"bad screen number" 
Author Message
 "bad screen number"

I can't get Wish to go.

The proximate symptom is an "application-specific initialization
failed: bad screen number '0'" diagnostic.  This amounts to a
report that

        ScreenCount(XOpenDisplay($DISPLAY))

is zero (!).  This is an incredible nuisance, 'cause, beside be-
ing up against the usual deadlines, the darn thing was doing
fine up until the latest re-generation.  I promise I didn't
change anything; I was just doing disk hygiene on my FreeBSD box,
and innocently commanded "make clean; make" in a directory that
shouldn't have changed.  The environment has been quiet, too; no
new versions of X, no new compilers, just a mostly-idle machine
that was happy one minute, and Screen-less the next.

Details:  I get the same symptom with Tk4.1, 4.2, and 8.0.  I'm
running 2.1.6-RELEASE of FreeBSD.  It looks most like some kind
of incompatibility between the Xlib.h and the -lX11 for the def-
inition of Display, but I know of no changes in those, and
certainly none that should affect my sources.  Binaries (xterm,
...) from before the cataclysm all do fine; I don't think it's
a misconfiguration of my X server.  I've tried the usual range
of possibilities for DISPLAY:  ":0.0"; "unix:0.0"; "$HOSTNAME:0.0";
...

Any clues?  Anyone?
--

Cameron Laird           http://www.*-*-*.com/ ~claird/home.html

                        +1 713 996 8546 FAX



Tue, 27 Jul 1999 03:00:00 GMT  
 "bad screen number"


Quote:
>I can't get Wish to go.

>The proximate symptom is an "application-specific initialization
>failed: bad screen number '0'" diagnostic.  This amounts to a

                        .
                        .
                        .
Quote:
>running 2.1.6-RELEASE of FreeBSD.  It looks most like some kind
>of incompatibility between the Xlib.h and the -lX11 for the def-
>inition of Display, but I know of no changes in those, and
>certainly none that should affect my sources.  Binaries (xterm,

                        .
                        .
                        .
Yup.  I was compiling with X11R4 (I think), and linking
with X11R6.  End of crisis.  I make a point of posting
this, though, because there are a couple of aspects to
my little drama that I think are of general interest.
More on those, in follow-ups of narrowed distribution.
--

Cameron Laird           http://starbase.neosoft.com/~claird/home.html

                        +1 713 996 8546 FAX



Thu, 29 Jul 1999 03:00:00 GMT  
 "bad screen number"


Quote:

>>I can't get Wish to go.
>                    .
>                    .
>                    .
>Yup.  I was compiling with X11R4 (I think), and linking
>with X11R6.  End of crisis.  I make a point of posting
>this, though, because there are a couple of aspects to
>my little drama that I think are of general interest.
>More on those, in follow-ups of narrowed distribution.

                        .
                        .
                        .
This note suggests changes to distribution sources.  I'm
forwarding a copy to Dr. Ousterhout.

First I'll describe how things are, then I'll itemize my
recommendations.

Tk sources come with a directory called xlib.  There's no
explanation of xlib in the READMEs or other documentation,
but xlib is mostly a historical artifact, from what I've
been told, a collection of declarations that facilitate
generation under MacOS and possibly elsewhere.  The headers
within xlib suggest to me that it's mostly X11R5, modified
for reasons I don't understand.

The xlib isn't particularly complete, as an X11R5 header
package.  It's not designed to be; it has just enough to be
useful in generating Tk.

There are definite errors in xlib (most obtrusively, that
Xutil.h declares "void XEmptyRegion()").  There are errors,
or at least considerable inconveniences, in Tk's configure,
and the way it processes --x--includes.  I've reported fixes
for some of these to Dr. Ousterhout before, but he found
them uncompelling.  What's the scoop, folks; does this stuff
interest a sufficient number of others to be worth getting
right?  Does it interest anyone else?  I recognize that it
makes far more sense to have X11 headers and libraries from
the same source, but I'll make the case that that presents
serious practical difficulties.

If the Sun development team does nothing about this, but I
receive enough expressions of interest from others, I'll
work up an explanation and perhaps some patches that I'll
publish in URL-land.
--

Cameron Laird           http://starbase.neosoft.com/~claird/home.html

                        +1 713 996 8546 FAX



Sat, 07 Aug 1999 03:00:00 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. bad screen distance "1 pixel" (fwd)

2. bad screen distance "1 pixel"

3. string.join(["Tk 4.2p2", "Python 1.4", "Win32", "free"], "for")

4. Printing "bad" font

5. "ep": Bad News

6. "Bad Managers" website

7. "bad file descriptor" from tcp sockets

8. "bad synchronous description" in Xilinx WebPack

9. Checking strings for "bad" characters

10. What is "BAD ASSUMPTION"?

11. bad option "identify"

12. bad option "_scrollWidget" when Freewrapped

 

 
Powered by phpBB® Forum Software