Re: [avr-gcc-list] Butterfly code ported to gcc with bugs...

[Desi]

Thanks Joerg, 

your comments was very useful, but
but now I need to know how to translate,
this:

#pragma vector = PCINT0_vect
__interrupt void PCINT0_interrupt(void)

or better this kind of code, I'm looking arround
and there is not to much info, and sorry for the
comments of the code (in spanish), was some kind
of dark joke because avery thing is in english, 
in the future, I'll put all the comments in english...

Thanks for all, and if somebody else 
want to join in this efford I will appreciate.

Regards...

Desi.
CICESE
Ensenada, Baja California, Mexico.


> "Desi" <address@hidden> wrote:
> 
> > I did port the butterfly code to GCC, 
> > but Still have a lot of warnings, no 
> > errors, and need help to debug this code
> > may be if some one of you are interested,
> > let me know, here is attached the code 
> > with the make file, to run just do make or 
> > make all
> 
> Sigh, i wish these Atmel folks were less sloppy about upper/lowercase
> filenames.  It took me quite some time to get the various #include
> statements match the actual filenames.
> 
> I think most of the work will be to convert the interrupt handlers
> from IAR style to avr-gcc style.
> 
> The »array subscript has type `char'« can certainly be ignored, though
> it might be worth the while to research whether the default signedness
> of the gcc "char" type is OK (i guess gcc will promote them to a
> signed int before actually applying them as an array index).
> 
> The »comparison is always false due to limited range of data type«
> warnings are more serious.  Perhaps this as well as the integer
> promotion mentioned before can be worked around using -funsigned-char
> in the gcc CFLAGS.  The clean solution though would be to explicitly
> use "signed char" vs. "unsigned char" in the code (resp. to use the
> appropriate types from the C standard header <inttypes.h>).
> 
> »built-in function `index' declared as non-function« could only be
> fixed in the gcc sources; `index' should not be a built-in function
> for avr-gcc.
> 
> Finally, the linker warning »size of symbol `LCD_character_table'
> changed from 2 to 108 in LCD_driver.o« should be resolved, but given
> that it's a single warning, this will probably be the last one.  This
> most likely corresponds to the compiler warning »array
> `LCD_character_table' assumed to have one element« which is just an
> indication of sloppy programming style.  Hmm, given that the comment
> on that variable in demo.c is in Spanish (which i can only partially
> understand), my guess is that you're missing an "extern" keyword
> there.
> -- 
> J"org Wunsch                                         Unix support 
engineer
> address@hidden        http://www.interface-
systems.de/~j/
> 
> _______________________________________________
> avr-gcc-list mailing list
> address@hidden
> http://www.avr1.org/mailman/listinfo/avr-gcc-list
> 
> 

--