[Matt Selsky <selsky@xxxxxxxxxxxx>]

> and I'd interpreted "s is null" to mean that s is null.  I guess that
> was an incorrect interpretation.  :-)

Sorry I wasn't more clear.

> Presumably you didn't mean "s = 0" in "s = 0 and length = 0", and meant
> "s should be set to string, but it's not since length is 0" rather
> than "...since s is null".
> 
> Unfortunately, at least some places appear to expect a null pointer to
> be returned for a zero-length string (the stuff that handles SASL
> negotiation), so changing "read_string()" is not necessarily the right
> answer; so, instead, I'll check in changes to check for a null pointer
> and display "(null)" if the pointer is null.

I just tried your patch and it works.  Thanks so much.