Re: [avr-gcc-list] Experimental WinAVR/avr-gcc update for

[Joerg Wunsch]

As "E. Weddington" <address@hidden> wrote:

>Jörg Wunsch, when he was writing the AVR to COFF converter, 
>pointed out that the COFF format is old and is very 
>unsuitable as a debugging format. Additionally, it cannot 
>support C++ features.

To their defense, they've probably chosen it simply because they
didn't know better, and (but) because they didn't research on a good
object format (otherwise they'd have chosen ELF right from the
beginning, because it's the logical successor of COFF).  Remember that
Windoze is using COFF itself (but not that cramped and closed-minded
COFF native debugging), so that might have influenced their decision.

>UBROF is proprietary to IAR. There is also a feature in AVR 
>Studio that, if you use IAR, you can printf values (such as 
>in debug code) and AVR Studio will automatically display 
>them in a windows.

That's not an UBROF feature.  It could be had on COFF as well, ICC and
CV do support it.  However, it's currently only available in AS3, not
yet in AS4, that's why i never made any attempts to provide a
framework for it.  This is btw. a lousy implementation: it's done by
specially-named function entry points where AVR Studio reads/modifies
known registers when reaching these functions.  The correct way would
be to have a real UART simulation (where you e. g. also see when
you've got UART timing problems), as VMLAB does.
-- 
J"org Wunsch                                           Unix support engineer
address@hidden        http://www.interface-systems.de/~j/