Page 2 of 2
Posted: Wed Mar 08, 2006 4:10 pm
by Wismuth
"DirectInput", another checkbox I hadn't noticed! It was off in my test. I turned it on and have the same mouse problem. I traced the dinput calls and it seems to make more of them, the new ones being keyboard related. In summary:
DirectInput=off uses direct input for mouse
DirectInput=on uses direct input for mouse+keyboard
Posted: Wed Mar 08, 2006 5:16 pm
by Myrd
Sounds like the Wine folks need to work on becoming more compatible with Windows' DirectInput mouse functionality.
EDIT: I did some searching, and the bug does seems to be in wine, and here's the source file: http://source.winehq.org/source/dlls/dinput/mouse.c
Basically, when Myth calls DirectInput's Acquire method, to get a pointer to the mouse so it can find its coordinates, Wine decides to center the mouse for some reason. Windows obviously doesn't work this way, so I'm not sure what functionality Wine is trying to emulate.
EDIT2: I submitted a bug report to WINE's bugzilla: http://bugs.winehq.org/show_bug.cgi?id=4777
Edited By Myrd on 1141858191
Posted: Wed Mar 08, 2006 6:26 pm
by Wismuth
I agree that it's a shortcoming in Wine. Here's more info about it:
http://wiki.winehq.org/DInput
Wine aims to be compatible with Windows, so of course you can sit and wait for them to get more compatible. Or... you can consider Wine to be a 3rd platform (alongside Mac and PC) which is like Windows but with bad DirectInput mouse support, and support that 3rd platform (more or less Myth on Linux).
In any case, I'm also curious about the original reasons why Melekor rewrote the input code in the first place, and I wanted to point out that Wine support is another reason to use the rewritten code, if you guys are weighing the options.
Posted: Wed Mar 08, 2006 10:10 pm
by Myrd
I'd rather make a patch for Wine. That particular problem seems pretty trivial after I looked at the code - compared to for example OpenGL / Direct3D working.