AVRDUDE reports incorrect signature... Is this the bootloader talking? ;)

[GoogleCodeExporter]

AVRDUDE included in Arduino 0022 was giving me fits when it reported the 
"wrong" signature for a ATmega88PA, when the signature it was reporting was for 
the '88!

I'm using an Optiboot bootloader compiled for use with an '88PA, but I think I 
must have "corrected" the AVRDUDE.CONF file to allow ISP programming.

Does the bootloader supply AVRDUDE with the "signature", based on the setting 
in the make file?  And if so, can they be corrected?

In the grand scheme of things, this is actually slightly preferable since it 
won't require the end-user to change the AVRDUDE.CONF, but it's still a little 
odd!  ;)

Original issue reported on code.google.com by awebu...@hotmail.com on 20 Oct 2011 at 11:10

[GoogleCodeExporter]

Just to clarify, the driving issue was it was reporting a signature mismatch, 
when -technically- none existed.  It just seems that the bootloader spits out 
whatever signature it was given.

Original comment by awebu...@hotmail.com on 20 Oct 2011 at 11:11

[GoogleCodeExporter]

Yes, optiboot happily lies when asked to provide a signature, getting the 
values it returns from the compiled code rather than the actual signature 
bytes.  As you say, this is normally a "feature"; permitting arduino-class IDEs 
and users from needing to worry about the differences between 88 and 88a (or 
328 and 328p) when they are operating at a code "level" that makes any 
differences irrelevant.  (actually reading the signature bytes would probably 
push optiboot over its 512byte size limit.)

Original comment by wes...@gmail.com on 30 Oct 2011 at 7:19

• Changed state: WontFix