Re: run-jobs test fails for bash-3.0, hangs in jobs1.sub and trap.tests

[Chet Ramey]

David Magerman wrote:
> Configuration Information [Automatically generated, do not change]:
> Machine: i686
> OS: linux-gnu
> Compiler: gcc
> Compilation CFLAGS:  -DPROGRAM='bash' -DCONF_HOSTTYPE='i686' 
> -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='i686-pc-linux-gnu' 
> -DCONF_VENDOR='pc' -DLOCALEDIR='/usr/local/products/bash/3.0/share/locale' 
> -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H  -I.  -I../bash-3.0 
> -I../bash-3.0/include -I../bash-3.0/lib   -g -O2
> uname output: Linux monstert24.rentec.com 2.6.8-24.5-bigsmp #1 SMP Wed Nov 17 
> 11:10:06 UTC 2004 i686 i686 i386 GNU/Linux
> Machine Type: i686-pc-linux-gnu
>
> Bash Version: 3.0
> Patch Level: 0
> Release Status: release
>
> Description:
>
> bash-3.0 fails the "run-jobs" test.  It hangs while running jobs1.sub.
> bash-3.0 also hangs if I try using the shell after a few commands.  If
> I kill the test manually, it subsequently fails on trap.tests.
>
> Repeat-By:
>
> Running "make tests" in the object directory after making bash-3.0.
>
> I'd be happy to help you debug this problem if you can give me some
> guidance.

This is probably the result of a bug in glibc.  I'll see if I can
dig up and forward you a patch submitted by Tim Waugh at Red Hat.

(The short story is that glibc defines WCONTINUED as a flag for
waitpid(), and bash tries to use it if available.  waitpid() rejects
it, and bash can never successfully wait for children.)

Chet
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
( ``Discere est Dolere'' -- chet )
                                                Live...Laugh...Love
Chet Ramey, ITS, CWRU    chet@po.cwru.edu 
http://tiswww.tis.cwru.edu/~chet/