enableTranslation fails in Win98.

Started by psiegel, July 28, 2003, 15:31:35

Previous topic - Next topic

psiegel

I'm using lwjgl 0.6, so if this is a bug that was already fixed in 0.7, just let me know.  I've been calling both Keyboard.enableBuffer() and Keyboard.enableTranslation in my code, and it seemed to be working fine.  At least, it was working under Windows 2000 and RedHat Linux 9.  Recently, however, we tested our code on a Windows 98 box, which I can grab any specs you want from.  

I discovered that on this machine, Keyboard.character is always garbage.  I'm not sure the exact character it's returning, or wether it changes, but it is certainly not the correct representation of Keyboard.key.  This has only appeared on one other machine we've tested on, which was also running Windows 98.  That seems to be the only common thread between the two.

I'm happy to provide any more info, or run any kind of tests you like.

Paul

elias

Interestingly, I alredy "fixed" AF/LWJGL on win9x by replacing toUnicode with toAscii in org_lwjg_input_Keyboard.cpp. But that might have broken in ther meantime as I have no win9x to test on. If you are able to build the win32 version of lwjgl, do try to tinker around with the translation code in that file, or else simply provide me with some info:

With translation enabled, try to translate a few standard ascii keys on both a working win2k and an win9x. E.g. try the keys 'a' through 'c' or something and dump the character values from java by casting the char to an int. My suspicion is that I have screwed the byte order or something like that.

Cas, is AF still working on win9x machines?

- elias

elias

.. And I'd still like you to upgrade to pre 0.7 to receive more testing of the extensive API changes done in that version. Or even better, the current CVS version ;) And I can't honestly remember if any keyboard fixes have been comitted from 0.6 to now.

- elias