Wireshark · Ethereal-dev: Re: [Ethereal-dev] Patch to fix possible crash in NCP dissector

[Guy Harris <guy@xxxxxxxxxxxx>]

On Sep 23, 2003, at 2:08 PM, Greg Morris wrote:

> On subsequent dissection it is possible that a reply packet is in the 
> trace that does not have a cooresponding request packet. The column 
> data update did not check for valid request data prior to updating. 
> This patch ensures that we check for valid request data prior to 
> updating the column data.

Checked in.

Does that code set the Info column differently depending on whether 
"ncp_tree" is non-null?  If so, that should be fixed, either by 
arranging to set it even if "ncp_tree" is null or by building a 
temporary protocol tree if "run_info_str" is set.