From: Kevin W. <kw...@co...> - 2012-06-01 00:18:57
|
On 5/31/12 7:56 PM, Ned Deily wrote: > Another data point, FWIW: I just tried the OP's test with Python 3.2.3 > linked with an X11 Tk 8.5 (a current MacPorts Python 3.2.3) rather than > the python.org 3.2.3 linked with A/S Aqua Cocoa Tk 8.5.11. While I was > easily able to reproduce the problem with the latter, I could not > reproduce with the former. My guess is that the symptoms the OP are reporting are related to this bug: http://sourceforge.net/tracker/?func=detail&aid=3166688&group_id=12997&atid=112997 I had originally said in the comments to that bug that it may simply be a difference in performance between Cocoa and Carbon, and Cocoa / X11, but I think it goes deeper than this, into some very low-level issues with the event loop integration between Cocoa and Tk, which affects things like screen redraw. I go into those issues here: http://sourceforge.net/mailarchive/message.php?msg_id=27868831 Bottom line, these issues are most likely insoluable without some significant re-architecting of the event loop integration between Cocoa and Tk--a task that is beyond my skills. The original author of the Tk-Cocoa port, Daniel Steffen, is no longer able to devote significant time to maintaining the port because of his employment at Apple, certainly not the level of time required to make changes to the event loop. As I mention in my archived message, it's sometimes possible to work around the screen redraw issues through judicious use of "after" and "update idletasks" commands, but that's often a matter of trial and error. I'm sorry I'm not able to provide further guidance on this or make a more substantial effort to tackle these bugs. As I've said before on this mailing list, patches are welcome, but no one has stepped forward to offer one in this particular case. Thanks, Kevin -- Kevin Walzer Code by Kevin http://www.codebykevin.com |