Re: [PATCH] Switch back to SIGUSR1 and use SIGWINCH on Linux 2.0

[Justin Erenkrantz]

On Mon, Sep 17, 2001 at 11:23:14AM -0700, Ryan Bloom wrote:
> Forcing people to use SIGUSR1on a platform that has co-opted that signal
> is broken.

Obviously, that wouldn't work.  My patch allowed SIGWINCH to be used on 
Linux 2.0 while switching the default to be SIGUSR1 on sane platforms.

Your complaint is that the signal must be the same on all platforms.
Therefore, I ask what you would like to do (should this be a vote?):

[ ]  Enforce SIGUSR1 across all platforms and drop support for any 
     platform that does not allow SIGUSR1 to be used for graceful
     restart.
[ ]  Allow the default to be SIGUSR1, but on platforms on which
     SIGUSR1 is not viable, use an alternative signal (such as 
     SIGWINCH on Linux 2.0/glibc 2.0).  This would also allow the
     site admin to customize the signal at configure time.

Since there is a veto against the use of SIGWINCH by default, the 
current code in CVS is broken.  I would like to resolve this 
before we tag and roll again.  -- justin