matplotlib-devel Mailing List for matplotlib

John Hunter wrote:
>>>>>>"David" == David Fraser <da...@sj...> writes:
>>>>>> 
>>>>>>
>
> David> This is fantastic, I hope it fixes the crashes I've been
> David> having which I finally seem to have traced down to
> David> matplotlib freeing images ... 
>
>Hi David,
>
>It probably won't fix it, because I didn't do anything that should
>effect memory handling, but we'll keep our fingers crossed. If you
>continue to experience problems and have any more information about
>the bug, pass it on.
> 
>
Actually it has seemed to have fixed it. I'm not going to go back and 
work out why though ... thanks for the update! Exactly when I needed it!
> David> One question though:
> David> FigureManagerBase's add_axes, add_subplot and
> David> set_current_axes methods have been removed (and this
> David> doesn't seem to be mentioned in API_CHANGES). What should I
> David> do to replace a call to add_axes?
>
>Sorry, I got lazy. I updated API_CHANGES on the web. I thought only
>the GUI backends were using add_subplot, set_current_axes, etc, and I
>ported the changes to all the GUIs. See my last post
>http://sourceforge.net/mailarchive/forum.php?thread_id=6207208&forum_id=36187
>for more information.
>
>Basically the figure class itself maintains the current axes - the
>motivation for the changes is in the post above. Call fig.add_axes,
>fig.add_subplot, fig.gca and fig.sca (set_current_axes).
>
>Hope this helps,
>JDH
> 
>
OK thanks, worked it out in the meantime but the list is slow :-)
David

David Fraser wrote:
> John Hunter wrote:
>
>> This is primarily a bug-fix release from 0.65, though a couple of
>> little features managed to sneak in
>> 
>>
> This is fantastic, I hope it fixes the crashes I've been having which 
> I finally seem to have traced down to matplotlib freeing images ...
> One question though: FigureManagerBase's add_axes, add_subplot and 
> set_current_axes methods have been removed (and this doesn't seem to 
> be mentioned in API_CHANGES). What should I do to replace a call to 
> add_axes?
Sorry, my stupid question, have now read the message on the list about this.
The answer is call figuremanager.canvas.figure.add_axes - but maybe this 
should be added to the API_CHANGES document
David

>>>>> "David" == David Fraser <da...@sj...> writes:
 David> This is fantastic, I hope it fixes the crashes I've been
 David> having which I finally seem to have traced down to
 David> matplotlib freeing images ... 
Hi David,
It probably won't fix it, because I didn't do anything that should
effect memory handling, but we'll keep our fingers crossed. If you
continue to experience problems and have any more information about
the bug, pass it on.
 David> One question though:
 David> FigureManagerBase's add_axes, add_subplot and
 David> set_current_axes methods have been removed (and this
 David> doesn't seem to be mentioned in API_CHANGES). What should I
 David> do to replace a call to add_axes?
Sorry, I got lazy. I updated API_CHANGES on the web. I thought only
the GUI backends were using add_subplot, set_current_axes, etc, and I
ported the changes to all the GUIs. See my last post
http://sourceforge.net/mailarchive/forum.php?thread_id=6207208&forum_id=36187
for more information.
Basically the figure class itself maintains the current axes - the
motivation for the changes is in the post above. Call fig.add_axes,
fig.add_subplot, fig.gca and fig.sca (set_current_axes).
Hope this helps,
JDH

John Hunter wrote:
>This is primarily a bug-fix release from 0.65, though a couple of
>little features managed to sneak in
> 
>
This is fantastic, I hope it fixes the crashes I've been having which I 
finally seem to have traced down to matplotlib freeing images ...
One question though: FigureManagerBase's add_axes, add_subplot and 
set_current_axes methods have been removed (and this doesn't seem to be 
mentioned in API_CHANGES). What should I do to replace a call to add_axes?
David