Perhaps we should include a modified libpcap source tree in
the ethereal distribution until wiretap can capture packets,
to make sure the Linux timeout bug and the lex-restart bug
are fixed for Ethereal.
I'm working on the wiretap capture routines right now, but they won't
be ready for the next few ethereal releases.
BTW, on my system, RedHat 6.0 Linux ix86, I was not able to
reproduce Roberto's problem.
--gilbert
----- Forwarded message from Roberto Machado <racm@xxxxxxxxxxxxxxx> -----
Received: from trex.centroin.com.br (trex.centroin.com.br [200.225.63.128]) by paat.pair.com (8.9.1/8.6.12) with ESMTP id PAA04734 for <gram@xxxxxxxxxx>; Wed, 1 Dec 1999 15:47:48 -0500 (EST)
X-Envelope-To: <gram@xxxxxxxxxx>
Received: from trex.centroin.com.br (trex.centroin.com.br [200.225.63.128])
by trex.centroin.com.br (8.9.3/8.9.3) with ESMTP id SAA05559
for <gram@xxxxxxxxxx>; Wed, 1 Dec 1999 18:47:25 -0200 (EDT)
Date: Wed, 1 Dec 1999 18:47:25 -0200 (EDT)
From: Roberto Machado <racm@xxxxxxxxxxxxxxx>
To: Gilbert Ramirez <gram@xxxxxxxxxx>
Subject: Re: [ethereal-users] Problem when starting ethereal
In-Reply-To: <19991201141408.C16989@xxxxxxxxxx>
Message-ID: <Pine.BSI.4.10.9912011825070.15234-100000@xxxxxxxxxxxxxxxxxxxx>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Hi, Gilbert!
The tests work ok. And I tried to make a capture under my own account
now, and it worked. Sorry to report a non existing problem. I'm
pretty sure this *was* happening, but I can't reproduce it right now.
I think this was happening with.... Wait a minute!
[a few minutes later]
Ok, what I saw might be related to the libcap's buffer flush error.
This is how you can, hopefuly, reproduce it:
- run ethereal
- open the capture windows
- fill the filter with an invalid entry
- parse error
- reopen the capture window
- remove the filter (delete)
There, capture goes ok, but there is a "The file /var/tmp/ether...
does not exist".
It happens as root, too. I was induced to error because when it
happened I always restarted etherealas root, and it was a fresh
session without any parsing error.
Regards, Roberto
On Wed, 1 Dec 1999, Gilbert Ramirez wrote:
> The code that creates the temporary capture file does this:
>
> old_umask = umask(0077);
> tmp_fd = mkstemp(namebuf);
> umask(old_umask);
>
> It changes the umask before creating the file. In your configuration,
> can you run this successfully from the shell?
>
> umask 0077; touch /var/temp/ether.test
>
> If so, what does `ls -dl /var/tmp` show?
>
> --gilbert
>
----- End forwarded message -----
