Wireshark · Ethereal-dev: Re: [Ethereal-dev] Re: packet-pgsql.c changes in 0.10.9

[Guy Harris <gharris@xxxxxxxxx>]

Abhijit Menon-Sen wrote:

> In case this caused any confusion, the problem was that code like this:
>
>     gint l = tvb_get_ntohl( ... );

When was there ever code like that?  There was never any code, at least 
in any version checked into SVN, that assigned the length in question to 
a signed integer, which was the real problem - given that it has always 
been unsigned, "l > 0" just means "l != 0", which means that it was 
treating a negative length, such as -1, as a large positive length.