[418] in Zephyr_Bugs
Re: [Dave Glowacki: Re: [Marc Horowitz: Re: zwgc vs. tvtwm] ]
daemon@ATHENA.MIT.EDU (Dave Glowacki)
Wed Oct 7 13:30:49 1992
From: dglo@CS.Berkeley.EDU (Dave Glowacki)
To: Marc Horowitz <marc@MIT.EDU>
Cc: John T Kohl <jtkohl@cs.berkeley.edu>, zephyr-bugs@MIT.EDU
In-Reply-To: Your message of "Tue, 06 Oct 92 22:43:30 CDT."
Date: Wed, 07 Oct 92 10:30:19 PDT
> >> It's quite probable that tvtwm behaves that way as well (after all,
> >> it's using the same base as vtwm.) Does 'RandomPlacement' affect
> >> this?
>
> yes, it does.
Well, actually 'RandomPlacement' only affects placement if
'UsePPosition' isn't set to either "on" or "non-zero" (I've got
it set to "non-zero".) Try setting 'UsePPosition' on vtwm
and it'll place the window for you.
> >> I guess I'd like to see zwgc default to PPosition. It seems like
> >> it's the window manager's fault if it ignores perfectly good
> >> position hints.
>
> That would immensely screw over every single user at MIT.
No, it would only screw over those users who were using a twm
derivative and didn't have 'UsePPosition' set, or those users
who had turned off PPosition in one of the other window managers.
> tvtwm is
> the only wm I've heard of which places the window itself by default if
> PPosition is specified.
Hmmm ... let me introduce you to some brand new window managers then:
mwm and olwm (or its virtual cousin, olvwm). All of these, BY
DEFAULT, place the window if PPosition is set.
> I don't have my ICCCM handy, but I seem to remember that it's pretty
> specific about when it the USPosition bit should be set, so an option
> might not be legal.
Here, let me quote the relevant section for you:
To indicate that the size and position of the window (when
mapped from the Withdrawn state) was specified by the user,
the client should set the USPosition and USSize flags,
which allow a window manager to know that the user
specifically asked where the window should be placed or how
the window should be sized and that further interaction is
superfluous.
To indicate that it was specified by the client without any
user involvement, the client should set PPosition and
PSize.
Note that the ICCCM *requires* user involvement before USPosition
is set. For 'zwgc' to be in compliance with this standard, there
would need to be an option set before the *current* behavior was
allowed.
I checked our entire X source tree and turned up 27 applications
which do anything with USPosition. Of those, only 3 set USPosition
by default. The rest default to PPosition and set USPosition if
the user specifies some option (usually geometry.)
According to both the ICCCM standard and to common programming
practice, what 'zwgc' does is clearly WRONG. I'm not asking that it
be fixed, merely that I have the *option* to make it behave
correctly.
