Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Date: Sat, 8 Jan 2005 21:53:37 -0500
From: Christopher Faylor <cgf-no-personal-reply-please AT cygwin DOT com>
To: cygwin AT cygwin DOT com
Subject: Re: Problem with 20050106 snapshot
Message-ID: <20050109025337.GG3713@trixie.casa.cgf.cx>
Reply-To: cygwin AT cygwin DOT com
References: <16864 DOT 26477 DOT 82750 DOT 561543 AT gargle DOT gargle DOT HOWL>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="nFreZHaLTZJo0R7j"
Content-Disposition: inline
In-Reply-To: <16864.26477.82750.561543@gargle.gargle.HOWL>
User-Agent: Mutt/1.4.1i

--nFreZHaLTZJo0R7j
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Sat, Jan 08, 2005 at 03:06:21PM -0800, David Rothenberger wrote:
Content-Description: message body text
>I'm having a problem with the 20050106 snapshot as well as my own
>DLL build from CVS HEAD on 20050107. I've also seen these problems
>in snapshots from late last year. I've just now isolated a small
>test case.
>
>The problem is with bash killing native processes. I have a bash
>script that starts a native process in the background, installs a
>signal handler for SIGINT, grabs the pid from the background
>process, and then issues wait. The signal handler tries to kill the
>pid that was captured.
>
>With 1.5.12 and a DLL built from CVS on 11/19/2004, this works just
>fine. With the latest snapshot, I often (but not always) get a "Bad
>file descriptor" error from kill.
>
>I've enclosed a C file with the source for a simple program that
>sleeps for the time specified on the command-line. I compiled this
>as a WIN32 application using
>
>  % gcc -o sleepMingw -mno-cygwin sleepMingw.c
>
>I've also enclosed the driver script sleepTest.sh.
>
>Here's the output from one run:
>

I modified the script as attached and ran it for about ten or fifteen
minutes and didn't see this behavior but, unless I'm missing something,
I don't see how this would ever work reliably, though, since, if you're
hitting CTRL-C in a console window, it will be propagated to all
subprocesses.  Cygwin processes know how to deal with this but mingw
processes don't and will just catch the CTRL-C and do what a windows
console process would normally do.  In this case what they'd normally do
is exit.  The "file not found" error code is a little odd in this case,
though.  I tightened up some of the errno setting and will be checking
in some new code to deal with some problems that Pierre found with
windows processes, soon.

The loop I used was:

while ./sleepTest.sh 60; do :; done

it never exited.

cgf

--nFreZHaLTZJo0R7j
Content-Type: application/x-sh
Content-Disposition: attachment; filename="sleepTest.sh"
Content-Transfer-Encoding: quoted-printable

#!/bin/bash

set -x

./sleepMingw $* &

enTrapSIGINT () { echo "killing process $pid"; { kill $pid && exit 0; }; ex=
it 1; }

pid=3D$!
trap enTrapSIGINT 2
x=3D$$; (exec kill -INT $x) || exit 1
wait


--nFreZHaLTZJo0R7j
Content-Type: text/plain; charset=us-ascii

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/
--nFreZHaLTZJo0R7j--