mingw-users

[Mingw-users] Problem with the kill builtin in MSYS bash

[Peter R. <pe...@ly...>]

Hi!

I found an automake test that didn't behave properly (cont5.test). It
had a success rate of about 50% for me on MSYS. I have reduced it to
the below simple test case, which fails much more often than the
original code.

I.e. start a process in the background, check if it is alive with
kill -0 (which it should be), and report failure if it isn't. But it
appears that kill fails if it is executed too soon after the new
process was created...

Now, I know that I'm not up to date, sorry if this is an old bug. But
it should be easy enough to test this for someone who is...

Cheers,
Peter


$ sh --version
GNU bash, version 3.1.0(1)-release (i686-pc-msys)
Copyright (C) 2005 Free Software Foundation, Inc.

$ uname -a
MINGW32_NT-5.1 PEDA-PC 1.0.11(0.46/3/2) 2009-07-11 17:46 i686 Msys

-----------8<------------
#! /bin/sh

sleep 10 &
pid=$!

# Uncomment the below sleep to reliably succeed.
#sleep 1

if kill -0 $pid; then
   echo success
else
   echo failure
fi
kill $pid
-----------8<------------

Re: [Mingw-users] Problem with the kill builtin in MSYS bash

[Earnie <ea...@us...>]

Peter Rosin wrote:
> Hi!
>
> I found an automake test that didn't behave properly (cont5.test). It
> had a success rate of about 50% for me on MSYS. I have reduced it to
> the below simple test case, which fails much more often than the
> original code.
>

http://www.mingw.org/Reporting_Bugs

-- 
Earnie
-- http://www.for-my-kids.com

Re: [Mingw-users] Problem with the kill builtin in MSYS bash

[Cesar S. <ces...@gm...>]

On 9/8/2010 04:00, Peter Rosin wrote:
> I found an automake test that didn't behave properly (cont5.test). It
> had a success rate of about 50% for me on MSYS. I have reduced it to
> the below simple test case, which fails much more often than the
> original code.

Thanks for the test case.

> I.e. start a process in the background, check if it is alive with
> kill -0 (which it should be), and report failure if it isn't. But it
> appears that kill fails if it is executed too soon after the new
> process was created...
>
> Now, I know that I'm not up to date, sorry if this is an old bug. But
> it should be easy enough to test this for someone who is...

I didn't heard of it before.

In my setup, which I believe is up to date, the test fails every single 
time. So, it seems to be a true bug. Please, feel free to submit a bug 
report to the bug tracker.

Thanks,
Cesar