[rkward-devel] Assertion failure when interrupting output

[Michel Lang]

Am Montag, 21. Mai 2007 14:49 schrieb Thomas Friedrichsmeier:
> I have removed the assertion failure, but I'm not sure, I actually fixed
> whatever caused the crash. Could you try to reproduce after an svn up?

Well, I also tried it a few times, no crashes this time. I'll report back when 
experiencing further crashes.


> I'm afraid we just don't have this sort of low-level control over the
> syntax highlighting. I've optimized the syntax hightlighting definition to
> be *slightly* faster on lots of output, but the effect is not too
> noticeable. The ratio is not quite as bad on my machine with roughly 20
> seconds inside RKWard vs. roughly 4 seconds in plain R. Right now, I have
> no good idea on how to optimize any further.

I'm still far away from 20 seconds. 
Could it be caused by my dual monitor setup or are there major speed 
improvements in katepart since kde-3.5.5?


> I think that since the command is cancellable (and hoping the crash is
> fixed), the problem is not too grave, and I'd rather not mess with any R
> options just for this issue. Do you disagree?

Agreed. But maybe we could include a configuration dialog, and set the 
R-options on startup if they differ from the defaults. To avoid confusion if 
the R developers are changing the defaults, I would suggest a checkbox and a 
text field for each option (and it's easier maintainable, too). 
Users touching these options hopefully know what they are doing ;)
Correct me if I'm wrong, but this should be safe enough.


Thanks for your efforts, Michel