Mice in X with xi2 send Axis events which are badly defined,
and carry basically useless information, as we also receive
proper mouse events. Notably, all mice input events are
"Rel something" but in fact they are absolute values (even
the wheel information is a counter increasing every time you
scroll).
This should not break any application as such axis events
carried only values with label ECORE_AXIS_LABEL_UNKNOWN.
This also fixes a leak when n == 0 (no "valuator" found
in the list, this used to be unlikely, now happens at every
mouse event).
This converts Evas_Axis or Ecore_Axis info arrays into basic
pointer data. Also marks those fields as set. All events need
to properly implement the value_has property (mark all bits
whenever a value is known).
This sets a bit whenever a callback listener is added.
I couldn't get any profiling data easily (too small for
valgrind).
Note: This removes the proper refcounting on the "move"
event listeners. I believe this is not a problem as most times
the move_ref goes to 0, it is because the object is deleted.
Worst case, we just trigger a callback_call with no listeners.
This adds 32 bits to each evas object private data.
This event should not be exposed at all, it's not necessary
anymore, EFL_EVENT_DEL already exists and should be good enough.
This does move the callback call a little bit ealier in the del
process, but at first glance, this shouldn't have any impact.
This moves MULTI events to those new finger event types,
and also sends a finger event for finger 0 (aka the pointer).
NOTE: This may require a separation between a mouse input and
an actual finger touch. To be defined, ie: do we let the app
check the input device info to decide whether the event is
actually the first finger of a multi touch device, or do
we want to send only actual finger events from multi touch
devices only?
@feature
This is getting trickier, as those events have a lot more
side effects and complexity than a simple wheel event...
Some code has been added that should be fixed in the following
commits.
So glibc has decided that readdir_r is hard to use safely and deprecated it
this summer. They recommand to use readdir, which was in the past unsafe to
use in a multi thread scenario, but is now on most system (and all system
we care, including our own implementation in evil). It is basically safe
as long the same DIRP is not accessed from another thread. This is true in
our code base, so we are fine to go with this.
For further reading: https://lwn.net/Articles/696474/
It has been a long journey, but here we are at last...
The infamous gl_cocoa engine has been migrated to the
gl_generic infrastructure. This should provide great
improvements and hopefully reduce side-channels b0rkage.
Fonts seems better, scrolling is smoother, expedite
does not segfault anymore... I haven't found a
regression with elementary_test, elementary_config,
terminology, expedite.
It's entirely possible that a system doesn't have a cpu 0, so
when we try to pin all our render threads onto processor 0 we
may fail.
This results in some very connfusing build breakage when
edje_cc hangs up because its render thread didn't start.
So, if starting the thread with affinity fails, let's try without
affinity.
(This is trivial to reproduce - just use sysfs to turn off cpu0
after boot.)
@fix
This fixes argb windows transparency in E software compositor.
My current problem is that I have no idea what changed, why this
is needed now, and how things could actually work before.
Fixes T4389
@fix
so ok - efreet crashed letting you know you have a missing mimedb...
return NULL instead fixes T4425 :) - rememebr to kill your efreetd's
to get a new mimedb - log out and in will do that.
The use of low-level interfaces such as Efl.Io.Reader and
Efl.Io.Writer are not that user-friendly as they can handle partial
data.
Classes such as Efl.Io.Copier makes them easy to use, but they need a
reader (source) or writer (destination) and in our examples we used
fixed buffers or some existing streams (stdin/stdout/stderr,
networking...).
However, if interactively we need to produce some data to be sent,
such as implementing some networking protocols, we'd have to write our
own Efl.Io.Reader and Efl.Io.Writer classes to handle the buffering.
Not anymore! With Efl.Io.Queue you can write stuff to it and it will
buffer to memory. Once stuff is read, it will automatically remove
those bytes from buffer.
We set the wayland surface to NULL in engine info, then destroy the wayland
surface later. The GL stack tries to render to the destroyed surface and
dies.
If we destroy the GL surface when we set the engine info (before we destroy
the wayland surface)it stops the GL stack from doing bad wayland calls.
This fixes a crash on exit, or when any window closes when using mali GL
drivers. Surprisingly, Mesa was tolerant of this.
So actually there is quite a big issue with semaphores
on OSX. We use (named) POSIX semaphores, but this was
a (my) mistake... I'll fix it later...
The real issue is that named semaphore are persistants:
when the program dies, it stays alive. This is pretty
bad with eina_debug_monitor because we create a semaphore
we never release, due to a wild thread...
This leak of semaphores went unnoticed before commit
4a40ff95de because the
name of the semaphore was unique per process, and
overriden when another process was launched. This
was very bad, but saved us from overflowing the
semaphore pool. It is now overflowed pretty fast when
building a lot EFL, because of Eolian that runs A LOT!
So that's one problem that still needs to be fixed,
by using OSX' own semaphores (see T4423).
Another big issue, which is now fixed is that the
buffer in which we generated the semaphore ID was
too small, and therefore we were reduced to one shared
semaphore for a whole process... This buffer has been
now set to 31 characters, which seems to be the maximum
length of a semaphore ID.
So now things are better, but still with a deadly issue.