> pylab.plot([1,2,3])
> [<matplotlib.lines.Line2D instance at 0x016FF828>]
>>>> fig.savefig('c:/temp/temp.pdf')
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> File "C:\Python25\Lib\site-packages\matplotlib\figure.py", line 759, in savefig
> self.canvas.print_figure(*args, **kwargs)
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_tkagg.py", line 188, in print_figu
> re
> **kwargs)
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_agg.py", line 497, in print_figure
> 
> printfunc(filename, dpi, facecolor, edgecolor, orientation, **kwargs)
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 1395, in print_figur
> e
> file.close()
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 401, in close
> self.writeFonts()
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 456, in writeFonts
> fontdictObject = self.embedTTF(filename)
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 508, in embedTTF
> widths = [ get_char_width(charcode) for charcode in range(firstchar, lastchar+1) ]
> File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 505, in get_char_wid
> th
> unicode = cp1252.decoding_map[charcode] or 0
> AttributeError: 'module' object has no attribute 'decoding_map'
> 
> 
> 
> 
> -------------------------------------------------------------------------
> 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
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

I'm having troubles saving figures as PDF.
Matplotlib version 0.90.1
Simple example below.
Cheers,
Alan Isaac
Python 2.5.1 (r251:54863, Apr 18 2007, 08:51:08) [MSC v.1310 
32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import pylab
>>> fig = pylab.figure()
>>> pylab.plot([1,2,3])
[<matplotlib.lines.Line2D instance at 0x016FF828>]
>>> fig.savefig('c:/temp/temp.pdf')
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
 File "C:\Python25\Lib\site-packages\matplotlib\figure.py", line 759, in savefig
 self.canvas.print_figure(*args, **kwargs)
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_tkagg.py", line 188, in print_figu
re
 **kwargs)
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_agg.py", line 497, in print_figure
 printfunc(filename, dpi, facecolor, edgecolor, orientation, **kwargs)
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 1395, in print_figur
e
 file.close()
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 401, in close
 self.writeFonts()
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 456, in writeFonts
 fontdictObject = self.embedTTF(filename)
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 508, in embedTTF
 widths = [ get_char_width(charcode) for charcode in range(firstchar, lastchar+1) ]
 File "C:\Python25\Lib\site-packages\matplotlib\backends\backend_pdf.py", line 505, in get_char_wid
th
 unicode = cp1252.decoding_map[charcode] or 0
AttributeError: 'module' object has no attribute 'decoding_map'
>>>

Michael Droettboom wrote:
> Bill Dandreta wrote:
>> This info may or may not be useful. If I save a plot with savefig()
>> without specifying a figsize, sometimes the figure gets clipped. What
>> gets saved is the portion of the image that show() displays. I
>> determined experimentally that figsize=(13,10) causes show() to open the
>> image at full screen size and so far using that size saves an unclipped
>> image.
>
> Is this with 0.90.1 or a SVN version? What backend are you using?
dev-python/matplotlib
 Installed versions: 0.90.1 (01:40:48 06/08/07) (doc examples gtk tk)
dev-lang/python
 Installed versions: 2.4.4-r5 (2.4) (19:12:16 09/23/07) (berkdb
-bootstrap -build doc examples gdbm ipv6 ncurses -nocxx -nothreads
readline ssl tk -ucs2)
$ pkg-config --version libpng
0.21
~/.matplotlib/matplotlibrc:
backend : GTKAgg
numerix : numpy
units : True
-- 
Bill
wjd...@at...
Gentoo Linux X86_64 2.6.20-gentoo-r8
Reclaim Your Inbox with http://www.mozilla.org/products/thunderbird/
All things cometh to he who waiteth as long as he who waiteth worketh like hell while he waiteth.

Hmmm. I'm very surprised that this change could cause that. All it 
does is add an additional metadata chunk to the PNG file, which 
shouldn't have any affect on the image data itself. simple_plot.py 
works fine for me in GIMP 2.0.5 both before and after this change. Can 
you verify that this plot was working before the change to save the 
resolution in the PNG file? If so, can you send me the source for your 
plot and the PNG file? Also, what version of libpng are you using? 
(pkg-config --version libpng should display this on most recent Linux 
distros).
Cheers,
Mike
Wayne E. Harlan wrote:
> Not quite. The figure opens at the correct size in the gimp but all I 
> see is background - no plot stuff.
> 
> Wayne
> 
> Michael Droettboom wrote:
>> This is now added in SVN, and only for the Agg backend(s). It won't 
>> work with some of the other ways to save PNG files, such as Gdk, Wx 
>> (non-Agg) and Cairo. If anyone was any thoughts about how to support 
>> saving the resolution in those backends, please share.
>>
>> Also note that resolution is saved in "dots per meter" in PNG files, 
>> so rounding error makes things slightly off -- for instance, 100 dpi 
>> is shows up as 99.999998 dpi in the GIMP.
>>
>> Cheers,
>> Mike
>>
>> Michael Droettboom wrote:
>>> I'll look into this. I actually made a similar fix on another 
>>> project I used to work on... It should be theoretically possible, 
>>> barring any roadblocks from how matplotlib is doing things.
>>>
>>> Cheers,
>>> Mike
>>>
>>> Wayne E. Harlan wrote:
>>>> I would like to follow up on my first response to Bill. It probably 
>>>> should be a new thread, but I'll start here. When png files are 
>>>> saved with a DPI=300 argument, and I open them in the Gimp, the dpi 
>>>> is only 72 (default ?). As I understand it, the dpi setting in the 
>>>> graphics file tells the application opening it how big to display 
>>>> it. When I use savefig with a DPI=300 and the plot figure I am 
>>>> saving was created with figsize=(6,4.5) I expect the figure to open 
>>>> in word or swriter at the size I specified (6" by 4.5") with the 
>>>> appropriate number of pixels. However, at present, that does not 
>>>> happen. The number of pixels in the figure is correct but I have to 
>>>> resize it manually to get the right size. Can this be fixed ?
>>>>
>>>> Thanks,
>>>>
>>>> Wayne
>>>>
>>>> Bill Dandreta wrote:
>>>>> I resolved the problem. It was unrelated to mpl. xv and gimp were
>>>>> complied w/o png support. Recompiling with png support resolved the 
>>>>> problem.
>>>>>
>>>>> 
>>>> ------------------------------------------------------------------------- 
>>>>
>>>> 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
>>>> Mat...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>
>>
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Not quite. The figure opens at the correct size in the gimp but all I 
see is background - no plot stuff.
Wayne
Michael Droettboom wrote:
> This is now added in SVN, and only for the Agg backend(s). It won't 
> work with some of the other ways to save PNG files, such as Gdk, Wx 
> (non-Agg) and Cairo. If anyone was any thoughts about how to support 
> saving the resolution in those backends, please share.
>
> Also note that resolution is saved in "dots per meter" in PNG files, 
> so rounding error makes things slightly off -- for instance, 100 dpi 
> is shows up as 99.999998 dpi in the GIMP.
>
> Cheers,
> Mike
>
> Michael Droettboom wrote:
>> I'll look into this. I actually made a similar fix on another 
>> project I used to work on... It should be theoretically possible, 
>> barring any roadblocks from how matplotlib is doing things.
>>
>> Cheers,
>> Mike
>>
>> Wayne E. Harlan wrote:
>>> I would like to follow up on my first response to Bill. It probably 
>>> should be a new thread, but I'll start here. When png files are 
>>> saved with a DPI=300 argument, and I open them in the Gimp, the dpi 
>>> is only 72 (default ?). As I understand it, the dpi setting in the 
>>> graphics file tells the application opening it how big to display 
>>> it. When I use savefig with a DPI=300 and the plot figure I am 
>>> saving was created with figsize=(6,4.5) I expect the figure to open 
>>> in word or swriter at the size I specified (6" by 4.5") with the 
>>> appropriate number of pixels. However, at present, that does not 
>>> happen. The number of pixels in the figure is correct but I have to 
>>> resize it manually to get the right size. Can this be fixed ?
>>>
>>> Thanks,
>>>
>>> Wayne
>>>
>>> Bill Dandreta wrote:
>>>> I resolved the problem. It was unrelated to mpl. xv and gimp were
>>>> complied w/o png support. Recompiling with png support resolved the 
>>>> problem.
>>>>
>>>> 
>>> ------------------------------------------------------------------------- 
>>>
>>> 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
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>

Bill Dandreta wrote:
> Wayne E. Harlan wrote:
>> I would like to follow up on my first response to Bill. It probably 
>> should be a new thread, but I'll start here. When png files are saved 
>> with a DPI=300 argument, and I open them in the Gimp, the dpi is only 72 
>> (default ?). As I understand it, the dpi setting in the graphics file 
>> tells the application opening it how big to display it. When I use 
>> savefig with a DPI=300 and the plot figure I am saving was created with 
>> figsize=(6,4.5) I expect the figure to open in word or swriter at the 
>> size I specified (6" by 4.5") with the appropriate number of pixels. 
>> However, at present, that does not happen. The number of pixels in the 
>> figure is correct but I have to resize it manually to get the right 
>> size. Can this be fixed ?
> This info may or may not be useful. If I save a plot with savefig()
> without specifying a figsize, sometimes the figure gets clipped. What
> gets saved is the portion of the image that show() displays. I
> determined experimentally that figsize=(13,10) causes show() to open the
> image at full screen size and so far using that size saves an unclipped
> image.
Is this with 0.90.1 or a SVN version? What backend are you using?
Cheers,
Mike
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Wayne E. Harlan wrote:
> I would like to follow up on my first response to Bill. It probably 
> should be a new thread, but I'll start here. When png files are saved 
> with a DPI=300 argument, and I open them in the Gimp, the dpi is only 72 
> (default ?). As I understand it, the dpi setting in the graphics file 
> tells the application opening it how big to display it. When I use 
> savefig with a DPI=300 and the plot figure I am saving was created with 
> figsize=(6,4.5) I expect the figure to open in word or swriter at the 
> size I specified (6" by 4.5") with the appropriate number of pixels. 
> However, at present, that does not happen. The number of pixels in the 
> figure is correct but I have to resize it manually to get the right 
> size. Can this be fixed ?
This info may or may not be useful. If I save a plot with savefig()
without specifying a figsize, sometimes the figure gets clipped. What
gets saved is the portion of the image that show() displays. I
determined experimentally that figsize=(13,10) causes show() to open the
image at full screen size and so far using that size saves an unclipped
image.
-- 
Bill
wjd...@at...
Gentoo Linux X86_64 2.6.20-gentoo-r8
Reclaim Your Inbox with http://www.mozilla.org/products/thunderbird/
All things cometh to he who waiteth as long as he who waiteth worketh like hell while he waiteth.

This is now added in SVN, and only for the Agg backend(s). It won't 
work with some of the other ways to save PNG files, such as Gdk, Wx 
(non-Agg) and Cairo. If anyone was any thoughts about how to support 
saving the resolution in those backends, please share.
Also note that resolution is saved in "dots per meter" in PNG files, so 
rounding error makes things slightly off -- for instance, 100 dpi is 
shows up as 99.999998 dpi in the GIMP.
Cheers,
Mike
Michael Droettboom wrote:
> I'll look into this. I actually made a similar fix on another project I 
> used to work on... It should be theoretically possible, barring any 
> roadblocks from how matplotlib is doing things.
> 
> Cheers,
> Mike
> 
> Wayne E. Harlan wrote:
>> I would like to follow up on my first response to Bill. It probably 
>> should be a new thread, but I'll start here. When png files are saved 
>> with a DPI=300 argument, and I open them in the Gimp, the dpi is only 72 
>> (default ?). As I understand it, the dpi setting in the graphics file 
>> tells the application opening it how big to display it. When I use 
>> savefig with a DPI=300 and the plot figure I am saving was created with 
>> figsize=(6,4.5) I expect the figure to open in word or swriter at the 
>> size I specified (6" by 4.5") with the appropriate number of pixels. 
>> However, at present, that does not happen. The number of pixels in the 
>> figure is correct but I have to resize it manually to get the right 
>> size. Can this be fixed ?
>>
>> Thanks,
>>
>> Wayne
>>
>> Bill Dandreta wrote:
>>> I resolved the problem. It was unrelated to mpl. xv and gimp were
>>> complied w/o png support. Recompiling with png support resolved the problem.
>>>
>>> 
>> -------------------------------------------------------------------------
>> 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
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

I'll look into this. I actually made a similar fix on another project I 
used to work on... It should be theoretically possible, barring any 
roadblocks from how matplotlib is doing things.
Cheers,
Mike
Wayne E. Harlan wrote:
> I would like to follow up on my first response to Bill. It probably 
> should be a new thread, but I'll start here. When png files are saved 
> with a DPI=300 argument, and I open them in the Gimp, the dpi is only 72 
> (default ?). As I understand it, the dpi setting in the graphics file 
> tells the application opening it how big to display it. When I use 
> savefig with a DPI=300 and the plot figure I am saving was created with 
> figsize=(6,4.5) I expect the figure to open in word or swriter at the 
> size I specified (6" by 4.5") with the appropriate number of pixels. 
> However, at present, that does not happen. The number of pixels in the 
> figure is correct but I have to resize it manually to get the right 
> size. Can this be fixed ?
> 
> Thanks,
> 
> Wayne
> 
> Bill Dandreta wrote:
>> I resolved the problem. It was unrelated to mpl. xv and gimp were
>> complied w/o png support. Recompiling with png support resolved the problem.
>>
>> 
> 
> -------------------------------------------------------------------------
> 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
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Okay, I got further. I needed to use CC for c++ files.
But now I see another error:
/opt/SUNWspro/bin/cc -DNDEBUG -O -Kpic -I/usr/local/include/python2.5
-c lib/matplotlib/enthought/traits/ctraits.c -o
build/temp.solaris-2.10-i86pc-2.5/lib/matplotlib/enthought/traits/ctraits.o
"lib/matplotlib/enthought/traits/ctraits.c", line 174: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 174: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 355: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 355: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 371: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 371: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 387: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 387: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 405: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 405: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 448: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 448: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error:
Unexpected type name "PyObject" encountered.
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error: Operand
expected instead of "class".
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error: Multiple
declaration for has_traits_object.
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error: ","
expected instead of "*".
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error: Use ";"
to terminate declarations.
"lib/matplotlib/enthought/traits/ctraits.c", line 502: Error: A
declaration was expected instead of ",".
"lib/matplotlib/enthought/traits/ctraits.c", line 503: Error: Multiple
declaration for PyObject.
"lib/matplotlib/enthought/traits/ctraits.c", line 503: Error: ","
expected instead of "*".
"lib/matplotlib/enthought/traits/ctraits.c", line 508: Error: A
declaration was expected instead of "if".
"lib/matplotlib/enthought/traits/ctraits.c", line 508: Error: No
direct declarator preceding "(".
"lib/matplotlib/enthought/traits/ctraits.c", line 510: Error: No
direct declarator preceding "(".
"lib/matplotlib/enthought/traits/ctraits.c", line 511: Error: obj is
not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 513: Error: value is
not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 515: Error: obj is
not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 517: Error: value is
not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 518: Error: result
is not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 518: Error: Expected
an expression.
"lib/matplotlib/enthought/traits/ctraits.c", line 518: Error:
Identifier expected instead of ",".
"lib/matplotlib/enthought/traits/ctraits.c", line 520: Error: result
is not defined.
"lib/matplotlib/enthought/traits/ctraits.c", line 560: Warning: String
literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 560: Warning: String
literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 1007: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 1054: Warning
(Anachronism): Using _object*(*)(_typeobject*,_object*,_object*) to
initialize extern "C" _object*(*)(_typeobject*,_object*,_object*).
"lib/matplotlib/enthought/traits/ctraits.c", line 1085: Error: Only a
function may be called.
"lib/matplotlib/enthought/traits/ctraits.c", line 1088: Error: Only a
function may be called.
"lib/matplotlib/enthought/traits/ctraits.c", line 2155: Warning:
String literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 2656: Warning:
String literal converted to char* in formal argument m in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 2656: Warning:
String literal converted to char* in formal argument format in call to
PyObject_CallMethod(_object*, char*, char*, ...).
"lib/matplotlib/enthought/traits/ctraits.c", line 3324: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 3325: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 3326: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 3335: Error:
Multiple declaration for trait_type.
"lib/matplotlib/enthought/traits/ctraits.c", line 3632: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 3632: Warning:
String literal converted to char* in initialization.
"lib/matplotlib/enthought/traits/ctraits.c", line 3743: Error: In this
declaration "getclassname" is of an incomplete type "void".
"lib/matplotlib/enthought/traits/ctraits.c", line 3743: Error:
Unexpected type name "PyObject" encountered.
"lib/matplotlib/enthought/traits/ctraits.c", line 3743: Error: Operand
expected instead of "class".
Compilation aborted, too many Error messages.
error: command '/opt/SUNWspro/bin/cc' failed with exit status 1

Trying to compile on Solaris using Sun Studio 12.
any ideas?
% python2.5 setup.py build
building for GTK requires pygtk; you must be able to "import gtk" in
your build/install environment
TKAgg requires TkInter
running build
running build_py
copying lib/matplotlib/mpl-data/matplotlibrc ->
build/lib.solaris-2.10-i86pc-2.5/matplotlib/mpl-datarunning build_ext
building 'matplotlib._agg' extension
/opt/SUNWspro/bin/cc -DNDEBUG -O -Kpic -Iagg23/include -Isrc -Iswig
-I/usr/local/include/python2.5 -c src/agg.cxx -o
build/temp.solaris-2.10-i86pc-2.5/src/agg.o
cc: No input file specified, no output generated
error: command '/opt/SUNWspro/bin/cc' failed with exit status 1

I would like to follow up on my first response to Bill. It probably 
should be a new thread, but I'll start here. When png files are saved 
with a DPI=300 argument, and I open them in the Gimp, the dpi is only 72 
(default ?). As I understand it, the dpi setting in the graphics file 
tells the application opening it how big to display it. When I use 
savefig with a DPI=300 and the plot figure I am saving was created with 
figsize=(6,4.5) I expect the figure to open in word or swriter at the 
size I specified (6" by 4.5") with the appropriate number of pixels. 
However, at present, that does not happen. The number of pixels in the 
figure is correct but I have to resize it manually to get the right 
size. Can this be fixed ?
Thanks,
Wayne
Bill Dandreta wrote:
> I resolved the problem. It was unrelated to mpl. xv and gimp were
> complied w/o png support. Recompiling with png support resolved the problem.
>
>