[avr-gcc-list] Has anyone gotten the Pololu software working on Mac OS X?

[David B. Gustavson]

There are instructions up to a point, at https://www.blogger.com/comment.g?blogID=5983110881018172452&postID=1421669160224186648&page=1 
 (pointed to by the Pololu web site), but having followed that path I  
get a strange failure that I don't know how to diagnose.


When I modify Makefile as suggested and compile OrangutanSV168Demo.c ,  
the result is
David-Gustavsons-Mac-Pro:OrangutanSV168Demo dbg$ make
/opt/local/bin/avr-gcc -mmcu=atmega168 -Wall -Os -std=gnu99 -funsigned- 
char -funsigned-bitfields -fpack-struct -fshort-enums -MD -MP -MT  
OrangutanSV168Demo.o -MF dep/OrangutanSV168Demo.o.d -c  
OrangutanSV168Demo.c
/opt/local/bin/avr-gcc -mmcu=atmega168 -Wl,-gc-sections -Wl,- 
Map=OrangutanSV168Demo.map OrangutanSV168Demo.o -lpololu -o  
OrangutanSV168Demo.elf
avr-objcopy -O ihex -R .eeprom OrangutanSV168Demo.elf  
OrangutanSV168Demo.hex
avr-objcopy -j .eeprom --set-section-flags=.eeprom="alloc,load" -- 
change-section-lma .eeprom=0 --no-change-warnings -O ihex  
OrangutanSV168Demo.elf OrangutanSV168Demo.eep || exit 0
avr-objdump -h -S OrangutanSV168Demo.elf > OrangutanSV168Demo.lss

text data bss dec hex filename
0 0 0 0 0 OrangutanSV168Demo.elf

and not surprizingly the hex file has only one byte in it or perhaps  
it means zero, not sure, and loading it into the OrangutanSV disables  
the Orangutan.

I also tried WinAVR and AVR Studio in XP via Parallels, and there the  
demo does compile correctly and produces the right hex file. However,  
I can't get the Pololu programmer to be recognized! But I can copy the  
hex file to the Mac file system, and use avrdude there.
But I really want to get the compile/link process described here to  
work. Where can I find help with diagnosing the problem? I'm guessing  
some little thing has been changed so an option somewhere needs to be  
set differently, but how to isolate the problem? Some clues from file  
sizes:
The .elf file sizes match, though their contents are very different  
(both unreadable by me, not text) but the .lss, .map, and .o files are  
less than half the size they should be.

Can anyone suggest how to track down the problem?
I'd think it must be in the linking stage or there'd be some error  
messages to clue me in. Even then I'm surprised it just generates a  
zero output.

Dave