Wrong logic for handling large BLOBs?

[movermeyer]

Shouldn't line 204 of fbcore.py be

while n != 2:

instead of

while n == 1:

I'm finding that the protocol will return a 0 instead of a 1 every 507873 bytes (strange number).
As a result of this logic, my BLOB is being truncated.

I haven't been able to prove that 0 is an OK response as I can't find the spec and reading the C++ code isn't my strong suit. But changing the Python to 'n != 2' seems to work.

Your input would help a lot.

P.S: I noticed that the same logic is also found in your Go version
I've filed a similar issue there.

[nakagami]

Thanks agin.