However before doing more tests and incorporating my patch into avrdude,
give me some more time. I'm currently working on integration of my code into
par.c (which should probably be renamed to bitbang.c or something like
this), because in the ende both drivers are doing the same - bit banging.
The only difference is the port accessed.
Some small changes to the logic of par.c are also needed, which might break
the support for parallel devices, but i doubt that. I will announce the new
patch when it is finished (probably later today).
It's an insane world, but i'm proud to be a part of it. -- Bill Hicks
> I'm currently working on integration of my code into par.c (which
> should probably be renamed to bitbang.c or something like this),
> because in the ende both drivers are doing the same - bit banging.
Well, renaming files in CVS is awkward, we'd lose the file's history.
Changing the comment on top might suffice.
However, just keep in mind that Win32 requires a completely different
serial handling than Posix. For the standard serially-attached
programmers (like the STK500), this has been encapsulated in two
different source files. It wouldn't be nice to `poison' par.c with
Posix-only source code.
IMHO it would be wise to break out the actual serial device handling
into separate files again, one for Posix, one for Win32. Unless
someone promises to hack the Win32 part really soon now, it also needs
to be #ifdef'ed so Win32 builds can be done without the serbb code.
(Perhaps it would be overall a good idea to have the various
programming adapters configurable at configure time.)