Wireshark-dev: Re: [Wireshark-dev] [Wireshark-commits] rev 21637: /trunk/gtk/ /trunk/gtk/: follow_dlg.c

From: Guy Harris <guy@xxxxxxxxxxxx>
Date: Wed, 02 May 2007 12:52:10 -0700

Joerg Mayer wrote:

That's the type of bugfix that doesn't make much sense. Just a dummy
read to squelch the warning without fixing the reason that the waring
was generated in the first place:

Yes, that's the wrong fix.

To state this differently: What happens, if the read doesn't return the
necessary amount of data?

Bad things - or, at least, not ideal things.

I've checked in changes in some places to check for read failures. Some of the RTP code doesn't have an immediately obvious way to fail out of the analysis operation on a read failure, so I left that as is for now, but it should be fixed.