Ah! Thanks for the wonderful research.
I'm not really sure why that code was deleted, and it certainly can't
be restored as is because changes in how the code works now, but I
added back the equivalent of the last three lines and that fixes full
screen in gqview, xine and feh (don't have gmplayer to test here), and
mplayer still works fine. Qiv is still broken, but qiv is broken even
in non-fullscreen, so the issue is somewhere else.
In any case, I pushed the fix to the hg tip, please test as it
probably has broken something given that I have no clue what I'm doing
:)
Best wishes and thanks again for the great research
uriel
P.S.: If you could find out when qiv started breaking it would be
great! *hint* *hint* ;)
On 10/2/07, Andre Kuehne <andre.kuehne_AT_gmx.net> wrote:
> Andre Kuehne wrote:
> > Uriel wrote:
> >
> >>> 5) support for apps native full screen modes sucks (qiv [drops behind
> >>> managed layer], gqview [window appears slightly offscreen], firefox
> >>> [titlebar doesn't come back when you return from fullscreen])
> >>
> >> This IMHO is the greatest issue, I know it was fixed at some point,
> >> sad to see it broken again, I hate X...
> >
> > In an earlier mail i identified one revision, which broke fullscreen
> > support for mplayer:
> >
> > http://www.suckless.org/pipermail/wmii/2007-June/003547.html
> >
> > Just to be sure i did some more tests.
> >
> > I built libixp from latest revision before the suspected wmii commit:
> >
> > changeset: 39:7c3b9ef60da1
> > user: Kris Maglione <jg_AT_suckless.org>
> > date: Tue Feb 20 13:25:44 2007 -0500
> > summary: Undoing last change. It was already done in LDFLAGS
> >
> > Then i build and run wmii from this revision:
> >
> > changeset: 1981:18c1736806e7
> > user: Kris Maglione <jg_AT_suckless.org>
> > date: Thu Mar 01 16:37:00 2007 -0500
> > summary: Fix flicker on keyboard focus change.
> >
> > These are the fullscreen test results:
> >
> > gqview: works
> > feh: works
> > mplayer: works
> > gmplayer: works
> > xine: works
> > firefox: not working
> >
> > And the same with the subsequent revision:
> >
> > changeset: 1982:65643b7a63d6
> > user: Kris Maglione <jg_AT_suckless.org>
> > date: Fri Mar 02 03:13:32 2007 -0500
> > summary: Implemented the Aspect hint, as well as some moderate cleanup.
> >
> > gqview: not working
> > feh: not working
> > mplayer: not working
> > gmplayer: not working
> > xine: not working
> > firefox: not working
>
> I looked over the changeset and found one part which looks like the culprit to me:
>
> --- a/client.c Thu Mar 01 16:37:00 2007 -0500
> +++ b/client.c Fri Mar 02 03:13:32 2007 -0500
> @@ -35,23 +34,16 @@ create_client(Window w, XWindowAttribute
> c->border = wa->border_width;
> c->rect.width = wa->width;
> c->rect.height = wa->height;
> +
> c->proto = win_proto(c->win);
> - update_client_name(c);
> -
> - c->fixedsize = False;
> - XGetTransientForHint(blz.dpy, c->win, &c->trans);
> - if(!XGetWMNormalHints(blz.dpy, c->win, &c->size, &msize) || !c->size.flags)
> - c->size.flags = PSize;
> - if(c->size.flags & PMinSize && c->size.flags & PMaxSize
> - && c->size.min_width == c->size.max_width
> - && c->size.min_height == c->size.max_height)
> - c->fixedsize = True;
> - if(c->rect.width == screen->rect.width
> - && c->rect.height == screen->rect.height)
> - c->fullscreen = True;
>
> But i am not familiar with the codebase so this is just a guess.
> Tomorrow is an official holiday in Germany, so maybe i find the time
> to catch up on this.
>
> Best wishes
> Andre
>
>
>
Received on Wed Oct 03 2007 - 03:02:05 UTC
This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:28:34 UTC