please note, not the return type, but the function.
When appending __attribute__((const)) to a function, the compiler is
told that its enough to call this function once in a function.
This is quite often happening when we are efl_data_scope_get and
efl_super in a function that is different from a implemented function.
The compiler now starts to remove the calls that aggressivly that we
need to ensure that these calls are not removed, which means, the static
function calls, and the eo init are now checking the return value of
these functions, to ensure that they are called.
Please note that you now have to be carefull when your app calls
eo_shutdown, if it does so, you *must* call it at the end of a function,
or never call class_get after that anymore.
Overall this improves elm test runs 0.1s which is fair i guess, the main
thing that is faster is textrendering, where is also the point where
this is the most beneficial.
Please note, this replaces 42 occurences of double _class_get() ... THAT
is a sign!
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Daniel Kolesa <daniel@octaforge.org>
Differential Revision: https://phab.enlightenment.org/D12057
Summary: Add AV1 image file loader to Evas
Test Plan: test executable with avif files found in libavif project
Reviewers: raster
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D12051
Make sure we check if pool->first_fill is NULL before passing it onwards
where it gets dereferenced. In case this happens return NULL instead of
garbage.
CID: 1293520
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11980
In case we fail make sure we error out to allow better debugging.
Also fix some surrounding indent.
CID: 1402693, 1402670
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11976
The former is a POSIX name, the latter is non-standard. I don't
know why the latter was used, considering glibc literally has
just #define LC_ALL __LC_ALL, but change it and unbreak build
on musl and other systems.
regions sometimes produce box sets like:
// +---+
// | |
// +---+ +-------+
// | | | |
// +---+ +-------+---------+
// | |
// | |
// +-----------------+
so the upper-left 2 boxes can be merged into 1 and they have the same
x coords and are flush-aligned one above the other. that is what
this does - find these and merge them to have fewer rects
@fix
we have used a fair bit of cpu to avoid gpu overhead and this tunes
the balance back a bit to throwing more at the gpu and less at the cpu
by reducing the number of pipes and max vertex counts per pipe by
default and only resetting pipe state vars if adding the first pipe
member.
@opt
Summary:
During shutdown we used to call engine_image_free() which was causing some deadlock.
as we have evas_cache which takes care of freeing all the images we just have to
delete the generic cache without freeing the image during shutdown.
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11915
this way we can use that list easily in the modules, and keep the list
for later module building.
Differential Revision: https://phab.enlightenment.org/D11897
libraries are split into deps, external deps, and pub deps.
Evas engines are refactored to use the predefined engine deps.
this is preparation work for efl-one.
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11806
Patch 09ab8c9967 added code to printf wl client animation
ticks, however the line that actually does the printing is commented
out (which is fine) but the variables used in that printf are not
needed if we are not printing things out
Adding it to all missing places so we can control it from the central
meson.build file.
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11855
Keep the backword compatibility for wayland ECORE_WL2_EVENT_DATA_SOURCE_END event handler.
If we don't use this handler, the dnd is not ended when the drop item is dropped.
@fix
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11846
Summary: .9.png is not 9 patch file, but a png file which name is .9
Test Plan: evas_object_image_file_set(image, "./.9.png", NULL);
Reviewers: Hermet, jsuya, herb
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11848
Summary: EAPI must be defined to dllexport when building DLL, and to dllimport when using these DLL. To achieve this, define EFL_BUILD for each library and module, and set DLL_EXPORT unconditionally. Static library are and will be not supported
Test Plan: compilation
Reviewers: zmike, raster, jptiz
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11834
Summary:
void* is an invalid type for calculating pointer offsets, so ensure that
this is always cast to something else (e.g., char*) in the few cases
where it's been misused
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11816
Summary:
There was a mistake caused by "2a0eeba evas: fix png regression issue".
The flag is_for_data was close_file before this commit.
_evas_image_load_file_internal_head_png closes file if it is called only for
header. So the close_file does not mean is_for_data.
This patch is changing is_for_data to is_for_head, and make
_evas_image_load_file_internal_head_png use EINA_FILE_SEQUENTIAL
for data decoding.
Reviewers: Hermet, jsuya, herb, raster
Reviewed By: raster
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11783
Summary:
Accidentally commit "382c580 evas: add support for .9.png file to PNG loader."
adding the 9 patch feature with small code refactoring made use of setjmp
incorrectly.
[Problem]
evas_image_load_file_data_png calls _evas_image_load_file_internal_head_png,
and _evas_image_load_file_internal_head_png calls setjmp and returns without
problem. And png_read_row calls longjmp. This causes jumping into a function
which was exited. Problematic png file will be attached.
[Solution]
Save calling environment i.e. call setjmp, after returning from
_evas_image_load_file_internal_head_png.
Test Plan:
Problematic png file
{F3876983}
And example code.
{F3876986}
Reviewers: Hermet, jsuya, herb
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11782
Since we moved to floating calculation for msaa, this comparsion
is not working for direct image drawing though it's doable.
To resolve this, we can compare floatings regardless of msaa enable.
we were losing cached chaders and overwriting the cache all the time
when apps finished compiling new shaders and at other points... when
they should have already had a populated shader cache that was all
fine and happy. this fixes that so once the cache is fully populated
it wont write anymore and it wont lose existing shaders in the cache
@fix
Summary:
since the existance of seleciton manager, the converter callbacks from
ecore_x are expecting custom struct pointers. However, enlightenment
never updated to use the elm dnd API for client side usages. Which
results in the fact that sometimes, when a client sents Notify, and
e_dnd is active, that this converter is executed with the wrong data.
With this commit the data passed in is ensured to have the correct magic
number. The proper solution for this would either be registering the
correct converters in enlightenment, or update to elm_drag. However,
since the continues cried river over the last 5 days has raised enough
hydro power to add these changes. This leaves us with just one question:
How was it possible to generate so many messages about a problem that
can be solved in a fraction of charaters that have been written?
fixes <a-issue-that-was-never-created>
Depends on D11700
Reviewers: zmike, stefan_schmidt, raster
Reviewed By: zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11701
Summary:
this is needed in order to tell the API user that this is not going to
be completed.
Depends on D11699
Reviewers: zmike, raster, stefan_schmidt
Reviewed By: zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11700
Summary:
ecore_x_dnd_send_status can be used to indicate if a item can be dropped
on a client or not. However, we should only indicate that this can be
dropped, if there is a object we signaled that a drop is in.
Long story short: there is no assertion that after indicating that
things can be dropped, that a notify for the data is sent. A drag
implementation should always listen to a mouse up event, and abort the
drag if no further operations are sent.
Depends on D11698
Reviewers: zmike, stefan_schmidt, raster
Reviewed By: zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11699
Summary:
this is needed in order to return the data the same way the legacy impl
did. This however has the annoying sideeffect that ecore_evas_x now
depends on efreet, and we had to change the build order.
Depends on D11696
Reviewers: zmike, stefan_schmidt, raster
Reviewed By: zmike
Subscribers: devilhorns, cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11697
Summary:
Static analysis tool reports passing a null pointer 'im->gc' to
_evas_gl_image_cache_add which directly dereferences it, so lets
be sure that 'im->gc' is valid before passing it to cache_add
Reviewers: Hermet, jsuya, herb, zmike
Reviewed By: zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11676
this will now result in the promise beeing rejected.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11653
this is just causing errors and nothing usefull, upower is not available
on macos.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11652
A long story in a few words: sometimes on macos modules will be compiled
into .dylib, sometimes, into .so suffix. We did not set the suffix
everywhere in our meson build instructions, hence our suffixes have been
differently, which resulted in random load fails on different maschines.
With this commit, we ensure that we suffix all modules the same way.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11650
memory is duplicated in eina_content_new.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11641
the structure "!strcmp(X, "foo", strlen("foo"))" is equal to
"eina_has_prefix(X, "foo")", and the later is nicer to read, hence this
replaces it.
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11620
this pointer needs to be freed if it's being copied
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11604
Coverity reports that 'fd' here is negative, and close() cannot accept
a negative parameter, so add a check to make sure 'fd' is not negative
before passing to close function.
Fixes CID1420318
write() cannot be passed a negative value. If
ecore_main_fd_handler_fd_get returns -1, then this is an issue. Check
for valid fd being returned from fd_handler_fd_get and if it is
negative, then cleanup and get out.
Fixes CID1420318
If we are going to be exiting this function without setting up the
ecore_main_fd_handler (and thus passing forign_slice), then we should
free forign_slice as it was previously calloc'd above.
Fixes CID1420322
I thought that i explicitly tested this. However, it seems i was wrong,
this way now this is handled correctly, and the image is stored and
coverted to a path.
With this you can now copy images from chromium and firefox to elm apps.
The change in elm_entry reverts parts of the rewrite of the elm
handling, as initially every path came *without* "file://" in front, so
we have to maintain that.
fixes T8625
enlightenment does not use the elm dnd operation implementations.
Therefore we have to ensure that we only handle these operations here,
when ecore_evas dnd API is used.
gc width/height indicates the main output size,
if the target surface(fbo) is larger than this, rect won't be drawn.
We should clip this with current context size.
@fix
Some methods were missing the "Drag" or "Selection" namespaces or the _Cb suffix.
Depends on D11219
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11426
Including Eina.Content
And a typo/bugfix in ecore_evas_x.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11204
with this commit you can do limited cnp for cocoa. You still cannot copy
and paste pictures or markup arround, only text is supported so far.
However, text on its own works quite stable and good.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11351
This adds cnp support, actions are right now only mapped to "ask",
further support can be added there, and synchronization can be added to
register more available actions. However, i did not find *any* wayland
implementation in gtk qt nor chromiumos that even use the action to
indicate anything. This here also has a slightly different behaviour to
X11 in terms of coordinates for motion,leave,enter. They can contain
negative coordinates (which is due to the fact that wl is CSD and X11 is
SSD. However, I did not want to fix this in any regard, as you might
want to use that, and it would be a none trivial amount of code to fix
that.
Reviewed-by: Chris Michael <cp.michael@samsung.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11329
Seats are not implemented, if there is a type mismatch promises are
going to be rejected. Most of this code is copied over from
selection_manager.
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11195
The idea of copy and paste here is:
- The user specifies the content he wants to have in the selection
buffer with a Eina_Content, these content pointer ownerships are
passed to the called. Internally ecore_evas code will memorieze the
pointer, and pass on function callbacks to the modules, which then do
not have to deal with the ownership.
- In case the module does not specify these APIs, the callback
implementation will be called, which only works for cnp *not* dnd.
- Action and mime types are handled as strings, which allows way better
custom organisations.
(The docs needs improvement)
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11192
Summary:
If use ein_str_split, have to free the first element of the array
and free the array.
Test Plan: N/A
Reviewers: Hermet, YOhoho, bu5hm4n, zmike
Reviewed By: zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11419
When we allocated s but fail to allocate l we need to make sure to free
the first allocation before erroring out.
CID: 1419874
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11409
The commit fixed a mem leak with freeing one element in the array, which
was never going to work, as only the array is a pointer you can
allocate.
This could have been discovered by simply running ninja test.
Hermet Park
Ali Alzyod
Marcel Hollerbach
Vincent Torri
Carsten Haitzler
Elyes HAOUAS
Christopher Michael
Stefan Schmidt
Daniel Kolesa
Subhransu Mohanty
Taehyub Kim
Shinwoo Kim
Mike Blumenkrantz
Xavi Artigas
junsu choi
Hermet Park