[comp.sys.apollo] Build Motif 1.1 under 10.3

rtdickerson@lescsse.jsc.nasa.gov (russel dickerson) (04/25/91)

What we need to know is which environments
we need to use to sucessfully build Motif 1.1 on
this machine : DN3000 runing 10.3 ?

We have the source for Motif1.1 from OSF.

The possible environments are:

1. From HP/Apollo, the PSKQ2-91 :
	Xdomain, xdm, and mwm (version unknown)

2. From MIT X11R4 :
	Xapollo, xdm

We have already built the X11 R4 from MIT.
We have built the Motif 1.1 with the PSKQ2-91 and found
that the majority of the demos would not run.  They 
get as far as widgets being managed onto the screen and 
then die with a "memory fault" (how informative).

Any help would be most appreciated!
--
Russell Dickerson          Internet: dickerson@vf.jsc.nasa.gov
Lockheed (LESC), A22       ===================================
SSE System Project         X Windows & Motif on Apollo/PC/Mac
Space Station Freedom      Phone +1 713 283 5193  

hanche@imf.unit.no (Harald Hanche-Olsen) (04/26/91)

In article <rtdickerson.672522451@node_25d97> rtdickerson@lescsse.jsc.nasa.gov (russel dickerson) writes:

   2. From MIT X11R4 :
	   Xapollo, xdm

From my experiences and the total lack of helpful responses from the
net when I asked for help, I would say that this is doomed to fail (on
an SR10.3 Apollo).  Send me mail if you want the gruesome details.

- Harald Hanche-Olsen <hanche@imf.unit.no>
  Division of Mathematical Sciences
  The Norwegian Institute of Technology
  N-7034 Trondheim, NORWAY

rxdk20@monarch.oakhill.uucp (Barry Spotts) (04/30/91)

Though some think it is doomed to fail, I have built MOTIF 1.1.1 on a DN4500.
It builds MOTIF applications okay, but Mwm does not run properly.  I have yet
had the time to go back a figure that one out.

Rule #1:  Don't use what Apollo gives you as far as X libraries go.  They
    simply won't work for software development.  Use MIT's version of 
    X-windows and make sure you get that to work properly. Make sure
    your have "#define HasNdbm YES" in your apollo.cf, and the following
    defines worked for me:

    DefaultCCOptions are -U__STDC__ -D_PROTOTYPES -s
    #define CppFullANSISupport 0
    #define Connection Flags -DTCPCONN
    #define miOptChoices -DAVOID_GLYPHBLT -DNEED_OLD_MFB_MASKS
    #define ServerDefines StandardDefines ExtensionDefines -DXDMCP miOptChoices

    Before building, make sure that you comment any of Apollo's global X 
    libraries found in /etc/sys.conf.

    I built X-windows and MOTIF using Apollo's CC 6.8 compiler.
    This provides a usable X-environment for building X applications although
    running applications might be less than ideal.  I have yet to find a decent
    X-server for Apollo.  You might want to use GNU malloc instead of the 
    malloc provided by Apollo. When I find the time to build MOTIF 1.1.2 I
    think I will use it.

Rule #2: Don't even think of building X11R4 under anything less than SR10.3.
         I've had lots and lots of problems, and I finally just gave up.
         
MOTIF BUILDING:
    OSF isn't much help on building MOTIF apollo.  They don't have one in house
to test even test MOTIF on.  I built MOTIF in a separate directory from my X
directory.  No since in corrupting my X tree. Just add softlinks from

 ./extensions => the extensions directory in your X build tree.
 ./lib/X => the lib/X in your X build tree.
 ./lib/Xt => the lib/Xt in your X build tree.

Copy your apollo.cf file used in building the X11R4 binaries to your 
config/apollo.cf in your MOTIF tree.

Add the following lines to the top of your apollo.cf file:

SYSLIBS=/usr/lib/libXt.a /usr/lib/libX11.a
MWM_DEFINES=-DSHAPE -DOPAQUE  #These may be wrong.  I haven't gotten MWM to
                               work properly. But, all MOTIF DEMOS run okay.

Change your DefaultCCOptins to be
#define DefaultCCoptions -U__STDC__ -D_NO_PROTO -DSYS_DIR -A nansi -s
I then did a make World.

Make sure that all Apollo's global libraries are removed from your 
/etc/sys.conf file.
I haven't had a chance to install the PSKQ2-91 patch yet.  I've seen it,
and have tried to port code and MOTIF using it to no avail.

IT MUST BE NOTED THAT MY GROUP USES THIS ENVIRONMENT FOR THE CREATION OF
X APPLICATIONS AND THEN TRY TO RUN IT UNDER APOLLO'S X-SERVER.  APOLLO'S
X-SERVER IS STILL BUGGY FOR SOME APPLICATIONS, BUT IT IS STILL FASTER THAN
WHAT COMES ON THE X11R4 DISTRIBUTION.  WE HAVE VERY LITTLE PROBLEMS
WHEN WE RUN ON X APPLICATIONS ON OUR APOLLOS AND DISPLAY ON OUR SUN.

===============================================================================
Motorola Incorporated                              Barry R. Spotts
Semiconductor System Design Technology             3501 Ed Bluestein Boulevard
                                                   Austin, Texas 78721
Internal Address: barry@monarch                    MailDrop: M-2
UUCP Address: barry@monarch.sps.mot.com            (512) 928-6630
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

--
===============================================================================
Motorola Incorporated                              Barry R. Spotts
Semiconductor System Design Technology             3501 Ed Bluestein Boulevard
                                                   Austin, Texas 78721
Internal Address: barry@monarch                    MailDrop: M-2
UUCP Address: barry@monarch.sps.mot.com            (512) 928-6630
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

hanche@imf.unit.no (Harald Hanche-Olsen) (04/30/91)

In article <RXDK20.91Apr29180013@monarch.oakhill.uucp> rxdk20@monarch.oakhill.uucp (Barry Spotts) writes:

   Though some think it is doomed to fail, I have built MOTIF 1.1.1 on a DN4500.

That is good news.  Just for the record, what I said is doomed to fail
is getting the MIT X11R4 SERVER (Xapollo) running properly.  The
client library is a different story altogether.  Sorry if I didn't
make that clear.

   IT MUST BE NOTED THAT MY GROUP USES THIS ENVIRONMENT FOR THE CREATION OF
   X APPLICATIONS AND THEN TRY TO RUN IT UNDER APOLLO'S X-SERVER.

And this explains why Barry does not share my "doom & gloom" feelings :-)

- Harald Hanche-Olsen <hanche@imf.unit.no>
  Division of Mathematical Sciences
  The Norwegian Institute of Technology
  N-7034 Trondheim, NORWAY