Bug 161955 – test262: TypedArray constructors length should be 3 and configurable

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Joseph Pecoraro]

test262: %TypedArray% constructors length should be 3

Things like:

    Float32Array.length
    Uint8ClampedArray.length

Should all be 3.

[Joseph Pecoraro]

Created attachment 288793 [details]
[PATCH] Proposed Fix

[Joseph Pecoraro]

Comment on attachment 288793 [details]
[PATCH] Proposed Fix

Err, this doesn't pass the test262 tests completely.

[Joseph Pecoraro]

The property should also be configurable (deletable).

[Joseph Pecoraro]

Created attachment 288797 [details]
[PATCH] Proposed Fix

[Joseph Pecoraro]

Comment on attachment 288797 [details]
[PATCH] Proposed Fix

And now this doesn't pass some of our layout tests that expected 0.

[Joseph Pecoraro]

Created attachment 288800 [details]
[PATCH] Proposed Fix

This magically fixes DataView.length as well.

[Joseph Pecoraro]

Created attachment 288801 [details]
[PATCH] Proposed Fix

Even more random LayoutTests.

[Joseph Pecoraro]

Created attachment 288802 [details]
[PATCH] Proposed Fix

[Build Bot]

Comment on attachment 288802 [details]
[PATCH] Proposed Fix

Attachment 288802 [details] did not pass ios-sim-ews (ios-simulator-wk2):
Output: http://webkit-queues.webkit.org/results/2071585

New failing tests:
js/dom/constructor-length.html

[Build Bot]

Created attachment 288805 [details]
Archive of layout-test-results from ews123 for ios-simulator-elcapitan-wk2

The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews123  Port: ios-simulator-elcapitan-wk2  Platform: Mac OS X 10.11.6

[Mark Lam]

Comment on attachment 288802 [details]
[PATCH] Proposed Fix

View in context: https://bugs.webkit.org/attachment.cgi?id=288802&action=review

Please also look into the ios-sim failure.  It seems to be related.  I'll r+ if I can find the spec that says that the length prop should be configurable.

> Source/JavaScriptCore/ChangeLog:7
> +        test262: TypedArray constructors length should be 3
> +        https://bugs.webkit.org/show_bug.cgi?id=161955
> +
> +        Reviewed by NOBODY (OOPS!).
> +

Can you quote the spec url here that says the length should be 3:
See: https://tc39.github.io/ecma262/#sec-typedarray-constructors

I'm still looking for the spec that says that the "length" property should be configurable.

[Joseph Pecoraro]

Take a look at the test262 test contents. It mentions the exact part of the spec, here it is completely non obvious sine it is "if not otherwise mentioned". I'd link you but I'm on my phone!

[Mark Lam]

Comment on attachment 288802 [details]
[PATCH] Proposed Fix

View in context: https://bugs.webkit.org/attachment.cgi?id=288802&action=review

r=me if ios-sim failures are resolved.

>> Source/JavaScriptCore/ChangeLog:7
>> +
> 
> Can you quote the spec url here that says the length should be 3:
> See: https://tc39.github.io/ecma262/#sec-typedarray-constructors
> 
> I'm still looking for the spec that says that the "length" property should be configurable.

Thanks for the tip to look in the test262 test file.  Here's the relevant spec:

https://tc39.github.io/ecma262/#sec-ecmascript-standard-built-in-objects
"Unless otherwise specified, the length property of a built-in Function object has the attributes { [[Writable]]: false, [[Enumerable]]: false, [[Configurable]]: true }."

Please add these spec details into the ChangeLog.  Thanks.

[Joseph Pecoraro]

Created attachment 288867 [details]
[PATCH] For Landing

Yep, I hand modified the expectations for ports and missed the Float*Arrays in a few of them. Updated the ChangeLog.

[WebKit Commit Bot]

Comment on attachment 288867 [details]
[PATCH] For Landing

Clearing flags on attachment: 288867

Committed r205932: <http://trac.webkit.org/changeset/205932>