platform/upstream/efl.git
6 years agoefl: remove _MSC_VER (Visual Studio macro) usage in source code
Vincent Torri [Wed, 3 Jan 2018 05:23:10 +0000 (06:23 +0100)]
efl: remove _MSC_VER (Visual Studio macro) usage in source code

6 years agobuild: commit also changes in Evil's Makefile.am
Vincent Torri [Wed, 3 Jan 2018 05:21:08 +0000 (06:21 +0100)]
build: commit also changes in Evil's Makefile.am

6 years agoevil: remove useless code
Vincent Torri [Wed, 3 Jan 2018 05:17:35 +0000 (06:17 +0100)]
evil: remove useless code

6 years agoelementary: alloca is already declared in eina_alloca.h
Vincent Torri [Wed, 3 Jan 2018 05:09:33 +0000 (06:09 +0100)]
elementary: alloca is already declared in eina_alloca.h

6 years agoevas: NEAR AND FAR are already defined on Windows
Vincent Torri [Wed, 3 Jan 2018 05:04:02 +0000 (06:04 +0100)]
evas: NEAR AND FAR are already defined on Windows

6 years agoecore: Efl_Core.h means using EFL_BETA_API_SUPPORT at the moment.
Cedric BAIL [Thu, 4 Jan 2018 19:44:32 +0000 (11:44 -0800)]
ecore: Efl_Core.h means using EFL_BETA_API_SUPPORT at the moment.

6 years agoeo: make efl_provider_find a @const function.
Cedric BAIL [Thu, 4 Jan 2018 19:41:49 +0000 (11:41 -0800)]
eo: make efl_provider_find a @const function.

6 years agoecore: fallback to use efl_provider_find if the passed object isn't an Efl.Loop_Consumer.
Cedric BAIL [Thu, 4 Jan 2018 19:21:36 +0000 (11:21 -0800)]
ecore: fallback to use efl_provider_find if the passed object isn't an Efl.Loop_Consumer.

6 years agofix evas test suite to now init both ecore and evas.... correctly.
Carsten Haitzler (Rasterman) [Thu, 4 Jan 2018 17:14:06 +0000 (02:14 +0900)]
fix evas test suite to now init both ecore and evas.... correctly.

see c8dcc4327b803e9b8ad2a0985e756c924946c442 - basicall evas depends
on ecore these days... thus requires ecore be initted THEN evas. ...
which in theory is an abi break for those using evas and ONLY evas
long ago from when efl was separate... but it''s how we're building
these days.

@fix

6 years agoecore evas init - init ecore then evas not the other way...
Carsten Haitzler (Rasterman) [Thu, 4 Jan 2018 17:10:26 +0000 (02:10 +0900)]
ecore evas init - init ecore then evas not the other way...

because as of... i don't know when, evas relies on ecore with
ecore_pipe_add to create the async fd... and if you init evas then
ecore this doesnt work. obviously. well now it isn't working. probably
due to new efl loop work. but the efl loop code is correct.
ecore_pipe_add should never work until you init ecore... it just
happesn to have managed to be gotten away with for a while.

@fix

6 years agoeocre timechanges - set obj to null before deletion on stop
Carsten Haitzler (Rasterman) [Thu, 4 Jan 2018 17:03:19 +0000 (02:03 +0900)]
eocre timechanges - set obj to null before deletion on stop

stops double-deletion due to callback nesting

6 years agoelm_popup: its not a focus layer
Marcel Hollerbach [Thu, 4 Jan 2018 17:23:56 +0000 (17:23 +0000)]
elm_popup: its not a focus layer

elm popup is only a contextual wrapper arround a elm_notify, so
implementing this thing here as focus layer is wrong.

6 years agoelm_widget_item_focus: use the correct parent
Marcel Hollerbach [Thu, 4 Jan 2018 17:18:21 +0000 (17:18 +0000)]
elm_widget_item_focus: use the correct parent

You need to use the parent here, otherwise the adapter will stay alive
for too long.

6 years agoecore: fix future -> promise doc
Andy Williams [Thu, 4 Jan 2018 17:02:52 +0000 (17:02 +0000)]
ecore: fix future -> promise doc

6 years agoemotion: unset DISPLAY when loading an engine under wayland
Mike Blumenkrantz [Thu, 4 Jan 2018 15:10:38 +0000 (10:10 -0500)]
emotion: unset DISPLAY when loading an engine under wayland

ensure that this doesn't trigger xwayland deadlocks during gst init

probably fixes T6418

@fix

6 years agoefl-loop: Don't use 'main' as a variable name
Chris Michael [Thu, 4 Jan 2018 14:26:28 +0000 (09:26 -0500)]
efl-loop: Don't use 'main' as a variable name

Gcc issues a warning here that 'main' is usually a function, so just
rename the variable to avoid the warning.

NB: No funtional changes

Signed-off-by: Chris Michael <cp.michael@samsung.com>
6 years agoevas: make tbm_surface api fields static
Marcel Hollerbach [Thu, 4 Jan 2018 07:51:45 +0000 (08:51 +0100)]
evas: make tbm_surface api fields static

When building statically this breaks build, since both symbols would be
defined globally.

6 years agodocs: Update elementary eo files for grammar and readability
Andy Williams [Thu, 4 Jan 2018 12:16:23 +0000 (12:16 +0000)]
docs: Update elementary eo files for grammar and readability

Author: Nate Drake

6 years agoefl_ui_popup_alert_text_scroll: do not calc size when need size calc flag is FALSE
JinYong Park [Thu, 4 Jan 2018 12:12:30 +0000 (21:12 +0900)]
efl_ui_popup_alert_text_scroll: do not calc size when need size calc flag is FALSE

Summary:
efl_ui_popup has needs_size_calc flag
to skip size calculation when it is not needed.
But efl_ui_popup_alert_text/scroll do size calc although that flag is FALSE

Test Plan:
1. elementary_test -to efl.ui.popup.alert.text
  or elementary_test -to efl.ui.popup.alert.scroll

2. resize window

3. watch _sizing_eval call

Reviewers: Jaehyun_Cho, herb, jpeg, cedric

Reviewed By: Jaehyun_Cho

Differential Revision: https://phab.enlightenment.org/D5720

6 years agopromise: Add even simpler helper for main loop promise creation
Andy Williams [Thu, 4 Jan 2018 11:56:01 +0000 (11:56 +0000)]
promise: Add even simpler helper for main loop promise creation

6 years agodoc generator: show params and return type for functions
Dave Andreoli [Thu, 4 Jan 2018 10:04:50 +0000 (11:04 +0100)]
doc generator: show params and return type for functions

6 years agoci: stay with 17.04 for Ubuntu on Travis builds
Stefan Schmidt [Thu, 4 Jan 2018 07:46:10 +0000 (08:46 +0100)]
ci: stay with 17.04 for Ubuntu on Travis builds

There is some glitch with 17.10 and systemd in our build. Roll back here
to keep the Travis build stable until this is investigated.

6 years agoefl-wl: add function to get the parent of an extracted surface
Mike Blumenkrantz [Wed, 3 Jan 2018 23:22:32 +0000 (18:22 -0500)]
efl-wl: add function to get the parent of an extracted surface

@feature

6 years agoecore_con: we should just stop using ecore_main_loop_get.
Cedric BAIL [Wed, 3 Jan 2018 20:49:10 +0000 (12:49 -0800)]
ecore_con: we should just stop using ecore_main_loop_get.

6 years agoeio: switch code to simpler pattern with efl_loop_promise_new.
Cedric BAIL [Wed, 3 Jan 2018 20:13:11 +0000 (12:13 -0800)]
eio: switch code to simpler pattern with efl_loop_promise_new.

6 years agoecore: introduce efl_loop_promise_new to simplify creation of Eina_Promise.
Cedric BAIL [Wed, 3 Jan 2018 20:12:48 +0000 (12:12 -0800)]
ecore: introduce efl_loop_promise_new to simplify creation of Eina_Promise.

6 years agoecore: add some documentation for efl_loop_future_scheduler_get.
Cedric BAIL [Wed, 3 Jan 2018 19:37:38 +0000 (11:37 -0800)]
ecore: add some documentation for efl_loop_future_scheduler_get.

6 years agoecore: efl_loop_future_scheduler_get actually should be considered a const method.
Cedric BAIL [Wed, 3 Jan 2018 19:35:08 +0000 (11:35 -0800)]
ecore: efl_loop_future_scheduler_get actually should be considered a const method.

6 years agoRevert "efl_loop: move scheduler_get to eo API"
Cedric BAIL [Wed, 3 Jan 2018 19:21:34 +0000 (11:21 -0800)]
Revert "efl_loop: move scheduler_get to eo API"

This reverts commit f910ba248e3f8f8390674e79cbbe49582eed861e.

The scheduler is meant to be used only in C, not by bindings so there isn't really
a use for it in the loop class. Now this patch was triggered due to complexity in
using future/promise, so will do a follow up patch to improve that.

6 years agoci: add forgotten Ubuntu 17.10 Dockerfile
Stefan Schmidt [Wed, 3 Jan 2018 15:57:50 +0000 (16:57 +0100)]
ci: add forgotten Ubuntu 17.10 Dockerfile

Oops, sorry for the Travis break due to this.

6 years agoPut in a new (experimental) doc generator built only by templates
Dave Andreoli [Wed, 3 Jan 2018 15:18:58 +0000 (16:18 +0100)]
Put in a new (experimental) doc generator built only by templates

This is (in my mind) meant to replace the current elua generator.

Currently the generated output is pratically identical to the elua
one, just some little difference here and there, some for thecnical
reasons and some just for my preference.

I consider this work just a starting point, extending the
templates we can now easily improve our docs. Whithout the need
to touch a single line of code.

Really I think this is a great improvements, and this are some
numbers to prove it:

Current elua implementation:
4185 lines of code in 7 lua files
generation time: ~ 7 seconds

New generator:
115 lines of python + 513 lines of templates
generation time: ~ 8 seconds (can be optimizd ALOT)

To generate the full Efl.* docs just run "./gendoc.py -v" in this folder.

...will wait for reviews (in particular from @andy and @q66)

6 years agodocs: Fix eo_add -> efl_add
Andy Williams [Wed, 3 Jan 2018 14:44:39 +0000 (14:44 +0000)]
docs: Fix eo_add -> efl_add

6 years agoci: add Archlinux Dockerfile and use it in Travis
Stefan Schmidt [Wed, 3 Jan 2018 14:13:26 +0000 (15:13 +0100)]
ci: add Archlinux Dockerfile and use it in Travis

6 years agoci: update travis distro builds for Ubuntu and Fedora to newer versions
Stefan Schmidt [Wed, 3 Jan 2018 11:22:15 +0000 (12:22 +0100)]
ci: update travis distro builds for Ubuntu and Fedora to newer versions

We already have the needed Dockerfiles for the latest releases and they
work fine. Make sure Travis runs use them.

6 years agoPyolian: quite complete documentation for usage and template syntax
Dave Andreoli [Wed, 3 Jan 2018 12:58:26 +0000 (13:58 +0100)]
Pyolian: quite complete documentation for usage and template syntax

The same content is available in phab wiki at:
https://phab.enlightenment.org/w/pyolian/

6 years agoefl_loop: move scheduler_get to eo API
Andy Williams [Wed, 3 Jan 2018 12:46:06 +0000 (12:46 +0000)]
efl_loop: move scheduler_get to eo API

6 years agoTypo in error message
Xavi Artigas [Tue, 2 Jan 2018 12:54:52 +0000 (13:54 +0100)]
Typo in error message

6 years agoTypo in example
Xavi Artigas [Fri, 22 Dec 2017 11:31:09 +0000 (12:31 +0100)]
Typo in example

6 years agodocs: Fix exmplanation for edje_file_iterator
Andy Williams [Wed, 3 Jan 2018 09:19:54 +0000 (09:19 +0000)]
docs: Fix exmplanation for edje_file_iterator

6 years agoeina debug - use pthread _sigmask instead of sigprocmask like elsewhere
Carsten Haitzler (Rasterman) [Wed, 3 Jan 2018 04:02:39 +0000 (13:02 +0900)]
eina debug - use pthread _sigmask instead of sigprocmask like elsewhere

elsewhere in efl we moved to pthread_sigmask but eina debug didn't, so
mirror the changes here too. at this point in time when we are
initting eina debug this shouldnt really matter much as we're single
threaded until this pthread_Create is called. after that tough...
we're not. signals + threads is a nightmare though... horrible
horrible...

6 years agoPyolian: fix segfault on shutdown
Dave Andreoli [Tue, 2 Jan 2018 20:52:57 +0000 (21:52 +0100)]
Pyolian: fix segfault on shutdown

also improve abit all_namespaces performance by using a set (hashable)

6 years agoPyolian: implement an utility class: Namspace
Dave Andreoli [Tue, 2 Jan 2018 14:43:15 +0000 (15:43 +0100)]
Pyolian: implement an utility class: Namspace

Update tests and generator to use this new feature

6 years agoecore_con: switch from 1234 to AF_UNSPEC
Marcel Hollerbach [Sat, 30 Dec 2017 10:30:38 +0000 (10:30 +0000)]
ecore_con: switch from 1234 to AF_UNSPEC

there are systems where sa_family_t is a enum, so assigning 1234 will
lead to compililation errors on picky systems

6 years agoci: travis: add new docker file for Fedora 27 release
Stefan Schmidt [Fri, 22 Dec 2017 15:05:10 +0000 (16:05 +0100)]
ci: travis: add new docker file for Fedora 27 release

6 years agobenchmarks: eine: include header for using time()
Stefan Schmidt [Fri, 22 Dec 2017 14:57:55 +0000 (15:57 +0100)]
benchmarks: eine: include header for using time()

../src/benchmarks/eina/eina_bench_sort.c: In function ‘eina_bench_sort_eina’:
../src/benchmarks/eina/eina_bench_sort.c:52:10: warning: implicit declaration of function ‘time’ [-Wimplicit-function-declaration]
    srand(time(NULL));

Found due to the nice quite build output in our meson feature branch.

6 years agoPyolian: implemented doc tokenizer
Dave Andreoli [Tue, 2 Jan 2018 12:57:22 +0000 (13:57 +0100)]
Pyolian: implemented doc tokenizer

6 years agoefl loop - rename ecore_main_loop_get to efl_main_loop_get
Carsten Haitzler (Rasterman) [Tue, 2 Jan 2018 06:41:46 +0000 (15:41 +0900)]
efl loop - rename ecore_main_loop_get to efl_main_loop_get

ecore_main_loop_get() is really a new "eo api" but it's using our old
ecore_* namespace, so move to the new efl namespace.

6 years agoEfl.Gfx.Color: add color_class_code.set/get{} method
Amitesh Singh [Fri, 29 Dec 2017 04:06:44 +0000 (13:06 +0900)]
Efl.Gfx.Color: add color_class_code.set/get{} method

6 years agoefl loop - make param as unused to remove warning
Carsten Haitzler (Rasterman) [Tue, 2 Jan 2018 06:07:12 +0000 (15:07 +0900)]
efl loop - make param as unused to remove warning

6 years agoPyolian: better equality test, new hierarchy prop and some tests
Dave Andreoli [Mon, 1 Jan 2018 07:35:39 +0000 (08:35 +0100)]
Pyolian: better equality test, new hierarchy prop and some tests

6 years agoPyolian: a fix and some new utils
Dave Andreoli [Mon, 1 Jan 2018 06:27:15 +0000 (07:27 +0100)]
Pyolian: a fix and some new utils

6 years agoPyolian generator: also make the classes visible in the template ctx
Dave Andreoli [Mon, 1 Jan 2018 06:01:10 +0000 (07:01 +0100)]
Pyolian generator: also make the classes visible in the template ctx

6 years agopyolian: make the library and the generator importable from another folder
Dave Andreoli [Fri, 29 Dec 2017 08:27:49 +0000 (09:27 +0100)]
pyolian: make the library and the generator importable from another folder

6 years agoFix and simpligy windows behaviour.
Andy Williams [Sun, 31 Dec 2017 19:57:48 +0000 (19:57 +0000)]
Fix and simpligy windows behaviour.

Author: Vincent Torri

6 years agoFix and simpligy windows behaviour.
Andy Williams [Sun, 31 Dec 2017 16:08:31 +0000 (16:08 +0000)]
Fix and simpligy windows behaviour.

Author: Vincent Torri

6 years agoelm_code: Fix usage of basename
Andy Williams [Sun, 31 Dec 2017 15:28:23 +0000 (15:28 +0000)]
elm_code: Fix usage of basename

6 years agoelm: fix memleak in combobox
Jean Guyomarc'h [Sat, 30 Dec 2017 09:14:12 +0000 (10:14 +0100)]
elm: fix memleak in combobox

@fix

6 years agoelm: fix typo in elm_list documentation
Jean Guyomarc'h [Sat, 30 Dec 2017 09:13:54 +0000 (10:13 +0100)]
elm: fix typo in elm_list documentation

6 years agoPyolian template-based generator.
Dave Andreoli [Thu, 28 Dec 2017 15:10:04 +0000 (16:10 +0100)]
Pyolian template-based generator.

This is a really powerfull tool that can be used to generate anything eolian
releted  just providing a template file. You can then render the template
with the wanted scope (class, namespace, enum, etc)

For example give a try at this (from the src/srcipts/pyolian folder):
./generator.py test_gen_class.template --cls Efl.Loop.Timer

or ./generator.py -h for the full help

Next step: maybe generate the new efl API doc using this tool?
@andy I think this will make your life much easier :)

6 years agoNew python bindings for eolian
Dave Andreoli [Thu, 28 Dec 2017 12:49:03 +0000 (13:49 +0100)]
New python bindings for eolian

This are manually written ctype bindings for eolian, that means they
run on the standard python library (nothing to install) and can run
without any build step (in fact ctypes just open the so/dll file at runtime)

Next step will be (soon) a template based generator for eolian that will
be a lot of fun :)

6 years agoremove elgacy ecore event usage in futures that limit to mainloop only
Carsten Haitzler (Rasterman) [Wed, 27 Dec 2017 16:19:52 +0000 (01:19 +0900)]
remove elgacy ecore event usage in futures that limit to mainloop only

also eina_procmis was not threadsafe so cannto use loops in different
threads at all until this was made safe. needed to disable the old
ecore_event using code in for ecore futures and create a new efl loop
message future and handler instead ... but now a quick experiment with
multiple loops in 10 threads plus mainloop have timers at least work.
i need to test more like fd handlers etc etc. but it's a step.

6 years agoEfl_Ui.h - remember to define EWAPI too
Carsten Haitzler (Rasterman) [Wed, 27 Dec 2017 05:23:28 +0000 (14:23 +0900)]
Efl_Ui.h - remember to define EWAPI too

6 years agoelm_priv.h - improve defines, EAPI EWAPI etc. because sub-includes need
Carsten Haitzler (Rasterman) [Wed, 27 Dec 2017 05:22:42 +0000 (14:22 +0900)]
elm_priv.h - improve defines, EAPI EWAPI etc. because sub-includes need

this also makes it build for windows

6 years agoecore/efl loop - use efl_loop_time_get on loop obj where needed
Carsten Haitzler (Rasterman) [Tue, 26 Dec 2017 16:40:07 +0000 (01:40 +0900)]
ecore/efl loop  - use efl_loop_time_get on loop obj where needed

stop using the legacy ecore_loop_time_get() func when it should be
coming from the loop object's loop time. also ecore_time_get should
never fall back on ecore_loop_time_get for similar reasons.

part of making the ecore/efl loop a non-global instance (allow loops
in threads)

6 years agoecore main loop - fix fd handler clearning
Carsten Haitzler (Rasterman) [Sat, 23 Dec 2017 12:07:23 +0000 (21:07 +0900)]
ecore main loop - fix fd handler clearning

so loop object destruction was clearing out fd handlers but those may
be later deleted by destructors of child objects. so leave legacy
fdh's and just remove them from the list

6 years agodocs: Group our API landing page by namespace
Andy Williams [Fri, 22 Dec 2017 17:30:52 +0000 (17:30 +0000)]
docs: Group our API landing page by namespace

6 years agoecore: Reduce the number of lines in our error log
Andy Williams [Fri, 22 Dec 2017 14:48:28 +0000 (14:48 +0000)]
ecore: Reduce the number of lines in our error log

And align for readability

6 years agoeolian: prevent unnecessary hash table allocs
Daniel Kolesa [Fri, 22 Dec 2017 14:34:53 +0000 (15:34 +0100)]
eolian: prevent unnecessary hash table allocs

6 years agoeolian: remove the toplevel information
Daniel Kolesa [Fri, 22 Dec 2017 14:29:23 +0000 (15:29 +0100)]
eolian: remove the toplevel information

6 years agoeolian: perform partial class validation every time
Daniel Kolesa [Fri, 22 Dec 2017 14:26:36 +0000 (15:26 +0100)]
eolian: perform partial class validation every time

This makes sure that duplicate method/part/etc checks are done on
every database update, removing the need for clunky toplevel
checks and improving reliability. It also sacrifices some
performance but it shouldn't be too bad (if a class is already
validated, some checks are avoided to speed things up).

6 years agoeolian: deferred parsing of type deps
Daniel Kolesa [Fri, 22 Dec 2017 11:53:51 +0000 (12:53 +0100)]
eolian: deferred parsing of type deps

6 years agoeolian: move inherit type checking to validation
Daniel Kolesa [Fri, 22 Dec 2017 11:41:29 +0000 (12:41 +0100)]
eolian: move inherit type checking to validation

6 years agodocs: Update evas canvas eo files for grammar and readability
Andy Williams [Fri, 22 Dec 2017 13:45:48 +0000 (13:45 +0000)]
docs: Update evas canvas eo files for grammar and readability

Author: Nate Drake

6 years agoRemove all ocurrences of SPANK from our user facing errors.
Andy Williams [Fri, 22 Dec 2017 11:18:46 +0000 (11:18 +0000)]
Remove all ocurrences of SPANK from our user facing errors.

This has been bugging me for some time but now we are triggering new errors internally
this is appearing to end users for problems they did not cause.

Additionally I was able to improve a couple of the errors by copying the
explanation from code comments into the error message.

Shorter error logs now too :)

6 years agodocs: Update efl interface eo files for grammar and readability
Andy Williams [Fri, 22 Dec 2017 10:41:11 +0000 (10:41 +0000)]
docs: Update efl interface eo files for grammar and readability

Author: Nate Drake

6 years agodocs: Update edje eo files for grammar and readability
Andy Williams [Fri, 22 Dec 2017 10:30:32 +0000 (10:30 +0000)]
docs: Update edje eo files for grammar and readability

Author: Nate Drake

6 years agoEfl.Ui.List: fix compiler warning.
Amitesh Singh [Fri, 22 Dec 2017 04:43:30 +0000 (13:43 +0900)]
Efl.Ui.List: fix compiler warning.

6 years agoEfl.Gfx.Color: add color_code{}
Amitesh Singh [Thu, 21 Dec 2017 05:48:41 +0000 (14:48 +0900)]
Efl.Gfx.Color: add color_code{}

color_code allows user to pass/get the color hex string
(e.g. efl_gfx_color_code_set(o, "#FFAA22CC"))

Also make this interface as mixin class.

6 years agointerface: add Efl.Gfx.Color interface for color related APIs
Amitesh Singh [Wed, 20 Dec 2017 07:34:01 +0000 (16:34 +0900)]
interface: add Efl.Gfx.Color interface for color related APIs

This would eventually turn into mixin class later after adding
more color related helper functions.

6 years agoelm_code: Cancel selection if clicked outside of the editor area
Andy Williams [Thu, 21 Dec 2017 22:05:28 +0000 (22:05 +0000)]
elm_code: Cancel selection if clicked outside of the editor area

6 years agoeolian: unify test directories
Marcel Hollerbach [Thu, 21 Dec 2017 19:31:39 +0000 (20:31 +0100)]
eolian: unify test directories

6 years agoecore/efl loop - refactor idle stuff to be less convluted when
Carsten Haitzler (Rasterman) [Thu, 21 Dec 2017 10:31:24 +0000 (19:31 +0900)]
ecore/efl loop - refactor idle stuff to be less convluted when

less jumping around the codebase and no need for a message exists
method on the loop as we can find out internally, so only the process
left.

6 years agoecore/efl loop. remove internal ecore_timer legacy api usage for eflloop
Carsten Haitzler (Rasterman) [Thu, 21 Dec 2017 07:34:05 +0000 (16:34 +0900)]
ecore/efl loop. remove internal ecore_timer legacy api usage for eflloop

efl.loop was still using legacy ecore_timer_* calls inside. of course
this is a big no-no if we are to allow multiple loops, so clean this
up and convert them to efl.loop.timers.

6 years agoefl-mono: Fix using right description pointer in event registration
Felipe Magno de Almeida [Wed, 20 Dec 2017 21:45:07 +0000 (19:45 -0200)]
efl-mono: Fix using right description pointer in event registration

6 years agoefl-mono: Fix removal of .exe extension in Linux systems
Felipe Magno de Almeida [Wed, 20 Dec 2017 21:44:22 +0000 (19:44 -0200)]
efl-mono: Fix removal of .exe extension in Linux systems

6 years agoeolian-mono: Add documentation generation support
Lauro Moura [Thu, 7 Dec 2017 00:03:55 +0000 (21:03 -0300)]
eolian-mono: Add documentation generation support

This commit adds the "documentation" generator, which gets the
documentation_def attribute of the given item and generates xml comments
to be exported by MCS.

For items requiring some customization of the generated comments (e.g.
functions and its parameters), the helpers to generate the preamble
(summary), body (paragraphs) and epilogue (currently just the @since
tag) were added.

Currently we do not support converting Eolian references into xmldoc
references.

As we explicitly generate Get/Set methods for properties, for now the
generator tries to get the get/set specific documentation first. If it
is not present, fallback to the common docs.

Later this could be changed to generate the common one as paragraphs of
the Get/Set.

Also some generated code like the wrappers for calling C# methods
from C can be private. This will cleanup the introspection results
and warnings when generating documentation.

Due to this visibility change, the binbuf tests had to be changed
to add redirect calls to the native methods instead of directly
calling the DllImport'd methods.

6 years agoeina: only execute that code if HAVE_MALLOC_USABLE_SIZE is defined
Marcel Hollerbach [Wed, 20 Dec 2017 17:50:17 +0000 (18:50 +0100)]
eina: only execute that code if HAVE_MALLOC_USABLE_SIZE is defined

thats why we check that ...

6 years agoCanvas image: fix endless call to 'load_error_get'
Daniel Hirt [Wed, 20 Dec 2017 14:40:23 +0000 (16:40 +0200)]
Canvas image: fix endless call to 'load_error_get'

Fixes a bug introduce with c8c4572d7005d19cf50f8c53bb7408606cf3ee40.
My guess is that the true intention was to query the Efl.File mixin
first, and the check the specific load_error.

Basically a call to 'efl_file_load_error_get' on an
Efl.Canvas.Image triggered this.

6 years agodocs: Update ecore_con eo files for grammar and readability
Andy Williams [Wed, 20 Dec 2017 17:30:29 +0000 (17:30 +0000)]
docs: Update ecore_con eo files for grammar and readability

Author: Nate Drake

6 years agodocs: Output correct signature for class method first parameter
Andy Williams [Wed, 20 Dec 2017 15:00:18 +0000 (15:00 +0000)]
docs: Output correct signature for class method first parameter

6 years agodocs: Note class scope for method listings
Andy Williams [Wed, 20 Dec 2017 14:33:54 +0000 (14:33 +0000)]
docs: Note class scope for method listings

6 years agodocs: Correct sig for eolian lua calls
Andy Williams [Wed, 20 Dec 2017 14:31:56 +0000 (14:31 +0000)]
docs: Correct sig for eolian lua calls

6 years agoelm - scroll manager - fix up momentum animator code to use real consts
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 12:49:18 +0000 (21:49 +0900)]
elm - scroll manager - fix up momentum animator code to use real consts

use real constnants and remove useless vasriables as well as fix up
formatting... so we don't get compilers complaining about non-const
values etc.

@fix

6 years agoecore_con: bug workaround SO_REUSEADDR and EADDRINUSE from bind (fix)
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 12:10:53 +0000 (21:10 +0900)]
ecore_con: bug workaround SO_REUSEADDR and EADDRINUSE from bind (fix)

what i'm seeing is this with local unix sockets:

1. server process not cleanly shut down (kill -9 for example).
2. run server process again and bind fails due to EADDRINUSE
3. we ARE doing setsockopt() with SO_REUSEADDR set to 1 ...

this just makes no sense because setsockopt() SHOULD allow use to
re-use... the previous efreetd process for example is gone. no such
process, yet socket is not re-usable. this should just not happen due
to SO_REUSEADDR, but it does. this has nasty consequences like efreetd
maybe never running because of stale sockets. this should never have
happened, but it does. odd. so a hacky workaround:

1. try bind.
2. if bind fails with EADDRINUSE and its a socket path AND
pd->unlink_before_bind is NOT set... then try a connect to the socket.
3. if connect succeeds then fail as normal (close socket and error on
bind'ing)
   if connect fails then we have a stale socket, so unlink it
forcibly. create the socket again and try bind again.

hacky but... fixes the core issue.

@fix

6 years agoelm build for osx - try another dotting of i's and crossing of t's
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 09:54:14 +0000 (18:54 +0900)]
elm build for osx - try another dotting of i's and crossing of t's

try this and fix T6448 ...

6 years agoeo: Update grammar and readability
Andy Williams [Wed, 20 Dec 2017 09:52:31 +0000 (09:52 +0000)]
eo: Update grammar and readability

6 years agoEo docs: update for grammar and readability
Andy Williams [Wed, 20 Dec 2017 09:41:41 +0000 (09:41 +0000)]
Eo docs: update for grammar and readability

6 years agoosx build - try again
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 09:04:05 +0000 (18:04 +0900)]
osx build - try again

try fix T6448 again

6 years agotry and fix osx build with symbols missing...
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 08:25:11 +0000 (17:25 +0900)]
try and fix osx build with symbols missing...

try and fix T6448 last comment. not sure if it does...

6 years agoremove all specific including of elm headers inside out src tree
Carsten Haitzler (Rasterman) [Wed, 20 Dec 2017 00:29:50 +0000 (09:29 +0900)]
remove all specific including of elm headers inside out src tree

... other than Elementary.h after d6e3ce3df422751cb1c4524f1ea0b16ce897410 etc.