Using Strictly OpenGL Core Profile — Handmade Hero — Episode Guide

0:16Run the game to recap and set the stage for the day, with a few words on using 3D APIs

4:44Determine to field debug messages and conditionally disable compatibility contexts

10:55glDebugMessageCallback¹

14:34glDebugMessageControl²

16:16GL_ARB_debug_output³

23:18win32_handmade.cpp: Pull in GLDEBUGPROC() and glDebugMessageCallbackARB() from corearb.h⁴

29:31handmade_opengl.cpp: Make OpenGLInit() call glDebugMessageCallbackARB()

30:13Run the game to successfully receive our callback and view an error

32:22handmade_opengl.cpp: Make OpenGLInit() perform the callback later

32:52Run the game and receive no errors

33:18win32_handmade.cpp: Switch to the OpenGL Core Profile and set WGL_CONTEXT_FORWARD_COMPATIBLE_BIT_ARB

34:42Run the game and get an error message

36:16handmade_opengl.cpp: Prevent OpenGLInit() from calling glTexEnvi()

36:24Run the game to find that glAlphaFunc() has been removed

36:51On why glAlphaFunc() is no longer necessary

38:23handmade_opengl.cpp: Make the FragmentCode conditionally apply blending and OpenGLRenderCommands() no longer call glAlphaFunc()

40:40handmade_opengl.cpp: Prevent OpenGLRenderCommands() from enabling GL_ALPHA_TEST

41:00Note that depth testing doesn't occur in the shader

42:14handmade_opengl.cpp: Prevent OpenGLRenderCommands() from enabling GL_TEXTURE_2D and the glMatrixMode() stuff

43:22Run the game and realise that we must create and bind a vertex array object^α

44:49handmade_opengl.cpp: Make OpenGLInit() call glGenVertexArrays()⁵ and glBindVertexArray()⁶

48:28Run the game and receive apparently the opposite error from before

52:13Note that we are using client-side memory and consider switching to GPU-mapped memory

57:33handmade_opengl.cpp: Make OpenGLInit() call glGenBuffers()⁷ to store the vertex array

1:01:32handmade_opengl.cpp: Make OpenGLRenderCommands() call glBufferData()⁸ for that GL_ARRAY_BUFFER

1:04:28A few words on how this buffer has improved the rendering pipeline

1:07:08handmade_opengl.cpp and win32_handmade.cpp: Pull in the necessary calls from corearb.h⁹

1:10:21Run the game to find that we have passed the glVertexAttribPointer error, and that the final error is relating to GL_QUADS

1:11:57On why there may not be a fix for this GL_QUADS error

1:15:08Blackboard: Vertex Sharing

1:17:25Blackboard: Specifying the vertices with an Index stream

1:22:31handmade_opengl.cpp: Re-enable GL_QUADS, turn off the debug messages, and run the game to find that we're not getting a working image

1:23:31win32_handmade.cpp: Turn off WGL_CONTEXT_FORWARD_COMPATIBLE_BIT_ARB, and run the game to find that it's now okay

1:24:17Blackboard: Our options regarding the WGL_CONTEXT_FORWARD_COMPATIBLE_BIT_ARB bit, GL_QUADS and GL_TRIANGLE_STRIP

1:31:41handmade_opengl.cpp: Change GL_QUADS to GL_TRIANGLE_STRIP in OpenGLRenderCommands(), and re-enable the debug message

1:32:59Run the game to see that this is not correct because the vertex order is wrong

1:33:24handmade_opengl.cpp: Re-enable the debug message and WGL_CONTEXT_FORWARD_COMPATIBLE_BIT_ARB

1:34:11Run the game to see that we're completely compliant, but just drawing wrong

1:34:21handmade_render_group.cpp: Switch PushQuad() to push the vertices in the correct order for GL_TRIANGLE_STRIP

1:35:18Run the game to see that it is now correct, and note that this won't work if we had a texture atlas

1:38:53Determine to get the orthographic projection back online for the debug view

1:43:25handmade_math.h: Enable OrthographicProjection() to use the near and far clip planes

1:45:16Run the game and note that we'll need to deal with the camera part of the transform for the debug UI

1:47:37handmade_debug.cpp: Consider how the SetCameraTransform() call in DEBUGStart() is now massively wrong

1:49:02handmade_debug.cpp: Make SetCameraTransform() use and scale the CameraX and CameraY

1:50:16Blackboard: Mapping the extremities of the screen to -1 to 1

1:52:36Run the game and assess our clip plane situation

1:54:14handmade_debug.cpp: Begin to consider what else in DEBUGStart() may be problematic

1:55:29Q&A

🗩

1:55:29Q&A

🗩

1:55:29Q&A

🗩

1:56:49ratchet_freak Where's owlbot? He's not online

🗪

1:56:49ratchet_freak Where's owlbot? He's not online

🗪

1:56:49ratchet_freak Where's owlbot? He's not online

🗪

1:59:19macielda About the QUAD thing, you might want to do something like this, maybe¹⁰

🗪

1:59:19macielda About the QUAD thing, you might want to do something like this, maybe¹⁰

🗪

1:59:19macielda About the QUAD thing, you might want to do something like this, maybe¹⁰

🗪

2:01:33alephant You could avoid geometry shaders by instancing a triangle strip that is just a quad

🗪

2:01:33alephant You could avoid geometry shaders by instancing a triangle strip that is just a quad

🗪

2:01:33alephant You could avoid geometry shaders by instancing a triangle strip that is just a quad

🗪

2:02:21hguleryuz What do you think about using a different set of program / shader for debug text, one which has no projection matrix at all? Would it be bad to switch programs for GUI text performance-wise?

🗪

2:02:21hguleryuz What do you think about using a different set of program / shader for debug text, one which has no projection matrix at all? Would it be bad to switch programs for GUI text performance-wise?

🗪

2:02:21hguleryuz What do you think about using a different set of program / shader for debug text, one which has no projection matrix at all? Would it be bad to switch programs for GUI text performance-wise?

🗪

2:05:40ratchetfreak There is a third way of setting the vertex data through ARB_separate_attrib_format (went to core in 4.3). This separates the vertex format from the buffer it reads the data from which is more optimal for the driver

🗪

2:05:40ratchetfreak There is a third way of setting the vertex data through ARB_separate_attrib_format (went to core in 4.3). This separates the vertex format from the buffer it reads the data from which is more optimal for the driver

🗪

2:05:40ratchetfreak There is a third way of setting the vertex data through ARB_separate_attrib_format (went to core in 4.3). This separates the vertex format from the buffer it reads the data from which is more optimal for the driver

🗪

2:06:32robrobby When the engine uses a deprecated function like GL_QUADS, what are the main issues that people would have to deal with playing this game, because it is running, right?

🗪

2:06:32robrobby When the engine uses a deprecated function like GL_QUADS, what are the main issues that people would have to deal with playing this game, because it is running, right?

🗪

2:06:32robrobby When the engine uses a deprecated function like GL_QUADS, what are the main issues that people would have to deal with playing this game, because it is running, right?

🗪

2:09:05soysaucethekid Is OpenGL 3.0 the target version?

🗪

2:09:05soysaucethekid Is OpenGL 3.0 the target version?

🗪

2:09:05soysaucethekid Is OpenGL 3.0 the target version?

🗪

2:11:14rizoma80 Any idea why one can have stuttering problems using OpenGL DWM stuff related?

🗪

2:11:14rizoma80 Any idea why one can have stuttering problems using OpenGL DWM stuff related?

🗪

2:11:14rizoma80 Any idea why one can have stuttering problems using OpenGL DWM stuff related?

🗪

2:14:01macielda What do you like better in OpenGL AZDO as opposed to Vulkan?

🗪

2:14:01macielda What do you like better in OpenGL AZDO as opposed to Vulkan?

🗪

2:14:01macielda What do you like better in OpenGL AZDO as opposed to Vulkan?

🗪

2:18:35pepevel Would it be interesting to implement OpenGL on top of OpenCL?

🗪

2:18:35pepevel Would it be interesting to implement OpenGL on top of OpenCL?

🗪

2:18:35pepevel Would it be interesting to implement OpenGL on top of OpenCL?

🗪

2:19:54robrobby Since Handmade Hero is using multithreaded processing of the game entities, how would the game even stutter when time wouldn't be enough 'til the next frame?

🗪

2:19:54robrobby Since Handmade Hero is using multithreaded processing of the game entities, how would the game even stutter when time wouldn't be enough 'til the next frame?

🗪

2:19:54robrobby Since Handmade Hero is using multithreaded processing of the game entities, how would the game even stutter when time wouldn't be enough 'til the next frame?

🗪

2:21:01Shut this down

🗩

2:21:01Shut this down

🗩

2:21:01Shut this down

🗩

Keyboard Navigation

Global Keys

Menu toggling

In-Menu and Index Controls

Quotes and References Menus and Index

Quotes, References and Credits Menus

Filter Menu

Filter and Link Menus

Credits Menu