[drl@vuse.vanderbilt.edu: Re: A new pretest]

[drl@vuse.vanderbilt.edu: Re: A new pretest]

[Richard Stallman]

Can someone investigate this X bug in the RC branch?

------- Start of forwarded message -------
Envelope-to: emacs-pretest-bug@gnu.org
Delivery-date: Sat, 12 Oct 2002 13:17:43 -0400
Date: Sat, 12 Oct 2002 12:17:38 -0500 (CDT)
From: "David R. Linn" <drl@vuse.vanderbilt.edu>
Reply-To: "David R. Linn" <drl@vuse.vanderbilt.edu>
Subject: Re: A new pretest
To: emacs-pretest-bug@gnu.org
Content-MD5: pVOlfw/63ZDQWq149/qsZw==
X-Spam-Status: No, hits=-2.9 required=5.0
	tests=QUOTED_EMAIL_TEXT,SPAM_PHRASE_00_01
	version=2.41
X-Spam-Level: 

>> There is a new pretest available in
>> 
>>   <ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-21.2.91.tar.gz>
>>   <ftp://alpha.gnu.org/gnu/emacs/pretest/leim-21.2.91.tar.gz>

Sorry about that last useless message - I clicked "send" when I meant to
click "delete".

Anyway, ... 
Under SunOS 5.9/Solaris 9 on SUN Ultra 5 (SPARC v9) with SunPro/Forte C
v6.0 and otherwise standard Solaris 9 tools/libs, I had no errors or
unexpected warnings building emacs v21.2.91.

However, installation failed because I do not (yet) have makeinfo on
this box.
- -----
(cd man; make  info)
cd /ufs/sys2/FSF/Emacs/21/91/emacs-21.2.91/man; makeinfo emacs.texi
sh: makeinfo: not found
*** Error code 1
make: Fatal error: Command failed for target `../info/emacs'
Current working directory /ufs/sys2/FSF/Emacs/21/91/s9/man
*** Error code 1
make: Fatal error: Command failed for target `info'
- -----

I'd like to suggest that we either ship the results of makeinfo in
the tarball or test for the presence of makeinfo in configure and
skip this step if makeinfo is not found.


In order to route around this problem, I put a symlink in ${HOME}/bin
from 'makeinfo' to '/bin/true' which allowed me to "complete" the 
installation process.

Execution of tty emacs (actually in a CDE dtterm with '-nw') worked
fine and more testing of new stuff will take place later.

Execution of X emacs (using the XWS DISPLAY directly) failed badly.
I got a window with a medium tan area above a medium purple area
(the latter is the default background color for CDE on Solaris9)
with no text, no graphics, and no keyboard or mouse capability.
The mouse cursor *did* change to the up-right-pointing arrow when
it went into what *should* have be the toolbar and reverted to the
vertical "i-beam" cursor when it was returned to what should have
been the normal text area.  I was unable to kill this (X) window
via keyboard or mouse (not even by trying the "Close/Alt-F4 action
in the window manager title bar) - I had to resort to sending
an interrupt via the parent shell's tty window.

I tried this again with the --no-init-file flag but that made no
difference.

Checking my build area, it appears I never actually built emacs v21.2.90
so these problems may not be new.

Obviously, debugging is needed but I'm not much of an X programmer.
What do I need to do next?
------- End of forwarded message -------

Re: [drl@vuse.vanderbilt.edu: Re: A new pretest]

[Jan D.]

> Can someone investigate this X bug in the RC branch?
> 

I don't have any access to Solaris 9, but one can speculate.
It sounds like Emacs is on another screen or visual.  I assume you
have set DISPLAY correctly (i.e. other apps work).

Can you try specifying visual to Emacs?  First do:
% xdpyinfo

and you will get some info (you can mail the output here).
 The first line is the DISPLAY value to use, recheck this just to be sure:
name of display:    :0.0

Then there are a list of visuals, like this:
  number of visuals:    2
  default visual id:  0x22
  visual:
    visual id:    0x22
    class:    TrueColor
    depth:    24 planes
    available colormap entries:    256 per subfield
    red, green, blue masks:    0xff0000, 0xff00, 0xff
    significant bits in color specification:    8 bits
  ...

Try each visual in turn by doing (see above, class-depth):
% emacs -xrm '*VisualClass: TrueColor-24'

and see if any of them give something useful.

Also, if you can, try running the same Emacs binary on another display
if you can.  For example, start X on a different machine, log in remotely
to the machine with problems, and run Emacs with DISPLAY set to the
first machine.  Does this give a working Emacs?

	Jan D.


> ------- Start of forwarded message -------
> Envelope-to: emacs-pretest-bug@gnu.org
> Delivery-date: Sat, 12 Oct 2002 13:17:43 -0400
> Date: Sat, 12 Oct 2002 12:17:38 -0500 (CDT)
> From: "David R. Linn" <drl@vuse.vanderbilt.edu>
> Reply-To: "David R. Linn" <drl@vuse.vanderbilt.edu>
> Subject: Re: A new pretest
> To: emacs-pretest-bug@gnu.org
> Content-MD5: pVOlfw/63ZDQWq149/qsZw==
> X-Spam-Status: No, hits=-2.9 required=5.0
> 	tests=QUOTED_EMAIL_TEXT,SPAM_PHRASE_00_01
> 	version=2.41
> X-Spam-Level: 
> 
> >> There is a new pretest available in
> >> 
> >>   <ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-21.2.91.tar.gz>
> >>   <ftp://alpha.gnu.org/gnu/emacs/pretest/leim-21.2.91.tar.gz>
> 
> Sorry about that last useless message - I clicked "send" when I meant to
> click "delete".
> 
> Anyway, ... 
> Under SunOS 5.9/Solaris 9 on SUN Ultra 5 (SPARC v9) with SunPro/Forte C
> v6.0 and otherwise standard Solaris 9 tools/libs, I had no errors or
> unexpected warnings building emacs v21.2.91.
> 
> However, installation failed because I do not (yet) have makeinfo on
> this box.
> - -----
> (cd man; make  info)
> cd /ufs/sys2/FSF/Emacs/21/91/emacs-21.2.91/man; makeinfo emacs.texi
> sh: makeinfo: not found
> *** Error code 1
> make: Fatal error: Command failed for target `../info/emacs'
> Current working directory /ufs/sys2/FSF/Emacs/21/91/s9/man
> *** Error code 1
> make: Fatal error: Command failed for target `info'
> - -----
> 
> I'd like to suggest that we either ship the results of makeinfo in
> the tarball or test for the presence of makeinfo in configure and
> skip this step if makeinfo is not found.
> 
> 
> In order to route around this problem, I put a symlink in ${HOME}/bin
> from 'makeinfo' to '/bin/true' which allowed me to "complete" the 
> installation process.
> 
> Execution of tty emacs (actually in a CDE dtterm with '-nw') worked
> fine and more testing of new stuff will take place later.
> 
> Execution of X emacs (using the XWS DISPLAY directly) failed badly.
> I got a window with a medium tan area above a medium purple area
> (the latter is the default background color for CDE on Solaris9)
> with no text, no graphics, and no keyboard or mouse capability.
> The mouse cursor *did* change to the up-right-pointing arrow when
> it went into what *should* have be the toolbar and reverted to the
> vertical "i-beam" cursor when it was returned to what should have
> been the normal text area.  I was unable to kill this (X) window
> via keyboard or mouse (not even by trying the "Close/Alt-F4 action
> in the window manager title bar) - I had to resort to sending
> an interrupt via the parent shell's tty window.
> 
> I tried this again with the --no-init-file flag but that made no
> difference.
> 
> Checking my build area, it appears I never actually built emacs v21.2.90
> so these problems may not be new.
> 
> Obviously, debugging is needed but I'm not much of an X programmer.
> What do I need to do next?

Re: [drl@vuse.vanderbilt.edu: Re: A new pretest]

[Jan D.]

<<< No Message Collected >>>