612336 - fix a number of webgl tests

Status: RESOLVED FIXED

(Core :: Graphics: CanvasWebGL, defect)

Description

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Here's a pile of small fixes for various webgl tests.

Comment 1

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix depth-clear value to return it as a float, not int

Comment 2

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix bind-uniform-location test

the underlying issue is that we do fake-attrib-0 handling, but when we try to restore, we didn't have the correct "default" values for the original attrib 0 since it wasn't modified.  So VertexAttribPointer raised a GL error, because size was invalid.  This sets the WebGLVertexAttribData object's initial params to the GL defaults.

Comment 3

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix the webgl context type

this (unfortunately large) patch renames the context interface from nsICanvasRenderingContextWebGL to nsIDOMWebGLRenderingContext, so that it ends up as "WebGLRenderingContext" in content.  The entire patch is a file rename and a bunch of search and replace.

Comment 4

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix invalid-passed-params test

glClear needs to check that only COLOR, DEPTH, STENCIL bits can be specified and nothing else.

Comment 5

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix webgl program-test

per ES 2.0, we can only have one shader of each type attached to a program, unlike desktop which can have an arbitrary number.  Enforce this here.

Comment 6

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix tex-image-invalid test

Comment 7

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: fix read-pixels test

this is kind of a silly fix.  The spec and man pages state that if /both/ format and type are invalid, then it's an INVALID_OPERATION.  But if one is valid and the other invalid, then it's an INVALID_ENUM.  *shrug*

Comment 8

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: update test suite to r12905

Comment 9

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 490627 [details] [diff] [review]
fix the webgl context type

r=me

Comment 10

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

http://hg.mozilla.org/mozilla-central/rev/56876e7bd590
http://hg.mozilla.org/mozilla-central/rev/53999baefb65
http://hg.mozilla.org/mozilla-central/rev/c00572802887
http://hg.mozilla.org/mozilla-central/rev/9d3796452887
http://hg.mozilla.org/mozilla-central/rev/929ce36f005b
http://hg.mozilla.org/mozilla-central/rev/25faa99c8fbe
http://hg.mozilla.org/mozilla-central/rev/775225b133b3
http://hg.mozilla.org/mozilla-central/rev/84776327a103