Xref: utzoo comp.sys.dec:1540 comp.windows.x:11754
Path: utzoo!censor!geac!jtsv16!uunet!cs.utexas.edu!usc!bloom-beacon!bu-cs!buengc!bph
From: bph@buengc.BU.EDU (Blair P. Houghton)
Newsgroups: comp.sys.dec,comp.windows.x
Subject: Re: DECWindows and twm
Keywords: 2
Message-ID: <3487@buengc.BU.EDU>
Date: 21 Jul 89 17:01:24 GMT
References: <5586@pt.cs.cmu.edu>
Reply-To: bph@buengc.bu.edu (Blair P. Houghton)
Followup-To: comp.sys.dec
Distribution: usa
Organization: Boston Univ. Col. of Eng.
Lines: 71

In article <5586@pt.cs.cmu.edu> te07@edrc.cmu.edu (Thomas Epperly) writes:
>I have a DECstation 3100 running DECwindows(Ultrix Worksystem Software 2.0
>I haven't gotten the patch yet), and I would like to replace dxwm with twm
>or awm.  After some work I got both twm and awm to compile, but neither
>works correctly.  They both have trouble with the Session manager window.
>awm complains "Unkown property 75".  The symptoms are kind of hard to
>describe, but I will try.  It is like the Session manager's window becomes
>part of the background.  If I move the Session manager window it moves
>away, but a copy remains where it used to be.  This copy cannot be
>selected, moved, or iconified.  This happens with both twm and awm.
>
>Please let me know how to fix this if it can be fixed.  When the
>advertising says that the DECstation uses X, I assumed that it could
>compile and run X software available on the networks(comp.sources.x).
>Also, it would be really nice if they included an appropriate imake in the
>distribution.

Two possible obvious problems:

	1.  You are running a different version of X.  Around here,
	we have various machines running X10R4 and X11R2, and
	in a month or so we may get X11R3.

	    Generally, the differences between R* versions is in the
	library, Xlib.  As I recall, Ultrix 2.0 comes with X10R4.
	Almost all of the currently snarfable code in the *.*.X groups
	is X11, probably R3.  I don't know if there's any reliable
	method of determining a program's requirements, other than
	trying to compile it and run it.

	    You'll need to install Ultrix 3.0, (and then the
	upgrade/revision Ultrix 3.1, since it's there to fix some bugs,
	and who needs those?), to get X11 installed and twm to
	run.

	    It's interesting that they compile but don't run, isn't
	it?  That's because a lot of the Xlib routines operate by
	interpreting data.  For instance, you can specify "expose"
	as an argument to a (ficticious) XDoSomethingToWin() function,
	or "iconify", and the like.  The XDoSomethingToWin() function
	must be able to recognize what "expose" and "iconify" mean.
	Since they are arguments to a function, there is no checking
	for their semantics when compiling, only that they have type
	(char *), so the program compiles.  Different versions of X may
	have different versions of XDoSomethingToWin(), in which one
	may not implement the "iconify" action.  When you compile a
	program written for a version of X different from yours, a
	program which does use "iconify" as an arg to XDoSomethingToWin(),
	and link it with your Xlib, your version of the
	XDoSomethingToWin() function may get confused.

	    Luckily, almost all Xlib routines have intelligent error
	messages, and a policy that implements them uniformly.  Otherwise
	you'd only get such garbola as "Segmentation Fault" and such,
	or "unknown error in XDoSomethingToWin", which is even worse,
	since I for one would be tempted to debug it.

	2.  I dunno.

An opinion (or four):
	
	twm is a pig.  We have it around here (it's X11 only, as far as
	I know) and it really sucks up virtual memory and cpu time.  I
	don't have to run finger to know when twm-lovers are on the
	system.  It has a few snazzy bells and whistles, but I still
	prefer the sleek efficacy of uwm.  awm never thrilled me,
	either.  dxwm is marginally slower and heavier than twm,
	though, so you can actually see an improvement.

				--Blair
				  "...but I hold forth..."