Re: [avrdude-dev] Re: STK500 version 2.0 protocol

[Erik Walthinsen]

Martin Thomas wrote:
> - is lenMSB=(len>>8)&0xff, lenLSB=len&0xff portable to every
> host (little-endian/big-endian)?
I think so, but not positive.  Anyone have more experience with 
big-endian machines (e.g. PPC) to be able to say for sure?

> Maybe you'd like to use parts of my preliminary "ReceiveMessage"-
> function since it is implemented as a state-machine as described in the 
> AppNote.
Hmm, ok, I didn't pay close enough attention to that section.  When I 
get back to work on it I'll see about integrating that.

One thing I'm wondering, and haven't gone through the appnote closely 
enough yet to know for sure whether it'll work or not, is if there's 
ever a case where a command is sent and a response is *not* received. 
If that never happens, I'd be very tempted to collapse the send and recv 
code into a single command function, and build all the retries and 
resync stuff into that, instead of duplicated in most functions as it is 
in the stk500.c code.