|Perl: the Markov chain saw|
Good question. If you can spread your processing in smaller bits and you can get the smaller bits called from the MainLoop, everything will work. The way I do it in one application is that I build my own MainLoop:
I did this, because my processing was not easily adjusted into the normal Tk::Eventloop processing and it needed to be called a lot ("live" 44Kz audio stream processing, with 100 samples per process_a_bit call)
If you want to go this route, you want to adjust the timing of the select() and process_a_bit() calls until your GUI and processing run fast enough, and it doesn't lock up the whole machine when there is no processing being done :-)
Alternatively, if you're able to run your processing based on file events or signals, you can take a look at the Tk::Eventloop documentation for registering extra events in the normal MainLoop.
Oh yes, and you want something like Tk::ProgressBar but you probably know that already :-)
In reply to Re: Keeping the user informed: A Tk/Threads question