Re: tcsh should be removed from hamm
>>>>> On 30 Mar 1998 09:05:29 -0500, Steve Dunham <dunham@cps.msu.edu> said:
[..]
SD> I've noticed it on Linux boxes but not on Solaris.
SD> It has something to do with dead/stale connections, but it hasn't
SD> happened frequently enough for me to nail it down.
Below there is a response from tcsh author to my bug report.
A copy of bug report has also been sent to tcsh.*deb maintainer.
--
piotr pogorzelski
From: christos@zoulas.com (Christos Zoulas)
Subject: Re: tcsh bug
To: piotr.pogorzelski@ippt.gov.pl
Date: Wed, 18 Mar 1998 08:54:37 -0500
Organization: Astron Software
On Mar 18, 10:24am, piotr.pogorzelski@ippt.gov.pl (piotr.pogorzelski@ippt.gov.pl) wrote:
-- Subject: tcsh bug
| hi,
|
| I'v just installed a new debian system, and found
| next day in the morning, that I have load of 3, caused
| by three tcsh shells.
| strace showed that each justs closes all file descriptors,
| then tries to write a job id "[1234]", gots error,
| closes all files,etc,etc.
| under gdb i've found out that it happens in
| sh.c, in function process.
| pendjob tries to write jobid (that process do not exists anymore),
| but user has already closed connection,
| so gets an error and sets haderr=1;
| later in the process there is a code
| if(haderr){
| if(!catch){
| sth
| }
| haderr=0;
| closem();
| contiune;
| }
| as catch=1
| it clears flags and returns to top.
|
| What happens if i change 'contiunue' into break,
| or remove it? Would it break tcsh or solve my problems?
| Or is it Linux/debian/hamm/sth.else specific error ?
|
I don't know. I've heard about a similar bug before, but the fix I attempted
caused other problems. You can try. I have not seen this problem outside
linux. This does not say that it is not a tcsh bug...
christos
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: