Commit Graph

231 Commits

Author SHA1 Message Date
Mike Blumenkrantz 9005332468 ensure that client eval correctly handles retries on visibility-pending clients
ref 2854352bc8
fix T3636
2016-05-13 11:25:06 -04:00
Mike Blumenkrantz 00e427b126 adjust some client eval parts to make manual placement work again
activating the window_move action doesn't require the client to successfully
be shown, and failing this check would cause the window_move action to be
deleted until the next restart
2016-05-12 12:43:49 -04:00
Mike Blumenkrantz 5cd256da78 ensure child windows are placed on screen
when centering a child over its parent, checks were not made to ensure that
by centering the child it had not gone offscreen
2016-05-12 12:42:21 -04:00
Mike Blumenkrantz dc0ec2339e feed mouse up events to internal wins before mouse out during action exec
ref 609276e12d

fix T3347
2016-05-12 12:34:00 -04:00
Mike Blumenkrantz dd07d01737 only set "dialog" border for transient clients if window type is unset
fix T3157

from https://specifications.freedesktop.org/wm-spec/wm-spec-latest.html :
_NET_WM_WINDOW_TYPE_DIALOG indicates that this is a dialog window.
If _NET_WM_WINDOW_TYPE is not set, then managed windows with
WM_TRANSIENT_FOR set MUST be taken as this type
2016-04-26 10:56:59 -04:00
Mike Blumenkrantz 203a118eb4 change client un/maximize_pre callback location to occur after safety checks 2016-04-26 09:59:17 -04:00
Mike Blumenkrantz b4f664b7bc make e_client_has_xwindow() more wayland-only friendly 2016-04-26 09:59:17 -04:00
Mike Blumenkrantz 12d1d70ca4 reject client maximize attempts using identical maximize params 2016-04-26 09:56:31 -04:00
Mike Blumenkrantz 783c35622d only attempt to rescue offscreen clients if they are completely offscreen 2016-04-26 09:50:26 -04:00
Mike Blumenkrantz bdc8078f87 feed mouse out to internal clients upon activating a mouse binding
this prevents internal windows from effectively processing pending
mouse-up events

ref T3347
2016-04-26 09:49:04 -04:00
Mike Blumenkrantz 56e96ce294 do not set initial changed state for new clients if they are ignored
this loops the client idler for no reason
2016-04-26 09:48:22 -04:00
Mike Blumenkrantz 39717a8a38 feed mouse-up events for all buttons on internal wins when activating a binding
bindings enforce compositor grabs, which will result in stuck canvas buttons and
break internal windows which have already received button presses

fix T3347
2016-04-26 09:48:06 -04:00
Mike Blumenkrantz 8d626e490b do not attempt to resize clients to 0x0 during res restore
ACK--
2016-04-26 09:44:06 -04:00
Mike Blumenkrantz af46fe8190 unify client e_hints window size setting in move/resize callbacks 2016-04-26 09:38:08 -04:00
Mike Blumenkrantz f009e7c752 do not decrement e_comp->new_clients for non-new clients during _e_client_eval
this overflows the int!
2016-03-09 12:10:03 -05:00
Mike Blumenkrantz 1a71696c3a pre-remove pixmap from x11 clients during del hook
this breaks the compositor!

ref 0ca200513e
2016-03-09 12:09:27 -05:00
Derek Foreman 1a6dac4677 Move pixmap free to client free from client destroy
Under wayland we need the pixmaps to exist until after the close
animations finish, even if clients can perform new rendering.

@fix
2016-03-09 12:08:43 -05:00
Mike Blumenkrantz 69ebda1df8 allow instant client mouse actions to persist until mouse up
if an action triggers on a window, the triggering mouse event should
not be passed to the window. the only way to determine this is if the
action object lives through the entire event
2016-02-02 13:36:08 -05:00
Mike Blumenkrantz f4c3d3eae8 unify client mouse action ending
the code for this in most places was identical, but in some places it was
not, leading to behavior inconsistencies and potentially dangling objects
2016-02-02 13:35:55 -05:00
Mike Blumenkrantz ac0cc26765 allow more client hooks to bypass delete check
failing to propagate these signals can result in a broken compositor
when a client is deleted before/during an operation
2015-11-24 21:42:22 -05:00
Mike Blumenkrantz 00252fad74 add safety checks to e_client_focus/raise_latest_set() fns
passing a null client here breaks the universe, so ensure that it's
easily catchable
2015-11-19 12:44:35 -05:00
Mike Blumenkrantz 64ceff507f update E_Client->desk member upon focusing a sticky client
a sticky window previously would always have the desk set for where
it was set as sticky, meaning that anything which tries to access it
will be reading wrong data here.
more useful information to provide is the last desk which the sticky
client was focused on, so update that upon focusing it
2015-11-12 17:38:58 -05:00
Mike Blumenkrantz b3d68ee828 use unsigned int for client focus tracking freeze state
there's no reason why this should ever be negative
2015-11-10 14:32:53 -05:00
Mike Blumenkrantz cc137f6cd7 move client iconic hint setting into x11 hide function
e_hints should probably be merged into the x11 compositor entirely
at some point
2015-11-07 13:06:23 -05:00
Mike Blumenkrantz d47252be4a simplify client SMART/EXPAND maximize
calling e_client_resize_limit() here was misleading since the values
were never used
2015-10-30 15:03:29 -04:00
Mike Blumenkrantz b9e51e1c0d Revert "when maximizing a wayland client with csd, use window geometry"
This reverts commit 6e5e51c2b8.

not a complete solution, makes client code somewhat unreadable
2015-10-29 14:34:32 -04:00
Mike Blumenkrantz 6e5e51c2b8 when maximizing a wayland client with csd, use window geometry
wayland clients which have csd must be resized according to window geometry,
not client (surface) geometry. this is somewhat tricky to handle because x11
clients which have csd work the exact opposite way and must continue to be
managed using client geometry

this is not my ideal solution for this issue, but I can't think of a
better one at this time which fully fixes wayland client maximization
2015-10-23 12:59:45 -04:00
Mike Blumenkrantz 7016c59c6f reject frame theme changes for clients which have CSD 2015-10-23 12:31:24 -04:00
Chris Michael 59ffe3bed2 enlightenment: Add missing field initializers for Eina_Rectangle
Signed-off-by: Chris Michael <cp.michael@samsung.com>
2015-10-14 10:19:43 -04:00
Mike Blumenkrantz 45b7224b42 explicitly disable screensaver for wayland urgency wakeups 2015-10-09 15:23:15 -04:00
Mike Blumenkrantz 277eb14ace make wake-on-urgent apply for all cases of window urgency
in the case that an urgent window is set as urgent again, wakeup should
still occur
2015-10-09 15:21:41 -04:00
Tom Hacohen 981e3f8f31 Client: fix potential out of bounds read.
This is an unsigned int underflow that could lead to an out
of bounds read.

CID 1291837

@fix
2015-10-09 09:19:53 +01:00
Mike Blumenkrantz 95b9770f08 when removing VERTICAL maximize state, also remove LEFT and RIGHT states
LEFT and RIGHT are supersets of VERTICAL which serve no purpose on their own,
so remove them when removing VERTICAL

fix T2770
2015-10-07 15:47:21 -04:00
Mike Blumenkrantz 64c4adeca2 unset E_Client->changes.need_unmaximize after calling unmaximize
...just in case
2015-09-28 14:37:57 -04:00
Mike Blumenkrantz f0fc7a12de remove client maximize states during unmaximize only if they are active
do not accidentally remove untested states using bad math operations
2015-09-28 14:36:59 -04:00
Mike Blumenkrantz 8a764901c6 check warp_client existence during pointer warp before dereferencing it
ref T2679
2015-09-25 16:26:50 -04:00
Mike Blumenkrantz 2b38eb9712 block border-affecting client functions for csd clients 2015-09-24 16:33:56 -04:00
Mike Blumenkrantz c9c880ec59 improve client exe_inst creation
ref T2679
2015-09-22 16:00:09 -04:00
Mike Blumenkrantz ae919fe716 unset E_Client->want/take_focus flags during client eval
ref 4a33cd7ba0
2015-09-22 12:03:09 -04:00
Mike Blumenkrantz 4a33cd7ba0 only set focus on clients which are visible during client eval
focusing a client will automatically uniconify and desk flip, so
setting focus on a hidden client should be avoided during eval since
these focus-sets are not "user triggered"

this fixes issues where clients could randomly grab focus from other
desks and also restores expected behavior when restarting e on an
empty vdesk
2015-09-21 19:19:01 -04:00
Mike Blumenkrantz 7056d598c4 do not return non-visible clients for e_client_under_pointer_get()
a hidden client should not be considered "under the mouse" since its
geometry is undefined until it becomes visible again

fix T1069
2015-09-21 17:50:48 -04:00
Mike Blumenkrantz e8a9ba34b0 always send CLIENT_SHOW event when clients are shown
I think it was the case for <E19 that uniconifying a client would
not send a show event, but this is not ideal and so the event should
still be sent
2015-09-17 19:08:27 -04:00
Mike Blumenkrantz f5657fc6c1 enforce gathering client moveinfo before beginning move/resize action 2015-09-17 16:54:00 -04:00
Carsten Haitzler b73364269b e winlist focus alt-tabbing - fix focus list changes while tabbing
this should fix T2700

@fix
2015-09-12 13:38:38 +09:00
Mike Blumenkrantz 8946e3e504 unshade clients when activating
fixes issue where an activated client on another vdesk which was shaded
would appear to be unshaded but was, in fact, shaded
2015-09-08 17:01:36 -04:00
Mike Blumenkrantz 5bd8706bac force-disable refocus during desk flip on client activation
this prevents an infinite focus loop where focus will be constantly
reapplied between multiple windows if the activated window is not the
refocus window
2015-09-08 16:55:21 -04:00
Mike Blumenkrantz fb79fff70d don't set E_Client->changes.visible on successful show/hide
iirc this was a fix for something else that was going on early in
e19 development, now it creates issues when attempting to hide clients
2015-08-19 14:00:52 -04:00
Mike Blumenkrantz 0730e3853a unset E_Client->want/take_focus flags in hide callback
deferred focus should no longer be valid if a client has been hidden
before the focus-set could be triggered

fixes super fun infinite loop with desk flips
2015-08-18 18:18:12 -04:00
Mike Blumenkrantz 68da56b430 do not modify E_Client->visible value in show/hide callbacks 2015-08-18 18:17:51 -04:00
Mike Blumenkrantz 442d63d8d6 track button mask for all drags, only end drag when mask has depopulated
previously, beginning a drag with the left button, then pressing and
releasing another button would result in the drag terminating without
the original button being released
2015-08-12 15:39:42 -04:00