| Age | Commit message (Collapse) | Author | Files | Lines | |
|---|---|---|---|---|---|
| 2017-04-07 | HACK: For performance evaluation only. Don't upstream.HEADmaster | Francisco Jerez | 4 | -4/+4 | |
| Uses global thread-unsafe memory pool. Might race in multithreaded environment. | |||||
| 2017-04-07 | retrace: Split retracing and parsing into separate threads for improved CPU ↵ | Francisco Jerez | 3 | -15/+34 | |
| utilization. | |||||
| 2017-04-07 | retrace: Implement lockless message queue for cross-thread communication. | Francisco Jerez | 1 | -0/+89 | |
| 2017-04-07 | trace: Switch trace representation to memory pool. | Francisco Jerez | 12 | -90/+97 | |
| 2017-04-07 | trace: Implement simple per-thread memory pool. | Francisco Jerez | 1 | -0/+185 | |
| 2017-04-07 | autotrim: Rework autotrim dependency analyzer to limit memory consumption ↵ | Francisco Jerez | 2 | -372/+1411 | |
| and improve accuracy. | |||||
| 2017-03-27 | retrace: add wglMakeContextCurrentARB() support | Brian Paul | 4 | -4/+41 | |
| 2017-03-27 | glws: add 'read' buffer parameter to makeCurrent() | Brian Paul | 7 | -13/+59 | |
| So that we can support the wglMakeContextCurrentARB() and glXMakeContextCurrent() functions. | |||||
| 2017-03-27 | gui: allow larger vertex buffer offsets in UI (1024 bytes) | Brian Paul | 1 | -1/+5 | |
| 2017-03-22 | travis: Add CentOS6 build via Docker. | Jose Fonseca | 1 | -8/+21 | |
| 2017-03-22 | cmake: Fix D3D11 include dir on dxerr. | Jose Fonseca | 1 | -1/+1 | |
| 2017-03-22 | thirdparty/dxsdk_aug2007: Bundle D3D8 headers from Aug 2007 DirectX SDK. | Jose Fonseca | 5 | -6/+3334 | |
| A bit of a grey area but common practice according to Microsoft: https://web.archive.org/web/20170322105210/https://blogs.msdn.microsoft.com/chuckw/2015/03/23/the-zombie-directx-sdk/ These headers are only used with MSVC. And they greatly simplify building apitrace for Windows, since modern Windows SDK include all necessary headers except D3D8's. | |||||
| 2017-03-22 | cmake: Don't check for unusued D3D headers/libraries. | Jose Fonseca | 3 | -86/+9 | |
| 2017-03-22 | compat: Don't redefine D3DERR_* macros. | Jose Fonseca | 2 | -21/+1 | |
| 2017-03-22 | cmake: Don't bother searching for D3DX headers/libraries. | Jose Fonseca | 1 | -24/+0 | |
| We never used them and never will. | |||||
| 2017-03-21 | cli: Write full path of where the Python scripts are expected to be. | Jose Fonseca | 1 | -1/+3 | |
| Should help diagnose issues like https://github.com/apitrace/apitrace/issues/511 | |||||
| 2017-03-21 | cli: Use APITRACE_PYTHON_EXECUTABLE. | Jose Fonseca | 1 | -1/+1 | |
| 2017-03-21 | glretrace: Initialize MetricWriter only when needed. | Alexander Trukhin | 5 | -12/+23 | |
| This object is only needed for profiling. Global initialization is unnecessary and it creates temporary file (with mmap_allocator.hpp), which can be avoided. | |||||
| 2017-03-21 | retrace: Unlink the .pbtmp immediately to not leave leftovers around. | Eric Anholt | 1 | -3/+1 | |
| If apitrace was to crash or get killed, the destructor wouldn't get called and the temp would get left on disk. Nothing else re-opens the file, so just unlink it immediately. | |||||
| 2017-03-21 | CMakeList: do not overwrite module path | Yann E. MORIN | 1 | -1/+1 | |
| Currently, the CMakeList.txt completely overwrites the CMAKE_MODULE_PATH variable. This is problematic when an upper-layer buildsystem wants to set its own module path to use custom modules. For example, Buldroot [0] provides a custom platform description [1] to fix cross-compilation issue. Overwriting the module path means that this custom platform description is not found: System is unknown to cmake, create: Platform/Buildroot to use this system, please send your config file to cmake@www.cmake.org so it can be added to cmake Providing such a custom platform description is what the upstream cmake devs suggest [2], quoting: If a toolchain file specifies CMAKE_SYSTEM_NAME such that a custom `Platform/MySystem.cmake` file is loaded then the latter can set them [*] as needed for the target platform. [*] offending settings causing RPATH issues during cross-compilation. So we need to append to the module path, rather than replace it blindly. [0] https://buildroot.org/ [1] https://git.buildroot.org/buildroot/tree/support/misc/Buildroot.cmake [2] http://public.kitware.com/pipermail/cmake/2017-February/065063.html Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> | |||||
| 2017-03-21 | retrace: fixed DDLOCK_DONOTWAIT flag clear mask | Simon Owen | 1 | -1/+1 | |
| This was unintentionally clearing all other lock flags, as well as potentially having both wait and do-not-wait set simultaneously. | |||||
| 2017-03-20 | Remove Windows XP support. | Jose Fonseca | 7 | -158/+47 | |
| I'm not aware of people using it, so I believe it's not worth the hassle as I'm trying to cut maintenance burden down. | |||||
| 2017-03-20 | compat: Avoid redefining macros. | Jose Fonseca | 1 | -0/+2010 | |
| Since newer version of MinGW headers define more and more of these. | |||||
| 2017-03-13 | retrace: support for dumping multiple snapshots (v3)windows-xp | Rob Clark | 7 | -28/+128 | |
| Usually if an app is using MRT, we want to dump (and diff) *all* the render targets to track down where things are going wrong. Also dumps depth and stencil buffers. When the --mrt (or -m) argument is specified, all render targets plus depth and/or stencil are dumped, with the suffix -mrtN/-z/-s. Otherwise the behavior is as before, only mrt0 is dumped with no suffix. Only implemented for GLDumper, since I don't know anything about D3D. | |||||
| 2017-03-09 | glws: Only process events from the main thread. | Jose Fonseca | 1 | -0/+4 | |
| Fixes https://github.com/apitrace/apitrace/issues/510 | |||||
| 2017-01-29 | cmake: Let CMAKE_OSX_DEPLOYMENT_TARGET take its default value. | Jose Fonseca | 1 | -1/+0 | |
| Should fix #505. | |||||
| 2017-01-25 | glstate: Avoid calling glGetProgramiv(GL_PROGRAM_BINARY_LENGTH). | Jose Fonseca | 1 | -2/+23 | |
| This should avoid unwanted side effects as described by emeakyl in https://github.com/apitrace/apitrace/issues/365#issuecomment-275029429 | |||||
| 2017-01-24 | glretrace: Allow to specify OpenGL vendor via environment variable. | Jose Fonseca | 1 | -5/+65 | |
| This enables on macOS one to do: VENDOR=Intel ./glretrace ... VENDOR=AMD ./glretrace ... VENDOR=NVIDIA ./glretrace ... VENDOR=Software ./glretrace ... | |||||
| 2017-01-24 | editorconfig: Specify indentation for Objective-C files. | Jose Fonseca | 1 | -1/+1 | |
| 2017-01-24 | cmake: Require Xcode 8.x; remove long build warning. | Jose Fonseca | 1 | -9/+7 | |
| Fixes https://github.com/apitrace/apitrace/issues/346 Fixes https://github.com/apitrace/apitrace/issues/503 | |||||
| 2017-01-24 | cmake: Target macos 10.11 and up; stop building 32bits executables. | Jose Fonseca | 2 | -10/+5 | |
| 2017-01-23 | glretrace: Try to enable automatic GPU switching on macOS. | Jose Fonseca | 4 | -0/+36 | |
| Minimal testing. | |||||
| 2017-01-12 | cmake: Use bundled libpng on macOS. | Jose Fonseca | 1 | -1/+1 | |
| cmake often picks up libpng from /usr/local which does not not include i386 architecture. | |||||
| 2017-01-12 | travis: Build with XCode 8.2 | Jose Fonseca | 1 | -1/+5 | |
| 2017-01-12 | glretrace: Avoid deprecated Cocoa enum warnings. | Jose Fonseca | 1 | -4/+4 | |
| 2017-01-12 | gui: Avoid deprecated QStyleOptionViewItemV4 warnings. | Jose Fonseca | 1 | -2/+2 | |
| 2016-12-09 | cmake: Use more than 2GB for 32 bits Windows processes where available. | Jose Fonseca | 1 | -1/+11 | |
| 2016-12-09 | specs: Recognize EXT_window_rectangles params. | Jose Fonseca | 1 | -0/+6 | |
| 2016-12-02 | snapdiff: Handle missing images more gracefully. | Jose Fonseca | 1 | -5/+8 | |
| 2016-12-02 | glretrace: Fix glViewportArrayv typo. | Jose Fonseca | 1 | -1/+1 | |
| 2016-11-14 | gltrace: Describe GL_MAX_VERTEX_ATTRIB_STRIDE. | Jose Fonseca | 1 | -1/+1 | |
| 2016-10-13 | docs: Add a status section. | Jose Fonseca | 1 | -0/+9 | |
| I've been very busy professionally/personally, which has left very little time to maintain Apitrace. So set the right expectations. | |||||
| 2016-10-05 | docs: Document the Snappy format too. | Jose Fonseca | 1 | -3/+25 | |
| 2016-10-05 | docs: Fix discrepancy and ambiguity in format spec. | Jose Fonseca | 1 | -7/+8 | |
| As pointed out by [bk]door.maus in issue #492. Also fix some spelling errors. | |||||
| 2016-09-21 | d3dstate: Port some of the recent D3D9 code to D3D8. | Jose Fonseca | 1 | -4/+169 | |
| 2016-09-21 | d3dretrace: Viewport Min/MaxZ are floats. | Jose Fonseca | 1 | -2/+2 | |
| 2016-09-21 | d3dretrace: Don't duplicate ATI1/ATI2 FourCC defines. | Jose Fonseca | 2 | -10/+2 | |
| "ATI1N"/"ATI2N" is how these texture formats are described in AMD's "Advanced DX9 Capabilities for ATI Radeon Cards" document. g test.sh | |||||
| 2016-09-21 | d3dretrace: Don't dump dummy rendertargets | Patrick Rudolph | 1 | -0/+10 | |
| Do not try to dump rendertargets that uses format D3DFMT_NULL. Found trying to dump Crysis renderstates. Signed-off-by: Patrick Rudolph <siro@das-labor.org> | |||||
| 2016-09-21 | d3dretrace: Dump d3d9 texture states | Patrick Rudolph | 1 | -0/+52 | |
| Dump all texturestagestates. Use seperate a seperate field for each stage. Signed-off-by: Patrick Rudolph <siro@das-labor.org> | |||||
| 2016-09-21 | d3dretrace: Dump viewport state for d3d9 | Tiziano Bacocco | 1 | -0/+21 | |
| Dump D3D9 viewport state. Rebased against current HEAD. Signed-off-by: Tiziano Bacocco <tizbac2@gmail.com> Signed-off-by: Patrick Rudolph <siro@das-labor.org> | |||||
 |
index : ~currojerez/apitrace | |
| Unnamed repository; edit this file 'description' to name the repository. | currojerez |
| summaryrefslogtreecommitdiff |