[413] in Zephyr_Bugs
Re: [Dave Glowacki: Re: [Marc Horowitz: Re: zwgc vs. tvtwm] ]
daemon@ATHENA.MIT.EDU (Marc Horowitz)
Sat Oct 3 04:37:18 1992
To: John T Kohl <jtkohl@cs.berkeley.edu>
Cc: dglo@CS.Berkeley.EDU (Dave Glowacki)
Cc: zephyr-bugs@MIT.EDU
In-Reply-To: [411] in Zephyr_Bugs
Reply-To: Marc Horowitz <marc@MIT.EDU>
Date: Sat, 03 Oct 92 04:36:53 EDT
From: Marc Horowitz <marc@Athena.MIT.EDU>
>> After Mr. Horowitz' message, it became a point of honor to find the
>> problem :-)
I don't even remember what I said :-) Could you send me a copy?
>> 'zwgc' is setting size and position hints for the window manager but
>> claiming that they're from the *user* instead of from a *program*.
>> Obviously, vtwm doesn't distinguish between the two but tvtwm does.
vtwm *does* distinguish between the two. empirically (determined by
playing around, not reading source), if the location is
user-specified, then the wm just puts the window there, and if it's
program-specified, then the wm rubber-bands for the location. That's
for a normal top-level window; transients may be different.
>> A more X-ish solution might be to have a resource, something subtle
>> like 'wm-is-spineless', which could be set if the window manager
>> didn't handle program-positioned windows. zwgc could check this
>> resource to decide which type of hint to give...
It's a little more complicated than that. One could claim that zwgc
should tell the absolute truth: always set PSize, and set USPosition
if the geometry of the zgram is derived from the user, via a resource
or desc file variable, or PPosition if it's just using the default.
But one could also claim that zgrams should always map w/o user
intervention, which would mean setting USSize|USPosition.
So, the real questions are:
1) what is tvtwm's behavior for different hints?
2) what behavior (independent of wm) do you want zwgc to have?
The answers to these questions (modulated by my whims as the
programmer ;-) determines if a new resource is appropriate.
Marc
