|
From: Ben A. <bpa...@ma...> - 2008-12-30 00:13:41
|
On Dec 29, 2008, at 6:05 PM, Ethan Merritt wrote: > On Monday 29 December 2008 14:36:53 Ben Abbott wrote: >> >> On Dec 29, 2008, at 12:15 AM, Ethan A Merritt wrote: >> >>> Well, if someone can figure out which bit of code in gplt_x11.c is >>> getting called erroneously in your original trials, I'd be happy to >>> help design a fix for it. But since I can't reproduce the problem >>> here, >>> I am entirely dependent on someone else to pin down the precise >>> source >>> of the error. >> >> >> I asked for some help on Octave's mail list. The relevant portion >> of a >> response is below >> >> On Dec 29, 2008, at 4:54 PM, Thomas Treichl wrote: >> >>> I started the debugger and did 'gnuplot> load "simple_example.gp"' >>> once again and I cannot see the problem. The reason now might be >>> that somewhere we have some kind of *timing* problem, ie. if the >>> commands are typed in per hand then this is 'slow' enough that the >>> window is not resized - the same happens if I run the program with >>> the debugger. The debugger needs some time for each line and that's >>> already enough so that the resizing problem doesn't appear. This >>> also means that we don't get any further with debugging as we cannot >>> see the problem if we use gdb. I'm currently out of ideas how we can >>> track for this problem... >> >> So the window grows when gnuplot form the command line and my example >> is loaded, but does not when run in the debugger. >> >> I checked Thomas' hypothesis by placing a "pause(0.1)" after each >> plot, and the window did not grow. > > Yuck. "timing problem" is a symptom, not a cause. > Nevertheless, please test the following one-line workaround. > If this fixes it then we can wrap it in some OSX-specific conditional > test and a comment explaining that it is just an empirical hack for > a timing problem. > > --- gnuplot/src/term.c 2008-12-12 13:06:13.000000000 -0800 > +++ gnuplot-cvs/src/term.c 2008-12-29 15:01:26.000000000 -0800 > @@ -822,6 +822,7 @@ term_end_multiplot() > * eventually statusline */ > mouse_setting.on = save_mouse_state; > UpdateStatusline(); > + usleep(100); > #endif > } Unfortunately, the window continues to grow. Ben |