[3807] in testers
Re: more xlogin stderr fun
daemon@ATHENA.MIT.EDU (Greg Hudson)
Fri Jul 17 18:54:43 1998
Date: Fri, 17 Jul 1998 18:54:33 -0400
From: Greg Hudson <ghudson@MIT.EDU>
To: Craig Fields <cfields@MIT.EDU>
Cc: ghudson@MIT.EDU, testers@MIT.EDU
In-Reply-To: "[3799] in testers"
>> I'm also very curious why dm redirects console output to a pipe
>> instead of just pointing it at consolefd (ttyp0).
> I'm pretty foggy on how that all works anymore and I don't know if
> this still applies. But at the time I think it had to do with what
> happens to /dev/console output while the console window is being
> restarted. Doing the pipe thing was to ensure that all the data
> would eventually show up somewhere.
I don't think it has that effect, though. When console is restarted,
a new pipe is created and console output is redirected there. Any
output which went to the old pipe would be lost, and any output which
went to /dev/console after the old pipe closes would presumably get
printed on the X screen.
On the other hand, if console output is directed to ttyp0, I expect
that all of it would make it to one console process or another.
Nothing's going to make output written to ttyp0 go away except reading
it, at least as long as dm is holding onto an open file descriptor.
So I would think the pipe hack would have the opposite of the intended
effect.
I'll experiment some for the 8.3 release.
