client_leave() served no real purpose, likewise no reason to handle
LeaveNotify events since an EnterNotify will process the next active
client (and we don't have anything important to process anyway), so
xev_handle_leavenotify() goes as well. Allows a simplification of
client_mtf() and client_cycle_leave() for clarity. While here, unify a
few client_current() checks.
No intended behaviour change.
queue check (removing the need for a server grab/ungrab) - if the client
is going away, let it fall all the way through to a DestroyNotify event.
There's no longer a need for us to manually destroy a client ourselves
(removing yet another server grab/ungrab). Instead, when the
UnmapNotify event is synthetic, simply set the state to Withdrawn (as
per ICCCM), else Iconic (in our case 'hidden').
Verified with test case from the 2009 race which was the original reason
for r1.35 of event.c.
XSetInputFocus() for clients that have the InputHint; latter fix
discovered by Valery Masiutsin with a PoC patch - solves keyboard input
focus loss for java apps.
triggering event was unmap (with pending destroy) log destroy; we only
need to grab/ungrab the server lock, set WithdrawnState and
XRemoveFromSaveSet clients coming from an unmap event - doing so for
clients coming from destroy are already gone, hence we generate errors.
BadValue) when the modifier is already AnyModifier . Instead alter
xu_btn_ungrab() to ungrab AnyButton/AnyModifier and call it only when a
client is coming into focus in client_setactive(), instead of iterating
over ignore mods - matches how we handle key grabs.
the cc->pwin days - don't redraw on every unhide, resize, and mouse
move/resize Expose event (note, all Expose events trigger a redraw
anyway).
Tested with some finicky X apps I could think of, though I'm sure others
will find more - if so, and they 'lose' the border, please report!
to grab keys in keybindingq. we don't need to ungrab/grab on every
addition to the queue, just once with a complete keybindingq; simplify
grabbing keys per screen (during init) and during a MappingNotify.
while here, change conf_grab_{kbd,mouse} to require only a Window.
- merge border/menu color structures/functions since they now both use Xft
- switch xu_xorcolor to operating on XftColor instead of just
XRenderColor (basically adding pixel)
- if color name allocation fails, revert back to default (this, along
with font validation should occur during config parse, but we don't
have screens setup yet - likely to change at some point)
such as those using sdl, attempt to manage the clients themselves when the
clients aren't fully ready. other wm's grab the xserver during the whole
client setup process, so match.
behavior found by jsg.
note that a re-exec of cwm will not rewrite the group number atom of
*existing* clients, so they will remain off-by-one until each client has
its atom updated, or of course a restart of X.
bwidth; this allows a client to be resized from a max state, which now
gets treated like a non-max'd client. based on a diff that does part of
this in a different way from Alexander Polakov.
i.e. horiz-max + vertmax = full maximisation. full - horiz = vertmax.
etc.
Martynas wrote something like this once, so I did okan, this version
seems to finally deal with the corner cases.
ok okan@.
keyword taken from a diff from Sviatoslav Chagaev to do the same thing,
but implemented in a completely way (based on some very old code from
mk@). default set to 0, so no behavior change.
ok oga@ (who would also like to take it further...)
any window, after which all move/resize requests will be ignored,
essentially freezing the window in place.
there's a possibility to merge this with the 'ignore' concept, pending
on how ignore+freeze should behave (really more ewmh stuff), but punting
for now since ponies are on the line.
requested and tested by thib at k2k11 with ponies, unicorns and rainbows.
'save the unicorns' todd@, ok oga@
that if it is not already inbounds, put it in the default location,
which happens to be inbounds.
behavior noticed by Thomas Pfaff while maximizing and un-maximizing a
window, leaving the pointer behind and causing client_cycle() to be a
bit lost.
ok oga@
remove screen_current() it was utterly bogus when nscreens > 1.
pass a fake client_ctx in the case where there's no client and the
kbfunc or mousefunc doesn't need a real one, it just contains the
current screen, modify these functions so that they pass down the screen
context to their callees.
make groups per screen, it's the only way it makes sense in this regard.
ok okan@.
the x property api doesn't let you remove one entry from an X property
array, so client_remove is kinda expensive, but there's no real way
around that..
ok okan@
