7c9c4cd435
gobj: devirtualize calls to GeomPatches::get_num_vertices_per_primitive()
2024-10-17 16:18:12 +02:00
777ab6746c
display: Fix double-precision build
...
Fixes #1688
2024-10-14 12:05:57 +02:00
bd4dc8a379
display: Fix ability to copy wdxGraphicsBuffer to RAM on DX9
2024-10-10 19:57:42 +02:00
30418158e8
dxgsg9: Allow getting alpha bits in buffer when host has no alpha
2024-10-10 19:57:14 +02:00
a59c857359
dxgsg9: Replace old A2W10V10U10 format with A32B32G32R32F
2024-10-10 19:56:41 +02:00
eac1cb57c4
dxgsg9: Add missing formats to DXTextureContext9::get_bits_per_pixel
2024-10-10 19:49:15 +02:00
05cf72e22f
dxgsg9: Fix texture clear color not working for 3D textures
2024-10-10 19:48:50 +02:00
7557826c8b
Merge branch 'release/1.10.x'
2024-10-10 19:48:37 +02:00
cd2e3dfde9
dxgsg9: Fix crash on window event after `close_window()`
2024-10-10 01:08:54 +02:00
beb49b3cd3
dxgsg9: Fix some format handling, support for r32f and rgba32f textures
2024-10-03 13:43:22 +02:00
0ea0d5fa2b
dxgsg9: Fix crash when copying inverted framebuffer to texture RAM
2024-10-03 13:36:02 +02:00
51876c6116
dxgsg9: Fix crash with buffer when there is no depth buffer
2024-10-03 13:36:02 +02:00
37799dc3ce
pipeline: Try to fix Windows compilation issue
2024-09-29 23:02:35 +02:00
89fcfa1fe9
Merge branch 'release/1.10.x'
2024-09-26 10:23:50 +02:00
c05a63f4ae
pipeline: Fix compile error without pipelining enabled
...
Closes #1684
[skip ci]
2024-09-21 10:55:05 +02:00
beff684af3
Further thread safety changes for Python 3.13 free threading
...
See #1683
2024-09-19 20:43:06 +02:00
4b2fa45578
Make use of PyImport_GetModule function
2024-09-19 20:42:31 +02:00
5106fc879f
Changes for compatibility with Python 3.13
...
Includes some preliminary work to make free-threading safe, see also #1683
2024-09-19 16:01:10 +02:00
b32459ffa1
Merge branch 'release/1.10.x'
2024-09-19 15:51:25 +02:00
cdafd81764
Fix compilation issues with Python 3.13
2024-09-19 15:47:50 +02:00
d7ede81aa3
cocoa: Fix compile error with MacOSX 10.6 SDK
2024-09-19 11:15:44 +02:00
a4faa802af
cocoa: Remove autorelease pool in process_events()
...
We now have one around the entire frame, in graphicsEngine.cxx
2024-09-19 00:25:51 +02:00
442558bb2d
Merge branch 'release/1.10.x'
2024-09-19 00:25:16 +02:00
9f493f588d
display: Fix significant memory leak on newer macOS versions
...
It seems the Metal-based OpenGL driver uses autorelease a lot for resources that last only a single frame, so we need to create an autorelease pool around the frame lest the resources will only get cleaned up at application exit.
2024-09-19 00:23:06 +02:00
756fe2e72b
cocoa: Fix a very minor memory leak in DisplayInformation
2024-09-18 23:51:09 +02:00
b337f2a3da
pipeline: Use a high-precision timer for Thread::sleep() on Win32
...
This is mainly used for clock mode M_limited, which is currently suffering from terrible imprecision on Windows. This fixes that, making it possible to lower the value of sleep-precision down to 1ms.
2024-09-16 21:43:42 +02:00
bb5e4454fe
putil: Add debug for oversleeping/undersleeping with M_limited clock
2024-09-16 21:41:52 +02:00
e3f580763a
pview: add shift+r keyboard bind to reload all textures ( #1678 )
2024-09-16 16:02:45 +02:00
f5401b19fa
Merge branch 'release/1.10.x'
2024-09-16 13:54:22 +02:00
e2fb2d1241
cocoa: Prevent eating keyUp when switching to FS from macOS' FS
2024-09-16 13:39:16 +02:00
c7bcec1ff5
cocoa: Exit macOS' fullscreen mode when switching fullscreen
...
Prevents black screen when switching to fullscreen
2024-09-16 13:39:16 +02:00
82b9a1266c
cocoa: Fix error when switching to FS while macOS' FS is active
2024-09-16 13:39:16 +02:00
e7604cb163
event: fix result refcount leak when awaiting AsyncFuture
...
Raising the StopIteration exception left the underlying
result with an elevated refcount which prevented it from
being cleaned up.
2024-09-14 15:59:05 -06:00
cfe3885c0e
event: use the persistent python wrapper when appending task object ( #1681 )
2024-09-14 20:25:08 +02:00
fc394e4c59
glgsg: Pass in osg_FrameNumber input via ShaderMatSpec mechanism
2024-08-23 21:08:50 +02:00
cd68287b6a
shader: Get all aggregate shader inputs (eg. lights) in one go
...
This means that p3d_LightSource[n] will be fetched once off the LightAttrib, written to the matrix cache, and then indexed into by the various ShaderMatSpec. This should be significantly more efficient, but the main aim is to prepare for a new binding system in the new shader pipeline
User structs are still an exception as of now
2024-08-23 16:29:46 +02:00
99225dfaef
shader: Unify ShaderPtrSpec and ShaderMatSpec mechanisms
2024-08-23 16:29:42 +02:00
99f4034701
shader: Don't store on ShaderMatSpec, use scratch matrix instead
2024-08-23 16:29:38 +02:00
2c5cebe7e0
display: Fix shader gen regression for more than 2 lights
2024-08-23 14:20:20 +02:00
76ffef38d1
display: Allow passing whole state into dispatch_compute()
...
This is mainly useful for unit testing
2024-08-23 12:34:57 +02:00
babd3a070d
display: Fix docstring formatting
2024-08-22 15:52:48 +02:00
9e87af9dd7
linmath: Fix compiler warning with gcc 14
2024-08-14 19:09:20 +02:00
8d1524b675
distort: Fix unprotected debug output
2024-08-14 19:09:08 +02:00
712300c9f1
Fix prebuilt bison cxx files includes to work with CMake build
...
Change makepanda to generate without the .yxx extension so that the include name is generated properly
2024-08-13 00:01:35 +02:00
4ca495d828
CMake: Add separate BUILD_TOOLS option to build binaries
...
This is separate from BUILD_PANDATOOL, which builds the pandatool tree
2024-08-12 16:39:20 +02:00
21b39da65d
pythonTask: fix refcount leak of non-panda Future::done
...
A missing Py_DECREF on the future's "done" method
caused both the bound method and the underlying
self instance of the future to be leaked when
awaiting non-panda futures (such as _asyncio.Future).
This change includes a simple new test addition
to catch this in the future.
2024-08-08 21:51:03 -06:00
23124dac07
x11: Fix missing _net_wm_name field due to bad merge
2024-08-07 23:39:20 +02:00
ad57762e9f
Merge branch 'release/1.10.x'
2024-08-07 22:34:08 +02:00
8c8cbeea98
linmath: For repr, use as much precision as needed for roundtrip
...
Will use as few digits as is necessary to ensure that round-tripping the same number with pstrtod will result in the same vector. This prevents eg. 3.3 formatting as 3.29999995 while still ensuring that two floats that are not equal (other than nan) are guaranteed to have a different string representation.
Uses a hacky pftoa function that can be abandoned as soon as we adopt C++17, which has to_chars that does what we need
Fixes #1671
2024-08-07 17:24:11 +02:00
122453811d
glgsg: Fix clear of offscreen buffer if back buffers requested
2024-06-30 11:53:17 +02:00
rdb
John C. Allwein