Bug 58467 – IndexedDB createIndex should fail if options arg is invalid

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Mark Pilgrim (Google)]

Two possible ways the options argument can be invalid: it's not a JavaScript object, or the object contains attributes other than keyPath and autoIncrement. In each case, Mozilla throws but WebKit does not. Attaching tests for each case.

[Mark Pilgrim (Google)]

Created attachment 89426 [details]
test case #1: options arg is not an object

[Mark Pilgrim (Google)]

Created attachment 89427 [details]
test case #2: options arg contains unknown attributes

[Mark Pilgrim (Google)]

Apologies, my original comment was slightly incorrect. The valid attributes for the options argument are unique and multirow. The test cases still properly demonstrate the problem of not throwing on unknown attributes.

[Joshua Bell]

Test case #2 is no longer valid - IndexedDB spec has changed to use the (new) WebIDL dictionary type for the options object, which ignores unknown attributes - http://dev.w3.org/2006/webapi/WebIDL/#es-dictionary

Test case #1 is still valid; in Chrome 15 I can pass a string as the options argument and no exception is thrown. Per WebIDL: "If Type(V) is not Object, then throw a TypeError."

[Eugene Girard]

Created attachment 125846 [details]
Patch

[Joshua Bell]

Comment on attachment 125846 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=125846&action=review

Test changes LGTM

> LayoutTests/ChangeLog:6
> +        https://bugs.webkit.org/show_bug.cgi?id=58467

Nit: Keep the specific URL under the bug subject, have the comment lines in a separate paragraph. Otherwise the commit queue may get confused.

[Tony Chang]

Comment on attachment 125846 [details]
Patch

As Joshua mentions, the bug URL should be directly below the bug title.  See the other ChangeLog entries as an example.

[Eugene Girard]

Created attachment 125875 [details]
Patch

[Eugene Girard]

(In reply to comment #7)
> (From update of attachment 125846 [details])
> As Joshua mentions, the bug URL should be directly below the bug title.  See the other ChangeLog entries as an example.

Thanks.  Fixed in this patch.

[WebKit Review Bot]

Comment on attachment 125875 [details]
Patch

Rejecting attachment 125875 [details] from commit-queue.

New failing tests:
perf/array-reverse.html
Full output: http://queues.webkit.org/results/11459034

[WebKit Review Bot]

Comment on attachment 125875 [details]
Patch

Clearing flags on attachment: 125875

Committed r106991: <http://trac.webkit.org/changeset/106991>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.