[COMMIT] gnuclient bug on Ben's part

[Mike Kupfer]

[putting xemacs-beta back into the conversation.]

>>>>> "Aidan" == Aidan Kehoe <kehoea at parhasard.net> writes:

>> > * keydefs.el (global-tty-map):
>> > 	"\C-z" is suspend-or-iconify-emacs on the TTY, not
>> > 	suspend-emacs.  Fixes problems that arose where pressing C-z in
>> > 	a gnuclient fram suspended the whole process.

Mike> Thank you!  This has been annoying me off and on for years (May 2000,
Mike> according to my notes), but I never got around to looking at it.

Aidan> Good. :-) . When I invoke it on my NetBSD machine, running fg
Aidan> requires that I type C-c for the frame to display, but on this
Aidan> Linux machine, that’s not the case. Do you have any problems
Aidan> with resuming gnuclient sessions after this patch was committed?

After looking at the patch more carefully, it occurred to me that I
could test this without installing and building 21.5.  Unfortunately, I
seem to have stumbled into a mass of confusion...

In 21.4.18, C-z is bound to suspend suspend-emacs-or-iconify-frame.
Neither that or suspend-or-iconify-emacs gives me my shell back.  But
suspend-emacs does.  So that seems to be backwards from Aidan's patch.

What's more confusing is that I'm suddenly having problems getting
"gnuclient -nw" to start up at all.  When I run it, instead of getting
an XEmacs frame, it hangs.  If I C-z it, I get

    ^Z[2] + Stopped (SIGTSTP)        /usr/local/bin/gnuclient -nw
    [1] - Stopped (SIGTTOU)        xemacs &

Putting job 1 (xemacs) into the foreground gives me my tty frame.

But then I can't get out with C-x 5 0.  The tty frame is cleared, but I
don't get my shell back.

This is on Solaris, a recent development build of 5.11, using ksh as my
shell.

I'll try to look into this further, since it could be a Solaris issue,
but it'll take me a few days.

mike