> -----Original Message-----
> From: Spitzer, Andy (BL60:9D30)
> Sent: Wednesday, October 01, 2008 10:26 AM
> To: Beeton, Carolyn (CAR:9D60); Lawrence, Scott (BL60:9D30)
> Cc: [EMAIL PROTECTED]; Dans, Raymond (CAR:9D30)
> Subject: Re: [sipX-dev] sipXportLib - osProcessLinux - change
> suggestions
>
> Woof!
>
> On Wed, 01 Oct 2008 08:52:42 -0400, Beeton, Carolyn
> (CAR:9D60) <[EMAIL PROTECTED]> wrote:
>
> > What Raymond and I are proposing is to modify the
> OsProcess::launch to
> > NOT ignore the SIGCHLD, and add a case into sig_routine in
> > sipXsupervisor to ignore it there. I don't see how this
> affects any
> > other process.
>
> As this is sipXportLib, (a generic portability library),
> changing it to allow SIGCHLD means any and all users have to
> be aware of their responsibility to catch SIGCHLD, and do a
> waitpid() eventually. This goes for all callers of
> OsProcess:launch() such as OsProcessMgr::startProcess(). I
> don't believe current users of
> OsProcessMgr::startProcess() call waitpid...it expects the OS
> to reap them. This would need to change.
>
> Another approach might be to add a flag to
> OsProcess::launch() that defaults to the old behavior of
> ignoring SIGCHLD, and allows new uses of the code to pass in
> if it wants to ignore SIGCHLD or not. That way we get both
> behaviors, and only callers who explicitly set the flag will
> need to perform the new requirements.
>
>
> --Woof!
>
OsProcessMgr will be removed once we have completed the supervisor
restructure.
I think we have no objection to using a flag to indicate whether to
ignore SIGCHLD. Also some clear instructions need to be added as using
wait without having set the flag on launch will cause the process to
hang. So I think we have agreement.
Carolyn
_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
