Hi.

When floating, requests from clients to cover the whole screen are
supposed to be honoured, and their borders aren't supposed to get in the
way.  That's my understanding, anyway.  However:

------------------------------------------------------------------------

When Firefox is floating and I switch it to fullscreen mode (by pressing
F11), the border remains [1] until I move the window with my mouse,
whereupon it disappears.

[1]: http://users.on.net/~panaceid/dwm-bug/firefox.png (notice how the
right and bottom edges have been cut off)

works: tile layout, c->isfloating False
       (no change)
BROKEN: tile layout, c->isfloating True
BROKEN: floating layout, c->isfloating False
BROKEN: floating layout, c->isfloating True

------------------------------------------------------------------------

The same problem (border displacing the window) can be seen [2] with
MPlayer, but triggering it is a little more complicated.

[2]: http://users.on.net/~panaceid/dwm-bug/mplayer.png

works: tile layout, config.h rule floating mplayer windows
BROKEN: floating layout, config.h rule floating mplayer windows
       (this is the only (slightly) annoying aspect of the bug)
BROKEN: tile layout, NO config.h rule
       (SOMETIMES *FULLSCREEN* when started with -fs option - race?)
BROKEN: floating layout, NO config.h rule
       (SOMETIMES works when started with -fs option - race?)

------------------------------------------------------------------------

A couple of observations: firefox doesn't set any interesting hints, and
mplayer seems to unmap and remap the window or something when
fullscreen-toggled (tag *and isfloating* data get reset).

An old but relevant changeset:
http://www.suckless.org/cgi-bin/hgwebdir.cgi/dwm/rev/722cffee3770

I'd spend a bit longer reading through the code to find out what's
wrong, but I'm probably too inexperienced and lazy to get anywhere.

Regards,
Tony Lainson

Reply via email to