On 29 May 2010 22:18, Sylvain Laurent <syll...@gmail.com> wrote: > 2010/5/29 Kris Maglione <maglion...@gmail.com>: >> On Sat, May 29, 2010 at 07:59:39PM +0100, Anselm R Garbe wrote: >>> >>> On 29 May 2010 19:17, Anselm R Garbe <garb...@gmail.com> wrote: >>>> >>>> On 29 May 2010 18:18, Kris Maglione <maglion...@gmail.com> wrote: >>>>> >>>>> On Sat, May 29, 2010 at 01:12:15PM +0100, Anselm R Garbe wrote: >>>>> The getfullscreen bit is probably not necessary in most cases. The rest >>>>> of the clientmessage function is a hack, because I don't know the dwm >>>>> sourcecode well enough to do it properly. It's just to show what's >>>>> required. >>>> >>>> I got a similar impression when reverting the changeset that the EWMH >>>> fullscreen handling was incomplete. >>>> >>>> Overall my aim is trying to prevent using too much EWMH magic, though as >>>> things look more clients seem to fully depend on EWMH nowadays for >>>> fullscreen handling, like chromium. >> >> I think the intention of the patch wasn't to support EWMH fullscreen so much >> as to tell apps that they were in fullscreen mode. Without it, chromium went >> fullscreen fine, but it still showed its tab bar, etc. A lot of apps (when >> the WM supports EWMH) only show their fullscreen UI when they have the EWMH >> fullscreen hint set on their window. It's actually a good thing in a lot of >> cases, because if you manually force the apps to fullscreen from the WM side >> they tend to hide their decorations properly. > > Yes, that's what I was thinking about EWMH, there is already monocle > mode to set full screen on the user's end, by not maximizing windows > when sending hint to the window it allows users to play with > monocle/floating/tiles mode.
I spend some time on the fullscreen issues and came up with a version that works for all cases I tested so far. It is committed into hg tip. It is a simplification of Kris' patch (basically I got rid of superflous checks that don't seem to be necessary), but with a more complicated clientmessage(). I also extended Client a bit to keep state and changed resize(). The changes I did are backwards compliant for any layout algorithms however. Please test hg tip and let me know any issues. Kind regards, Anselm
