Re: [avrdude-dev] patch #5655, again

[Joerg Wunsch]

As Windell H. Oskay wrote:

> I was able to fix it by applying patch #5655, which is apparently
> still open after three years.  I've been recommending this patch to
> people since 2006...  and it feels silly to keep doing so.  Any way
> to close it up?

Yes, the patch is rather a hack, see my comment which nobody
bothered to reply to within those 2.5 years:

"I'd really like to see some more analysis for the actual cause
of this problem before adding a hack that simply disables the
sign-on command."

So this would require someone with a Mac willing to actually *do*
some analysis first.

The hack gratitously removes the sign-on command when talking to an
AVRISPmkII, which is obviously working for about everyone else on
earth.  Why is that not working for MacOS users (and apparently, even
only for MacOS/Intel users)?  Can you get me a communication trace
(-vvvv) for both (patched/unpatched) cases?

Note that the JTAGICE mkII is explicitly taken out from the sign-on
message since the STK500v2 protocol there is only wrapped up within
the JTAG ICE mkII protocol, so the sign-on to the JTAG ICE happens at
an earlier stage in a different way.  When applying the hack, talking
to an AVRISPmkII would issue no sign-on command whatsoever, thus
violating the protocol described by Atmel.

Btw., the patch no longer applies at all, due to other structural
changes.  Did it apply for you?

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)