Issue metadata
Sign in to add a comment
|
PresentationRequest interface prototype not named "PresentationRequestPrototype" |
||||||||||||||||||||||||
Issue descriptionVersion: 53.0.2785.8 unknown (64-bit) OS: Linux What steps will reproduce the problem? (1) Check out and setup W3C Web platform tests here: https://github.com/w3c/web-platform-tests (2) Run the test presentation-api/idlharness.html What is the expected output? ALl tests pass What do you see instead? Failures like the following: assert_equals: class string of Presentation.prototype expected "[object PresentationPrototype]" but got "[object Presentation]" at IdlObject.<anonymous> (http://localhost:8000/resources/idlharness.js:1002:9) at Object.Test.step (http://localhost:8000/resources/testharness.js:1403:25) at test (http://localhost:8000/resources/testharness.js:501:18) at IdlObject.IdlInterface.test_self (http://localhost:8000/resources/idlharness.js:906:5) at IdlObject.IdlInterface.test (http://localhost:8000/resources/idlharness.js:731:14) ******* It appears that the generated blink bindings from e.g. PresentationRequest.idl are not setting the name of the interface prototype object per the WebIDL specification [1]. e.g. > PresentationRequest.prototype.toString() => "[object PresentationRequest]" instead of "[object PresentationRequestPrototype]". This is causing these tests to fail. I looked at the blink extended attributes for WebIDL and did not see any way to customize this property of the interface prototype object. Perhaps I missed something? [1] https://heycam.github.io/webidl/#interface-prototype-object (last sentence of 4.6.3) Note this is causing failures in other test suites that use idlharness, such as gamepad, webstorage, etc. Perhaps I'm missing some IDL annotation? Philip, I believe you've done some IDL maintenance in the past?
,
Jul 12 2016
,
Jul 13 2016
The current status of Blink about what the class string for the prototype objects should be is that we intentionally use the interface name without "Prototype". We know the spec, but we decided to keep this as is. So, this is WontFix. We don't have a plan to change this for the time being.
,
Feb 1 2017
There is more discussion at Issue 643712 so marking this as a dupe. (We clearly can't WontFix this since it is blocking idlharness from being usable.)
,
Feb 1 2017
re: "blocking idlharness from being usable" - for IndexedDB and other test suites we have expectations with the necessary FAIL lines. Annoying to rebase when idlharness changes, but catches regressions. |
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by mfo...@chromium.org
, Jul 12 2016