matplotlib-users Mailing List for matplotlib

>>>>> "andrea" == andrea gavana <and...@ti...> writes:

    andrea> No matter which is my choice, the window crashes and I get
    andrea> the usual XP message about sending error messages to
    andrea> Microsoft and so on. It is a strange exception, if I
    andrea> resize the window VERY slowly, there is no exception
    andrea> (usually).

There is a known bug in Agg on Windows XP that appears on repeated
draws (resizing does this) that causes a segmentation fault.  I've
only seen this in the context of draw markers (eg

  plot(x,y,'o') 

where 'o' is a plot marker symbol.  Do you have markers in your plot?

Unfortunately, even though I think I know where the bug is occurring,
I don't yet have a fix for it.  If it is too annoying, you may want to
roll back to 0.71 until I figure this one out.


    andrea> Problem 2: Considering always the keypress_demo.py, if I
    andrea> run it (without resizing the window ;-) ) and I press the
    andrea> key "g" (for the grid), the grid appears and suddenly
    andrea> disappears. I opened the keypress_demo.py and I commented
    andrea> out the draw() command (that follows the grid() command),
    andrea> and now the grid does not disappear. What could be the
    andrea> problem with draw()?!? I have to use it in my application,
    andrea> but I am not able to display the grid...


Oh, I get it :-)

the 'g' key is now a default part of matplotlib -- eg pressing 'g'
toggles the grid on any axes.  I wrote the keypress_demo before making
this behavior default, so what you see in the demo is that you toggle
grid two times in keypress_demo, once in the default axes handling,
and once in the demo.  And so it ends up turning it on and back off.
Time to update the demo!

    andrea> Thank your for every suggestion, and sorry for the long
    andrea> post.

Bug reports are always welcome -- thanks!

JDH

>>>>> "Eli" == Eli Glaser <eg...@se...> writes:

    Eli> Thanks...Originally I opened up the PS file in a text editor
    Eli> and saw lines of hex data - I assumed it was a raster version
    Eli> of the plot.  Now I see that hex data is describing the
    Eli> fonts, and there are standard vector descriptors at the end
    Eli> of the file.

That's right -- to support mathtext and font compatibility between the
raster formats and ps output, we embed the truetype font information
directly into the file.  Unfortunately, we haven't yet succeeded in
getting just the characters we need embedded in the file, and have to
dump the whole font file.  We have a couple of approaches to fix this,
and so the file size issue should get better in future releases.

The only rasters we embed in postscript are images you create with
imshow or figimage -- everything else is vector graphics.  

JDH

Hi all,

I've built a new matplotlib package for OS-X. Bob Ippolito has
graciously hosted it at:

http://pythonmac.org/packages/

This is an upgrade to matplotlib 0.72.1, and it should work with Agg, Wx
and Tk. I have not included PyGTK, as you need fink and/or darwinports 
to run PyGTK, and if you're running those, you can get matplotlib 
through them (I think). This version should work on an stock OS-X 10.3.*

You, or course, need Numeric or numarray (available on the same site), 
and TK or wxPython, if you want to use those back ends.

-Chris


-- 
Christopher Barker, Ph.D.
Oceanographer
                                     		
NOAA/OR&R/HAZMAT         (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chr...@no...