Commit Graph

19447 Commits

Author SHA1 Message Date
Ryan C. Gordon 67f89d0a2b openslES: Patched to compile.
(whoops.)

(cherry picked from commit 7323104f97)
2025-09-01 11:10:33 -04:00
Ryan C. Gordon f6de8877ec openslES: OpenSL ES on Android only supports two formats, limit to that.
(Three with the floating point extension, which we use.)

This is according to:

https://developer.android.com/ndk/guides/audio/opensl/opensl-for-android

Previously, this would accept a request for Sint8 or Sint32 and disaster
would ensue.

Fixes #13779.

(cherry picked from commit 0b2a003a35)
2025-09-01 10:31:05 -04:00
Ozkan Sezer 7c189b1f17 cmake: revert commit 5d1bbd9 and change sse4.2 test
clang doesn't support -mcrc32 until version 14.0, therefore drop -mcrc32
use in sse 4.2 test, and change it to check _mm_cmpgt_epi64() instead of
_mm_crc32_u32(). without this, sse4.2 check was failing with clang <= 13

Fixes https://github.com/libsdl-org/SDL/pull/12223

(cherry picked from commit 875653658a)
2025-08-31 03:20:40 +03:00
Sam Lantinga a7947080c0 Fixed warnings building with Visual Studio
(cherry picked from commit 7c8df4ea52)
2025-08-30 10:16:11 -07:00
Sam Lantinga 08c82a3781 Fixed build
(cherry picked from commit 3cf2350f21)
2025-08-30 10:15:48 -07:00
Jordan Saunders 589aea50c2 Fix two uninitialized variables
Found when running in valgrind looking at another issue.

- RenderPass' depth_stencil_target
    Tripped in SDL_BindGPUFragmentSamplers when not binding a DS target

- VulkanCommandBuffer's swapchainRequested
    Tripped in VULKAN_Submit for the end transition barrier when
    creating an image. The field is only reset when reused, not on first
    use

(cherry picked from commit 265236d952)
2025-08-30 06:05:43 -07:00
Ozkan Sezer 72acd7c9df CI: add OpenBSD workflow
(cherry picked from commit c0e5fd55d2)
2025-08-30 14:22:24 +03:00
Frank Praznik 1a70d1158e fs: Fix OpenBSD path retrieval
Set the path start pointer to point to a valid string.

(cherry picked from commit fdfde42db1)
2025-08-30 14:21:24 +03:00
Frank Praznik ae915c6516 wayland: Add a sigtimedwait() implementation for OpenBSD
sigtimedwait() is an optional part of POSIX.1-2001, and OpenBSD doesn't implement it. Add a replacement implementation based on https://comp.unix.programmer.narkive.com/rEDH0sPT/sigtimedwait-implementation

(cherry picked from commit 1049426a76)
2025-08-30 14:21:00 +03:00
Xander 54d573332e Add SDL_SENSOR_COUNT to SDL_SensorType
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit dc7a3a1219)
2025-08-29 07:31:25 -07:00
Matthew Zavislak 0204a69f6b Add elf alignment check for Android artifacts (#13818)
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 4f11feb708)
2025-08-28 16:31:25 -07:00
Susko3 b2689ff76a Avoid checking for HAVE_POSIX_SPAWN on Android
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
This makes the CMake build closer to SDL_build_config_android.h
as both will now use SDL_PROCESS_DUMMY.

(cherry picked from commit cd21bbe796)
2025-08-28 04:10:16 -07:00
Ozkan Sezer 611940fb80 cmake: fall back to -Wconversion if -Wfloat-conversion isn't available
... during checks for valid isinf/isinff/isnan/isnanf macros.

(cherry picked from commit 98bed62259)
2025-08-27 22:20:37 +03:00
Nintorch 16c0329a2a Add Emscripten joystick rumble support
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
Adds support for Emscripten (Web) joystick rumble support via EM_ASM_INT macros and HTML5's Gamepad API.

(cherry picked from commit 1fbed16cb0)
2025-08-25 12:59:15 -07:00
Nintorch 2743716132 Allow Android to ignore unnecessary joysticks
Previously, SDL_ShouldIgnoreJoystick wasn't being called for Android, and fingerprint sensors were recognized as joysticks.

(cherry picked from commit 1af7dfb0a7)
2025-08-25 11:23:34 -07:00
Colin Kinloch a744eee993 testffmpeg: avutil queue family version check
The `AVVulkanDeviceQueueFamily` struct was introduced by libavutil
59.34.100

(cherry picked from commit 81920b5db7)
2025-08-25 11:10:31 -07:00
Evan Hemsley 0e24267eb5 gpu: D3D12 only requires feature level 11_0 with Resource Binding Tier 2. (#13782)
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
We previously thought this wasn't possible because constant buffer offsets and
partial updates were unavailable, but we were reading the wrong table - this is
only the case for D3D11...

https://learn.microsoft.com/en-us/windows/win32/direct3d11/overviews-direct3d-11-devices-downlevel-intro

... while 12 doesn't list this feature at all:

https://learn.microsoft.com/en-us/windows/win32/direct3d12/hardware-feature-levels

We double checked and Jesse Natalie confirmed that this feature is required for
D3D12 even for 11_0 drivers. (Thanks Jesse!)

Additionally, D3D12 requires that UAVs are accessible from all shader stages,
meaning Tier 2 is enough to support the number of UAVs we need. Tier 1 could be
a property to lower the requirements, but that can be done later.
2025-08-22 14:07:58 -07:00
Wilson Jallet f4d6043aa4 GPU: Fix Vulkan indexing error for resolve attachment refs (#13768)
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 03b14f5211)
2025-08-19 12:13:23 -07:00
Petar Popovic 9fb9bd31a3 emscripten tests: fix warning: uninitialized variable
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 45feacf608)
2025-08-13 17:01:31 -07:00
ChaseKnowlden c3828bb0a3 Keep MSVC Flags Consistent across CMake runs
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 10478c59db)
2025-08-12 19:39:44 -07:00
Sam Lantinga 40371f0907 Support the "ambient" value for SDL_HINT_AUDIO_CATEGORY
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
Fixes https://github.com/libsdl-org/SDL/issues/13732

(cherry picked from commit 4725213eef)
2025-08-12 09:54:43 -07:00
Sylvain 94f9434564 Fixed bug #13493: Assertion failure at SDL_AddTouch with Android API 28
Java touch id should be -1 because it's reserved for internal SDL
synthetic events.
It should also not be 0, because this is SDL invalid value.

(cherry picked from commit 970c0bfe96)
2025-08-12 09:44:04 -07:00
Chase Knowlden bfaf247d30 Update NDK version to 28 (#13729)
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
* Update NDK version to 28 and add 16kb page size linker flags to x86_64

* Remove Android Linker Options

16kb page size is now the default since NDK r28c

* Update Android CI to use NDK 28

(cherry picked from commit 6e422e5ff2)
2025-08-12 07:46:27 -07:00
Beyley Cardellio 2287c43b59 GPU: Hold submit lock before waiting for device idle
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit e699f3dca1)
2025-08-10 07:22:19 -07:00
Petar Popovic 15cc0f5f91 linux/SDL_syshaptic.c:SDL_SYS_HapticStopAll(): Fix return on error
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 43f3991398)
2025-08-09 09:28:02 -07:00
Anonymous Maarten 68e0108b1c release: build aarch64 libraries with 16kiB page size
[ci skip]

(cherry picked from commit 7017fbaa8e)
2025-08-09 05:23:43 +02:00
Mohamed Shazan 7b3796bc39 SDL_TriggerBreakpoint() will default to __debugbreak() on MinGW toolchain on windows
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit f4c124e4bf)
2025-08-08 16:21:15 -07:00
Petar Popovic b5bc6d2cc4 SDL_SendJoystickVirtualSensorDataInner(): Fix max_sensor_events increment
(cherry picked from commit d9c20cfd0a)
2025-08-08 12:52:03 -07:00
Petar Popovic 0d1d4ea1f9 SDL_SetRenderDrawBlendMode(): Remove redundant param check
(cherry picked from commit b63c32e790)
2025-08-08 14:06:27 -04:00
Pino Toscano 1a48f897f2 ime: fcitx: use SDL_GetExeName() in GetAppName()
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
Use the existing SDL_GetExeName(), available for all the UNIX
platforms, in the internal GetAppName(); this has few advantanges:
- SDL_GetExeName() (and SDL_GetAppID() that builds on top of it) are
  used in various places already; since it caches the executable name,
  this may remove one extra read of the application name
- SDL_GetExeName() has a non-dummy implementation in more OSes than
  GetAppName(), thus providing a small improvement for this IME

As drive-by change: since SDL_GetExeName() provides a constant string,
there is no more need to allocate a new string in GetAppName(), which
is used as constant string anyway. Hence, return a constant string in
GetAppName() too.

(cherry picked from commit 248bcf6b29)
2025-08-07 13:58:27 -07:00
Simon McVittie 2b42789de6 Fix some typos detected by Debian's lintian QA tool
I assume the demoninator is a typo, rather than an indication that
someone has been playing too much Doom :-)

Signed-off-by: Simon McVittie <smcv@collabora.com>

(cherry picked from commit d83503f80e)
2025-08-07 13:55:44 -07:00
Simon McVittie b9c1da10ba hints: Rephrase documentation to improve grammar
"This thing allows to do something" is not really grammatically correct.
The closest rephrasing would be "allows one to do something" or "allows
the user to do something", but I think the passive voice reads more
naturally here.

Detected by Debian's lintian QA tool.

Signed-off-by: Simon McVittie <smcv@collabora.com>
(cherry picked from commit 40b941c826)
2025-08-07 13:55:44 -07:00
cosmonaut 56f5a76696 GPU: Fix uninitialized value in Vulkan command buffer structure
(cherry picked from commit fe6b2161bf)
2025-08-07 13:42:16 -07:00
Sam Lantinga c7e8977e60 Revert "Support Google Play 16 KB Page Size Requirement (#13470)"
This reverts commit edef6e66e9

If you need to support the Google Play 16 kiB page size requirement, the recommendation is to use NDK r28c or newer, which automatically aligns binaries correctly.
2025-08-07 11:41:43 -07:00
Pino Toscano e6ba3612db SDL_endian.h: extend Linux way for GNU libc
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
The currently used way to determine the endianness (i.e. include
<endian.h> and use the __BYTE_ORDER macro) is provided in general by
GNU libc. Thus, extend that to any platform/OS based on GNU libc.

(cherry picked from commit 561c99ee11)
2025-08-06 16:13:44 -07:00
Frank Praznik 5e77fb314c win32: Use the current flags to determine if NCCALCSIZE is required
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
SDL_GetWindowFlags() also ORs in pending flags, whereas the current state is needed here, particularly when creating/showing a window.

(cherry picked from commit cd0c660dea)
2025-08-05 18:49:41 -04:00
Frank Praznik 1a2841deb1 win32: Use STYLE_BORDERLESS when showing a pending fullscreen window
In addition to hiding the border on bordered windows that will immediately become fullscreen, The combination of flags used in STYLE_BORDERLESS_WINDOWED will still show the borders on borderless windows if the initial window size exactly matches the desktop, so STYLE_BORDERLESS must be used instead.

(cherry picked from commit 90a023007f)
2025-08-05 15:59:06 -04:00
Mathieu Eyraud 688637eca8 Fix condition for setting HDR properties
(cherry picked from commit a05aca51ec)
2025-08-05 09:26:12 -07:00
Frank Praznik 85d1d70ca1 cocoa: Wait for fullscreen spaces transitions to complete if switching to an exclusive mode
If attempting to switch to an exclusive mode while a fullscreen spaces transition is active, wait until the transition is complete before trying to apply the changes, or the window can wind up in a weird, broken state if a mode switch occurs while in a fullscreen space.

(cherry picked from commit f44a98729c)
2025-08-05 12:17:01 -04:00
Frank Praznik 09356c709a Revert "cocoa: Don't re-enter a fullscreen space if leaving to enter an exclusive mode"
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
This reverts commit 07b9e86d02.

It turns out that the problem is elsewhere, related to needing to block mode changes until spaces transitions are complete.
2025-08-04 23:51:47 -04:00
Frank Praznik 07b9e86d02 cocoa: Don't re-enter a fullscreen space if leaving to enter an exclusive mode
Doing so can leave the window in a weird, offset state.

(cherry picked from commit ee8f2861e7)
2025-08-04 22:16:22 -04:00
Sam Lantinga 8e46e5d8b4 Updated to version 3.2.21 for development
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
2025-08-04 11:54:34 -07:00
Sam Lantinga 96292a5b46 Updated to version 3.2.20 for release release-3.2.20 2025-08-04 09:58:55 -07:00
Petar Popovic ffa618c00b SDL_LoadWAV_IO(): On error, set *audio_buf to NULL and *audio_len to 0
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
(cherry picked from commit 23e08f7807)
2025-08-03 13:44:37 -04:00
Petar Popovic 291b9b3c82 SDL_enabled_assert(): Use NULL istead of 0 to explicity initialize the pointer members of SDL_AssertData
(cherry picked from commit faf3bd9991)
2025-08-03 13:36:14 -04:00
Frank Praznik ec45117f0f x11: Filter mouse wheel events from "Master" devices
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
Discard wheel events from "Master" devices to avoid duplicates, as wheel events are stateless and can't be deduplicated.

(cherry picked from commit 51ce3f8c8d)
2025-08-02 18:57:44 -04:00
Frank Praznik 0d7aff9c56 GPU: Vulkan backend flags command buffer for cleanup when swapchain is requested
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
When skipping presentation due to the window being hidden, presentDataCount is not incremented on the command buffer, and subsequently the submitted command buffers will not be cleaned up as long as the window is hidden. This results in a lag spike when showing the window due to all previously submitted command buffers suddenly being cleaned up at once, and lag at shutdown due to an equivalent number of fences needing to be destroyed.

Instead of relying on presentDataCount to determine whether a command buffer should be cleaned up, use a flag, which is set under the appropriate circumstances.

(cherry picked from commit 42463569d5)
2025-08-02 09:24:21 -07:00
Petar Popovic d31b239288 Fix double-free warning in src/hidapi/linux/hid.c
(cherry picked from commit ae5ce258cf)
2025-08-02 07:09:12 -07:00
capehill b1cadf0e99 Fix SDL_BlitSurfaceScaled crash
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
SDL_BlitSurfaceScaled could crash when passed large coordinates, due
to final_dst.w or final_dst.h getting negative values.

(cherry picked from commit 1c5c3b1479)
2025-08-01 09:22:38 -07:00
Frank Praznik c1f67585e6 wayland: Cleanup an outdated conditional and comment
Build (All) / Create test plan (push) Has been cancelled
Build (All) / level1 (push) Has been cancelled
Build (All) / level2 (push) Has been cancelled
There may have been a reason for not attaching a null buffer when destroying a popup at some point in the past, but that is unnecessary now, as is the comment about ShowWindow crashing, as ShowWindow assures that a null buffer is attached before (re)creating the window.

(cherry picked from commit 550d95e04f)
2025-08-01 10:40:32 -04:00