[comp.windows.x] R2- experiences

jip@nih-csl.UUCP (john powell) (03/08/88)

Environment: SUN 3/50 OS 3.4, X11R2 via epos.mit.edu FTP.

X11 server, libs and clients all build with no problems, following
releasenotes and server/ddx/sun/README files.

Observations:

1.  I have not discovered the relevant documentation concerning
    setting up of .Xdefaults and other files for clients which use
    the toolkit.  It seems that the '.' changes to a '*' for clients
    (like xterm) which use the toolkit.  Although MAN pages suggest 
    that the format is *[lowercase] it seems that both upper and lower
    case work.

2.  Also it is not clear whether the -geometry switch
    wants an '=' sign or just the 111x111+0+0.  For example
    xclock says
    Usage:
		 [-g =[<width>][x<height>][<+-><xoff>[<+-><yoff>]]]

    but it accepts -g <width>][x<height><+-><xoff>[<+-><yoff>.

3. If you bring up xterm, get both vt and tek windows showing,
   reverse video on vt and both reverse (OK).  However if you
   bring up xterm (NO tek window showing, just vt) and then
   select reverse video, now when tek window is shown it is
   still normal video and nothing shows up on the tek display.

4. xterm windows do not seem to want to DIE.  The menu kill doesn't
   seem to do anything (signalInhibit is not in .Xdefaults and default
   should be false according to man).

5. Running xterm on a SUN (via rsh) to a server on IBM RT, the xterm
   comes up reverse video.  xterm from R1 comes up correctly.

6. The spaceout (or space) program posted a while back, which runs
   correctly on RTs and uVAXs ... still displays a black screen on
   the SUN X11R2 as it did on R1.  Although I have not yet tried andrew
   on R2 (still FTPing :)), trying the pre-release 2 andrew 
   - cursors inside andrew windows are invisible as they were in R1
   (included here because I suspect its the same problem.)

7. x10tox11 - is very nice - to bad Mark Poeppings Menuwm will
   not run thru it ... maybe CMU will provide an overlapping
   CWM - it would be NICE.  (I am grateful for cmenu)



				John Powell (301) 496-2963
				National Institutes of Health
				DCRT, CSL
				Bldg. 12-A, Rm 2031
				Bethesda, Md. 20892

	uucp -			uunet!ncifcrf.gov!nih-csl!jip
	Internet -		jip%nih-csl@ncifcrf.gov

jeff@polyslo.UUCP (Jeff Weinstein) (03/08/88)

In article <315@nih-csl.UUCP> jip@nih-csl.UUCP (john powell) writes:
>
>6. The spaceout (or space) program posted a while back, which runs
>   correctly on RTs and uVAXs ... still displays a black screen on
>   the SUN X11R2 as it did on R1.  Although I have not yet tried andrew
>   on R2 (still FTPing :)), trying the pre-release 2 andrew 
>   - cursors inside andrew windows are invisible as they were in R1
>   (included here because I suspect its the same problem.)
	I worked on the IBM servers while on co-op.  The spaceout
	program helped to find many bugs in the servers.  It might
	be instructive for other server implementors to try it,
	especially Sun.  As I recall the problem was with stipples.
>
>7. x10tox11 - is very nice - to bad Mark Poeppings Menuwm will
>   not run thru it ... maybe CMU will provide an overlapping
>   CWM - it would be NICE.  (I am grateful for cmenu)
>
	CWM and cmenu were done by the folks at the IBM cambridge
	scientific center in support of the Andrew on X work going
	on at CMU and Palo Alto.

	Jeff Weinstein
	Computer Systems Lab
	Cal Poly State Univ.
	San Luis Obispo, CA
	jeff@polyslo.uucp
	ucbvax!voder!polyslo!jeff

DISCLAIMER:  I no longer work for IBM, and I don't speak for them, or
	Cal Poly for that matter.

RWS@ZERMATT.LCS.MIT.EDU (Robert Scheifler) (03/09/88)

    Date: 7 Mar 88 20:23:27 GMT
    From: elsie!nih-csl!jip@cvl.umd.edu  (john powell)

    1.  I have not discovered the relevant documentation concerning
	setting up of .Xdefaults and other files for clients which use
	the toolkit.

Actually, it is for all clients using the resource manager services of
Xlib, not just toolkit ones.  Relevant documentaion can be found in the
Xlib manual, section 10.11.

			It seems that the '.' changes to a '*' for clients
	(like xterm) which use the toolkit.

That is a good first approximation; lots of things probably want the
weak binding of '*'.  However, you will still want to use '.' for
various things.  For example, in xmh, you probably want to specify
.borderWith instead of *borderWidth, unless you want to affect the
borders of all the internal buttons, etc.

					     Although MAN pages suggest 
	that the format is *[lowercase] it seems that both upper and lower
	case work.

Case is significant in resource names, typically distinguishing instance
from class.  In simple toolkit clients either will work because there is
only one instance.

    2.  Also it is not clear whether the -geometry switch
	wants an '=' sign or just the 111x111+0+0.

The leading = is not required; we allow it for backward compatibility
of a sort.

    3. If you bring up xterm ...

If you find bugs, submit reports to xbugs@expo.lcs.mit.edu.  Use the
template provided with the release.

    4. xterm windows do not seem to want to DIE.

Works fine for me.  If you find bugs ...  and provide us with sufficient
context (OS, etc).

    5. Running xterm on a SUN (via rsh) to a server on IBM RT, the xterm
       comes up reverse video.  xterm from R1 comes up correctly.

If you find bugs ...

    7. x10tox11 - is very nice - to bad Mark Poeppings Menuwm will
       not run thru it

If you find bugs ...