I noticed that if the window handle is invalid, the resulting rect will be uninitialized. This change makes it fail immediately.

Review URL: http://codereview.appspot.com/2009042

git-svn-id: https://angleproject.googlecode.com/svn/trunk@386 736b8ea6-26fd-11df-bfd4-992fa37f6226

TIntermBinary::promote() was incorrectly marking the type of result as const in some cases. The result can only be const if both operands are const. Also cleaned up the function to remove redundant/repeated checks.
BUG=24
TEST=OpenGL ES 2.0 Conformance tests
Review URL: http://codereview.appspot.com/1938047

git-svn-id: https://angleproject.googlecode.com/svn/trunk@385 736b8ea6-26fd-11df-bfd4-992fa37f6226

This fixes the crash on lost context for vista and windows 7 if suitable drivers are available. It now uses D3D9Ex when available and this only reports lost contexts for hardware failures and suchlike.

Testing procedure was:
- run simple_vertex_shader without this change.
- ctrl+alt+del to lock machine.
- return to desktop.
- observe that simple_vertex_shader crashes.
- repeat with this change and check that simple_vertex_shader is still animating.
- simulate the code path that an XP machine would take using the debugger and check that old behavior is preserved.

I decided to load D3D9.DLL at runtime for a couple of reasons. First, I didn't want to assume that older implementations of D3D9 would have the Direct3DCreate9Ex entry point. Second, it might be advantageous for some applications to not have a load time dependency on D3D9. I didn't address this for D3DX9 because it's a little harder - there's no clear way to determine which D3DX9_x.DLL library to open. At least D3DX9_x.DLL does not seem to import D3D9.DLL and D3DX is separately redistributable so there is still some advantage to be had.
Review URL: http://codereview.appspot.com/1951044

git-svn-id: https://angleproject.googlecode.com/svn/trunk@384 736b8ea6-26fd-11df-bfd4-992fa37f6226

Review URL: http://codereview.appspot.com/1980041

git-svn-id: https://angleproject.googlecode.com/svn/trunk@383 736b8ea6-26fd-11df-bfd4-992fa37f6226

The crash occurs when createDevice attempts to dereference mD3d9, which is NULL.


Review URL: http://codereview.appspot.com/1956042

git-svn-id: https://angleproject.googlecode.com/svn/trunk@382 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=18
Review URL: http://codereview.appspot.com/1936041

git-svn-id: https://angleproject.googlecode.com/svn/trunk@381 736b8ea6-26fd-11df-bfd4-992fa37f6226

Some compilers do not support standard STL allocator interface. Also removed _Charalloc function only needed by VC++6.0, which we do not support. Submitted by Eagle.Lu.
BUG=19
Review URL: http://codereview.appspot.com/1913048

git-svn-id: https://angleproject.googlecode.com/svn/trunk@380 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=15
Review URL: http://codereview.appspot.com/1916046

git-svn-id: https://angleproject.googlecode.com/svn/trunk@379 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=13
Review URL: http://codereview.appspot.com/1895051

git-svn-id: https://angleproject.googlecode.com/svn/trunk@378 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=16
Review URL: http://codereview.appspot.com/1868056

git-svn-id: https://angleproject.googlecode.com/svn/trunk@377 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=14
Review URL: http://codereview.appspot.com/1850054

git-svn-id: https://angleproject.googlecode.com/svn/trunk@376 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=22
Review URL: http://codereview.appspot.com/1855057

git-svn-id: https://angleproject.googlecode.com/svn/trunk@375 736b8ea6-26fd-11df-bfd4-992fa37f6226

We were trying to change a const_iterator which was rightly caught by SunStudio compiler. Submitted by Eagle.Lu.
BUG=19
Review URL: http://codereview.appspot.com/1879049

git-svn-id: https://angleproject.googlecode.com/svn/trunk@374 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=19
Review URL: http://codereview.appspot.com/1910047

git-svn-id: https://angleproject.googlecode.com/svn/trunk@373 736b8ea6-26fd-11df-bfd4-992fa37f6226

Adding separate license file for the preprocessor code and removing the ANGLE boilerplate license from those files.

git-svn-id: https://angleproject.googlecode.com/svn/trunk@372 736b8ea6-26fd-11df-bfd4-992fa37f6226

git-svn-id: https://angleproject.googlecode.com/svn/trunk@371 736b8ea6-26fd-11df-bfd4-992fa37f6226

Trac #13050, Issue=21
Original Author: Vladimir Vukicevic
Signed-off-by: Daniel Koch

git-svn-id: https://angleproject.googlecode.com/svn/trunk@370 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #13077
 - clamp scissor to rendertarget size
 - ensure textures are renderable when required
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@369 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #13074
Also adds initializations for missed Texture members.
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@368 736b8ea6-26fd-11df-bfd4-992fa37f6226

Review URL: http://codereview.appspot.com/1907043

git-svn-id: https://angleproject.googlecode.com/svn/trunk@367 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12714
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@366 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12660

Automatically resize the D3D swap chain when the eglSwapBuffers is called and the presentation window has changed size.
Also change D3D device init to be performed once when the Display is created using a hidden 1x1 window.

Signed-off-by: Nicolas Capens
Signed-off-by: Daniel Koch

Author:    Andrew Lewycky

git-svn-id: https://angleproject.googlecode.com/svn/trunk@365 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12496

Derive Renderbuffer, Texture and Buffer from RefCountObject.
This class keeps a reference count for all objects that need cross-context
reference counting, and also the object id.
Restructure Renderbuffers to create a wrapper object and a storage object.
Use BindingPointer for all binding points instead of binding by object id.

Signed-off-by: Shannon Woods
Signed-off-by: Daniel Koch

git-svn-id: https://angleproject.googlecode.com/svn/trunk@364 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12498

Signed-off-by: Shannon Woods
Signed-off-by: Daniel Koch

Author:    Andrew Lewycky

git-svn-id: https://angleproject.googlecode.com/svn/trunk@363 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12565
Signed-off-by: Andrew Lewycky
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@362 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12571
Automatically passes default framebuffer, and fails
framebuffers which have no color attachments.
Signed-off-by: Andrew Lewycky
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@361 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12493
The resource manager class is now in charge of allocation &
management of objects which may be shared by multiple contexts.
Signed-off-by: Andrew Lewycky
Signed-off-by: Nicolas Capens
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@360 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12580

Wrong because it should be checking active attributes, not inactive attributes.
Unnecessary because when a buffer is deleted, all the vertex attrib client state bindings are cleared.

Signed-off-by: Shannon Woods
Signed-off-by: Daniel Koch

Author:    Andrew Lewycky

git-svn-id: https://angleproject.googlecode.com/svn/trunk@359 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12336
Signed-off-by: Andrew Lewycky
Signed-off-by: Daniel Koch

Author:    Shannon Woods

git-svn-id: https://angleproject.googlecode.com/svn/trunk@358 736b8ea6-26fd-11df-bfd4-992fa37f6226

git-svn-id: https://angleproject.googlecode.com/svn/trunk@357 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #11594
Signed-off-by: Daniel Koch

Author:    Nicolas Capens

git-svn-id: https://angleproject.googlecode.com/svn/trunk@356 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #11594
Signed-off-by: Andrew Lewycky
Signed-off-by: Daniel Koch

Author:    Nicolas Capens

git-svn-id: https://angleproject.googlecode.com/svn/trunk@355 736b8ea6-26fd-11df-bfd4-992fa37f6226

TRAC #12547

ValidateDevice checks for correct rendering state as well, which we don't need.

Signed-off-by: Nicolas Capens
Signed-off-by: Daniel Koch

Author:    Andrew Lewycky

git-svn-id: https://angleproject.googlecode.com/svn/trunk@354 736b8ea6-26fd-11df-bfd4-992fa37f6226

De-allocate memory in ShFinalize(). There still seems to be a bug here if the translator is used in a multi-threaded environment. It seems memory local to only one thread is being deallocated - the thread from which ShFinalize() is called. I will look into it closely in the next CL.
Review URL: http://codereview.appspot.com/1878044

git-svn-id: https://angleproject.googlecode.com/svn/trunk@353 736b8ea6-26fd-11df-bfd4-992fa37f6226

Encapsulate expressions with ternary operators with brackets. This preserves the order of precedence when ternary expressions are used in a compound expression.
BUG=20
Review URL: http://codereview.appspot.com/1894041

git-svn-id: https://angleproject.googlecode.com/svn/trunk@352 736b8ea6-26fd-11df-bfd4-992fa37f6226

Refactored the way symbol tables are initialized and stored. This was done in response to the addition of EShSpec.
Symbol table entries depend on three things - language, spec (not now but may eventually), and built-in resources.
We used to build two global symbol-tables - one for each language. During each compile, one of the symbol table was
copied and resource-specific stuff was added. I have moved the symbol table to TCompiler that gets initilized when
compiler is created and reused for each compile. This makes it much cleaner and extensible in case a spec requires
special entries to be added to the symbol table.

PS: Sorry for the long CL, but all of it needed to be done in one CL. I have verified that everything still compiles
and passes all conformance tests.
Review URL: http://codereview.appspot.com/1864044

git-svn-id: https://angleproject.googlecode.com/svn/trunk@351 736b8ea6-26fd-11df-bfd4-992fa37f6226

I found that when I switch desktops, GetRenderTargetData will return D3DERR_DEVICELOST. It is unclear to me whether ReadPixels should report a GL error in this case and if so, which one. I went with GL_OUT_OF_MEMORY. Let me know if that's wrong.

This change is not sufficient to make eglGetError report EGL_CONTEXT_LOST. For a program doing exclusively offscreen rendering, ReadPixels might be the only place where D3DERR_DEVICELOST can be detected. The GLES code can't call egl::setCurrentError though. I'm not sure how to fix that.




Review URL: http://codereview.appspot.com/1692053

git-svn-id: https://angleproject.googlecode.com/svn/trunk@350 736b8ea6-26fd-11df-bfd4-992fa37f6226

Review URL: http://codereview.appspot.com/1842046

git-svn-id: https://angleproject.googlecode.com/svn/trunk@349 736b8ea6-26fd-11df-bfd4-992fa37f6226

BUG=9
Review URL: http://codereview.appspot.com/1665050

git-svn-id: https://angleproject.googlecode.com/svn/trunk@348 736b8ea6-26fd-11df-bfd4-992fa37f6226

Check for reserved prefixes (gl_, webgl_, and _webgl_) in struct and field names. GLES2 and WebGL both require this for all identifiers - variable, function, struct, and field names. ANGLE was only validating variable and function names.
BUG=11
Review URL: http://codereview.appspot.com/1856046

git-svn-id: https://angleproject.googlecode.com/svn/trunk@347 736b8ea6-26fd-11df-bfd4-992fa37f6226