Mike Blumenkrantz [Mon, 27 Apr 2015 18:36:42 +0000 (14:36 -0400)]
fix evry mouse event detection on wayland
Mike Blumenkrantz [Mon, 27 Apr 2015 18:32:19 +0000 (14:32 -0400)]
more accurately determine whether a mouse grab is active in some parts of e
Mike Blumenkrantz [Mon, 27 Apr 2015 18:32:05 +0000 (14:32 -0400)]
add comp mouse grab detection function
Mike Blumenkrantz [Mon, 27 Apr 2015 18:31:52 +0000 (14:31 -0400)]
add e_grabinput_mouse_win_get()
Mike Blumenkrantz [Mon, 27 Apr 2015 18:26:33 +0000 (14:26 -0400)]
only allow efm key actions if key grab does not exist
Mike Blumenkrantz [Mon, 27 Apr 2015 18:21:33 +0000 (14:21 -0400)]
add comp util function for determining whether the keyboard has been grabbed
Mike Blumenkrantz [Mon, 27 Apr 2015 18:11:05 +0000 (14:11 -0400)]
fix MANAGER context keybinds to work with rootless compositors
Mike Blumenkrantz [Mon, 27 Apr 2015 17:56:56 +0000 (13:56 -0400)]
add internal win trap to set override property on windows
Mike Blumenkrantz [Mon, 27 Apr 2015 17:55:47 +0000 (13:55 -0400)]
set internal win icccm name/class/title during win trap show
wl clients have no concept of icccm name, so this is the only way to set it
also, setting it here makes it available sooner for other functionalities
Mike Blumenkrantz [Mon, 27 Apr 2015 17:20:52 +0000 (13:20 -0400)]
set focus on comp autoclose objects
fix T2114
Mike Blumenkrantz [Mon, 27 Apr 2015 17:04:53 +0000 (13:04 -0400)]
do not pass key events through autoclose handler
Mike Blumenkrantz [Mon, 27 Apr 2015 16:24:59 +0000 (12:24 -0400)]
ensure client max w/h props are set before applying them to client geom
Carsten Haitzler (Rasterman) [Mon, 27 Apr 2015 07:54:15 +0000 (16:54 +0900)]
e auto-placement - don't include shelves from other screens
unbeknownst to anyone e was placing around shelvs on OTHER screens...
this ended up with some .. odd results every now and again.
@fix
maxerba [Sun, 26 Apr 2015 18:52:30 +0000 (20:52 +0200)]
Updating italian translation
Mike Blumenkrantz [Fri, 24 Apr 2015 18:58:00 +0000 (14:58 -0400)]
ensure evry key events do not get erroneously rejected in grab mode
ref T2030
Mike Blumenkrantz [Fri, 24 Apr 2015 18:57:05 +0000 (14:57 -0400)]
don't set e_comp->root for wl compositors
root is the root window. wl compositors do not have root windows.
Mike Blumenkrantz [Fri, 24 Apr 2015 18:10:48 +0000 (14:10 -0400)]
fix comp init failure case
Mike Blumenkrantz [Fri, 24 Apr 2015 18:10:25 +0000 (14:10 -0400)]
fix comp fake layer init for wl compositors
fix T2363
Mike Blumenkrantz [Thu, 23 Apr 2015 17:47:05 +0000 (13:47 -0400)]
remove rest of ecore-drm usage from wl-drm randr info grabbing
this was not an optimal solution since only connected outputs are
tracked/available to ecore-drm, so now the module uses libdrm for everything
Mike Blumenkrantz [Thu, 23 Apr 2015 17:37:11 +0000 (13:37 -0400)]
fix wl-drm install rule
Mike Blumenkrantz [Thu, 23 Apr 2015 16:35:22 +0000 (12:35 -0400)]
fix internal win pixmap id int size in show trap
Mike Blumenkrantz [Thu, 23 Apr 2015 12:42:09 +0000 (08:42 -0400)]
fix tasks gadget parenting
gc->o_container is NULL for desktop gadgets
ref T2336
Mike Blumenkrantz [Thu, 23 Apr 2015 12:33:20 +0000 (08:33 -0400)]
fix typo when setting event state in e_randr2_screen_refresh_queue()
a non-event should not unset this variable
Mike Blumenkrantz [Thu, 23 Apr 2015 12:27:45 +0000 (08:27 -0400)]
simplify comp_x screen setup
we only care about the first root now, no point getting all the roots
Mike Blumenkrantz [Thu, 23 Apr 2015 12:23:13 +0000 (08:23 -0400)]
check module name existence before reading from it
CID 1295427
Mike Blumenkrantz [Thu, 23 Apr 2015 01:15:12 +0000 (21:15 -0400)]
add read-only randr iface for wl drm output module
mostly just reads stuff right out of libdrm for now. seems to display mostly as expected.
no applying yet, and only connected+active monitors will display since that's all ecore-drm
tracks for now
try using ecore-drm, they said. it'll be easier than using libdrm, they said.
Mike Blumenkrantz [Thu, 23 Apr 2015 01:04:38 +0000 (21:04 -0400)]
add wl drm module phony make targets
Mike Blumenkrantz [Wed, 22 Apr 2015 23:36:21 +0000 (19:36 -0400)]
use const lists for drm devices in signal handlers
Mike Blumenkrantz [Wed, 22 Apr 2015 23:27:36 +0000 (19:27 -0400)]
move canvas init after wl init in output modules
ensure that randr screens have been created
Mike Blumenkrantz [Wed, 22 Apr 2015 23:24:33 +0000 (19:24 -0400)]
move comp_x screen setup code into e_randr
this was not x-specific, so move it for use in wl compositors
Mike Blumenkrantz [Wed, 22 Apr 2015 23:16:32 +0000 (19:16 -0400)]
resize comp canvas on e_comp_canvas_init()
remove unnecessary call in wl drm output module
Mike Blumenkrantz [Wed, 22 Apr 2015 23:09:19 +0000 (19:09 -0400)]
fix module blocking for wl modules
these should never be saved to or loaded from the config
Mike Blumenkrantz [Wed, 22 Apr 2015 22:28:33 +0000 (18:28 -0400)]
enable conf_randr module unconditionally by default
this no longer uses/requires ecore-x
Mike Blumenkrantz [Wed, 22 Apr 2015 22:28:23 +0000 (18:28 -0400)]
remove dead conf_randr files
Mike Blumenkrantz [Wed, 22 Apr 2015 22:20:47 +0000 (18:20 -0400)]
init randr during wayland init...always
Mike Blumenkrantz [Wed, 22 Apr 2015 22:19:52 +0000 (18:19 -0400)]
change comp type check at start of e_comp_wl_init()
at present, this value will be set to either TYPE_NONE if no valid
compositor has been created or TYPE_X if running a wl compositor inside
an x11 compositor
Mike Blumenkrantz [Wed, 22 Apr 2015 22:18:39 +0000 (18:18 -0400)]
use randr for screen setup in wl-x11 output module
enabling this module requires that e is linked against ecore-x to
allow screen management
Mike Blumenkrantz [Wed, 22 Apr 2015 22:18:01 +0000 (18:18 -0400)]
always add e_randr2.c to e build
randr no longer requires x11
Mike Blumenkrantz [Wed, 22 Apr 2015 22:14:55 +0000 (18:14 -0400)]
add a small screen management interface to E_Comp, remove hardcoded x11 randr calls
this interface is extremely basic and should allow easy additions for other display
backends
Mike Blumenkrantz [Wed, 22 Apr 2015 22:12:22 +0000 (18:12 -0400)]
create wl-x11 canvas using comp_x canvas creation method
Mike Blumenkrantz [Wed, 22 Apr 2015 22:09:35 +0000 (18:09 -0400)]
break out comp_x gl/sw canvas creation into external function
Mike Blumenkrantz [Wed, 22 Apr 2015 22:06:26 +0000 (18:06 -0400)]
move comp_x randr init to screen init function
Mike Blumenkrantz [Wed, 22 Apr 2015 21:46:43 +0000 (17:46 -0400)]
ensure valid strings are send when sending wl output geometry
Mike Blumenkrantz [Wed, 22 Apr 2015 21:30:44 +0000 (17:30 -0400)]
fix wl-x11 configure detection/check
previous check would not allow compiling against ecore-x which would
prevent such a compositor from being able to manage its own xserver
Mike Blumenkrantz [Wed, 22 Apr 2015 21:29:53 +0000 (17:29 -0400)]
fix ecore-x cflags/libs inclusion for e binary
these variables will always exist
Mike Blumenkrantz [Wed, 22 Apr 2015 21:01:53 +0000 (17:01 -0400)]
always register E_EVENT_RANDR_CHANGE
failing to register this will cause event handlers for it to fail
on creation
Mike Blumenkrantz [Wed, 22 Apr 2015 20:58:00 +0000 (16:58 -0400)]
create e_comp during e_comp_init, set comp_type automatically
this simplifies comp creation slightly
Mike Blumenkrantz [Wed, 22 Apr 2015 20:57:47 +0000 (16:57 -0400)]
add E_PIXMAP_TYPE_NONE
null type currently only used during comp init
Mike Blumenkrantz [Wed, 22 Apr 2015 20:16:05 +0000 (16:16 -0400)]
move all x11-related randr code into e_comp_x_randr.c
Mike Blumenkrantz [Wed, 22 Apr 2015 20:07:58 +0000 (16:07 -0400)]
_config_screen_find() -> e_randr2_config_screen_find()
Mike Blumenkrantz [Wed, 22 Apr 2015 19:56:01 +0000 (15:56 -0400)]
move acpi lid status tracking from e_randr to e_acpi
Mike Blumenkrantz [Wed, 22 Apr 2015 20:01:48 +0000 (16:01 -0400)]
move e_randr internal screen changing timer -> e_randr2_screen_refresh_queue()
Mike Blumenkrantz [Wed, 22 Apr 2015 19:50:05 +0000 (15:50 -0400)]
add e_acpi_lid_is_closed() for returning cached state of computer lid
Mike Blumenkrantz [Wed, 22 Apr 2015 18:54:49 +0000 (14:54 -0400)]
update wl shell parent setting code to use clients instead of pixmaps
ref T2354
Chris Michael [Wed, 22 Apr 2015 14:29:42 +0000 (10:29 -0400)]
e-comp-wl: Init randr if we are not wayland only
This is needed for the wl-x11 module
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Chris Michael [Wed, 22 Apr 2015 14:42:33 +0000 (10:42 -0400)]
wl-drm: We need to set the compositor pointer values After wl_comp_data has been created else we crash
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Chris Michael [Wed, 22 Apr 2015 14:41:13 +0000 (10:41 -0400)]
Revert "wl-drm: Set initial compositor pointer values base on ecore_evas_pointer_position"
This reverts commit
16e9126302c3a95f84921b62dbe547b1c112cd2f.
Reverting this due to crash
Chris Michael [Wed, 22 Apr 2015 13:41:23 +0000 (09:41 -0400)]
wl-drm: Set initial compositor pointer values base on ecore_evas_pointer_position
This addresses the issue of initial compositor pointer values not
being proper. They are 0,0 on startup, however things like ecore_drm
will center the mouse on startup. The e_comp pointer values should
reflect this.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Stefan Schmidt [Wed, 22 Apr 2015 11:54:51 +0000 (13:54 +0200)]
module/geolocation: Stop setting gadget icon from start and stop callbacks
We keep track of this now by monitoring the InUse property which works also
if other clients use the service.
Stefan Schmidt [Wed, 22 Apr 2015 11:54:51 +0000 (13:54 +0200)]
modules/geolocation: Follow InUse property change to give UI indication
The InUse property of the manager object is available in newer GeoClue2
versions. It shows if the service is in use by any application at the
moment. We use this to indicate the usage in the UI.
A blue geolocation icon means the service is in use, the grey icon means
it is inactive.
Stefan Schmidt [Wed, 22 Apr 2015 11:54:51 +0000 (13:54 +0200)]
modules/geolocation: Make module work with newer GeoClue2
In newer GeoClue2 version, tested with 2.1.10, we need to set
the DesktopId and AccurancyLevel to get a location object.
Stefan Schmidt [Wed, 22 Apr 2015 11:54:51 +0000 (13:54 +0200)]
modules/geolocation: Only display a valid altitude value
If our backends does no know the altitude it sets it to -DBL_MAX as 0
is a valid altitude value. Check for it and display N/A if it is not valid.
Carsten Haitzler (Rasterman) [Wed, 22 Apr 2015 08:03:44 +0000 (17:03 +0900)]
e screenlock config diloag - note insecureness for personal pw/pin
these store pin/pw in your user config files - it may be primitively
hashed to obscure it, but it's there. it never pretended to have
secure storage and even saved cleartext until e19. make sure people
are aware
Stefan Schmidt [Wed, 22 Apr 2015 07:43:16 +0000 (09:43 +0200)]
po: Remove unused e_int_config_randr.c file from POTFILES.in to pass distcheck
This file is still around while not really used. It is also not distributed and
thus fails if po tries to update during distcheck with this file still in list.
Mike Blumenkrantz [Tue, 21 Apr 2015 20:19:30 +0000 (16:19 -0400)]
unadjust frame coords when calculating wl surface placement during commit
ec->y -> unadjust -> ec->client.y
Mike Blumenkrantz [Tue, 21 Apr 2015 19:57:18 +0000 (15:57 -0400)]
use pointer hash for wl pixmaps
Mike Blumenkrantz [Tue, 21 Apr 2015 19:41:10 +0000 (15:41 -0400)]
pass surface resource when creating wl surface pixmaps
Mike Blumenkrantz [Tue, 21 Apr 2015 19:31:21 +0000 (15:31 -0400)]
Revert "build: Remove generated eldbus files during normal clean and not only maintainer-clean"
This reverts commit
95d701207783bcf58034797468024c05fa59ca03.
do want in tree
Mike Blumenkrantz [Tue, 21 Apr 2015 19:31:04 +0000 (15:31 -0400)]
add explicit includes for elm color class header in colorclass cfd
Mike Blumenkrantz [Tue, 21 Apr 2015 19:21:19 +0000 (15:21 -0400)]
disable client color editor for non-X clients for now
Mike Blumenkrantz [Tue, 21 Apr 2015 19:19:20 +0000 (15:19 -0400)]
wl client pixmaps now use a different method for creating the pixmap id
Mike Blumenkrantz [Tue, 21 Apr 2015 19:17:56 +0000 (15:17 -0400)]
add wl client event callbacks at lower priority
this ensures that bindings will activate before events are passed to clients
Mike Blumenkrantz [Tue, 21 Apr 2015 19:05:25 +0000 (15:05 -0400)]
use client geom for comp object centering functions if ec->new_client is set
Chris Michael [Tue, 21 Apr 2015 13:36:00 +0000 (09:36 -0400)]
e-comp-wl: Don't crash if we fail to create a new e_client
This "should" fix T2344 crashing, however I suppose a better question
is ... why is e_client_new returning NULL when it Does find the client
already in the hash...
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Conrad Meyer [Tue, 21 Apr 2015 12:11:56 +0000 (13:11 +0100)]
e_auth: Add suid helper for lokker own-pw checking on FreeBSD
Summary:
PAM on FreeBSD, unlike on Linux, does not allow users to check their own
password. Instead, we need a suid helper to do it for us. Add such a
helper on FreeBSD.
For now, it is limited to checking users in the local password database
(traditional Unix passwd file). This could and should be extended to use
PAM in a later patch.
Test Plan:
Tested empty pw, wrong pw, correct pw at lock screen; observed correct behavior
in each instance.
Reviewers: q66, zmike
Reviewed By: q66, zmike
Subscribers: cedric, seoz
Differential Revision: https://phab.enlightenment.org/D2355
Tom Hacohen [Tue, 21 Apr 2015 09:45:01 +0000 (10:45 +0100)]
Auth: try harder when clearing passwords.
Based on http://www.daemonology.net/blog/2014-09-04-how-to-zero-a-buffer.html
Thanks again to Matthew Garrett, for pointing this out on twitter.
Tom Hacohen [Tue, 21 Apr 2015 09:29:58 +0000 (10:29 +0100)]
Module teamwork: Stop using a deprecated function.
Tom Hacohen [Tue, 21 Apr 2015 09:07:42 +0000 (10:07 +0100)]
E auth: improve clearing out passwords from memory.
Optimising compilers (like gcc/clang with -O1 or above) were optimising
out the memset(). Until link time optimisations are good enough, this
will prevent them from doing so. The best solution would be to use
memset_s() (c11), though it's not readily available yet. This is the
first step towards using memset_s() with a fallback for systems who
don't have it. A better solution, is to put it in Eina, to prevent LTO
completely. This will have to be done after the EFL release.
Even this is not entirely safe though, but at least it protects us from
some memory disclosure issues.
This doesn't solve the fact that we may store a copy of the password in
other places, like the input system. We need to address that too.
Thanks to Matthew Garrett for pointing this out or Twitter.
Carsten Haitzler (Rasterman) [Tue, 21 Apr 2015 02:00:24 +0000 (11:00 +0900)]
e - e_client - fix added printf
665caad451c7184bb85b85067c491d15d3dfc1ea fix added a printf i didn't
notice - fix.
Carsten Haitzler (Rasterman) [Tue, 21 Apr 2015 00:36:33 +0000 (09:36 +0900)]
e client - desk wait - ref it properly and clean up delfn
delfn's on desks being referred to as profile.wait_desk were never
cleared when client was freed. als the desk itself never got an obj
reference - fix that and ref/unref it.
Carsten Haitzler (Rasterman) [Mon, 20 Apr 2015 23:52:19 +0000 (08:52 +0900)]
e - xsettings - calloc buffer since we leave 3 of 4 bytes unset
the first 4 bytes is a msb/lsb thing for xsetttings - we ignore 3 of
the 4 bytes. thus we should just calloc the whole buffer to ensure its 0
Derek Foreman [Mon, 20 Apr 2015 22:17:55 +0000 (18:17 -0400)]
Fix wayland compositor on 32 bit systems
Summary:
_e_elm_win_trap_show() treats Ecore_Window as a 64bit data type when
it's actually the same size as the system's pointer type.
#24HourFullPowerNoToiletNoShower
Reviewers: devilhorns, zmike
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2388
fix T2194
Mike Blumenkrantz [Mon, 20 Apr 2015 22:08:07 +0000 (18:08 -0400)]
fix submenu visibility during scroll
sometimes while quickly moving the cursor through menu items with submenus
the visibility flag would toggle more quickly than the menu object's visibility.
this is no longer the case
Mike Blumenkrantz [Mon, 20 Apr 2015 21:24:48 +0000 (17:24 -0400)]
attempt to decode notification image uris with efreet before showing them
fix T1939
Jean Guyomarc'h [Mon, 20 Apr 2015 21:03:22 +0000 (17:03 -0400)]
e_backlight_main: fix compiling when not having Eeze
Summary:
EINA_UNUSED is defined in Eina. Since we are not using the program at all,
let's void main()'s arguments instead of trying to ignore them.
Test Plan: tested only on OSX, unbreaks the build.
Reviewers: zmike
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2387
Mike Blumenkrantz [Mon, 20 Apr 2015 20:04:48 +0000 (16:04 -0400)]
clear app menu cache any time a desktop is edited internally
ref T2271
Mike Blumenkrantz [Mon, 20 Apr 2015 20:04:31 +0000 (16:04 -0400)]
add function for clearing app menu cache
ref T2271
Mike Blumenkrantz [Mon, 20 Apr 2015 19:31:41 +0000 (15:31 -0400)]
swap key -> keyname when attempting to bind lparen/rparen keys
this is impossible to grab on X, so ensure the generated binding is usable
fix T2059
Mike Blumenkrantz [Mon, 20 Apr 2015 19:17:05 +0000 (15:17 -0400)]
fix clock popup widget parenting
Mike Blumenkrantz [Mon, 20 Apr 2015 18:31:27 +0000 (14:31 -0400)]
use client geometry for starting unmaximize calculations
the resize operations in this function operate off client geometry, so
stop using frame geometry
fix T2170
Chris Michael [Mon, 20 Apr 2015 15:44:33 +0000 (11:44 -0400)]
wl-drm: Use ecore_evas_new with an engine name rather than direct ecore_evas_drm_new calls
Signed-off-by: Chris Michael <cp.michael@samsung.com>
maxerba [Mon, 20 Apr 2015 14:23:33 +0000 (16:23 +0200)]
Make some randr messages translatable
Simon Lees [Mon, 20 Apr 2015 04:53:00 +0000 (13:53 +0900)]
Wizard: Update icon theme search list
Summary:
@fix The list of icon themes in the search list is quite out of date
it doesn't include the icon themes used by gnome3 or kde4/5 this means
for a lot of users at first boot e has alot of missing icons.
The Icon sets that have been added are as follows
"Oxygen", /* KDE 4 */
"Adwaita", /* Gnome 3 */
"Breeze", /* KDE 5 */
"HighContrast"
This change will cause T1732 and T1923 to not occur for almost all users
but it does not fix them properly the code should be modified to fall back
to pick up icons from fallback locations such as the hicolor theme if no
icon theme is set.
It would be nice if someone could backport this to e19 I don't know how to
Reviewers: bu5hm4n, zmike, raster
Reviewed By: raster
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2377
Jean-Philippe ANDRE [Sat, 18 Apr 2015 08:19:00 +0000 (17:19 +0900)]
PackageKit: Fix a crash properly
This reverts and tries to actually fix:
commit
b30892056e0a527545c8a022cd2b3cdef0ec85af
Author: Jean-Philippe Andre <jp.andre@samsung.com>
Date: Fri Apr 17 10:30:06 2015 +0900
PackageKit: Fix crash on error
Sometimes a "normal" error occurs and then the callback is
called with no context data. Then crash boom sigsegv happens.
Mike Blumenkrantz [Fri, 17 Apr 2015 18:12:27 +0000 (14:12 -0400)]
remove support for wayland builds on 32bit
this doesn't work for some reason and there are no plans to fix it anytime soon.
ref some infinite number of tickets
Jean-Philippe Andre [Fri, 17 Apr 2015 01:30:06 +0000 (10:30 +0900)]
PackageKit: Fix crash on error
Sometimes a "normal" error occurs and then the callback is
called with no context data. Then crash boom sigsegv happens.
Mike Blumenkrantz [Thu, 16 Apr 2015 20:02:29 +0000 (16:02 -0400)]
fix unoptimized stringshare usage in keybindings cfd
Mike Blumenkrantz [Thu, 16 Apr 2015 19:16:50 +0000 (15:16 -0400)]
prune no_display desktops from evry apps plugin
probably fix T2256
Mike Blumenkrantz [Thu, 16 Apr 2015 19:04:42 +0000 (15:04 -0400)]
fix evry app listing list leak
Mike Blumenkrantz [Thu, 16 Apr 2015 18:34:54 +0000 (14:34 -0400)]
fix pixmap image clearing based on pixmap type
!@%!@$#!# borkerhorns!!!!!!
invalidates T2291