[comp.windows.x] X on Apollo

mjc@me.brunel.ac.uk (Mike Crawley) (12/15/87)

I have fixes 1-92 installed and running nicely on a Sun 3/110
and I'm trying to get it all compiled again on an Apollo DN3000.
When I first copied X11R1 to the Apollo I had a few problems
with imake, Xtk/Load.c, oldXtk/Load.c et cetera.
Most of these are mentioned in the server/apollo/README.

If you are lucky enough to have a Sun you can use imake there :-

  mv Sun.macros Sun.macros.moved
  ln -s Apllo.macros Sun.macros
  make Makefiles
  rm Sun.macros
  mv Sun.macros.moved Sun.macros

  now copy/tar the lot to the Apollo

I had one or two problems with the original version which STILL
exist in the current version.

I had to #define GLYPHPADBYTES 4 in server/include/servermd.h
and #define DEFAULTGLPAD 4 in fonts/compiler/fc.h
before I could get all of the fonts to work.
The worst culprit was variable.snf.
(xterm -fn variable would just hang)
Once I had done the change it was ok, and the font files compared
exactly with those on the Sun (as would be expected).

xmh would not build since the author decided to access internal structures
in <stdio.h>. This is not a wonderful idea since on every Unix system
I've ever used I have noticed that they are always slightly different.
I think he should have used the fileno() macro instead of f->_file or f->_fd

I also got many Warnings out of the Apollo compiler - some just
the usual "value computed but result never used" type of thing,
but some more worrying things like "variable x used before set".
This is the sort of thing lint is supposed to find. I've noticed
that some of these have now gone away after fixes 1-92 but I still have
one or two of these "worrying" warnings left.
If someone out there would like a complete LOG from "make -k >& LOG" 
I will email it to you.

The Imakefile in server seems to have a typo near APOLLOOBJS = ...
There is a blank line in mine between APOLLOOBJS and /sys/traits/io_traits
I think it should mention ddx/apolloGC.o and ddx/apollo/gpr_tmgr.o
somewhere since it won't link/bind without these two.

How many people are trying to use the Apollo version ?
The net seems to have lots of messages about Suns but very few about Apollos.
We are hoping to be exhibiting our software (works fine on Sun so far)
at European exhibitions in '88 on both Suns and Apollos and it would be nice
for us (and Sun and Apollo) if we could show everything working.

My application uses the oldXtk and oldXrm libraries and creates
a lot of buttons, windows and bitmaps. At the moment it is falling
over in XrmAtomToQuark called from XtGetResources called from XtCommandCreate.
On the Sun it works fine - any ideas anyone.

Mike Crawley                           email: mjc@ahl.co.uk
Abstract Hardware Limited                Tel: 44 895 58501

alan@arc.UUCP (Alan Mimms) (12/23/87)

	How many people are trying to use the Apollo version ?
	The net seems to have lots of messages about Suns but very few about Apollos.
	We are hoping to be exhibiting our software (works fine on Sun so far)
	at European exhibitions in '88 on both Suns and Apollos and it would be nice
	for us (and Sun and Apollo) if we could show everything working.

We're running our X.V11R1 on Apollos and Suns.  Since Suns are scarcer here,
we use the Apollos as Xwindows terminals most of the time, starting a window
manager on the Suns with "rsh".  Apollos (with a few minor tweaks to the
X.V11R1 release of the server -- if you're interested, drop me a line) make
pretty nice terminals.

We're attempting to do portable development of a toolkit (which is actually a
port of our old VAX/VMS/UIS toolkit) to run on VMS, Apollos, and Suns.  Right
now, the development team is using Apollos exclusively because of a (minor)
difficulty with Sun's C compiler accepting things like

	x = 'ABCD';
	
The people at Sun argue that 'ABCD' is an erroneous constant value, whereas
EVERY other compiler I've tried (including the original VAX/4.2BSD compiler)
accepts it just fine.

	Alan Mimms							texsun!arc!alan
	ADVANSOFT Research Corp.			Telephone (408) 727-3357

jkh@VIOLET.BERKELEY.EDU (12/23/87)

I think Sun's right. Multiple character char constants are completely
weird.

x = 'ABCD'??

Bogus! Evil!


					Rev. Hubbard

hamid@lims01.lerc.nasa.gov (HAMID SAMADANI) (12/24/90)

	I'm porting a database system from a

		Sun 3/60 OS 4.0 (BSD 4.2 Unix)
		INGRES DBMS
		written in C
		supports X Windows as optional U/I

	to an Apollo DN 2500 SR10.2 (supports BSD 4.2).

	Questions:

	What do I need on the Apollo to support the X Windows interface?
	Would you recommend X as the user interface?  Or should I 
	pursue another alternative?  The system was written using Sun's
	cursor manipulation package 'curses' which in turn uses 'termcaps' 
	for its original user interface.  I'm told that the system also 
	supports X Windows.

		Hamid