]> git.cworth.org Git - apitrace/log
apitrace
11 years agoMerge branch 'opengl-4.3'
José Fonseca [Thu, 9 Aug 2012 10:49:32 +0000 (11:49 +0100)]
Merge branch 'opengl-4.3'

11 years agoRecognize IDirect3DSwapChain9::Present as end of frame
José Fonseca [Wed, 8 Aug 2012 17:50:47 +0000 (18:50 +0100)]
Recognize IDirect3DSwapChain9::Present as end of frame

11 years agoTweak new state/entrypoints.
José Fonseca [Tue, 7 Aug 2012 23:11:09 +0000 (00:11 +0100)]
Tweak new state/entrypoints.

11 years agoMinimal OpenGL 4.3 support.
José Fonseca [Tue, 7 Aug 2012 21:08:29 +0000 (22:08 +0100)]
Minimal OpenGL 4.3 support.

11 years agoUpdate khronos headers.
José Fonseca [Tue, 7 Aug 2012 20:46:04 +0000 (21:46 +0100)]
Update khronos headers.

11 years agoCleanup strings constness.
José Fonseca [Tue, 7 Aug 2012 20:07:22 +0000 (21:07 +0100)]
Cleanup strings constness.

Allows to distinguish between a const char *, and char * const.

11 years agoUse CString shorthand more places.
José Fonseca [Tue, 7 Aug 2012 19:00:18 +0000 (20:00 +0100)]
Use CString shorthand more places.

11 years agoAllow to pipe profile output directly to profileshader.py
José Fonseca [Mon, 6 Aug 2012 16:35:06 +0000 (17:35 +0100)]
Allow to pipe profile output directly to profileshader.py

For example

  ./glretrace -pgpu foo.trace | ./scripts/profileshader.py

11 years agoMerge remote-tracking branch 'jbenton/master'
José Fonseca [Mon, 6 Aug 2012 15:35:14 +0000 (16:35 +0100)]
Merge remote-tracking branch 'jbenton/master'

11 years agoCheck GL_QUERY_COUNTER_BITS before profiling.
James Benton [Mon, 6 Aug 2012 15:05:58 +0000 (16:05 +0100)]
Check GL_QUERY_COUNTER_BITS before profiling.

11 years agoCorrectly handle profiling of glBegin/glEnd.
James Benton [Mon, 6 Aug 2012 10:44:23 +0000 (11:44 +0100)]
Correctly handle profiling of glBegin/glEnd.

11 years agoCheck for GL_EXT_timer_query and GL_ARB_occlusion_query.
James Benton [Mon, 6 Aug 2012 10:40:28 +0000 (11:40 +0100)]
Check for GL_EXT_timer_query and GL_ARB_occlusion_query.

11 years agoMerge branch 'egl-image'
José Fonseca [Sun, 5 Aug 2012 09:24:28 +0000 (10:24 +0100)]
Merge branch 'egl-image'

Conflicts:
wrappers/egltrace.py

11 years agoRecreate egl surface when glViewport for watching framebuffer in qapitrace's surface...
Huang Dongsung [Sat, 23 Jun 2012 07:03:26 +0000 (16:03 +0900)]
Recreate egl surface when glViewport for watching framebuffer in qapitrace's surface tab.

It is quirk of mesa egl 1.4 with Nvidia-dri. It is ok with intel-dri.

After resizing XWindow, mesa egl does not know new size of the surface.
If you call eglQuerySurface(currentDisplay, currentSurface, EGL_WIDTH, width),
width is the width at the time to create the egl surface, not current width.
However, glx can know what the size of window well.

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoProcess window events on d3dretrace.
José Fonseca [Fri, 3 Aug 2012 18:56:23 +0000 (19:56 +0100)]
Process window events on d3dretrace.

11 years agoResize window on IDirect3DDevice9::Reset.
José Fonseca [Fri, 3 Aug 2012 15:47:54 +0000 (16:47 +0100)]
Resize window on IDirect3DDevice9::Reset.

11 years agoRemove old unused function declaration.
James Benton [Fri, 3 Aug 2012 13:25:25 +0000 (14:25 +0100)]
Remove old unused function declaration.

11 years agoAdd missing comma in last commit.
José Fonseca [Thu, 2 Aug 2012 19:05:19 +0000 (20:05 +0100)]
Add missing comma in last commit.

11 years agoRecognize for context creation/destruction/binding functions (issue #90).
José Fonseca [Thu, 2 Aug 2012 19:00:28 +0000 (20:00 +0100)]
Recognize for context creation/destruction/binding functions (issue #90).

Thanks to gregory38 for spotting this.

11 years agoAvoid inserting 'inf' and friends into the json output.
Carl Worth [Mon, 30 Jul 2012 22:01:15 +0000 (15:01 -0700)]
Avoid inserting 'inf' and friends into the json output.

The JSON-generating code was previously carefult to avoid emitting NaN
into its output, but it would still write "inf" for an infinite value,
(which is invalid JSON).

Fix by calling writeNull for any value for which std::isfinite returns
false, (rather than the previous test for merely (n != n)).

11 years agoRemoved arbitary limit on active queries.
James Benton [Thu, 2 Aug 2012 10:37:44 +0000 (11:37 +0100)]
Removed arbitary limit on active queries.

This previously existed due to a suspected limit on queries on nvidia drivers,
however that bug was actually due to context switching which has now been fixed.

11 years agoCleaned up trace profiler output.
James Benton [Tue, 31 Jul 2012 13:11:14 +0000 (14:11 +0100)]
Cleaned up trace profiler output.

Updated profileshader.py to reflect these changes.

11 years agoFixed bug in CPU profiling.
James Benton [Tue, 31 Jul 2012 13:10:35 +0000 (14:10 +0100)]
Fixed bug in CPU profiling.

11 years agoCheck if GL_ARB_timer_query is supported before using it.
James Benton [Tue, 31 Jul 2012 11:48:27 +0000 (12:48 +0100)]
Check if GL_ARB_timer_query is supported before using it.

11 years agoEmit a fake glTexImage for glEGLImageTargetTexture2DOES.
José Fonseca [Thu, 2 Aug 2012 07:42:24 +0000 (08:42 +0100)]
Emit a fake glTexImage for glEGLImageTargetTexture2DOES.

11 years agoegl: EGL image trace support
Imre Deak [Mon, 4 Jun 2012 11:14:24 +0000 (14:14 +0300)]
egl: EGL image trace support

Getting the dimension / format of EGL images is not straitghforward.
One way to get these would be through glGetTexLevelParameter, but it's
not supported by GLES. Another way is to attach the EGL image to a
renderbuffer and call glGetRenderbufferParameter on it, but at least
mesa doesn't support attaching images in RGBA_REV format. And that
format happens to be an important one in applications we care about
(like the Android browser). We could also check the parameters of the
underlying native buffer, but these are not exposed through any API (for
either Xpixmaps or Android native buffers)

So the only way I found is to try a non-destructive glCopyTexSubImage2D
with sizes starting from GL_MAX_TEXTURE_SIZE down to 1 and determine the
correct size based on whether this returns an error or not.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agoRetrace should info correct image's format to qapitrace.
Huang Dongsung [Sat, 23 Jun 2012 07:38:57 +0000 (16:38 +0900)]
Retrace should info correct image's format to qapitrace.

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoAdd UCHAR.
José Fonseca [Wed, 25 Jul 2012 19:31:30 +0000 (20:31 +0100)]
Add UCHAR.

11 years agoDon't includes windows.h inside extern "C" {...}
José Fonseca [Wed, 25 Jul 2012 19:03:34 +0000 (20:03 +0100)]
Don't includes windows.h inside extern "C" {...}

11 years agoEnumerate a few more D3DUSAGE values.
José Fonseca [Wed, 25 Jul 2012 19:03:08 +0000 (20:03 +0100)]
Enumerate a few more D3DUSAGE values.

11 years agoBump windows version.
José Fonseca [Wed, 25 Jul 2012 18:59:34 +0000 (19:59 +0100)]
Bump windows version.

11 years agoAdd a script to analyse profile output per shader.
James Benton [Wed, 25 Jul 2012 12:52:14 +0000 (13:52 +0100)]
Add a script to analyse profile output per shader.

11 years agoImproved profiling capabilities.
James Benton [Wed, 25 Jul 2012 12:48:20 +0000 (13:48 +0100)]
Improved profiling capabilities.

Logs which shader is active to allow per shader profiling.
Merged cpu profiling code from retrace to trace_profiler.
Allow choice of profiling for gpu / cpu / pixels drawn.

11 years agoAdded support for recording GL_SAMPLES_PASSED.
James Benton [Mon, 23 Jul 2012 17:45:46 +0000 (18:45 +0100)]
Added support for recording GL_SAMPLES_PASSED.

11 years agoFixed gpu profiling under multiple contexts.
James Benton [Mon, 23 Jul 2012 17:12:27 +0000 (18:12 +0100)]
Fixed gpu profiling under multiple contexts.

Now flushes the results for all queries on context switch.

11 years agoImproved GPU profiling code.
James Benton [Mon, 23 Jul 2012 14:26:54 +0000 (15:26 +0100)]
Improved GPU profiling code.

Now only allows a certain number of active queries,
this fixes results with applications that have many draw calls.

11 years agoAdd support for basic GPU profiling of draw calls to retrace.
James Benton [Fri, 20 Jul 2012 17:56:19 +0000 (18:56 +0100)]
Add support for basic GPU profiling of draw calls to retrace.

Currently only supports opengl using glBegin/EndQuery and glQueryCounter.

Based on code from Yuanhan Liu timing-trace branch.

11 years agoMerge commit '5c298db8fe117e0a445af051335aa0da91c3a31b'
José Fonseca [Thu, 12 Jul 2012 12:44:59 +0000 (13:44 +0100)]
Merge commit '5c298db8fe117e0a445af051335aa0da91c3a31b'

11 years agoEliminate destroyContext.
José Fonseca [Wed, 11 Jul 2012 11:14:31 +0000 (12:14 +0100)]
Eliminate destroyContext.

Same as releaseContext.

11 years agoFix build on MSVC.
José Fonseca [Mon, 9 Jul 2012 10:16:50 +0000 (11:16 +0100)]
Fix build on MSVC.

11 years agoFix wglDeleteContext tracing.
José Fonseca [Mon, 9 Jul 2012 09:36:01 +0000 (10:36 +0100)]
Fix wglDeleteContext tracing.

wglGetCurrentContext must be called before.

11 years agoFix CGL.
José Fonseca [Mon, 9 Jul 2012 09:47:16 +0000 (10:47 +0100)]
Fix CGL.

11 years agoFix build. Add copyright header to new files.
José Fonseca [Sun, 8 Jul 2012 12:16:48 +0000 (13:16 +0100)]
Fix build. Add copyright header to new files.

11 years agoRelocate tracing (dis/en)enablement.
José Fonseca [Sat, 7 Jul 2012 11:54:09 +0000 (12:54 +0100)]
Relocate tracing (dis/en)enablement.

Out of common -- as it is tracing specific

11 years agoAndroid: add usage instructions for tracing zygote applications
Imre Deak [Fri, 30 Mar 2012 12:44:39 +0000 (15:44 +0300)]
Android: add usage instructions for tracing zygote applications

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agoAndroid: add support for dynamically enable/disable tracing
Imre Deak [Fri, 30 Mar 2012 12:46:26 +0000 (15:46 +0300)]
Android: add support for dynamically enable/disable tracing

To trace applications started by the Dalvik VM we have to wrap the main
Dalvik process zygote (or app_process) which is started at boot time and
never stopped afterwards. We would still want to restrict tracing to a
single process which will be forked by zygote.

To achieve this we'll use the fact that the forked process image is
contained in the same file as zygote, pointed to by /proc/self/exe. So
if this is 'app_process' we know it's a zygote process. To distinguish
between zygote processes we'll check /proc/self/cmdline which will in
turn be set to a unique application specific string (for example
com.android.settings). The user has to set the debug.apitrace.procname
to this string in order to enable tracing.

For non-zygote processes tracing will be always enabled.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agoMerge pull request #88 from gregory38/image-alpha-channel-I84-v2
zackr [Sat, 7 Jul 2012 00:10:22 +0000 (17:10 -0700)]
Merge pull request #88 from gregory38/image-alpha-channel-I84-v2

Image alpha channel i84 v2

11 years agoMerge pull request #87 from gregory38/multiple-line-errors
zackr [Sat, 7 Jul 2012 00:08:30 +0000 (17:08 -0700)]
Merge pull request #87 from gregory38/multiple-line-errors

glsl edit: grep multiline error/warning messages

11 years agoglsl edit: grep multiline error/warning messages
gregory [Fri, 6 Jul 2012 19:48:59 +0000 (21:48 +0200)]
glsl edit: grep multiline error/warning messages

Useful for GL shader errors which spawn multiple lines error message like that:
    >> 2135: warning: Vertex shader failed to compile with the following errors:
    >> ERROR: 1:1: error(#132) Syntax error: 'sdfsafasf' parse error
    >> ERROR: error(#273) 1 compilation errors.  No code generated

11 years agogui: Fix editing of the shaders
Zack Rusin [Fri, 6 Jul 2012 19:53:29 +0000 (15:53 -0400)]
gui: Fix editing of the shaders

Editing shaders was broken because we were adding extra values to
the arguments array. This fixes it, meaning that if you right click
on glShaderSource you can edit the shader source and the rest of the
trace will properly catch the changes. An incredibly useful feature.

11 years agogui: remove a depth check box that sneak into alpha code
gregory [Sun, 1 Jul 2012 16:57:10 +0000 (18:57 +0200)]
gui: remove a depth check box that sneak into alpha code

11 years agogui: new checkbox to visualize alpha channel of image
gregory [Sun, 1 Jul 2012 15:18:04 +0000 (17:18 +0200)]
gui: new checkbox to visualize alpha channel of image

11 years agoHandle windows when retracing CreateAdditionalSwapChain
José Fonseca [Fri, 22 Jun 2012 18:36:31 +0000 (19:36 +0100)]
Handle windows when retracing CreateAdditionalSwapChain

11 years agoTrace IDirect3DSwapChain9Ex interface too.
José Fonseca [Fri, 22 Jun 2012 16:27:56 +0000 (17:27 +0100)]
Trace IDirect3DSwapChain9Ex interface too.

11 years agoAvoid walking off array if given a negative value for array size.
Carl Worth [Mon, 4 Jun 2012 21:47:33 +0000 (14:47 -0700)]
Avoid walking off array if given a negative value for array size.

Obviously, a negative size is bogus, but we want apitrace to be robust
enough to capture even bogus usage of the API, (so that the user can
investigate and fix the bogosity in the application).

In this case, if the user specifies a negative size for an array,
simply don't write anything out for the array content.

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoSkip glEnable/Disable(GL_DEBUG_OUTPUT_SYNCHRONOUS_ARB) (issue #83).
Gregory Hainaut [Mon, 11 Jun 2012 12:58:22 +0000 (13:58 +0100)]
Skip glEnable/Disable(GL_DEBUG_OUTPUT_SYNCHRONOUS_ARB) (issue #83).

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agomt trace: wgl: track GL context on a per-thread basis
Imre Deak [Wed, 23 May 2012 07:57:12 +0000 (10:57 +0300)]
mt trace: wgl: track GL context on a per-thread basis

Wrap WGL context handling calls and keep our own per-thread basis track
of the context object.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agomt trace: glx: track GL context on a per-thread basis
Imre Deak [Wed, 23 May 2012 07:56:57 +0000 (10:56 +0300)]
mt trace: glx: track GL context on a per-thread basis

Wrap GLX context handling calls and keep our own per-thread basis track
of the context object.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agomt trace: egl: track GL context on a per-thread basis
Imre Deak [Wed, 23 May 2012 07:56:35 +0000 (10:56 +0300)]
mt trace: egl: track GL context on a per-thread basis

Wrap EGL context handling calls and keep our own per-thread basis track
of the context object.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agocheck for error in eglMakeCurrent wrapper
Imre Deak [Mon, 21 May 2012 10:55:40 +0000 (13:55 +0300)]
check for error in eglMakeCurrent wrapper

In case of error we shouldn't update our book-keeping.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agomt trace: cgl: track GL context on a per-thread basis
Imre Deak [Wed, 23 May 2012 07:56:06 +0000 (10:56 +0300)]
mt trace: cgl: track GL context on a per-thread basis

Wrap CGL context handling calls and keep our own per-thread basis track
of the context object.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agomt trace: add helper to track GL contexts on a per-thread basis
Imre Deak [Mon, 7 May 2012 06:10:26 +0000 (09:10 +0300)]
mt trace: add helper to track GL contexts on a per-thread basis

This is needed for muti-threaded tracing to work properly.

Tracing multi-threaded GL apps requires that we track the lifetime of
the GL context objects and store thread specific state like the currently
active GL context on thread-local memory area.

Also add 'retain_count' to the context needed by the CGL retain/release
context calls.

Needed by the 4 upcoming GL API specific patches.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agomt trace: create gltrace_state.cpp, move getContext into it
Imre Deak [Thu, 3 May 2012 14:33:13 +0000 (17:33 +0300)]
mt trace: create gltrace_state.cpp, move getContext into it

There is no need to keep this function in a python script and the
following patches will extend the functionality around this code, so
move the function to a .cpp file.

Signed-off-by: Imre Deak <imre.deak@intel.com>
11 years agotrace: add support for tracing GL_OES_draw_texture calls
Imre Deak [Mon, 23 Apr 2012 13:18:39 +0000 (16:18 +0300)]
trace: add support for tracing GL_OES_draw_texture calls

These are present at least in the Android EGL implementation.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoExport undocumented Android gl*PointerBounds entrypoints.
José Fonseca [Thu, 17 May 2012 21:06:11 +0000 (22:06 +0100)]
Export undocumented Android gl*PointerBounds entrypoints.

Used in Dalvik.

11 years agoCompute the vertex buffer sizes in term of vertex count.
José Fonseca [Thu, 17 May 2012 20:33:14 +0000 (21:33 +0100)]
Compute the vertex buffer sizes in term of vertex count.

As opposed to max index, given that max index can not represent zero count.

11 years agofilter based on thread_id
Imre Deak [Tue, 8 May 2012 12:20:43 +0000 (15:20 +0300)]
filter based on thread_id

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agomake GLES/glext.hpp guard macro name unique
Imre Deak [Fri, 11 May 2012 16:32:21 +0000 (19:32 +0300)]
make GLES/glext.hpp guard macro name unique

We want to include both GL/glext.hpp and GLES/glext.hpp but they are
guarded using the same macro name. Make them unique.

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoPickle Repr nodes too.
José Fonseca [Thu, 17 May 2012 07:38:04 +0000 (08:38 +0100)]
Pickle Repr nodes too.

11 years agoadd debug option for debug builds
Imre Deak [Fri, 20 Apr 2012 13:34:31 +0000 (16:34 +0300)]
add debug option for debug builds

Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
11 years agoUse appropriate number of digits when dumping floating point numbers.
José Fonseca [Mon, 14 May 2012 19:20:44 +0000 (20:20 +0100)]
Use appropriate number of digits when dumping floating point numbers.

11 years agoTry to disassemble d3d10 shaders too.
José Fonseca [Sun, 13 May 2012 14:08:37 +0000 (15:08 +0100)]
Try to disassemble d3d10 shaders too.

11 years agoRename d3dshader.* -> d3d9shader.*
José Fonseca [Sun, 13 May 2012 10:41:08 +0000 (11:41 +0100)]
Rename d3dshader.* -> d3d9shader.*

11 years agoFactor out object swizzling.
José Fonseca [Sun, 13 May 2012 09:22:07 +0000 (10:22 +0100)]
Factor out object swizzling.

11 years agoSplit retrace swizlling helpers.
José Fonseca [Sun, 13 May 2012 09:04:19 +0000 (10:04 +0100)]
Split retrace swizlling helpers.

11 years agoMove d3d window creation to a separate source file.
José Fonseca [Sun, 13 May 2012 08:45:04 +0000 (09:45 +0100)]
Move d3d window creation to a separate source file.

11 years agoDescribe a few more parameters.
José Fonseca [Sun, 13 May 2012 07:56:10 +0000 (08:56 +0100)]
Describe a few more parameters.

12 years agoDump a few more parameters.
José Fonseca [Sat, 12 May 2012 11:19:44 +0000 (12:19 +0100)]
Dump a few more parameters.

12 years agoAdd multiarch support.
Michal Suchanek [Mon, 7 May 2012 14:07:16 +0000 (16:07 +0200)]
Add multiarch support.

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
12 years agoAvoid glCreateShaderProgramv when dumping satate so that the shader
José Fonseca [Fri, 11 May 2012 13:55:50 +0000 (14:55 +0100)]
Avoid glCreateShaderProgramv when dumping satate so that the shader
source can be recovered.

12 years agoTake in consideration the pipeline when dumping shaders/uniforms.
Gregory Hainaut [Fri, 11 May 2012 13:54:18 +0000 (14:54 +0100)]
Take in consideration the pipeline when dumping shaders/uniforms.

A new member for JSON is created for each stage to separate easily the
different stages.

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
12 years agoTry active shader program before current program.
Gregory Hainaut [Fri, 11 May 2012 13:49:10 +0000 (14:49 +0100)]
Try active shader program before current program.

With separate shader program, one can't query anymore current program
with glGetIntegerv(GL_CURRENT_PROGRAM, &program), so glUniform* fails to
store the correct location during retracing.

Technically glActiveShaderProgram must be used to set the active program
for Uniform. But on my system (AMD) it works without it, I don't know if
it is valid or not.

This change uses the active shader (would be 0 if the users don't call
glActiveShaderProgram) instead of current program (0 in pipeline mode)

Signed-off-by: José Fonseca <jose.r.fonseca@gmail.com>
12 years agoPass INFO_LOG warnings in glCreateShaderProgramv/glCreateShaderProgramEXT.
José Fonseca [Fri, 11 May 2012 13:16:49 +0000 (14:16 +0100)]
Pass INFO_LOG warnings in glCreateShaderProgramv/glCreateShaderProgramEXT.

12 years agoFix auto completion typo.
Gregory Hainaut [Thu, 10 May 2012 17:57:08 +0000 (18:57 +0100)]
Fix auto completion typo.

12 years agoEnsure alloca is declared.
José Fonseca [Thu, 10 May 2012 16:28:31 +0000 (17:28 +0100)]
Ensure alloca is declared.

12 years agoSupport dumping uniform buffer objects (issue #75).
José Fonseca [Thu, 10 May 2012 12:18:38 +0000 (13:18 +0100)]
Support dumping uniform buffer objects (issue #75).

Based on Gregory Hainaut's prototype.

12 years agoDon't attempt to dump built-in uniforms.
José Fonseca [Thu, 10 May 2012 08:36:59 +0000 (09:36 +0100)]
Don't attempt to dump built-in uniforms.

12 years agoSilent bogus warnings.
José Fonseca [Wed, 9 May 2012 20:57:10 +0000 (21:57 +0100)]
Silent bogus warnings.

12 years agoFix D3D11 tracing with D3D11_CREATE_DEVICE_DEBUG flag.
José Fonseca [Wed, 9 May 2012 20:12:30 +0000 (21:12 +0100)]
Fix D3D11 tracing with D3D11_CREATE_DEVICE_DEBUG flag.

12 years agoCleanup return value wrapping.
José Fonseca [Wed, 9 May 2012 18:33:33 +0000 (19:33 +0100)]
Cleanup return value wrapping.

12 years agoPrevent hangs in XCloseDisplay.
José Fonseca [Wed, 9 May 2012 18:33:15 +0000 (19:33 +0100)]
Prevent hangs in XCloseDisplay.

That surfaced with precedent commit.

12 years agoAdd ability to run multiple traces in one time.
Si Chen [Wed, 9 May 2012 17:16:30 +0000 (18:16 +0100)]
Add ability to run multiple traces in one time.

12 years agoSuppress warning when running in benchmark mode.
Si Chen [Wed, 9 May 2012 17:15:42 +0000 (18:15 +0100)]
Suppress warning when running in benchmark mode.

12 years agoFind FXC executable.
José Fonseca [Wed, 9 May 2012 08:23:38 +0000 (09:23 +0100)]
Find FXC executable.

12 years agoStub d3d state dumping.
José Fonseca [Tue, 8 May 2012 22:41:35 +0000 (23:41 +0100)]
Stub d3d state dumping.

12 years agoRefactor d3d9 lock tracking code further.
José Fonseca [Tue, 8 May 2012 22:23:38 +0000 (23:23 +0100)]
Refactor d3d9 lock tracking code further.

12 years agoFix scripts absolute earch path (issue #76).
José Fonseca [Tue, 8 May 2012 14:02:51 +0000 (15:02 +0100)]
Fix scripts absolute earch path (issue #76).

Wrongly copy'n'pasted from wrappers code.

Probably never actually used, as relative search paths where successful.

Thanks to hramrach for spotting it.

12 years agoFix typos in earlier commit.
José Fonseca [Tue, 8 May 2012 11:09:42 +0000 (12:09 +0100)]
Fix typos in earlier commit.

Should have actually looked at the build output before pushing...

12 years agoSet d3d10 sideeffect=False tags.
José Fonseca [Tue, 8 May 2012 08:07:27 +0000 (09:07 +0100)]
Set d3d10 sideeffect=False tags.

12 years agoFix tracediff.py's --width argument conversion (issue #77).
José Fonseca [Tue, 8 May 2012 08:05:53 +0000 (09:05 +0100)]
Fix tracediff.py's --width argument conversion (issue #77).

Thanks for wrabbit-revisited spotting problem and proposing fix.