On Friday, October 27, 2006 9:52 AM Camm Maguire wrote:
>
> Can anyone see what this might suggest in the sman code?
>
> Waldek Hebisch <[EMAIL PROTECTED]> writes:
>
> Tim Daly wrote:
> > >
> > > try starting axiom with
> > >
> > > axiom -noclef
> > >
> > > and see if the problem still happens.
> > >
> >
> > Reproducible (using build-improvements) both with and without
> > clef.
When the command 'axiom -noclef' is run, the process tree
looks like this:
$ axiom -noclef
user i/o
sman . . . /\ /\
. . ./ \
. . / . \
. . \/ . \/
graphics====AXIOMsys ===== hyperdoc
/\
where ==== represents a unix socket and \/ represents
serial pty and . . . is a sub-process.
Right?
The version with clef in place, I suppose would look like
this:
$ axiom
user i/o
sman . . /\ /\
. . . / \
.. . ./ \
. . . \/ . \
. . clef . \
. . /\ . \
. . \/ . \/
graphics====AXIOMsys ========= hyperdoc
> > When AXIOMsys is run directly ^D exits it.
Does ^D send an end-of-file condition to the pty interface?
I suppose that when you say AXIOMsys run directly you
mean the following single process:
$ AXIOMsys
User i/o
|\
|
\|
AXIOMsys
> When AXIOMsys is run via clef ^D\n exits it.
I am not sure how to create this situation. What options
to the 'axiom' script did you use?
> With sman present Axiom crashes.
My impression was the sman is always present except when
running AXIOMsys directly since it's purpose is to start
the other sub-processes in the tree and to assist with
setting up sockets etc.
Any help understanding this would be great.
Regards,
Bill Page.
