Hi Dietmar,
nice to see you here
Quote:
I don't see why GoldED needs to be ported. I'm using UAE and it runs faster under the emulator than many native Windows editors, courtesy of AmigaOS. I'd be very surprised if a Pegasos would have trouble running it quickly, considering that a PPC appears to be better suited to emulate an m68k than a x86 (same byte order).
I agree with you here. Especially if your code has SAS/C dependencies, to modify it to compile with e.g. a flavour of gcc, is not very tempting when compared to the result - possibly a speed increase of a few "felt" percent - of a lengthy adaption.
However, what you could do is adapt GoldED to some API-advancements of MorphOS. This is no problem even from within 68K-code, it's just as if new functions and/or tags respectively had been added to the old 68K-library. For starters, an adaption to the skinning functions of MOS Intuition would be a welcome extra (this mostly affects how you handle borders and titlebars and sliders in them).
Furthermore, if there are more CPU-intense tasks to do which would profit from a PPC-native version, you could take an approach that is similiar to mine on MorphOS: encapsulate the CPU-intensive routines in seperate source files or make sure "superfluous" routines are defined out, adapt this code to work with e.g. gcc and/or vbcc, put it in a seperate binary (server), launch it at startup and then communicate between the main application and the server via Exec messages (which contain IDs for functions, function parameters, etc).
--
Felix Schwarz
