Re: [AD] Bug in readkey() with numeric keypad?

[ Thread Index | Date Index | More lists.liballeg.org/allegro-developers Archives ]


Elias Pschernig wrote:

On Fri, 2005-12-02 at 09:29 +0100, Evert Glebbeek wrote:
Probably an oversight, as from 4.1.12 to 4.2.0 there were several keyb
changes,
Possibly. Is the behavior inconsistent with the documented behavior (can't check right now)?

eventually i decided to write my own as the changes/bugs kept cropping up, and i thought, well there will be
more bugs in the future (most likely, and today is that day).
Well, cheers. Any chance you could help out fix those bugs you're talking about? As for changes, why does that matter? As long as the documented behavior does not change what difference does it make?


His is not using DirectX AFAIK. Simply using the windows messages indeed
has several advantages:

- Translation to unicode is much easier, since Windows does it all for
us

- Nothing is done in the input thread, it simply is windows messages
received in the main windows loop.

- Some of the mentioned problems about the current one wouldn't exist,
for example, Alt+number sequences, or Numkey behavior, would be exactly
like in Windows itself.

So, I guess, we should switch to the non-DX driver if you submit a
patch.


I use the WM_ messages, but my implmentation is not complete, i simply detect only the ones i need. I also use a C++ threadsafe template FIFO buffer, which i guess will not be very useful for allegro, however pulling it to pieces and making it C code is still very do-able. Im not convinced my code is 100% bug free yet.. the alt-key sometimes gets stuck down, i suspect something to do with those bothersome weird WM_ msg combos involving F10, but all the other keys seem to work.

aj.




Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/