[SCM] Samba Shared Repository - branch master updated

[Michael Adam]

Hi Andrew,

Andrew Bartlett wrote:
> The branch, master has been updated
>        ...
>        via  2b47afa selftest: Avoid being run over by armies of the undead
>        ...
>       from  fcb90f4 s3-printing: Add forward declaration for dcerpc_binding_handle.
> 
> http://gitweb.samba.org/?p=samba.git;a=shortlog;h=master
> 
> 
> - Log -----------------------------------------------------------------
> commit 2b47aface434000b29aa29f4ff6348cc147ae757
> Author: Andrew Bartlett <abartlet at samba.org>
> Date:   Tue Jul 26 17:04:40 2011 +1000
> 
>     selftest: Avoid being run over by armies of the undead
>     
>     Ignore SIGCHILD to reap zombies
>     
>     Andrew Bartlett
>     
>     Signed-off-by: Andrew Tridgell <tridge at samba.org>
> 
> 
> -----------------------------------------------------------------------
>  
> diff --git a/selftest/selftest.pl b/selftest/selftest.pl
> index 78627c8..5cbb686 100755
> --- a/selftest/selftest.pl
> +++ b/selftest/selftest.pl
> @@ -176,6 +176,8 @@ sub pipe_handler {
>  
>  $SIG{PIPE} = \&pipe_handler;
>  
> +$SIG{CHILD} = 'IGNORE';
> +
>  sub find_in_list($$)
>  {
>  	my ($list, $fullname) = @_;

Ok, there is a typo here that annihilates the intended effect of this patch:

It is "SIGCHLD" (sometimes also spelled SIGCLD), but _not_ SIGCHILD...

The only effect of this patch is the additional warning printed
by make test:

>>>> No such signal: SIGCHILD at ../selftest/selftest.pl line 179.

If I change the corrponding line to read "$SIG{CHLD} = 'IGNORE';
instead, then make test fails completely (at least on the ubuntu
10.04 and 10.10 platforms I am working on). This is the effect:

>>>> Can't ignore signal CHLD, forcing to default.
>>>> Traceback (most recent call last):
>>>>   File "/home/obnox/samba-build/source3/selftest/tests.py", line 23, in <module>
>>>>     from selftesthelpers import *
>>>>   File "/home/obnox/samba-build/selftest/selftesthelpers.py", line 49, in <module>
>>>>     if subprocess.call(perl + ["-e", "eval require Test::More;"]) == 0:
>>>>   File "/usr/lib/python2.6/subprocess.py", line 480, in call
>>>>     return Popen(*popenargs, **kwargs).wait()
>>>>   File "/usr/lib/python2.6/subprocess.py", line 1170, in wait
>>>>     pid, sts = _eintr_retry_call(os.waitpid, self.pid, 0)
>>>>   File "/usr/lib/python2.6/subprocess.py", line 465, in _eintr_retry_call
>>>>     return func(*args)
>>>> OSError: [Errno 10] No child processes
>>>> Error creating recipe at ../selftest/selftest.pl line 705.
>>>> 
>>>> ALL OK (0 tests in 0 testsuites)
>>>> 
>>>> A summary with detailed information can be found in:
>>>>   ./st/summary
>>>> SELFTEST FAILED
>>>> make: *** [test] Error 1

So, no child processe are created at all (or they are killed immediately).
This puzzles me slightly, because reading from the documentation, it sounds
like this setting should be possible on newer systems with exaclty the
effect intended by your original commit, and it is mentioned that this does
not work on some older systems.

So I was not able to fix the typo easily. Maybe we have to install a SIGCHLD
handler that waits for the child processes?
(This might be useful anyways to stay portable.)

Any clues?

Cheers - Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 206 bytes
Desc: not available
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20110822/d7d38264/attachment.pgp>