In addition to lots of bug fixes, here are the major changes in this release:
General:
New version numbering scheme, similar to GLib and Flatpak.
An even number in the minor version (second component) indicates a production-ready stable release such as 2.24.0, which would have been 2.0.24 under the old system.
The patchlevel (micro version, third component) indicates a bugfix-only update: for example, 2.24.1 would be a bugfix-only release to fix bugs in 2.24.0, without adding new features.
An odd number in the minor version indicates a prerelease such as 2.23.0. Stable distributions should not use these prereleases.
The patchlevel indicates successive prereleases, for example 2.23.1 and 2.23.2 would be prereleases during development of the SDL 2.24.0 stable release.
Added SDL_GetPointDisplayIndex() and SDL_GetRectDisplayIndex() to get the display associated with a point and rectangle in screen space
Added SDL_bsearch(), SDL_crc16(), and SDL_utf8strnlen() to the stdlib routines
Added SDL_CPUPauseInstruction() as a macro in SDL_atomic.h
Added SDL_size_mul_overflow() and SDL_size_add_overflow() for better size overflow protection
Added SDL_ResetHint() to reset a hint to the default value
Added SDL_ResetKeyboard() to reset SDL's internal keyboard state, generating key up events for all currently pressed keys
Added the hint SDL_HINT_MOUSE_RELATIVE_WARP_MOTION to control whether mouse warping generates motion events in relative mode. This hint defaults off.
Added the hint SDL_HINT_TRACKPAD_IS_TOUCH_ONLY to control whether trackpads are treated as touch devices or mice. By default touchpads are treated as mouse input.
The hint SDL_HINT_JOYSTICK_HIDAPI_JOY_CONS now defaults on
Added support for mini-gamepad mode for Nintendo Joy-Con controllers using the HIDAPI driver
Added the hint SDL_HINT_JOYSTICK_HIDAPI_COMBINE_JOY_CONS to control whether Joy-Con controllers are automatically merged into a unified gamepad when using the HIDAPI driver. This hint defaults on.
The hint SDL_HINT_JOYSTICK_HIDAPI_SWITCH_HOME_LED can be set to a floating point value to set the brightness of the Home LED on Nintendo Switch controllers
Added the hint SDL_HINT_JOYSTICK_HIDAPI_JOYCON_HOME_LED to set the Home LED brightness for the Nintendo Joy-Con controllers. By default the Home LED is not modified.
Added the hint SDL_HINT_JOYSTICK_HIDAPI_SWITCH_PLAYER_LED to control whether the player LED should be lit on the Nintendo Joy-Con controllers
Added support for Nintendo Online classic controllers using the HIDAPI driver
Added the hint SDL_HINT_JOYSTICK_HIDAPI_NINTENDO_CLASSIC to control whether the HIDAPI driver for Nintendo Online classic controllers should be used
Added support for the NVIDIA Shield Controller to the HIDAPI driver, supporting rumble and battery status
Added support for NVIDIA SHIELD controller to the HIDAPI driver, and a hint SDL_HINT_JOYSTICK_HIDAPI_SHIELD to control whether this is used
Added functions to get the platform dependent name for a joystick or game controller:
SDL_JoystickPathForIndex()
SDL_JoystickPath()
SDL_GameControllerPathForIndex()
SDL_GameControllerPath()
Added SDL_GameControllerGetFirmwareVersion() and SDL_JoystickGetFirmwareVersion(), currently implemented for DualSense(tm) Wireless Controllers using HIDAPI
Added SDL_JoystickAttachVirtualEx() for extended virtual controller support
Added joystick event SDL_JOYBATTERYUPDATED for when battery status changes
Added SDL_GUIDToString() and SDL_GUIDFromString() to convert between SDL GUID and string
Added SDL_HasLSX() and SDL_HasLASX() to detect LoongArch SIMD support
Added SDL_GetOriginalMemoryFunctions()
Added SDL_GetDefaultAudioInfo() to get the name and format of the default audio device, currently implemented for PipeWire, PulseAudio, WASAPI, and DirectSound
Added HIDAPI driver for the NVIDIA SHIELD controller (2017 model) to enable support for battery status and rumble
Added support for opening audio devices with 3 or 5 channels (2.1, 4.1). All channel counts from Mono to 7.1 are now supported.
Rewrote audio channel converters used by SDL_AudioCVT, based on the channel matrix coefficients used as the default for FAudio voices
SDL log messages are no longer limited to 4K and can be any length
Fixed a long-standing calling convention issue with dynapi affecting OpenWatcom or OS/2 builds
Windows:
Added initial support for building for Windows and Xbox with Microsoft's Game Development Kit (GDK), see docs/README-gdk.md for details
Added a D3D12 renderer implementation and SDL_RenderGetD3D12Device() to retrieve the D3D12 device associated with it
Added the hint SDL_HINT_WINDOWS_DPI_AWARENESS to set whether the application is DPI-aware. This hint must be set before initializing the video subsystem
Added the hint SDL_HINT_WINDOWS_DPI_SCALING to control whether the SDL coordinates are in DPI-scaled points or pixels
Added the hint SDL_HINT_DIRECTINPUT_ENABLED to control whether the DirectInput driver should be used
Added support for SDL_GetAudioDeviceSpec to the DirectSound backend
Linux:
Support for XVidMode has been removed, mode changes are only supported using the XRandR extension
Added the hint SDL_HINT_VIDEO_WAYLAND_MODE_EMULATION to control whether to expose a set of emulated modes in addition to the native resolution modes available on Wayland
Added the hint SDL_HINT_KMSDRM_DEVICE_INDEX to specify which KMSDRM device to use if the default is not desired
Added the hint SDL_HINT_LINUX_DIGITAL_HATS to control whether to treat hats as digital rather than checking to see if they may be analog
Added the hint SDL_HINT_LINUX_HAT_DEADZONES to control whether to use deadzones on analog hats
macOS:
Bumped minimum OS deployment version to macOS 10.9
Added SDL_GL_FLOATBUFFERS to allow Cocoa GL contexts to use EDR
Added the hint SDL_HINT_MAC_OPENGL_ASYNC_DISPATCH to control whether dispatching OpenGL context updates should block the dispatching thread until the main thread finishes processing. This hint defaults to blocking, which is the safer option on modern macOS.
(From OE-Core rev: f54faf3df0c1252b965c96ec6dbf71574a01e9a3)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building libsdl2-native, and I met a do_configure error as below.
| CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
| Please set them or make sure they are set and tested correctly in the CMake files:
| XINERAMA_LIB
This error appears on hosts with libxinerama-dev installed.
Looking at the CMakeLists.txt file, I got:
set(SDL_X11_OPTIONS Xcursor Xdbe Xinerama XInput Xfixes Xrandr Xscrnsaver XShape Xvm
We an see neither Xinerama nor Xvm is removed. So we should add back
these options removed by a previous commit.
(From OE-Core rev: 2c5d78153614fcdf30dbe3e41d88f18e9ae06684)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Support for XVM and Xinerama has been removed
(From OE-Core rev: 09092685fdc8ffe6384bea441394ce02d6ff7e66)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
by default libsdl2 is build with code from src/video/yuv2rgb, which
is licensed under BSD-2-Clause.
Additional by default hidapi is build, which is licensed under
GPL3 | BSD-2-Clause | HIDAPI license, pick the least restrictive
and best matching BSD-2-Clause.
(From OE-Core rev: a97f7975edaf0ed833950f87c3d880a8897c67df)
Signed-off-by: Konrad Weihmann <kweihmann@outlook.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libunwind-native currently cannot be added to the class-native DEPENDS
due to the following problem:
https://autobuilder.yoctoproject.org/typhoon/#/builders/91/builds/5624/steps/12/logs/stdio
And without libunwind-native, libsdl2-native config fails errors like:
| -- Checking for one of the modules 'libunwind'
| CMake Error at [...]/build/tmp/work/x86_64-linux/libsdl2-native/2.0.20-r0/recipe-sysroot-native/usr/share/cmake-3.22/Modules/FindPkgConfig.cmake:890 (message):
| None of the required 'libunwind' found
| Call Stack (most recent call first):
| CMakeLists.txt:1367 (pkg_search_module)
Ideally, the libunwind issue would be fixed, and then this could be added
to the libsdl2 recipe:
DEPENDS:append:class-native = " libunwind-native"
However, the libunwind problem is not trivial to fix, and requires a
specific setup to reproduce. As an alternative workaround, turn off
libunwind support in the native SDL2 build by disabling the unwind.h
check in the CMake script (this subsequently turns off all the other
libunwind checks). The main user of libsdl2-native is qemu-native, which
does not need libunwind support in SDL2.
(From OE-Core rev: ecd044b09f8715353cbf20b39ce2b1fe9e620631)
Signed-off-by: Carlos Rafael Giani <crg7475@mailbox.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
virtual/libgl may not be implemented by some non-mesa graphics in
embedded space, this ensures that libsdl2 can build on such stacks
(From OE-Core rev: 26f511bb21c01a371722caa739f4f6d079211cc2)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit 1b114fa2d221c9810852526f8b45a3d2c856a502.
Unfortunately this breaks on cento7 as libunwind-native doesn't appear to build there.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fixes this CMake configuration error:
| -- Checking for one of the modules 'libunwind'
| CMake Error at [...]/build/tmp/work/x86_64-linux/libsdl2-native/2.0.20-r0/recipe-sysroot-native/usr/share/cmake-3.22/Modules/FindPkgConfig.cmake:890 (message):
| None of the required 'libunwind' found
| Call Stack (most recent call first):
| CMakeLists.txt:1367 (pkg_search_module)
This error happens even if libunwind is installed complete with its C
headers and development .so files (in Ubuntu, this means libunwind-dev
is installed).
(From OE-Core rev: 1b114fa2d221c9810852526f8b45a3d2c856a502)
Signed-off-by: Carlos Rafael Giani <crg7475@mailbox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
0001-Fix-build-against-wayland-1.20.patch
removed since it is included in 2.0.20
License-Update: year updated to 2022.
Changelog:
=========
General:
SDL_RenderGeometryRaw() takes a pointer to SDL_Color, not int.
You can cast color data in SDL_PIXELFORMAT_RGBA32 format
(SDL_PIXELFORMAT_ABGR8888 on little endian systems) for this parameter.
Improved accuracy of horizontal and vertical line drawing when using OpenGL
or OpenGLES
Added the hint SDL_HINT_RENDER_LINE_METHOD to control the method of line
drawing used, to select speed, correctness, and compatibility.
Fixed hotplug controller detection, broken in 2.0.18
(From OE-Core rev: 21de64185beb799aee04dc12ae84e3a60cc671a3)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
clang provides libunwind.h and cmake adds a check to
find libunwind when this header is detected, which was
not the case with automake. The check however is expecting
specific unwinder implementation which provides libunwind-generic
solib, this is not a standard library that all implementations
will provide, therefore make this check optional.
(From OE-Core rev: fb450807774d100b9b568364b014ac46f5642b7e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We're seeing various failures where the X11 headers are found on the native
system but not all libraries are present and hence autoconfiguration of the
X11 subcomponents fails.
We don't list any of these X11 subcomponents as a dependency so disable
them by default. Configuration and dependencies can be added if people need
them.
(From OE-Core rev: f30ebc0b82b10f56f250a3a9c4f1f2fe9fb281b3)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Wayland 1.20.0 is released!
This release contains the following major changes:
- FreeBSD support has been entirely upstreamed and has been added to
our continuous integration system.
- The autotools build system has been dropped. Meson has replaced it.
- A few protocol additions: wl_surface.offset allows clients to update
a surface's buffer offset independently from the buffer,
wl_output.name and description allow clients to identify outputs
without depending on xdg-output-unstable-v1.
- In protocol definitions, events have a new "type" attribute and can
now be marked as destructors.
- A number of bug fixes, including a race condition when destroying
proxies in multi-threaded clients.
Also backport libsdl2 patch to fix building against wayland-1.20.0
(From OE-Core rev: d81c8d137895e1fb70bb56c61d53502670d7bc43)
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sdl-dlopen option is no longer supported.
(From OE-Core rev: 202c138a9b6905440deb2f86a3b650ae79da1bcd)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is the result of automated script conversion:
scripts/contrib/convert-overrides.py <oe-core directory>
converting the metadata to use ":" as the override character instead of "_".
(From OE-Core rev: 42344347be29f0997cc2f7636d9603b1fe1875ae)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Since oe-core 543e39ad "bitbake.conf: handle cmake -dev files packaging
with default rules" (June 2018) there's no need for recipes to add
${libdir}/cmake or ${datadir}/cmake to FILES_${PN}-dev themselves.
(From OE-Core rev: e6f62b8e639a79626d95568c070a410c24bce25e)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
libsdl2 currently does not handle the failures that
occur due to shared memory in case of remote hosts
where it should ideally switch to socket based writing
when a failure is seen with shared memory.
Consider a scenario where a qemu build is done with
graphics support on a build machine and then accessed
remotely using ssh. If the remote host's X presents
MIT-SHM as an extension, launching qemu fails with
runqemu - ERROR - Failed to run qemu: X Error: BadValue
Request Major code 130 (MIT-SHM)
Request Minor code 3 ()
This is most often seen when the remote machine doing
ssh is a Ubuntu 20.04. The libsdl2 native are mainly
used for qemu at this time so it is a major usecase.
A report of such a failure was also presented at
https://lists.yoctoproject.org/g/poky/topic/78854857
(From OE-Core rev: d12306a6331238a25db52816ad74a775c026d7da)
Signed-off-by: Awais Belal <awais_belal@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
PACKAGECONFIG[directfb] already adds directfb to the build
dependencies. But the automatically generated runtime
dependencies do not add the directfb package to the runtime
dependencies. Most likely this happens because libsdl2 does not link
against directfb but uses dlopen() or similar. Thus, the runtime
dependency to directfb needs to be declared explicitly.
(From OE-Core rev: 62d3cbe7c8261c1c875ff6da11572409262aaf02)
Signed-off-by: Mark Jonas <toertel@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When building on Raspbery Pi OS/64-bit, libsdl2 fails with
the error:
.../tmp/work/aarch64-linux/libsdl2-native/2.9.12-r0/SDL2-2.0.12/src/video/raspberry/SDL.rpivideo.h:47:5:
error: unknown type name 'EGL_DISPMANX_WINDOW_T'
EGL_DISPMANX_WINDOW_T dispman_window;
Disable video-rpi to avoid the problem since video suppport is
not needed for rpi and this configuration option is harmless when
building on x86-64 hosts.
(From OE-Core rev: 51009ed1b88aa391aae819fd800b907badc2be8c)
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Refactoring of SDL2 internal API has broken SDL_RenderFillRect for
DirectFB. The problem has already been fixed upstream.
(From OE-Core rev: a7c8dfc1f9beebeb9da7f61b323d85fba82ec1cb)
Signed-off-by: Mark Jonas <toertel@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build of libsdl2 with directfb is broken due to a spurious '}' and a
missing 'E' since version 2.0.12. The upstream is already fixed.
(From OE-Core rev: 8963daba093c3c5e2c60e1e4e057862971b84cb0)
Signed-off-by: Mark Jonas <toertel@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This allows virgl support in qemu with the SDL frontend
(From OE-Core rev: 0e0113c350e599f4da00eda384570e0db04f3b5d)
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* checked all hunks: backported patches can go
* for machines with neon in TUNE_FEATURES enable new configure option
--enable-arm-neon. If enabled, license must be extended to MIT
* license checksum changed by copyright year
(From OE-Core rev: aaa1c3ee0fe782fb08a58f100bc04686824de815)
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When you bitbake libsdl2-native and your host had installed fcitx
or ibus, it may cause you bitbake failed in do_compile.
Disable fcitx and ibus check with EXTRA_OECONF in bb file.
-------------snip---------------------
CC build/SDL_sysfilesystem.lo
CC build/SDL_syshaptic.lo
CC build/SDL_systimer.lo
/poky/build/tmp/work/x86_64-linux/libsdl2-native/2.0.10-r0/SDL2-2.0.10/src/core/linux/SDL_fcitx.c:52:5:
error: unknown type name ‘SDL_DBusContext’
SDL_DBusContext *dbus;
^~~~~~~~~~~~~~~
/poky/build/tmp/work/x86_64-linux/libsdl2-native/2.0.10-r0/SDL2-2.0.10/src/core/linux/SDL_fcitx.c:121:8:
error: unknown type name ‘DBusHandlerResult’
static DBusHandlerResult
^~~~~~~~~~~~~~~~~
/poky/build/tmp/work/x86_64-linux/libsdl2-native/2.0.10-r0/SDL2-2.0.10/src/core/linux/SDL_fcitx.c:122:20:
error: unknown type name ‘DBusConnection’;
------------snip----------------------
(From OE-Core rev: eb286172e413e9d6449dc13ccec1b78c04481eb5)
Signed-off-by: Yanfei Xu <yanfei.xu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixes a race where it tries to build version res file in build directory
before it has even been created. Prevents intermittent errors on
autobuilder when building for mingw:
| /bin/bash ../SDL2-2.0.10/build-scripts/updaterev.sh
| /bin/bash ../SDL2-2.0.10/build-scripts/mkinstalldirs build
| mkdir -p -- build
| x86_64-pokysdk-mingw32-windres --include-dir=/home/pokybuild/yocto-worker/meta-mingw/build/build/tmp/work/x86_64-nativesdk-mingw32-pokysdk-mingw32/nativesdk-libsdl2/2.0.10-r0/recipe-sysroot/opt/poky/3.0/sysroots/x86_64-pokysdk-mingw32/us
| /include ../SDL2-2.0.10/src/main/windows/version.rc build/version.o
| x86_64-pokysdk-mingw32-windres: build/version.o: No such file or directory
| Makefile:692: recipe for target 'build/version.o' failed
| make: *** [build/version.o] Error 1
| make: *** Waiting for unfinished jobs....
| touch build/.created
| WARNING: exit code 1 from a shell command.
(From OE-Core rev: 1cb3c64b70e0b16aabd2fe9cd6012564165ac476)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
License-Update: Change in Copyright Year.
Refresh the patch to apply on configure.ac instead of .in.
(From OE-Core rev: f00b9aad7c3e2e1ec41597539c4fe2307043b6cf)
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Just like libsdl, we can rip out the sdl2-config script from the sysroot as
pkg-config is preferrable.
(From OE-Core rev: 8d93fc236a72cb1bb0e713c4fc5b73e65ede6f53)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Drop 0001-GLES2-Get-sin-cos-out-of-vertex-shader.patch as it had been
merged upstream.
(From OE-Core rev: b46dbcad31c990b5556d61357e0a976948a5dede)
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Explicitly disable sndio to avoid inconsistent result on different
hosts having or not having libsndio.
This must be done if we take into consideration of eSDK.
I built eSDK on Ubuntu 14.04 which has libsndio installed, and then
installed the eSDK on Fedora 27, which does not have libsndio. In
fact, on Fedora 27, I even cannot find libsndio in its repo.
The problem happens when trying to use `devtool runqemu'. The qemu
binary built on Ubuntu 14.04 needs to link to libsndio, and thus
the following error.
runqemu - ERROR - Failed to run qemu: /path/to/qemu-system-x86_64:
error while loading shared libraries: libsndio.so.6.1:
cannot open shared object file: No such file or directory
So explicitly disable sndio for libsdl2 to avoid the above problem.
(From OE-Core rev: 72c37b06b034e97e7dd7e1a20fb0ff5f96088440)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The patch should increase performance for libsdl2 on GLES2 too.
(From OE-Core rev: 52f9659f2bb44affec2f67935df01f13b6ff3e02)
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There's an occasional build race from headers being generated in parallel with
other files which include the headers being compiled. Solve this by adding more
dependencies.
[ YOCTO #12815 ]
(From OE-Core rev: 177f4782e1ffca1eed3c9b102d910239a3dceea4)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These dependencies are in the x11 PACKAGECONFIG and will magically transform for
nativesdk builds.
(From OE-Core rev: 5e55efce4f7ecb3fcb219a15dcbc031f0f233cb1)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* Drop linkage.patch, this version already links .lo object.
* Remove backported patch fix-build-failure-on-ppc.patch
* Update LIC_FILES_CHKSUM since the year is changed.
(From OE-Core rev: 2702506f48a1b7e72e1786d24bd0113396dbc85d)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch modify the way we manage wayland path issues.
Instead of patching each recipe to make it work with wayland and its protocols,
it is better to patch wayland to fix its path issues.
So wayland-scanner.pc, wayland-client.pc and wayland-protocols.pc are patched to change paths.
Then we can drop the following workaround:
WAYLAND_PROTOCOLS_SYSROOT_DIR=${RECIPE_SYSROOT}
in:
- gtk+3
- libsdl2
- xserver-xorg
- gstreamer1.0-plugins-bad
- weston-2.0.0
We also dropped libsdl2 patches which fix wayland paths.
(From OE-Core rev: 14c0d992152ce27ee616558cafb408ed008d936e)
Signed-off-by: Fabien Lahoudere <fabien.lahoudere@collabora.co.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch is comparatively large and invasive. It does only do one thing, switching the
system to build using recipe specific sysroots and where changes could be isolated from it,
that has been done.
With the current single sysroot approach, its possible for software to find things which
aren't in their dependencies. This leads to a determinism problem and is a growing issue in
several of the market segments where OE makes sense. The way to solve this problem for OE is
to have seperate sysroots for each recipe and these will only contain the dependencies for
that recipe.
Its worth noting that this is not task specific sysroots and that OE's dependencies do vary
enormously by task. This did result in some implementation challenges. There is nothing stopping
the implementation of task specific sysroots at some later point based on this work but
that as deemed a bridge too far right now.
Implementation details:
* Rather than installing the sysroot artefacts into a combined sysroots, they are now placed in
TMPDIR/sysroot-components/PACKAGE_ARCH/PN.
* WORKDIR/recipe-sysroot and WORKDIR/recipe-sysroot-native are built by hardlinking in files
from the sysroot-component trees. These new directories are known as RECIPE_SYSROOT and
RECIPE_SYSROOT_NATIVE.
* This construction is primarily done by a new do_prepare_recipe_sysroot task which runs
before do_configure and consists of a call to the extend_recipe_sysroot function.
* Other tasks need things in the sysroot before/after this, e.g. do_patch needs quilt-native
and do_package_write_deb needs dpkg-native. The code therefore inspects the dependencies
for each task and adds extend_recipe_sysroot as a prefunc if it has populate_sysroot
dependencies.
* We have to do a search/replace 'fixme' operation on the files installed into the sysroot to
change hardcoded paths into the correct ones. We create a fixmepath file in the component
directory which lists the files which need this operation.
* Some files have "postinstall" commands which need to run against them, e.g. gdk-pixbuf each
time a new loader is added. These are handled by adding files in bindir with the name
prefixed by "postinst-" and are run in each sysroot as its created if they're present.
This did mean most sstate postinstalls have to be rewritten but there shouldn't be many of them.
* Since a recipe can have multiple tasks and these tasks can run against each other at the same
time we have to have a lock when we perform write operations against the sysroot. We also have
to maintain manifests of what we install against a task checksum of the dependency. If the
checksum changes, we remove its files and then add the new ones.
* The autotools logic for filtering the view of m4 files is no longer needed (and was the model
for the way extend_recipe_sysroot works).
* For autotools, we used to build a combined m4 macros directory which had both the native and
target m4 files. We can no longer do this so we use the target sysroot as the default and add
the native sysroot as an extra backup include path. If we don't do this, we'd have to build
target pkg-config before we could built anything using pkg-config for example (ditto gettext).
Such dependencies would be painful so we haven't required that.
* PKDDATA_DIR was moved out the sysroot and works as before using sstate to build a hybrid copy
for each machine. The paths therefore changed, the behaviour did not.
* The ccache class had to be reworked to function with rss.
* The TCBOOTSTRAP sysroot for compiler bootstrap is no longer needed but the -initial data
does have to be filtered out from the main recipe sysroots. Putting "-initial" in a normal
recipe name therefore remains a bad idea.
* The logic in insane needed tweaks to deal with the new path layout, as did the debug source
file extraction code in package.bbclass.
* The logic in sstate.bbclass had to be rewritten since it previously only performed search and
replace on extracted sstate and we now need this to happen even if the compiled path was
"correct". This in theory could cause a mild performance issue but since the sysroot data
was the main data that needed this and we'd have to do it there regardless with rss, I've opted
just to change the way the class for everything. The built output used to build the sstate output
is now retained and installed rather than deleted.
* The search and replace logic used in sstate objects also seemed weak/incorrect and didn't hold
up against testing. This has been rewritten too. There are some assumptions made about paths, we
save the 'proper' search and replace operations to fixmepath.cmd but then ignore this. What is
here works but is a little hardcoded and an area for future improvement.
* In order to work with eSDK we need a way to build something that looks like the old style sysroot.
"bitbake build-sysroots" will construct such a sysroot based on everything in the components
directory that matches the current MACHINE. It will allow transition of external tools and can
built target or native variants or both. It also supports a clean task. I'd suggest not relying on
this for anything other than transitional purposes though. To see XXX in that sysroot, you'd have
to have built that in a previous bitbake invocation.
* pseudo is run out of its components directory. This is fine as its statically linked.
* The hacks for wayland to see allarch dependencies in the multilib case are no longer needed
and can be dropped.
* wic needed more extensive changes to work with rss and the fixes are in a separate commit series
* Various oe-selftest tweaks were needed since tests did assume the location to binaries and the
combined sysroot in several cases.
* Most missing dependencies this work found have been sent out as separate patches as they were found
but a few tweaks are still included here.
* A late addition is that extend_recipe_sysroot became multilib aware and able to populate multilib
sysroots. I had hoped not to have to add that complexity but the meta-environment recipe forced my
hand. That implementation can probably be neater but this is on the list of things to cleanup later
at this point.
In summary, the impact people will likely see after this change:
* Recipes may fail with missing dependencies, particularly native tools like gettext-native,
glib-2.0-native and libxml2.0-native. Some hosts have these installed and will mask these errors
* Any recipe/class using SSTATEPOSTINSTFUNCS will need that code rewriting into a postinst
* There was a separate patch series dealing with roots postinst native dependency issues. Any postinst
which expects native tools at rootfs time will need to mark that dependency with PACKAGE_WRITE_DEPS.
There could well be other issues. This has been tested repeatedly against our autobuilders and oe-selftest
and issues found have been fixed. We believe at least OE-Core is in good shape but that doesn't mean
we've found all the issues.
Also, the logging is a bit chatty at the moment. It does help if something goes wrong and goes to the
task logfiles, not the console so I've intentionally left this like that for now. We can turn it down
easily enough in due course.
(From OE-Core rev: 809746f56df4b91af014bf6a3f28997d6698ac78)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* add sysroot prefix to wayland core protocols
* do not use pkg-config to find wayland-scanner
(From OE-Core rev: a8e7c5b415b99973c39a7ddd57cae45695fb0119)
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
