[avrdude-dev] Re: init/exit-cleanup (was: Support for Futurlec AVR development boards)

[Brian Dean]

On Thu, Jan 29, 2004 at 02:50:27AM +0100, Jan-Hinnerk Reichert wrote:

> Another way would be to implement programmer-specific options.
> If main finds an option that it doesn't know, it just calls 
> pgm->parse_option().
> 
> There is one issue here, either we insist in "-E" appearing after 
> "-c" (change of behaviour) or we need to buffer the options until we 
> know the programmer.
> 
> This would also allow us to easily implement other programmer-specific 
> stuff (e.g. the STK500-stuff like "set voltage").
> 
> BTW: Programmers that don't have programmer-specific options, would 
> not have to be changed. We could just add a default-functions in 
> "pgm.c" that returns -1 (i.e. "can't parse this option").
> 
> > Note that we lost this capability at some point and we don't
> > currently honor the exit specs, so this is not something that is
> > working at the moment.
> 
> I just tested the exitspecs on Linux and they are fine except for the 
> case mentioned in my last post.

Ah, OK.  I see that after we do the pgm->save() in main, we then apply
the exitspecs directly to pgm->ppidata.  This is probably one of those
things that didn't get abstracted very well back when I converted from
a purely PPI programmer to supporting multiple programmers.  I
certainly didn't do a very good job of data hiding :-)

However, I could have sworn I got a report of the exitspecs not
working.  I must have imagined it.  Sorry about the confusion.

Regarding the per-programmer options for options that main() doesn't
know about, that sound's pretty useful, as you mentioned, in the case
of the STK500 since it can be configured in several ways.  Sounds like
a good idea to me.

-Brian