| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
comboboxes are still an issue, and there's experimental code to try to fix them, but it doesn't work...
|
|
now. If I change my mind, it's a documentaiton change =P
|
|
window, documented it was undefined (Windows and GTK+ behave reasonably but at least Windows is not documented here; Mac OS X shows a sheet attached to invisible where the titlebar should be and then considers the window closed), and added the panic() checks for uncreated Windows.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
implementations will come next.
|
|
flag for maximum safety. Also added more future plans.
|
|
leave it to the implementation.
|
|
|
|
|
|
DefWindowProc().
|
|
|
|
they no longer return errors.
|
|
since I'm going to separate creating and opening (Open() will call Create()).
|
|
situation /is/ exceptional, after all (and we need to just get rid of all errors sometime).
|
|
correctly... No idea why that TODO was there; I know I added it when I was too tired to think though.
|
|
|
|
change the behavior I can do it later.
|
|
|
|
proposal is too risky.
|
|
|
|
application modal! The real issue is window transience... and that's where platforms differ greatly.
|
|
resolved.
|
|
TODO about the 32-bit build being unstable.
|
|
converting existing code... Now to test on real Windows.
|
|
|
|
|
|
|
|
went against what we're trying to do with strings in Windows.
|
|
|
|
|
|
previous commit. This has to be first, otherwise a few other things will fail to compile...
|
|
handling. I'll convert the existing code to use it shortly.
|
|
(GetKeyState() vs. GetAsyncKeyState()) is indeed what we want to do (the former).
|
|
entries in the Modifiers table table just to be safe.
|
|
|
|
decided it would be safest to let event handlers propagate, as I did with Areas earlier.)
|
|
documenting that we don't do it.
|
|
modality to the other windows in the program. Does not affect resizes.
|
|
|
|
|
|
on/document/verify vertical alignment of Labels.
|
|
gtkcalls_unix.go.
|
|
|
