Bug 61743

Summary: [Qt] LayoutTests/fast/canvas/webgl/renderbuffer-initialization.html fails for Qt based webkit
Product: WebKit Reporter: Idrees <sidreesshah>
Component: WebGLAssignee: Nobody <webkit-unassigned>
Status: RESOLVED WONTFIX    
Severity: Normal CC: bfulgham, eric, hausmann, jamesr, kbr, kling, robert, tonikitoo
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: PC   
OS: All   
Attachments:
Description Flags
Test Case
none
Patch for the fix.
kling: review-
Another patch with indentation fixed.
none
Patch for the fix. kbr: review-

Idrees
Reported 2011-05-30 13:27:09 PDT
Created attachment 95365 [details] Test Case We are not initializing the color, depth and stensil buffer upon the resizing of webgl canvas. As a result the test case is failing.
Attachments
Test Case (2.94 KB, text/html)
2011-05-30 13:27 PDT, Idrees
no flags
Patch for the fix. (3.61 KB, patch)
2011-05-30 13:43 PDT, Idrees
kling: review-
Another patch with indentation fixed. (3.56 KB, patch)
2011-05-31 10:18 PDT, Idrees
no flags
Patch for the fix. (3.56 KB, patch)
2011-06-15 11:39 PDT, Idrees
kbr: review-
Idrees
Comment 1 2011-05-30 13:43:41 PDT
Created attachment 95367 [details] Patch for the fix. This patch will initialize the buffers upon resizing.
Andreas Kling
Comment 2 2011-05-31 05:56:22 PDT
Comment on attachment 95367 [details] Patch for the fix. View in context: https://bugs.webkit.org/attachment.cgi?id=95367&action=review > Source/WebCore/platform/graphics/qt/GraphicsContext3DQt.cpp:554 > + One newline is enough. > Source/WebCore/platform/graphics/qt/GraphicsContext3DQt.cpp:596 > + if (m_attrs.depth) { > + glClearDepth(clearDepth); > + glDepthMask(depthMask); > + } > + if (m_attrs.stencil) { > + glClearStencil(clearStencil); > + stencilMaskSeparate(GraphicsContext3D::FRONT, stencilMask); > + } > + if (isScissorEnabled) > + glEnable(GraphicsContext3D::SCISSOR_TEST); > + if (isDitherEnabled) > + glEnable(GraphicsContext3D::DITHER); > + Wrong indentation.
Idrees
Comment 3 2011-05-31 10:18:33 PDT
Created attachment 95448 [details] Another patch with indentation fixed.
Simon Hausmann
Comment 4 2011-05-31 15:07:07 PDT
I wish we were using GraphicsContext3DOpenGL.cpp with Qt and get rid of this duplicated code :(
Idrees
Comment 5 2011-05-31 16:49:09 PDT
(In reply to comment #4) > I wish we were using GraphicsContext3DOpenGL.cpp with Qt and get rid of this duplicated code :( I second your wish. What is stopping us from doing so?
Simon Hausmann
Comment 6 2011-05-31 22:25:25 PDT
(In reply to comment #5) > (In reply to comment #4) > > I wish we were using GraphicsContext3DOpenGL.cpp with Qt and get rid of this duplicated code :( > > I second your wish. What is stopping us from doing so? Nothing conceptually AFAIK. Do you want to give it a try? :)
Idrees
Comment 7 2011-06-01 10:31:06 PDT
(In reply to comment #6) > > Nothing conceptually AFAIK. Do you want to give it a try? :) I will but am not sure when.
Idrees
Comment 8 2011-06-02 10:28:47 PDT
Getting issues resolved with qtwebkit webgl should be more important. We can do the restructuring of the code later on. Using GraphicsContext3DOpenGL.cpp wont fix all the issues with qtwebkit webgl.
Alexis Menard (darktears)
Comment 9 2011-06-15 07:25:21 PDT
Comment on attachment 95448 [details] Another patch with indentation fixed. View in context: https://bugs.webkit.org/attachment.cgi?id=95448&action=review > Source/WebCore/ChangeLog:8 > + LayoutTests/fast/canvas/webgl/renderbuffer-initialization.html The test description goes after the explanation of the change.
Idrees
Comment 10 2011-06-15 11:39:38 PDT
Created attachment 97335 [details] Patch for the fix. Fixing the position of texts in changelog.
Robert Hogan
Comment 11 2011-08-10 13:13:30 PDT
Comment on attachment 97335 [details] Patch for the fix. View in context: https://bugs.webkit.org/attachment.cgi?id=97335&action=review This is shared code with GraphicsContext3DOpenGL. Seems fair enough. Let me know what you think of the comment below and we'll bug one of the reviewers! > Source/WebCore/platform/graphics/qt/GraphicsContext3DQt.cpp:593 > + if (isScissorEnabled) > + glEnable(GraphicsContext3D::SCISSOR_TEST); > + if (isDitherEnabled) > + glEnable(GraphicsContext3D::DITHER); > + No need for the glDisable() equivalents if false?
James Robinson
Comment 12 2011-09-12 15:35:17 PDT
(In reply to comment #11) > (From update of attachment 97335 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=97335&action=review > > This is shared code with GraphicsContext3DOpenGL. Seems fair enough. Let me know what you think of the comment below and we'll bug one of the reviewers! > By shared, do you mean copy-pasted?
Robert Hogan
Comment 13 2011-09-13 04:27:54 PDT
(In reply to comment #12) > By shared, do you mean copy-pasted? Yes.
Kenneth Russell
Comment 14 2011-09-13 10:50:01 PDT
Comment on attachment 97335 [details] Patch for the fix. This patch is out of date. GraphicsContext3DInternal was renamed to GraphicsContext3DPrivate in https://bugs.webkit.org/show_bug.cgi?id=67172 . Further, if the Qt port uses GraphicsContext3DOpenGL.cpp at all, then it should be able to share the implementation of GraphicsContext3D::reshape in that file. Please both update the patch and try to reuse this code.
Brent Fulgham
Comment 15 2014-01-09 20:55:33 PST
The QT port is no longer part of WebKit.
Note You need to log in before you can comment on or make changes to this bug.