Re: [dwm] Java Saga Continues...

On Thu, Dec 21, 2006 at 12:44:28PM +0100, Anselm R. Garbe wrote:
> On Thu, Dec 21, 2006 at 12:41:03PM +0100, Anselm R. Garbe wrote:
> > On Thu, Dec 21, 2006 at 12:15:42PM +0100, Sander van Dijk wrote:
> > > On 12/21/06, Anselm R. Garbe <arg_AT_suckless.org> wrote:
> > > >I fear there is no simple solution unless Sun don't fixes the
> > > >JDK properly. Introducing frames for each client would mean much
> > > >more resource consumption of X, and the dwm codebase would get
> > > >more complex - one would need to change nearly all
> > > >resize-related stuff, manage() and unmanage(), besides extending
> > > >Client struct with Client->fwin.
> > >
> > > Just a (probably silly) idea: would reparenting to the parent help? I
> > > haven't read Sun's sources myself, but from what I understand these
> > > java progs discard all configurenotify's and resizes if they are not
> > > reparented (or rather, _until_ they are reparented). If one would
> > > reparenting such a window to what already is its parent, this java
> > > stuff might still react to the ReparentNotify this generates. For the
> > > rest nothing would change, so dwm's codebase wouldn't need to change
> > > any further. If this works, a nice "/* stupid hack to work around java
> > > issue */" above the code would suffice I think.
> > >
> > > Note that I don't really have a clue about this stuff, so this might
> > > all be nonsense; decided to share it anyway in case it isn't garbage
> > > though :-)
> >
> > I checked this already a while ago, but no that won't help. In
> > the Java Odyssey discussion I pointed to the Insets method, which
> > performs an XQueryTree(3X11) to check the window hierarchy.
> > We need a real parent window which is not the root window.
> >
> > One ugly trick which comes to my mind is, using a parent window
> > for all, which is not the root window. This could be a solution,
> > but I'm not sure about its impacts yet. Will investigate further
> > during 23c3 into this issue.
>
> Nah nah nah, when remembering how the insets were calculated,
> that's not a solution... This stupid Java crap would determine a
> window decoration of the rest screen space which is not covered
> by a client window in such a situation.
>

I don't think dwm should be adjusted to the fact that Java developers
did such a poor job at writing the backend for X. Complaints should be
directed to Sun. That's of course all fine as a principle but in the
meantime some sort of practical solution would be nice (although for now
continuing with Java5 is just fine).

I'll try and experiment a bit more over the holliday period.

cheers,
Anthony
Received on Thu Dec 21 2006 - 14:22:18 UTC