[comp.unix.sysv386] Looking for SCO Unix X11R4 implementation

mjung@leuze-owen.de (Marc Jung) (09/13/90)

I had already a look on several news groups to find 
someone who did the job of implementing the MIT tape
X11R4 under SCO Unix V3.2, but there seem to be noone....

Has somebody done this job before ?

We feel that Xsight or Open Desktop are too expensive
to buy, when the X11R4 tape is available, but we did not
succeed in compiling until now....


=======================================================================
snail-mail:
Marc Jung                         email   : mjung@leuze-owen.de                   
Leuze Electronic, Abt. TDV          old   : uunet!unido!leuze!mjung
In der Braike 1                           : mjung@leuze.UUCP
D7311 Owen / Teck                 voice   : +49 7021 573215 fax: 573200
West - Germany                    ampr-net: dl5saq@db0sao
=======================================================================

jcc@adesign.uucp (Jean Christophe) (09/21/90)

>I had already a look on several news groups to find 
>someone who did the job of implementing the MIT tape
>X11R4 under SCO Unix V3.2, but there seem to be noone....

>Has somebody done this job before ?
Yep, I did it for the libraries, clients & demos for now.
I am currently working on the server (but not for VGA).
I had to use TLI (streams) instead of sockets and the code had to be
compilable with vanilla SCO cc (MSC 5.1 I think...).

This was a real pain either because the STREAMSCONN code given with X11R4
tapes is very buggy AND because SCO is a damn sh*t (std disclaimer :
this is my personal opinion only).

Now the point is that I'm doing this work for my company (in fact : for a
customer of my company) so I can't make the code public (sorry guys).
But I can give you some technical hints :

#1 - When building with STREAMSCONN

	- Get rid of the Xdaemon as soon as possible : You don't need it if you
	  have no intents to support multiple type of LAN.

	- Correct WriteTliStream in Xstreams.c so that it can handle message
	  longer than 4096 bytes (t_snd will fail if you send more than 4K in
	  one single packer). You have to test and split if necessary.

#2 - Generally

	- A lot of header files are to be corrected.

	- Some identifier (principally in Xext) are longer than 32 chars and
	  there are some collisions (MbufGetMultiBufferAttributes*, 
	  GetSelectedExtensionEvents*)

>We feel that Xsight or Open Desktop are too expensive

Not only that, it's terribly buggy and you don't have Athena Widgets!

>to buy, when the X11R4 tape is available, but we did not
>succeed in compiling until now....
Hope this will help.
Don't hesisate to contact me if you have any specific questions

-- Jessie

PS: If I had enough spare time, I would be very willing to port X11R4 on SCO
    on my own (for VGA) and make it PD....
    But I can and will gladly help anybody working on this.

	jcc

------------------------------------------------------------------------------
jcc@adesign.uucp	| "An artificial intelligence is better than none!"
..!inria!adesign!jcc	| "Artificial intelligence matches natural stupidity!"
Collet Jean-Christophe	| "Objets inanimes avez-vous donc une ame ?"
------------------------------------------------------------------------------
Axis Design			|
119, rue de Flandres		|
75019 Paris			|    <this space left intentionaly blank>
France				|
Tel: +33 (1) 40 35 20 20	|
------------------------------------------------------------------------------

ghost@robecdc.UUCP (William Sneed) (09/22/90)

Maybe by Christmas?  

This is an interesting port.


-- 
-William A. Sneed	 | All words are from me not my employer. All errors
-ghost@dc.robec.COM      | are mine, all the glory is mine, all the flames are
-(703) 631-4800 x114	 | mine, good, bad and ugly it's all mine. :-{> :-{> :-)
-------------------------| God grant me the Serenity....