Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
| 1. Revert to the original grafinfo -- the change didn't help in this
| case, and made the failure mode worse at times

Adding the MEMORY string for VGA did not help me.

| 2. Disable VGA IRQ in BIOS setup; test

My BIOS affords no such option.

| 4. Try "defbootstr clock.disable_short_timers=1"; test

Made no difference.

| 7. Comment on all the steps you took so we learn what was really
| relevant...

If I run 'startx' when logged in as root or any other user on the system
except 'jpr', exiting X presents no problem.

But I usually run 'startx' when logged in as 'jpr', and that's when I get
the hang-ups upon trying to exit X "normally" (meaning other than using
Alt-PrtScrn).

User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
appears, and I do not have any problem exiting X. The .startxrc:

# following script zaps existing mozilla and its locks, relaunches mozilla
/u/jpr/bin/mz &

xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
scoterm -n JPR -title jpr -geometry 156x25+0+100 &
scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
-e su - news &

# ... more scoterms with assorted '-e su other_users'

win &
exec pmwm

--
JP

On Wed, Oct 08, 2003 at 08:39:21PM -0400, Jean-Pierre Radley wrote:
> Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
> | 1. Revert to the original grafinfo -- the change didn't help in this
> | case, and made the failure mode worse at times
>
> Adding the MEMORY string for VGA did not help me.
>
> | 2. Disable VGA IRQ in BIOS setup; test
>
> My BIOS affords no such option.
>
> | 4. Try "defbootstr clock.disable_short_timers=1"; test
>
> Made no difference.
>
> | 7. Comment on all the steps you took so we learn what was really
> | relevant...
>
> If I run 'startx' when logged in as root or any other user on the system
> except 'jpr', exiting X presents no problem.
>
> But I usually run 'startx' when logged in as 'jpr', and that's when I get
> the hang-ups upon trying to exit X "normally" (meaning other than using
> Alt-PrtScrn).
>
> User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
> appears, and I do not have any problem exiting X. The .startxrc:
>
> # following script zaps existing mozilla and its locks, relaunches mozilla
> /u/jpr/bin/mz &
>
> xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
> xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
> scoterm -n JPR -title jpr -geometry 156x25+0+100 &
> scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
> scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
> scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
> scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
> -e su - news &
>
> # ... more scoterms with assorted '-e su other_users'
>
> win &
> exec pmwm

FWIW, i, too have a .startxrc, but alls i have in it is
xterm &
exec pmwm

--- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ---
-Joe Chasan- Magnatech Business Systems, Inc.
joe@magnatechonline.com Hicksville, NY - USA
http://www.MagnatechOnline.com Tel.(516) 931-4444/Fax.(516) 931-1264

Joe Chasan typed (on Wed, Oct 08, 2003 at 08:52:29PM -0400):
|
| FWIW, i, too have a .startxrc, but alls i have in it is
| xterm &
| exec pmwm

A bit off the topic of this thread, but why would you use xterm instead
of scoterm?

--
JP

Joe Chasan wrote:

> On Wed, Oct 08, 2003 at 08:39:21PM -0400, Jean-Pierre Radley wrote:
> > Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
> > | 1. Revert to the original grafinfo -- the change didn't help in this
> > | case, and made the failure mode worse at times
> >
> > Adding the MEMORY string for VGA did not help me.
> >
> > | 2. Disable VGA IRQ in BIOS setup; test
> >
> > My BIOS affords no such option.
> >
> > | 4. Try "defbootstr clock.disable_short_timers=1"; test
> >
> > Made no difference.
> >
> > | 7. Comment on all the steps you took so we learn what was really
> > | relevant...
> >
> > If I run 'startx' when logged in as root or any other user on the system
> > except 'jpr', exiting X presents no problem.
> >
> > But I usually run 'startx' when logged in as 'jpr', and that's when I get
> > the hang-ups upon trying to exit X "normally" (meaning other than using
> > Alt-PrtScrn).
> >
> > User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
> > appears, and I do not have any problem exiting X. The .startxrc:
> >
> > # following script zaps existing mozilla and its locks, relaunches mozilla
> > /u/jpr/bin/mz &
> >
> > xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
> > xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
> > scoterm -n JPR -title jpr -geometry 156x25+0+100 &
> > scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
> > scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
> > scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
> > scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
> > -e su - news &
> >
> > # ... more scoterms with assorted '-e su other_users'
> >
> > win &
> > exec pmwm
>
> FWIW, i, too have a .startxrc, but alls i have in it is
> xterm &
> exec pmwm

WIW would be more if you'd tested what JPR mentioned: that moving aside
his .startxrc -- or logging in as a different user who doesn't have one
-- corrects the problem for him.

So it looks like each of you three may have completely different
underlying problems. Roger Cornelius has a Dell which (so far) may have
been fixed by giving the video BIOS access to standard VGA video memory
ranges, and an IBM which has some serious BIOS conflict with how we
execute the BIOS. JPR has issues with processes shutting down properly
on the way out of X. We don't know too much about your situation yet.

>Bela<

Jean-Pierre Radley wrote:

> Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
> | 1. Revert to the original grafinfo -- the change didn't help in this
> | case, and made the failure mode worse at times
>
> Adding the MEMORY string for VGA did not help me.
>
> | 2. Disable VGA IRQ in BIOS setup; test
>
> My BIOS affords no such option.
>
> | 4. Try "defbootstr clock.disable_short_timers=1"; test
>
> Made no difference.

Ok, so revert all of those...

> | 7. Comment on all the steps you took so we learn what was really
> | relevant...
>
> If I run 'startx' when logged in as root or any other user on the system
> except 'jpr', exiting X presents no problem.
>
> But I usually run 'startx' when logged in as 'jpr', and that's when I get
> the hang-ups upon trying to exit X "normally" (meaning other than using
> Alt-PrtScrn).
>
> User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
> appears, and I do not have any problem exiting X. The .startxrc:
>
> # following script zaps existing mozilla and its locks, relaunches mozilla
> /u/jpr/bin/mz &
>
> xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
> xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
> scoterm -n JPR -title jpr -geometry 156x25+0+100 &
> scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
> scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
> scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
> scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
> -e su - news &
>
> # ... more scoterms with assorted '-e su other_users'
>
> win &
> exec pmwm

Well, clearly it must be due to something in your .startxrc. You could
conduct a series of experiments, commenting out various bits until you
know which bits cause it. I would start by commenting out all the
scoterms, xclock & xeyes, and observing that the problem still occurs,
because it's going to be an interaction between either Mozilla or Merge,
and pmwm.

Then comment out one or the other of those, test some more.

Once you've narrowed it down to Mozilla, you can try something like
manually killing Mozilla (be sure to get all of its associated
processes) before exiting X.

Then try your normal .startxrc, exit X, take a look at what processes
are still running (do this from a network login from a different
display). You'll probably see one of Mozilla's outrigger processes
still hanging around [or possibly Merge's]. If you kill that, X
shutdown will probably complete.

At which point you have to figure out how to convince Mozilla [or Merge]
to gracefully die all the way when the X server sends it an "I'm going
away now" message. Having narrowed it down that much, you'll be able to
do things like kill the main M* process and observe whether it carries
the message along to its various subprocesses, or in fact leaves itself
in exactly the same contorted state as when you abort the X server. The
details will suggest the next step.

>Bela<

On Thu, Oct 09, 2003 at 01:44:28AM +0000, Bela Lubkin wrote:
> Joe Chasan wrote:
>
> > On Wed, Oct 08, 2003 at 08:39:21PM -0400, Jean-Pierre Radley wrote:
> > > Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
> > > | 1. Revert to the original grafinfo -- the change didn't help in this
> > > | case, and made the failure mode worse at times
> > >
> > > Adding the MEMORY string for VGA did not help me.
> > >
> > > | 2. Disable VGA IRQ in BIOS setup; test
> > >
> > > My BIOS affords no such option.
> > >
> > > | 4. Try "defbootstr clock.disable_short_timers=1"; test
> > >
> > > Made no difference.
> > >
> > > | 7. Comment on all the steps you took so we learn what was really
> > > | relevant...
> > >
> > > If I run 'startx' when logged in as root or any other user on the system
> > > except 'jpr', exiting X presents no problem.
> > >
> > > But I usually run 'startx' when logged in as 'jpr', and that's when I get
> > > the hang-ups upon trying to exit X "normally" (meaning other than using
> > > Alt-PrtScrn).
> > >
> > > User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
> > > appears, and I do not have any problem exiting X. The .startxrc:
> > >
> > > # following script zaps existing mozilla and its locks, relaunches mozilla
> > > /u/jpr/bin/mz &
> > >
> > > xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
> > > xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
> > > scoterm -n JPR -title jpr -geometry 156x25+0+100 &
> > > scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
> > > scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
> > > scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
> > > scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
> > > -e su - news &
> > >
> > > # ... more scoterms with assorted '-e su other_users'
> > >
> > > win &
> > > exec pmwm
> >
> > FWIW, i, too have a .startxrc, but alls i have in it is
> > xterm &
> > exec pmwm
>
> WIW would be more if you'd tested what JPR mentioned: that moving aside
> his .startxrc -- or logging in as a different user who doesn't have one
> -- corrects the problem for him.
>
> So it looks like each of you three may have completely different
> underlying problems. Roger Cornelius has a Dell which (so far) may have
> been fixed by giving the video BIOS access to standard VGA video memory
> ranges, and an IBM which has some serious BIOS conflict with how we
> execute the BIOS. JPR has issues with processes shutting down properly
> on the way out of X. We don't know too much about your situation yet.

this is true, maybe i should have waited to tell you of it existence
until i was in front of the machine and could test w/it moved aside.
will post back in the A.M.

--- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ---
-Joe Chasan- Magnatech Business Systems, Inc.
joe@magnatechonline.com Hicksville, NY - USA
http://www.MagnatechOnline.com Tel.(516) 931-4444/Fax.(516) 931-1264

On Wed, Oct 08, 2003 at 09:34:39PM -0400, Jean-Pierre Radley wrote:
> Joe Chasan typed (on Wed, Oct 08, 2003 at 08:52:29PM -0400):
> |
> | FWIW, i, too have a .startxrc, but alls i have in it is
> | xterm &
> | exec pmwm
>
> A bit off the topic of this thread, but why would you use xterm instead
> of scoterm?

i'm sure there must've been some compatibility issue with some application
or another at some point, or some other preference, but i've no idea
anymore why, date on the file is pretty old.

--- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ---
-Joe Chasan- Magnatech Business Systems, Inc.
joe@magnatechonline.com Hicksville, NY - USA
http://www.MagnatechOnline.com Tel.(516) 931-4444/Fax.(516) 931-1264

On Thu, Oct 09, 2003 at 01:44:28AM +0000, Bela Lubkin wrote:
> Joe Chasan wrote:
>
> > On Wed, Oct 08, 2003 at 08:39:21PM -0400, Jean-Pierre Radley wrote:
> > > Bela Lubkin typed (on Wed, Oct 08, 2003 at 08:39:36AM +0000):
> > > | 1. Revert to the original grafinfo -- the change didn't help in this
> > > | case, and made the failure mode worse at times
> > >
> > > Adding the MEMORY string for VGA did not help me.
> > >
> > > | 2. Disable VGA IRQ in BIOS setup; test
> > >
> > > My BIOS affords no such option.
> > >
> > > | 4. Try "defbootstr clock.disable_short_timers=1"; test
> > >
> > > Made no difference.
> > >
> > > | 7. Comment on all the steps you took so we learn what was really
> > > | relevant...
> > >
> > > If I run 'startx' when logged in as root or any other user on the system
> > > except 'jpr', exiting X presents no problem.
> > >
> > > But I usually run 'startx' when logged in as 'jpr', and that's when I get
> > > the hang-ups upon trying to exit X "normally" (meaning other than using
> > > Alt-PrtScrn).
> > >
> > > User 'jpr' has a .startxrc; if I put it aside, then a normal desktop
> > > appears, and I do not have any problem exiting X. The .startxrc:
> > >
> > > # following script zaps existing mozilla and its locks, relaunches mozilla
> > > /u/jpr/bin/mz &
> > >
> > > xclock -update 1 -hd magenta -hl magenta -geometry 110x110+0+1012 &
> > > xeyes -geometry 120x80+510+1120 -center yellow -outline magenta &
> > > scoterm -n JPR -title jpr -geometry 156x25+0+100 &
> > > scoterm -fg cyan -bg black -n JPR -title jpr -geometry 80x47+1600+0 &
> > > scoterm -fg magenta -bg white -n JPR -title jpr -geometry 80x47+1700+100 &
> > > scoterm -fg yellow -bg blue -n JPR -title jpr -geometry 80x47+2364+224 &
> > > scoterm -fg red -bg cyan -n NEWS -title news -geometry 80x47+7164+224 \
> > > -e su - news &
> > >
> > > # ... more scoterms with assorted '-e su other_users'
> > >
> > > win &
> > > exec pmwm
> >
> > FWIW, i, too have a .startxrc, but alls i have in it is
> > xterm &
> > exec pmwm
>
> WIW would be more if you'd tested what JPR mentioned: that moving aside
> his .startxrc -- or logging in as a different user who doesn't have one
> -- corrects the problem for him.
>
> So it looks like each of you three may have completely different
> underlying problems. Roger Cornelius has a Dell which (so far) may have
> been fixed by giving the video BIOS access to standard VGA video memory
> ranges, and an IBM which has some serious BIOS conflict with how we
> execute the BIOS. JPR has issues with processes shutting down properly
> on the way out of X. We don't know too much about your situation yet.

no, moving out the .startxrc had no effect. i am now 99% sure it is
tied in to mozilla (which i note JP had in his startuprc file).

removing the startxrc and not removing the startxrc, i can get in and
out ok - if i start mozilla (scohelp, using mozilla), i get stuck leaving
X with the 'dead screen', whether my startxrc was there or not.

--- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ---
-Joe Chasan- Magnatech Business Systems, Inc.
joe@magnatechonline.com Hicksville, NY - USA
http://www.MagnatechOnline.com Tel.(516) 931-4444/Fax.(516) 931-1264

Bela Lubkin typed (on Wed, Oct 08, 2003 at 10:19:47PM -0400):
| Well, clearly it must be due to something in your .startxrc. You could
| conduct a series of experiments, commenting out various bits until you
| know which bits cause it. I would start by commenting out all the
| scoterms, xclock & xeyes, and observing that the problem still occurs,
| because it's going to be an interaction between either Mozilla or Merge,
| and pmwm.
|
| Then comment out one or the other of those, test some more.
|
| Once you've narrowed it down to Mozilla, you can try something like
| manually killing Mozilla (be sure to get all of its associated
| processes) before exiting X.
|
| Then try your normal .startxrc, exit X, take a look at what processes
| are still running (do this from a network login from a different
| display). You'll probably see one of Mozilla's outrigger processes
| still hanging around [or possibly Merge's]. If you kill that, X
| shutdown will probably complete.
|
| At which point you have to figure out how to convince Mozilla [or Merge]
| to gracefully die all the way when the X server sends it an "I'm going
| away now" message. Having narrowed it down that much, you'll be able to
| do things like kill the main M* process and observe whether it carries
| the message along to its various subprocesses, or in fact leaves itself
| in exactly the same contorted state as when you abort the X server. The
| details will suggest the next step.

It's Mozilla.

Doesn't matter if Mozilla is started by a .startxrc, from a Desktop icon,
or from a command line in a scoterm.

Doesn't matter if I choose "Quit" from Mozilla's "File" menu, or if I
kill
/opt/mozilla/lib/mozilla-bins
which then kills its parent
/opt/mozilla/lib/run-mozilla.sh /opt/mozilla/lib/mozilla-bin,
I cannot then exit X with hanging up the console except via the
Alt-PrtScr route.

And I see no other processes that are the children of those two.

--
JP

Jean-Pierre Radley wrote:

> It's Mozilla.
>
> Doesn't matter if Mozilla is started by a .startxrc, from a Desktop icon,
> or from a command line in a scoterm.
>
> Doesn't matter if I choose "Quit" from Mozilla's "File" menu, or if I
> kill
> /opt/mozilla/lib/mozilla-bins
> which then kills its parent
> /opt/mozilla/lib/run-mozilla.sh /opt/mozilla/lib/mozilla-bin,
> I cannot then exit X with hanging up the console except via the
> Alt-PrtScr route.
>
> And I see no other processes that are the children of those two.

Nevertheless. Run a `ps -ef > /tmp/before` before you start X; start X
(sans .startxrc, to keep things very simple); start Mozilla; `ps -ef >
/tmp/during`; then blow out of X, `ps -ef > /tmp/after`; send 'em all to
me. There must be something hanging out causing the trouble...

(direct email; I can post my findings, but no need to post hundreds of
lines of `ps` gibberish...)

>Bela<