[webkit-reviews] review granted: [Bug 205520] ANGLE: Fix WebGL conformance tests for EXT_texture_filter_anisotropic : [Attachment 386424] Fix one more expectations file

[bugzilla-daemon at webkit.org]

Dean Jackson <dino at apple.com> has granted James Darpinian
<jdarpinian at gmail.com>'s request for review:
Bug 205520: ANGLE: Fix WebGL conformance tests for
EXT_texture_filter_anisotropic
https://bugs.webkit.org/show_bug.cgi?id=205520

Attachment 386424: Fix one more expectations file

https://bugs.webkit.org/attachment.cgi?id=386424&action=review




--- Comment #7 from Dean Jackson <dino at apple.com> ---
Comment on attachment 386424
  --> https://bugs.webkit.org/attachment.cgi?id=386424
Fix one more expectations file

View in context: https://bugs.webkit.org/attachment.cgi?id=386424&action=review

>> Source/WebCore/html/canvas/WebGL2RenderingContext.cpp:-1372
>> -	ENABLE_IF_REQUESTED(EXTTextureFilterAnisotropic,
m_extTextureFilterAnisotropic, "WEBKIT_EXT_texture_filter_anisotropic",
enableSupportedExtension("GL_OES_texture_float"_s));
> 
> What mitigates the risk of removing this? Is this possibly incompatible with
WebKit-specific content, such as content embedded in iOS apps, that does not
work with other web engines?

I think it is safe to remove this. Mozilla hasn't supported their equivalent
for a long time (I tried searching in their commits but failed, although I see
a warning of imminent removal in 2013). Blink removed this about the same time
(again, searching through their history is difficult because of file renames
after the fork). A quick search on github shows three.js uses it, but only as a
fallback from EXT_texture_filter_anisotropic. The WebGL conformance suite uses
it in the same manner (we should remove that too).

This was one of the first ratified extensions, so it would have to be extremely
old content, written when Chrome supported this.