[Webkit-unassigned] [Bug 58738] Port Mozilla's IndexedDB tests: versionchange

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=58738





--- Comment #2 from David Grogan <dgrogan at chromium.org>  2011-04-18 11:10:09 PST ---
(From update of attachment 89937)
View in context: https://bugs.webkit.org/attachment.cgi?id=89937&action=review

> third_party/WebKit/LayoutTests/storage/indexeddb/mozilla/versionchange.html:42
> +    // db1 will be open when we call db2.setVersion, which will trigger this

Put this in a debug("").

Copy/pasted comment from jorlow that's applicable: "Stuff like this should be in shouldBe/testPassed/debug()'s.  The idea is that you should be able to follow along via the on-screen output."

> third_party/WebKit/LayoutTests/storage/indexeddb/mozilla/versionchange.html:62
> +        shouldBe("event.target.version", "'1'");

Puzzled at the nested quotes - do we need them to pass the shouldBe assert?  Have I just never noticed them elsewhere in our tests?

> third_party/WebKit/LayoutTests/storage/indexeddb/mozilla/versionchange.html:70
> +    request.onsuccess = postSetVersion;

Can you add a listener for request.onblocked?  UnexpectedBlockCallback or function() { testPassed("db2 received blocked event") } would be sufficient.

I think we will fire a blocked event at db2 in spite of the call to db1.close, but I remember the spec being unclear on this point.  Do you know what firefox does?

> third_party/WebKit/LayoutTests/storage/indexeddb/mozilla/versionchange.html:89
> +    // this will block because db2 is still open

Yeah, put this in debug("") inside the onblocked function

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.