I lack the visibility to see how "next" is a replacement for "break" i.e. "last" seems a far more more natural replacement to end a loop. "last" certainly stopped any more bindtags being looked at but also took the interpreter out at the same time suggesting that the loop is the main message loop. It never occured to me to try and continue that loop.
Its a tcl mistake, there should have been an actual Tk API call to break the chain, thats how all the other GUI toolkits do it; what Tk team chose to do was peek behind the curtain ... and then Tkx.pm guys didn't catch this corner case like other guys did
For example, this is how the other guys handle it http://grep.cpan.me/?q=_TK_BREAK_
Where does Tkx stick it? I dare not look :) bored with that one
But I do remember that you can always write stuff in tcl, unofficially Tkx::eval('tcl source code');
Once again thanks. BTW is there a best place/file I should be looking at to get your level of understanding re Tkx? Perlmonks? :)
Maybe this list archived at http://www.nntp.perl.org/group/perl.tcltk/?
Seriously though, you just about have all my secrets ... all GUIs work like this , one event loop that calls stuff... peek at the perl sources to see holes/opportunities ... super search for author [zentara][Anonymous Monk] to find an example ... write a short test case and then bug the maintainer ... repeat
pretty much everything is possible, its just a matter of how much blind poking you want to do, the more things are used the less holes they have...
tk tips like this... tk scoping issue, tk closures people, Re: TK Submenus (Tk::Menu , global variables/ spirit of strict), tk scoping function argument passing, can't share Tk mainwondow between threads
|