[comp.sys.ultrix] problems with X11R4/Sequent/DECstation 3100

pauld@cs.washington.edu (Paul Barton-Davis) (04/24/91)

[ Sorry if this seems like a repost, but things have gotten worse
  since I mentioned this about a week ago ]

We have just recently started having problems with X11R4 on our
sequent. We are using the "vanilla"/athena distribution and to the
best of my knowledge, have not been suffering from these problems
until a few weeks ago. 

Basically, running an X client on the sequent, with the X server on a
DECstation 3100 causes X errors if the client tries to redraw the
whole screen (too fast, too large, who knows ?). It doesn't happen
predictably in the sense that you can't say exactly which redraw
request will cause the server to return an error, but it is
repeatable: you just fire up an xterm, zoom the screen (tho' even 24
lines will do sometimes), and do "cat /etc/termcap". Next thing you
know, you'll get a message like this:

X Error of failed request:  BadLength (poly request too large or internal Xlib length error)
  Major opcode of failed request:  58 (X_SetDashes)
  Minor opcode of failed request:  0
  Resource id in failed request:  0xd00002
  Serial number of failed request:  3586
  Current serial number in output stream:  5277

This one is from emacs running in its own window - I visited
/etc/termcap, and pressed Ctrl-v a few times. If you wait for screen
updates to happen, instead of just typing ahead (space bar in more(1),
or ^v in emacs), the X error rarely happens.  Alternative behaviour is
to freeze the screen: xterm will recover from this, emacs will not,
although it will still respond to input. Seems very X-related to me.

Anyone have any ideas about this ? 

-- 
Paul Barton-Davis			<pauld@cs.washington.edu>
UW Computer Science Lab		``to shatter tradition makes us feel free''