487916 - [c3d] Generated functions with constant checks based on the spec

Status: RESOLVED WONTFIX

(Core :: Graphics: CanvasWebGL, enhancement)

Description

[Ilmari Heikkinen]

Attachment: Functions with constant checks

Replaces GL_SAME_METHOD macro-generated functions with functions that check their arguments against constants, generated with http://github.com/kig/canvas3d-tests/tree/master/apigen

The constant lists are copy-pasted from GLES 2.0 man pages at http://www.khronos.org/opengles/sdk/docs/man/ 

The one change I can remember is accepting GL_MULTISAMPLE for glEnable/glDisable.

The patched version passes canvas3d-tests (except for the disabled getImageData), but it's possible that there are bugs with the constant checks as I don't have tests for them (and even if I did, they'd be based on the same autogen :I

Comment 1

[Benoit Jacob [:bjacob] (mostly away)]

Wow, I can't believe that this patch has been sitting in the bugzilla for more than a year. What a waste :( I guess that your code came in a little bit early since the draft spec was really far from final at that time.

Hint: to (hugely) increase your chances of getting your patches reviewed, when you attach them, set a "review" to "?" and enter Vlad's email address (or mine).

There is stuff in your patch that we _still_ don't have in our code, so still have to do, for example the StencilFunc enums validation. However we are doing things a bit differently to improve code reusability: we implement the checks in separate helper functions in WebGLContextValidate.cpp.

Comment 2

[Benoit Jacob [:bjacob] (mostly away)]

Taking care of this item...

Comment 3

[Benoit Jacob [:bjacob] (mostly away)]

I just filed bug 573541 with an updated patch doing very similar things; sorry to say I couldn't use your patch because 1) the code has changed a lot since, and 2) I wanted to factor code through helper functions.

Comment 4

[Benoit Jacob [:bjacob] (mostly away)]

of course, in the meanwhile, a lot of these checks had already been added...

Comment 5

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

Indeed; let's close this and instead Benoit can pull out the checks that we're missing into Validate functions.

Comment 6

[Benoit Jacob [:bjacob] (mostly away)]

I don't think that we're still missing any enums check.