i got a whole mountain of eo invalid obj complaints from deskmirror
even on startup. the backtraces were long, but they all ended in
comp_object being invalid or null. it seesm deskmiror wasnt properly
tracking the deletion of comp_object outside and that led to this. i
simply set it up once and deleted it where it is no longer referenced
and all is good now. this may possibly in theory have led to odd bugs
but thanks to eo - unlikely.
@fix
Summary:
Moving _xdg_data_dirs_augment call to earlier in
initalisation process. Currently first-run will
leave XDG_MENU_PREFIX broken, thus the application
menu will be empty in E.
Seems eio_init might be the culprit.
Moved to above eio_init.
Test Plan:
* Launch E from X from a console (startx)
* Check application menu (empty).
* Apply patch.
* Launch E from X from a console (startx)
* E's Application menu should be populated.
Related to https://phab.enlightenment.org/D7534
Reviewers: #committers, raster, cedric, zmike, devilhorns
Reviewed By: #committers, zmike
Tags: #enlightenment-git
Differential Revision: https://phab.enlightenment.org/D7535
Wayland mouse acceleration uses different values (via libinput) than
X11 does. As such, we need to check the compositor type when creating
the Mouse Input config dialog and adjust values accordingly.
ref T7534
@fix
it wasn't working. first response may not be a 200 ident so keep
looking for them. also send a bit more than 1 byte to be sure, and
chekc the response is what we sent to be sure. also enforce a timeout
(10sec here) where we give up so it doesn't hang possibly forever.
all in all l2ping in e_sys works again. now. in the process i added a
timeout param too.
@fix
we had multiple drag resistance values here - unify with 1 and make it
work, now it gets it right deciding between dnd and a long press menu
etc. etc. ...
while also still using the new menu system.
This binding check/action handler was removed, because I am assuming
that the presumption was that the only binding/action that was handled
was showing a menu? This was actually not the case. This
binding/action handler handled resizing, moving, dragging, etc... on
gadgets on the desktop and in the gadget bar. Without it using gadgets
was near impossible on the desktop and inconvenient on the gadget bar.
The longpress menu was only getting cancelled if a drag occured at a
distance of least 25 pixels. This is due to the code checking for horizontal drag
distance + vertical drag distance >= 25. I believe the intent here was
to cancel drag if >= 5 vertically or >= 5 horizontally, not both. Most
drags wouldn't be 25 pixels in a single gadget such as pager, and a 25
pixel drag would not happen quick enough to offset longpress. This
commit also lowers the drag cancelling threshold to 3 pixels, not 5.
hold mode is where if u hold mosue down menu stays only as long as
mouse is held down... then dismissed on up. doesnt work well when youa
re trying to overload a single click with longpresses and so on -
optionally turn it off. used in gadgets.
now right click on any gagdte in bryce and they ALL have a menu that
allows removal of the gagdte bar or the gadget as well as access to
gadget settings AND the ability for gadgets to extendthis menu like
lunhcer does per icon. now it's standard behavior everywhere which is
much easier to use and discover. it also removes code from every
gadget to do their own "button 3" handling as its handled centrally
making the code in gadgets simpler.
this is part of my effort to improve usability (mostly discoverability
and accessibility of settings/features).
also long press left mouse gets u gadget right click menu
this has to move many modules/gadgets actions to mouse up instead of
mouse down so the bryce has a chance to trap the events first and set
hold flags. but now long press for 0.5 sec and bryce menu come sup
(with left mouse .. so touch friendly).
in addition move context menu hanbdling to e_gadget instead of in
bryce and in e_gadget. a context callback is called so different
systems can still do different things. this should probably change to
always pop up a mnenu and simple call populate callbacks for site
owner specific content.
all in all it makes the new gagdtes more consistent, easier to use
(without a right mouse button), doesn't need special action bindings
etc. etc.
bryce was missing the ability to espose orientation to the child
items, so they were the same irrespective of orientation.
i also notice that orientaiont is a simple horiz/vert ... so no
ability to special case corners etc. in the theme... :( not sure if
this should be changed.
also fix the aspect calculation to round up correctly to avoid
off-by-1 pixel gaps i noticed with the pager - necessary for the
styling in the flat theme to be right.
but is not explicitly linking against it. Previously this was not
discovered due to a wrong flag in elementaries pkgconfig. However - the
new .pc file of elementary does not contain dl anymore (as no library
_needs_ to link against it when using elm). So we need to link this here
explicitly
Differential Revision: https://phab.enlightenment.org/D7150
Summary:
Prevent wayland clients from being able to destroy the compositor's
singleton keymap by making individual copies for each client.
Reviewers: zmike, devilhorns
Reviewed By: zmike, devilhorns
Subscribers: cedric
Tags: #enlightenment-git
Differential Revision: https://phab.enlightenment.org/D6861
Summary:
evas functions
This fixes an issue when quickly mousing through menus can cause a
segfault in Enlightenment due to menu->comp_object being NULL
ref T7030
Reviewers: zmike
Reviewed By: zmike
Subscribers: cedric
Tags: #efl, #enlightenment22
Maniphest Tasks: T7030
Differential Revision: https://phab.enlightenment.org/D6641
elm wont switch profile on the fly if ELM_PROFILE is set this is
considered a custom override thus it stops working... so don't set it
- the elm config files should contain the right profiles to use.
follow on form 4c7b798b45 - really
remove from the alias hash. the alias id is different and should ave
been stored in the pixmap and be deleted when pixmap is freed. i had
it right to remove from the aliases hash too, but i used the wrong id
- i used the "core" pixmap id, not alias. this tracks and uses that
alias.
this means internal windows are reliable now and dont crash...
@fix.
on pixmap free only the pixmaps entry was deleted not the pixmaps hash
one. this led to lookup of stale pixmaps in the aliases hash... and
then a crash.
also use the correct local type with the correct byte order as well.
this has probably been an issue for a while but now internal windows
should work much better without crashes.
@fix
so a new bug in a gpu driver. if dpms is enabled, wakeup doesn't
happen. mouse doesn't do anything, rendering doesn't happen... i can
use the keyboard and ctrl+alt+enmd restart e or killall -HUP
enlightenment and e restarts and renders ... but nothing appears.
interestingly if i let it timeout again and wake it up a second
time... things render (but e is confused it seems and mouse input
doesnt work until i restart e). it's some kind of xorg/driver bug here
with this dpms - no dpms and all is fine. all e does with regards to
dplms is enable or disbale it (and set the timeouts) so e isn't doing
antyhign special otherwise with dpms on vs off ... so somethnig deeper
down the stack here, but to get a desktop that works at least for now,
add an option to not use dpms.
Summary:
in the case where deletion is deferred to preserve a window animation this
codepath may be triggered by a deleted client, at which point no render update
should occur in order to avoid compositor errors
ref f78eb3c108
fix T5203
Reviewers: ManMower, devilhorns
Reviewed By: devilhorns
Subscribers: netstar, cedric
Tags: #enlightenment-git
Maniphest Tasks: T5203
Differential Revision: https://phab.enlightenment.org/D6367
Summary:
these are all harmless but will trigger error messages from efl
ref T7030
Depends on D6315
Reviewers: ManMower, devilhorns
Reviewed By: devilhorns
Subscribers: cedric
Tags: #enlightenment-git
Maniphest Tasks: T7030
Differential Revision: https://phab.enlightenment.org/D6316
e was not properly handling the opacity hint in its 0-0xffffffff
range. in one case it converted e's color value to this range but just
with << 24 which is wrong as it then ignors the next 24 lower value
bits, so it should fill the next 3 bytes with repeats of the same
value to do this right, but far worse is on READING the value it just
used the value as-is as if it were a 0-ff (0-255) alpha value that we
use in evas and didnt "thunk it down" with val >> 24. this resulted in
renoise menus being blank as renoise set the opacity value on its
menu windows and e happily made them transparent thanks to this.
this fixes that.
not to peolpe fro the above. bitshifting DOWN is ok, but bitshifting
UP leaves the lower bits all 0 and you should fill this range with
repetitions of the value to properly scale in integert space with
bitshifts. :)
@fix
The symbol table fix on Linux doesn't translate well on BSD.
Adding code to use the older behaviour with the BSD systems
and retaining the new preferred behaviour when using dlopen(3)
on Linux.
so bluez itself wont go rfkill block or unblock the bt device. you
have to rfkill unblock separately if it's blocked. this is needed to
make bluez at a minimum work right. possibly should be added to
connman too. but this is the first step.
so i spent a few days lopening and closing the lids of a few laptops,
plugging and unplugging external screens in, plugging and unplugging
ac power and doing lots of combinations of these. this led to a whole
slew of interrealted issues that were pretty hard to detangle into
just single issues, so this is all one blob.
this fixes:
1. if a lid close gets a monitor unplug from x or e's randr wants to
unplug then this lead to slow unsuspend or lid open times as e was
reconfirguring the screens entireluy on lid open. dont do that. just
keep things as is, unless we have an external display, then reconfigure.
2. we had 2 systems monitoring for a wake. a poller and a systemd
event. this was redundant and lead to interesting things in the debug
output, so clean that up and on systemd systsems use the systemd wake
events
3. some systems dont shut down their screens on lid close, so they
stay open until screensaver timeouts kick in. bad. so also force the
screen to go off on lid close (if the lid screen is the only one we
have).
4. x seems to have a bug where if you force dpms off etc. to turn the
screen on, it still thinks it's of and wont dpms off again after that
until you eother give some input to have the wake event make the dpms
system in x think its now time to go on, or you toggle dpms on and
off, so i found toggling tricks x into thinking the right thing.
this makes some debugging also be consistent with printfs.
all in all i have a pretty well supported laptop doing swimmingly with
e and a not so well designed (acpi dsdt - missing many events like
acpi battery ones, ac power change acpi events, missing logic to power
off a closed screen in firmware or hardware and leaving it to sw...
not this laptop has a tocuh panel and extra fun points awarded since
the touch panel doesnt shut off on lid close... AND it reprots touch
events when it closes as it touches the keys/case... hooray! that has
another work-around local to my system, as efl has no mechanism to do
this).
@fix
Summary:
bluez4 support is now basically dead. nothing ships it anymore. bluez5
is a new api that is rather different so new code. also a new gui with
more complete features etc.
not everything is done as i'd like. need:
1. many more icons for device types (60-70 maybe?)
2. a few specific custom icons for some action buttons (like
pair/unpair)
3. icons for group headers
4. gadget status - the gagdte itself displays zero status. it's a
button to display a popup. that's all.
Reviewers: zmike!
Subscribers: devilhorns, cedric
Tags: #enlightenment-git
Differential Revision: https://phab.enlightenment.org/D6148
Under wayland, if a client declares itself iconified before its first
commit, we should wait until it's no longer iconified before sending it
a configure. This way it can avoid uselessly rendering buffers that
will never be displayed.
ref T6834
this fixes when you drag a wqindow to screen edge and it flips if you
use alt+drag as opposed to just dragging titlebar around without alt.
it fixes it by updating your config to allow "any modifier" for edge
flips which is actually correct.
@fix
handle quotes, escapes and so on for the shell command so you can do
things like Exec="ls" and it still work. this will also allow
executables with spaces in them as well with quotes or escapes as
backslash.
this fixes T6787
@fix
is e crashes, catches it and restarts while locked you end up
unlocked. this lets enlightenment_start know this lock down state and
it sets an env var to ensure locking happens on restart after recovery.
Turns out this is called too late in the render pass and a single frame
can occur with the old border setting. This can lead to nasty flicker
artifacts.
this moves to the newer mimeapps.list file in ~/.config from the older
defaults.list. it also provides a migration (copy file ovr if target
doesnt exist on e startup).
this fixes T6784
@fix
if no pam support detected at build time, dont even lock to begin
with. this should fix T6779
yes - even better is making pam a requirement for building and to be
optionallhy disabled, but then even if disabled we still need this check.
this has been a long long long standing bug in e. i think this is why
we had some reports of "any password unlocks e" - because it was built
without pam support and before auth was always succeeding without pam.
now it is always FAILING instead (which is probably better).
@fix
this should fix T6211 ensuring no drivers can cause a segfault at exit
time. this also happens to remove the enlightenment_sys -z option for
openbsd and unifies all the passwd checking into the single
enlightenment_ckpasswd binary util (that has ifdefs for openbsd,
freebsd and linux pam in it).
this simplifies code removing a mess of auth being done in multiple
places, removes special fork vs run 1 exe or a different exe in
different cases making it more maintainable. yes - this requires
enlightenment_ckpasswd to be setuid root, but it already was when it
was built.
@fix
If you have edje icons installed in your system (efreet smartly searches
the candidate icon based in your Icon= entry), they works correctly in
the menus, bar, etc... but doesn't works in the window borders icon,
this is because the code try's to add it without dealing with the .edj
extensions, probably attempting to load multiple image types and
returning some errors/warning, resulting in an empty icon in the window
borders
Fixes T6723
for objects without animations, it's important to instantly perform
state transitions in order to preserve intended visibility behaviors of
these objects and avoid race conditions due to asynchronous signal
processing
fix T6722
in the case where a window does not place with the smart placement algorithm,
clamp position to the top-left available position within the useful geometry
so that the titlebar will always be visible
this seems to happen to @manmower in T6619. it may just be timing
changes that changed the start loop time to be at init of ecore maybe ...
but it was doing a forced hide and not a clean "done"
that animates/fades. so fix this along with deferring show until the
first render pre.
this fixes T6619
a remaining client from before a restart which is maximized does not need its
position adjusted; this will cause it to go offscreen when unmaximized
ref T6414
Summary:
Drop deprecated Encoding key from desktop files
The Encoding key is no longer required, all desktop files are assumed to
be UTF-8 encoded. See details at:
https://standards.freedesktop.org/desktop-entry-spec/1.1/apc.html
Fix various typos and misspellings
lintian, Debian's package checker, uses strings to check for common typos
in compiled binaries. This change fixes the ones it identified in 0.22.1.
Reviewers: zmike!
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D5585
In EFL git head, this function has changed to be just
evas_object_coords_inside_get. This patch wraps the two different
function names around an #ifdef so we can check EFL version that we
are building against and call the correct function accordingly.
Fixes T6500
Signed-off-by: Chris Michael <cp.michael@samsung.com>
several override redirect windows are being forcibly re-placed by e.
they should not be. ever. this affects scim's IM panel i see and
chromium's menus. it probably affects more than this. this fixes both
of these cases for sure. probably more.
@fix
setting desk after e_client_new() completes is too late, as a desk will have
been automatically set by this time and any existing positioning data will
have been mangled in the process
fix T6317
previously this resulted in attempting to move the comp object before it
was created, but evas does not throw any errors when null is passed and so
it was never noticed
fix T6317
ref T6326
this was incorrectly referenced from weston's x11 compositor, which directly
uses xcb events and sends them as wl_pointer events. efl translates all input
events to use the same button scheme numbering, coincidentally enabling the use
of this code for all backends
this works around an apparently atexit() handler in nvidia drivers
that if you fork() a parent and do NOT exec() then the child process
will end up messing with the parent processes access to the nvidia
devices/display/whatever i believe via an atexit() handler because
moving from exit() to _exit() solves the issue. fixes T6197
@fix
Currently e_comp_util_has_x() will still return true after the X
machinery is shut down - this can lead to crashes on logout under
wayland when xwayland apps are running. Clearing the root window
id resolves this.
@ref T5593
XWayland likes to set a buffer on the cursor surface then delete it before
we release it. I'm pretty sure when a client does that we're within spec
to just kill it, but users will likely find this response ungratifying.
So, instead, just gracefully fail to render the undefined surface.
@ref T5593
cancel may destroy the selection source, resulting in invalid access
==10735== Invalid write of size 8
==10735== at 0x87C8095: wl_list_remove (wayland-util.c:56)
==10735== by 0x2EE745: _e_comp_wl_data_device_selection_set (e_comp_wl_data.c:506)
==10735== by 0x2EF241: _e_comp_wl_data_device_cb_selection_set (e_comp_wl_data.c:714)
==10735== by 0x1021F037: ffi_call_unix64 (in /usr/lib/x86_64-linux-gnu/libffi.so.6.0.4)
==10735== by 0x1021EA99: ffi_call (in /usr/lib/x86_64-linux-gnu/libffi.so.6.0.4)
==10735== by 0x87C70C0: wl_closure_invoke (connection.c:935)
==10735== by 0x87C13C8: wl_client_connection_data (wayland-server.c:406)
==10735== by 0x87C4483: wl_event_source_fd_dispatch (event-loop.c:90)
==10735== by 0x87C4DE4: wl_event_loop_dispatch (event-loop.c:423)
==10735== by 0x85AE346: _cb_create_data (ecore_wl2_display.c:399)
==10735== by 0x946857A: _ecore_call_fd_cb (ecore_private.h:347)
==10735== by 0x946AE51: _ecore_main_fd_handlers_call (ecore_main.c:2015)
==10735== by 0x946B823: _ecore_main_loop_iterate_internal (ecore_main.c:2403)
==10735== by 0x946935E: ecore_main_loop_begin (ecore_main.c:1308)
==10735== by 0x151139: main (e_main.c:1088)
==10735== Address 0x222ca980 is 16 bytes inside a block of size 136 free'd
==10735== at 0x4C2CE1B: free (vg_replace_malloc.c:530)
==10735== by 0x2F076F: e_comp_wl_clipboard_source_unref (e_comp_wl_data.c:1291)
==10735== by 0x214C6A09: _xwayland_send_cancelled (dnd.c:149)
==10735== by 0x2EE71D: _e_comp_wl_data_device_selection_set (e_comp_wl_data.c:504)
==10735== by 0x2EF241: _e_comp_wl_data_device_cb_selection_set (e_comp_wl_data.c:714)
==10735== by 0x1021F037: ffi_call_unix64 (in /usr/lib/x86_64-linux-gnu/libffi.so.6.0.4)
==10735== by 0x1021EA99: ffi_call (in /usr/lib/x86_64-linux-gnu/libffi.so.6.0.4)
==10735== by 0x87C70C0: wl_closure_invoke (connection.c:935)
==10735== by 0x87C13C8: wl_client_connection_data (wayland-server.c:406)
==10735== by 0x87C4483: wl_event_source_fd_dispatch (event-loop.c:90)
==10735== by 0x87C4DE4: wl_event_loop_dispatch (event-loop.c:423)
==10735== by 0x85AE346: _cb_create_data (ecore_wl2_display.c:399)
==10735== by 0x946857A: _ecore_call_fd_cb (ecore_private.h:347)
==10735== by 0x946AE51: _ecore_main_fd_handlers_call (ecore_main.c:2015)
==10735== by 0x946B823: _ecore_main_loop_iterate_internal (ecore_main.c:2403)
==10735== by 0x946935E: ecore_main_loop_begin (ecore_main.c:1308)
==10735== by 0x151139: main (e_main.c:1088)
==10735== Block was alloc'd at
==10735== at 0x4C2DC05: calloc (vg_replace_malloc.c:711)
==10735== by 0x2F0520: e_comp_wl_clipboard_source_create (e_comp_wl_data.c:1231)
==10735== by 0x214C71C9: _xwl_selection_notify (dnd.c:286)
==10735== by 0x946130C: _ecore_call_handler_cb (ecore_private.h:331)
==10735== by 0x94626BC: _ecore_event_call (ecore_events.c:629)
==10735== by 0x946B83E: _ecore_main_loop_iterate_internal (ecore_main.c:2408)
==10735== by 0x946935E: ecore_main_loop_begin (ecore_main.c:1308)
==10735== by 0x151139: main (e_main.c:1088)
so if you closed the lid ANd didn't have external screens ANd had ac
plugged in ... e would suspend even if youdidn't have "suspend on ac"
checked in blanking dialog. respect this and handle it properly.
@fix
side note... i've notced acpid no longer reporting lid events on one
of my systems. this is very sucky. you now have to open
/dev/input/event2 (this may vary from device to device) to get lid
events... it's an input device. this means elput needs to start doing
this as we cant get power button or lid events anymore .... and we may
have to do this for x11 too.
Carsten Haitzler
Alastair Poole
Christopher Michael
Stephen Houston
Marcel Hollerbach
Derek Foreman
Chris Michael
Mike Blumenkrantz
Stephen Houston
Mike Blumenkrantz
Thanatermesis
Ross Vandegrift
Amitesh Singh