This reverts commit 7a98f617e9.
this commit breaks compiling things against efl badly. i'm on holiday
and haver only a mini screen and keyboard wwith me so any serious wok
in hunting these issues isn't going to be fun or easy. as this causes
bad build brreaks this is worth a revert IMHO. please re-submit wwhen
you've tested against efl by building things against it. also our
public headers just should not have such ifdefs/if's that change api
presented based on how efl is cnfigured. we expose the same api and
macros and types regardless of internal config.
Summary:
Realization of audio data playback through WASAPI (Windows Audio Session API).(minimal)
WASAPI model:
1. Find a playback device (in the default system)
2. Register the client on the playback device.
3. Create a playback format for audio data.
4. Initialize the client with the created format, and access mode, ....
5. Take the object-rendering from the client for play the data stream.
6. Play data stream
ecore_audio model:
1. create a object for play the data stream(out object)
2. create a object to receive the data stream(in object)
3. register in the out-object the in-object
4. play data stream
Necessary:
Realize the ecore_audio object to play the data stream using the WASAPI model.
How implemented:
1. The object ecore_audio_out_wasapi is implemented
2. object ecore_audio_out_wasapi - the object constructor is find a playback device (in the default system)(WASAPI)
3. _ecore_audio_out_wasapi_ecore_audio_out_input_attach - register in the out-object the in-object
3.1 Register the client on the playback device.(WASAPI)
3.2 Create a playback format for audio data.(WASAPI)
3.3 Initialize the client with the created format, and access mode, ....(WASAPI)
4. _write_cb - play data
4.1 Take the object-rendering from the client for play the data stream.(WASAPI)
4.2 Play data stream(WASAPI)
Reviewers: cedric, vtorri, raster, an.kroitor, NikaWhite, FurryMyad, rimmed, t.naumenko, Jaehyun, bowonryu
Reviewed By: vtorri, NikaWhite
Subscribers: artem.popov, cedric, jpeg
Tags: #windows
Differential Revision: https://phab.enlightenment.org/D5029
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
This backend has received no patch and maintenance from anyone who could
actually test it over the last few years. After talking with KaKaRoTo it
is best to remove it. If anyone want to take over its maintenance, you
are welcome to revert this patch.
the previous method of forcing this to be enabled for dist builds caused
breaks when the original configure disabled examples, as the little-known
DISTCHECK_CONFIGURE_FLAGS variable would need to also be set to disable
examples even though the user would think they were disabled based on configure
output
this would previously break if:
* cxx bindings were disabled
* elua was disabled
* base configure disabled examples and dist build disabled examples
* base configure disabled examples and dist build enabled examples
it still breaks if:
* base configure disables examples and dist build enables examples
This check for drmModeAtomicCommit is no longer necessary as we have
moved to using static_libs/libdrm during compile time.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
Well, the build is broken again on BSD and Windows.
efl_debugd is full of lots lovely Linuxisms. Just
don't compile.
This unbreaks the build for FreeBSD and others.
Reviewers: cedric, raster, jpeg
Reviewed By: jpeg
Subscribers: JackDanielZ, jpeg
Differential Revision: https://phab.enlightenment.org/D4946
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Changing ck_assert_ptr_nonnull() to ck_assert_ptr_ne() in order to
require "only" check >= 0.9.10. ck_assert_ptr_nonnull() was
introduced in 0.11.0. ck_assert_ptr_ne() is already used a lot
in the test suite so a recent version of check is required.
For this one I feel like I earned an archaeologist medal. :-)
This is related to the commit raster made in aa92cddb to accommodate for older
freetype version. The TT_INTERPRETER_VERSION_35 define was actually introduced
in the release 2.5.0 which was released in June 2013 which means it had almost
4 years to get picked up. I would consider this safe to be a minimum dependency
for current efl versions.
The trick here is that the release version does not match the library version
spit out by pkg-config. 16.2.10 equals the freetype 2.5.0.1 release in this case.
Better make a note about it in configure for the next poor soul trying to
understand this.
Thanks again to llelectronics for reporting this.
Fix T5437
@fix
This releases offers the gnutls_certificate_set_x509_trust_dir() API we use in
ecore_con. This gnutls version was release in July 2014 which gave it enough
time to be picked up by distributions.
Thanks a lot to llelectronics for reporting and tracking down the problem.
ref T5437
@fix
We have a report where the use of epoll breaks such systems. Disabling it for
now to make them work again. A deeper analysis is underway to understand this
better and maybe have epoll support later.
Not sure _ecore_fd_valid() is all that useful anymore, as the
commit that introduced it said it would be removed "before release"
a long time ago - it's a debug assist that probably doesn't need
to be in release builds.
(I'm counting syscalls on rpi3 - still, calling this an optimization
seems like a bit of a stretch.)