[comp.windows.news] Non-dying psterm's for non-root users? ; screen dumps

earle@MAHENDO.JPL.NASA.GOV (Greg Earle) (12/01/87)

Ever since I got NeWS, I'd been using it as `root'.  Then someone badgered
me to show them how to use NeWS, so I logged in as them, did my little
alias trick (set the environment, run the server, and do a clear_colormap
afterwords to clean up the garbage), and away we went.  First thing I discover
is that psterms don't go away when the shell inside is exited!  The shell
goes away (window blanks if one does a `Redisplay'), but the psterm window
stays around (it goes away as root).  You have to manually `Zap' the window,
and I discovered if you accidentally do this to the Console psterm, when you
exit NeWS the console hangs (can't get back to login shell).  Remotely killing
the psterms still lying around causes the login shell to die as well!  This
seems like a permissions problem of some sort, but I tried the various
combinations of setuid for the server and psterm, all to no avail.  I certainly
don't recall seeing this in the manuals anywhere (I installed the one binary
patch), so I'm completely at a loss ...

Whilest on the subject of screen dumps, does anyone have a general method
for doing them (under SunView, NeWS, or X)?  I wrote a shell script called
`dumpscreen' that would do `screendump', pass it to `rasfilter8to1' (anyone
notice that the 3.4 version of this does not match the manual page?  Try
the `choose dithering' option) if `/dev/cgtwo0' exists, then pass it to the
TranScript `pssun' with various option tweaks to get it almost full size (and
Landscape) on a LaserWriter.  On the color machines, it becomes hit or miss
whether the end product will be correct, and this has happened under all three
window systems.  Sometimes the symptoms are a `censored' all-black bar (or
bars) across certain parts of the output; other times in addition to the above
I get little replicated squished versions of the whole screen peppered across
the top of the `normal' screendump (along with the `censored' bars).  Once
in a while it works totally normally, but more often than not it doesn't.
This isn't really a NeWS question, but since it happens under NeWS, and
involves screen dumping to a PostScript filter, it's tangentially relevant ...
(I hope (-: )

	Greg Earle		earle@mahendo.JPL.NASA.GOV
	Gainfully unemployed	earle%mahendo@jpl-elroy.ARPA	[aka:]
				earle%mahendo@elroy.JPL.NASA.GOV
	Los Angeles, CA		...!cit-vax!elroy!jplgodo!mahendo!earle

david@sun.uucp (David DiGiacomo) (12/02/87)

In article <8712011210.AA02315@mahendo.Jpl.Nasa.Gov> earle@MAHENDO.JPL.NASA.GOV (Greg Earle) writes:
>I wrote a shell script called
>`dumpscreen' that would do `screendump', pass it to `rasfilter8to1' (anyone
>notice that the 3.4 version of this does not match the manual page?  Try
>the `choose dithering' option)

This is a known bug in the 3.4 upgrade script.  To fix it:

	cd /usr/bin
	foreach i (clear_colormap rasfilter8to1 rastrepl screenload)
		rm $i
		ln -s screendump $i
	end

> if `/dev/cgtwo0' exists, then pass it to the
>TranScript `pssun' with various option tweaks to get it almost full size (and
>Landscape) on a LaserWriter.  On the color machines, it becomes hit or miss
>whether the end product will be correct, and this has happened under all three
>window systems.  Sometimes the symptoms are a `censored' all-black bar (or
>bars) across certain parts of the output; other times in addition to the above
>I get little replicated squished versions of the whole screen peppered across
>the top of the `normal' screendump (along with the `censored' bars).

RTFM screendump(1) !!!

	SCREENDUMP(1)             USER COMMANDS             SCREENDUMP(1)

	BUGS
	     The output file or the screen may be corrupted if the  frame
	     buffer contents are modified while the dump is in progress.

	Sun Release 3.4   Last change: 16 January 1987                  2

Make sure nothing touches the frame buffer while the screendump is running
(e.g. no clocks).

-- 
David DiGiacomo, Sun Microsystems, Mt. View, CA  sun!david david@sun.com

barnett@steinmetz.UUCP (Bruce Barnett) (12/02/87)

> On the color machines, it becomes hit or miss whether the 
> [screendump] will be correct, and this has happened under all three
>window systems.  Sometimes the symptoms are a `censored' all-black bar (or
>bars) across certain parts of the output; other times in addition to the above
>I get little replicated squished versions of the whole screen peppered across
>the top of the `normal' screendump (along with the `censored' bars). 

This will happen is there is any change of the screen while the dump
is made. If the clocktool's minute hand moves, the mouse cursor, etc.

--
Bruce G. Barnett  <barnett@ge-crd.ARPA> <barnett@steinmetz.UUCP>