On Fri, 02 Aug 2002 10:55:02 -0700
Gordon Prieur <Gordon.Prieur_at_Sun.COM> wrote:
>
> Dan Espen wrote:
>
> > Gordon Prieur <Gordon.Prieur_at_sun.com> writes:
> >
> >>Hi,
> >>
> >> I just looked at the Java bug report and they experimented with a
> >>fix for Java 1.4 and decided its a fvwm bug. The comment in the bug
> >>evaluation says that fvwm ignores gravity when handling
> >>ConfigureRequests. Perhaps one of the current fvwm developers could
> >>take a look at this and see if this is true and if it fixes the
> >>problem. If its not the problem you might try contacting Sun about
> >>bug # 4401846. Perhaps the 2 sets of developers (fvwm and Java) can
> >>get this resolved together.
A golden opportunity. Let's grab it!
This bug report is really a restatement of bug # 4102292, the great-grand-daddy
of these reports (open since Jan 07, 1998, with a long list of public comments
including a Happy Birthday message to the bug). In a nutshell, the problem is
that Java insists on placing windows at 0,0 if the program hasn't specified
placement coordinates. It correctly calculates the offsets for the window
decorations for relatively simple WM's - e.g., twm - and adjusts accordingly. It
calculates them incorrectly for WM's that do double (or more) reparenting of the
top level window.
There are actually over 40 bug reports (that we can see on the public list,
anyway) over the past 4.5 years (since around jdk 1.1.5) describing what is
substantially the same problem. Some reference other window managers (e.g., the
GNOME and KDE WM's); some describe the symptoms a bit differently, and so on;
but they all boil down to the same misfeature.
The infuriating thing about the 1.4 release is that we had developed a hack
that worked - setting the property on the root window that made the JVM think
the mwm was running, and it seemed to be happy enough to let mwm decide where to
place the windows without interfering. We've now had to work on other hacks to
try to get around the problem.
The fundamental question I see in all this is: what earthly purpose is served
by trying to force the top level window positions to 0,0? What would it take to
convince the Sun developers that this is genuinely a Bad Thing?
> I'll see if I can get authorization to send the full evaluation to
> this list (it shouldn't be hard to get). I doubt its intentionally
> not on the external page but suspect the external (ie, viewable from
> the java web pages) just isn't as up-to-date as it should be.
That would be great, if you can!
> Its still open but I doubt if the bug will get serious consideration
> again unless somebody can convince the JDK developer their evaluation
> is wrong (this is what I'm hoping to get the fvwm developers to do :-)
We'll certainly try.
Cheers,
Bob
--
Visit the official FVWM web page at <URL: http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm" in the body of a
message to majordomo_at_fvwm.org.
To report problems, send mail to fvwm-owner_at_fvwm.org.
Received on Fri Aug 02 2002 - 14:48:59 BST