If a class (e.g Evas_Object) is expected as a base class but is not
present in the existing expected file (e.g evas_object.eo), it will not
fail.
This patch fixes it by checking the existence of the class into the
database after the parsing of the file.
@fix
This function lets you hook at the end of eo_add and override it for a
class. This is essentially the first step towards killing custom
constructors. Instead of having a custom constructor, you should just
do:
eo_add(CLASS, parent, a_set(3), b_set("eou"));
eo_constructor is called at the beginning for pre-init things.
eo_finalize is called at the end, for actually finalizing and doing
things. This cleans up the API and possibly saves a lot of things that
would have been stupid and slow in the past, like loading an elm widget
with an existing theme, and then changing the theme.
** This breaks Eo ABI, please recompile elementary and everything else that
creates eo objects.
@feature
Summary:
This enables textblock to support more values other than 1.0.
For 0 <= ellipsis < 1.0, it splits the text such that it fits the
textblock's width. The ellipsis is relatively position according to the
ellipsis value, and characters are removed also relatively.
For example, a value of 0.5 will position the ellipsis right in the
center of the textblock's width, while removing characters equally right
and left from the center.
Basic approach to this feature was to do some work before the layout
process. We calculate the expected total width of the items, and by how
much we exceed from the textblock's width. Afterwards is it just some
careful work to set the boundaries of the width we want to cut, and
deciding which characters we need to removed to fulfill this requirement.
The rest is splitting the text and visually-removing the part we
need to cut.
This is all handled before any logical lines are created, so the
_layout_par function remains almost intact. A designated _ellip_prev_it
field in the Paragraph struct instructs after which item we place the
ellipsis item, if at all.
Note that we keep the fast path for ellipsis 1.0, as heavier work needs
to be done for the other values.
Added tests to evas_suite for a range of ellipsis values.
Also, multiline is unsupported at the moment.
@feature
Test Plan: Anything that uses Evas Textblock (single-line, please)
Reviewers: tasn, id213sin, raster
CC: cedric, JackDanielZ, raster
Differential Revision: https://phab.enlightenment.org/D905
The function eolian_implement_information_get was returning strings for
the class and the function. It was written in this way at the beginning
because it was not needed to verify the correctness of the class and
the function.
Now that we have the namespace feature, this function must check it,
meaning that the class and the function are now known.
So we can return them instead of returning the strings.
The generators had to find the class from the classname. It is no more
needed.
The C++ generator has been adapted to this new API.
Summary:
Key event string encoding was LATIN1 instead of UTF-8.
This unifies the handling with the x backend.
Reviewers: raster
Reviewed By: raster
CC: cedric
Differential Revision: https://phab.enlightenment.org/D912
Summary: Add 3 new APIs function for full work of the picking nodes.
Reviewers: Hermet, cedric, raster
CC: cedric
Differential Revision: https://phab.enlightenment.org/D889
Summary:
Support union, subtract, intersection, equal(comparison) between tilers.
@feature
Test Plan: Test with added test case(src/tests/eina/eina_test_tiler.c) and the example(src/examples/eina/eina_tiler_02.c)
Reviewers: gwanglim, devilhorns, raster, zmike, cedric
CC: cedric
Differential Revision: https://phab.enlightenment.org/D880
On Windows, '\' are used to separate the names of a path. It was not
supported.
This patch fixes that. Thanks to Vincent Torri for catching it (and good
luck in Windows ;-))
Summary: The comparison dc with NULL is not necessary. So remove the unnecessary conditional expression.
Reviewers: Hermet
Reviewed By: Hermet
CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D908
Summary: Pointer fd is used after being deallocated.
Reviewers: Hermet
Reviewed By: Hermet
CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D906
this massively improves edje performance when using groups, which previously would continue calculating their parts even when their parent object was hidden
CPU usage in my test case went from 20-30% to 1%.
@fix
Until now, the functions giving access to class information were taking
the class name as parameter.
Except the fact that we needed to search into a hash table for the internal
class structure, no flexibility is possible.
This change consists in modifying most of the APIs using the class name
with a new Eolian_Class type and adapt the code of the C and C++
generators accordingly.
It includes parser updates, database fill and tests on basic and complex
types.
One can define types in this way:
type Evas_Coord: int; /* Simple type definition */
type List_Objects: Eina_List * <Eo *>;
@feature
If classes A and B are stored in a same file, when A inherits from B,
and A needs to be generated, the generator, even if B is parsed, will
not search the class in the database but will look for some b.eo.
This patch fixes that issue by checking the existence of the classes
into the database before looking for a corresponding file.
Summary:
text_props.len indicates visual length from harfbuzz logic when it is available.
But, input pram "idx" is logical length. So, ellipsis works abnormally in some case.
@fix
Test Plan: D870
Reviewers: tasn, woohyun
CC: cedric, herdsman
Differential Revision: https://phab.enlightenment.org/D871
Match function names when the API pointer is out of range.
Reviewed by TAsn and modified according to his comments :)
Differential Revision: https://phab.enlightenment.org/D876
Summary:
the draggable.event field (in edje) is pointing on the id of the part it is assigned to.
The main problem is about deleting parts.
After part being deleted, all other parts (that are below the deleted part) have their id shifted,
but event doesnt shift.
For example,
1) draggable.event at some part pointing at part with id=7,
2) some other part is being deleted
3) part with id=7 have new id now (id=6)
4) draggable.event DOESN'T change it's pointer and still points at part with id=7 (which is now has id=6)
5) if there is no parts with id=7 anymore, then it could give you SEGFAULT someday (or not, depending on luck).
This commit contains next changes:
1) after deleting part, draggable.event also shifts (if required)
2) after deleting part that is being pointed with event by someone, it will remove all event points at it
@fix
Reviewers: cedric, raster, seoz
CC: reutskiy.v.v, cedric
Differential Revision: https://phab.enlightenment.org/D878
Signed-off-by: Cedric Bail <cedric.bail@free.fr>
Engines are stored in a lib/ folder, while the main DLL files
are in the bin/ folder, so the engine would never be found.
A solution was to add the proper checkme file in the share
folder, but since this is necessary only for Windows, we
can simply use ../lib instead of using the full eina_prefix
detector.
Thanks vtorri for the review.
Fixes a mistake in commit 78c9a82a14
The authority part would get placed in the struct member "path",
and vice versa.
Lesson learned: Don't rewrite your patch late at night after
arc has messed it up.
The syntax to define a parameter of a property as const is as following:
set {
data: const;
}
When no space was present between ':' and "const", a syntax error was
occurring.
We check _current_table for NULL, and then populate it (it's a global)
through another function, but we don't really check it's not NULL before
using it, we just assume because of an indirect other variable.
This confused coveritiy, can confuse humans too, and in general risky
(if something changes).
CID 1039419
Coverity found the issue, but it's an issue we've already fixed in the
past. I don't know how it got lost, but it seems like someone did a bad
merge. Probably when migrating to Eo2.
This reverts commit 831c20464d.
In particular, ellipsis is -1 by default in Evas, but at this
point (first layout calc), the parameters used for recalc are
incomplete and ellipsis would then be 0 by default (calloc).
As a consequence, Edje will call ellipsis_set(0) enabling
ellipsis even on objects that force "ellipsis: -1".
Solution: set all the parameters before entering text/tb calc.
I believe the other changes are only color and image padding
and should not affect recalc_single.
Summary:
To correcly display the added pats using the edje edit api, update
init the filds of Edje_Real_Part. Initialize the next filds: type and
typedata. It filds initialized accordingly to Edje_Part_Type value.
@fix
Reviewers: raster, cedric, seoz, Hermet
CC: cedric
Differential Revision: https://phab.enlightenment.org/D854
Signed-off-by: Cedric Bail <cedric.bail@free.fr>
Text objects declared in Edje will see their padding added twice,
as the Evas_Object_Text itself contains the padding already.
This WILL break some EDC files. It's a bug nonetheless.
Should this be backported?
erroneous variable.
@bugfix: Since later functions which process the ecore key event will
end up subtracting X's broken keycode values from This original
keycode value, we had better send the proper code with the offset
already added.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
I am concerned that we do duplicate the color format of a texture insteed
of reusing the colorspace enum from Evas (that already exist). I believe
this is a side effect of not using Evas_GL_Image for texture.
This let me think that I wont allow this API out of BETA until we reduce
duplication by using Evas_GL_Image as it seems it will also impact the
public API of Evas_3D.
Summary: This patch gets rid of a "defect" pointed out by Coverty Scan -- as per jackdanielz request.
CID 1211985.
Reviewers: cedric, smohanty, JackDanielZ
CC: felipealmeida, cedric
Differential Revision: https://phab.enlightenment.org/D840
Signed-off-by: Cedric Bail <cedric.bail@free.fr>
bugfix@: As we have support in the Ecore_Event structure for storing
keycode on key press/release, let's utilize it. This will simplify
wayland compositor code as we won't have to parse key names to get the
keycode.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
The type of the virtual functions implemented in a class was let unresolved
and led to inconsistency when working with them.
In fact, the type can be resolved easily after all the classes that the
specific class inherits from are parsed and stored in the database.
Fix T1228
Summary:
This patch adds 'eolian_cxx' -- a C++ bindings generator --
to the EFL tree. Eolian Cxx uses Eolian API to read .eo files and generate
.eo.hh. It relies/depends on Eo Cxx and Eina Cxx (both non-generated
bindings).
src/bin/eolian_cxx: The eolian_cxx program.
src/lib/eolian_cxx: A header-only library that implements the C++ code
generation that binds the .eo classes.
=Examples=
src/examples/eolian_cxx/eolian_cxx_simple_01.cc: The simplest example,
it just uses some "dummy" generated C++ classes.
src/examples/eolian_cxx/eolian_cxx_inherit_01.cc: Illustrates how
pure C++ classes inherit from .eo generated classes.
src/examples/evas/evas_cxx_rectangle.cc: More realistic example using
the generated bindings Evas Cxx. Still a bit shallow because we don't
have full fledged .eo descriptions yet, but will be improved.
=Important=
The generated code is not supported and not a stable API/ABI. It is
here to gather people interest and get review before we set things in
stone for release 1.11.
@feature
Reviewers: cedric, smohanty, raster, stefan_schmidt
CC: felipealmeida, JackDanielZ, cedric, stefan
Differential Revision: https://phab.enlightenment.org/D805
Signed-off-by: Cedric Bail <cedric.bail@free.fr>
Depending on the "filled" flag, adjust how the image is filtered.
Filled: take padding into account and zoom out original image if needed
Non filled: we're tiling anyways, so padding doesn't have any meaning
This adds filter support to Image objects as well.
The exact same filters can run on Text and on Images
(provided some colorspace limitations are respected).
This basically adds:
- Support for RGBA input buffer
- Eo entry points for Image filter support
- Implement basic filter support in Evas_Image
It was a pretty stupid idea to write a parser for a custom language
when we already have Lua as a dependency and it's so beautiful and
easy.
There is a fallback function to allow for compatibility with legacy
filters. But that broken syntax is not recommended. I'll probably
remove it soon.
All the test cases I have in my example app work fine with this
compatibily layer.
Before this change, all the .eo files of the directories given with -I
option were parsed. Most of this information was not necessary at all,
since only the classes belonging to the inheritance of the class given
as parameter were needed.
Now, during the parsing of the given class, the inherits classes are
searched and parsed.
A condition is needed to make it work well. To find a filename for a
class, we consider the lowercase of the class name as the filename we
have to parse.
e.g, Elm_Button -> elm_button -> elm_button.eo
It considerably reduces the generation time.
A fix in the tests was needed.
this is enabled for all scripts within a group, and it should only be used if you:
1) know what you are doing
2) know why this is unsafe (T905)
@feature
It includes too fixes for the complex type. If I had listened to Tasn,
I would have detected them a long time ago.
But he didn't insist enough. He just said:
"Write your tests, ?#@*&%! french!"
the declared t3d_scene api names are not matched exactly between header and code.
these name should be just "3d_scene"
and still there was a Evas_3D.h reference in evas Makefile.
Enable 3D features using --enable-evas-3d=yes when configuring.
APIs are exposed through Evas_3D.h.
Currently, evas-3d is being supported only on gl_x11 engine.
Conflicts:
src/lib/evas/Evas_Eo.h
Due to some invalid geometry considerations, there was a
1 pixel distortion in images, varying with the lz4 compressed
macro-block size.
Anyhow, I couldn't wrap my head around Cedric's code. So I rewrote
the whole thing instead, fixed it and improved the block size
selection (based on the image size, to optimize lz4 compression).
Summary: Input type for offset value must be Integer in the "edje_edit_rel1(2)_offset_x(y)_set" functions
Reviewers: cedric, Hermet
CC: cedric, Hermet, raster
Differential Revision: https://phab.enlightenment.org/D562
Summary:
Applications are stuck when handler of pipe made nested loop.
In the nested loop, _ecore_pipe_read() tried to read new data based on previous information which is not cleared yet.
Spotted by gyuyoung.kim, sy302.park.
Related webkit bug is https://bugs.webkit.org/show_bug.cgi?id=129294
Reviewers: cedric, seoz, raster
CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D790
evas_object_text_clear was no more defined in the headers. Since the
function code is not generated by Eolian, the declaration in the headers
has to be done manually too.
Summary:
Before "open_done" callback, play unset is ineffective.
is it intentional logic?
if we set play and unset play again before "open_done" callback, the video will be played.
@fix
Reviewers: raster, cedric
CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D786
what a huge, colossal cock-up of a clusterfuck. it's a good thing nobody ever uses ellipses or edje. otherwise we'd probably get complaints about this kind of thing.
this should always be applied, not just when min_x isn't set; failure to do so causes erroneous ellipsizing in some cases where it should be explicitly prohibited
@fix
define scope keywords 'protected' and 'public'
disallow scope keywords as property or method name
use move_ts action to move token start after scope modifier if found
use end_property_scope and end_method_scope actions to handle it
this fixes and eo2 problem where when callstack grows (or shrinks)
and realloc nas to relocate memory, the frame ptrs like fptr become
invalid and all sorts of hell ensues.
this uses mmap so blowing the stack will segv, not scribble over
memory, also its separated from malloc heap, and now big enough to not
need to size ... ever (1024 entries).
@fix
Summary:
In emotion_engine_instance_new(), if given module name is wrong or if loding is failed, fallback module loading should be triggered.a (previously, it check just whether name is NULL or not)
In case of T1025, "vlc" was given name. Therefore, fallback was not triggered.
Fixes T1025
Test Plan: enlightenment -> efm -> mouse over multimedia files -> check whether preview is shown properly or not
Reviewers: raster, cedric
CC: seoz, cedric, zmike
Maniphest Tasks: T1025
Differential Revision: https://phab.enlightenment.org/D773
Add a new flag in EDC files to specify ETC1 compression
should be enabled. It follows the same rules as the
current LOSSY flag for JPEG compression.
@feature
There were a few critical issues:
- Invalid pointer arithmetics on the input data (char vs. int)
- Invalid logic in the pixel duplication code
All of these due to bad copy and paste :(
Also, use LZ4HC instead of LZ4 when compression is enabled.
ETC1 encoding is so damn slow you won't see the difference between
LZ4 and LZ4HC compression times.
Summary:
Changed uses of EINA_MAIN_LOOP_CHECK_RETURN for EINA_MAIN_LOOP_CHECK_RETURN_VAL
for functions that doesn't have void return types.
These only error out when compiling with --with-profile=debug
@fixed
Reviewers: raster, cedric, smohanty
CC: cedric
Differential Revision: https://phab.enlightenment.org/D765
Summary:
Due to unnecessary magic checking, there was freezing in /proc in efm.
Proc file's st_size is zero, but, it's readable. therfore, it takes unnecessary time in magic checking. And, there is no need to check magic in case of 0 sized regular files as well.
Therefore, skip magic check in case of st_size is zero.
Fixes T1173
Test Plan: enlightenment -> file browser (efm) -> get int /proc --> check whether efm freezes or not
Reviewers: raster, cedric, zmike
CC: seoz, cedric
Maniphest Tasks: T1173
Differential Revision: https://phab.enlightenment.org/D764
Signed-off-by: Cedric Bail <cedric.bail@free.fr>
The _ecore_wl_cb_handle_data callback only has the flag for ECORE_FD_READ
set while the callback funtion also monitors and acts on writes. I wonder
if and how that worked before.
the comments say it all. this stops adding massive full window updates
when clippers lose all children (and are visible) or gain a first
child while visible as all code hides these clip rects (or shows them)
in these cases and other usages just make no sense.
due to recent changes a lot of objects are now NULL (correctly) and eo
complains on access of them. it's simply too noisy adding too many
if's all through code, so let's just make eo be sensible here.
@fix: There is no use for an input cursor surface if the seat does not
have the pointer ability (for kiosk cases, or touch-only cases where
a pointer is not created).
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
This problem occurred due to xkb_keysym_t value of libxkbcommon by e wl server.
e wl server should pass keycode from evdev input device on to wl client.
In order that e wl server receives valid keycode Ecore_Event_Key should have
an extended data member. This patch should be applied with server side patch.
@fix
Test Plan: run e wl server -> create wl client -> type keys
Reviewers: raster, devilhorns, zmike
CC: cedric
Differential Revision: https://phab.enlightenment.org/D712
It's now completely valid to do:
a = eo_do(obj, a_get());
or:
b = eo_do(obj, a_set(1), b_get());
Also, the default return value for eo2 functions is now also returned
when the object is invalid, not just when the object does not match
class.
It's a small refactor that fixed both issues at once.
@feature
@fix
They are not needed on normal build. Two flags have been added to
enable/disable INF and DBG.
Two warnings have been removed since they are not printing interesting
and relevant information.
num was changed to size_t which makes it unsigned while count is a normal
(signed) int. Comparing them does not always work as expected but in this
case we already checked if count is < 0 before and would have returned if
that would be true. Thus we can safely cast count to unsigned here as it
will be greater 0 anyway.