Interpret first byte as length if command doesn't match expected

[cellularmitosis]

Multiple "read" commands in a row seem to fail using the existing code.

A work-around seems to be that if the command byte wasn't as expected, to reinterpret the first byte as the message length, which seems to work.

[cellularmitosis]

Hmm, but this makes my '*IDN?' loop messed up. The first character is missing:

b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'
b'EWLETT-PACKARD,34401A,0,10-5-2'

Perhaps there is an off-by-one error somewhere in the buffer handling of this code?

[sirmo]

Hey there @cellularmitosis this is Muxr from eevblog forums. I have a PR that could be of use to you. You can check it out from my fork here: https://github.com/sirmo/ugsimple-usb-gpib/tree/fixes1

Or see it here: #7

I have implemented a new ask() method which resolved all the issues I was having with this lib. In fact I was able to backup the NVRAM contents from my hp 3458a with it.

Maybe the maintainer can take a look and figure out how to fix the read() method which has desyncing issues for me too. I didn't want to overwrite it since I am not 100% sure how it's supposed to work. Instead I just use the new ask() method in that Pull Request (7).

p.s. Anyways, hope that's useful to you, since it beats paying couple of hounded of dollars for another GPIB interface. (better to use that on low tempco resistors :P)

[cellularmitosis]

@sirmo awesome! 😀