quote:
Originally posted by eli:
<STRONG>I played with it a bit more and adding the endian conversions after the fread() calls fixed the graphics on PowerPC.The text input is also a bit wacko, I partly fixed it by taking out the cast to char on the Unicode value, but the virtual keycode lookup seems problematic. You can probably get away with it, but the table looks wrong, apparently OS X uses the keycodes from the good old Apple Extended Keyboard 2 as described here keycodes from OS X.
All of my changes were limited to enabler_osx.cpp, let me know if you'd like the new file.</STRONG>
Yeah, the text input is a bit wacko... It's been a while since I've done regular Mac coding, and a lot has changed. I was trying to just get from the mac key codes to windows, to keep as much of the port work in enabler_osx rather than scattered elsewhere through the code. (I certainly don't write cross-platform code this way, but in this instance, it seemed better to try and isolate the mac stuff that start telling Toady how to rearrange everything.
But hey, another Mac-head lookin' at my code is just fine.
EDIT: Oh, and on the endianness issue... Doing the porting work on an Intel mac meant I could leave that to last... since I try not to mix sources of potential problems. Once that was finished, I had planned to go back and fix then endian stuff for PPC. I haven't yet looked to see how Toady was reading
things in... as text, structured, or as binary blobs dumped into structures. Each requires different endian-ness handling, and it's not difficult stuff, just needs to be looked at.
[ December 29, 2007: Message edited by: mattmoss ]