Bug 7392 – GMAIL: XMLHttpRequest does not correctly report "Interactive" state on receipt of load data

• Status: RESOLVED
• Resolution: FIXED
• Component: DOM

[Eric Seidel (no email)]

GMAIL: XMLHttpRequest does not correctly report "Interactive" state on receipt of load data

Simple fix (and cleanup) attached.

This blocks GMail chat in Safari.

[Eric Seidel (no email)]

Created attachment 6627 [details]
simple fix (and a little cleanup)

[Darin Adler]

Comment on attachment 6627 [details]
simple fix (and a little cleanup)

I think it's awkward to have changeState have a side effect of always calling the listener if the passed state is Interactive. Seems like the wrong name for a function that will both change state and have this particular side effect.

I'd suggest instead doing something different at the caller in XMLHttpRequest::slotData; maybe code that detects the state and calls callReadyStateChangeListener explicitly if we're already in "interactive" state?

Also, I think that putting this code into a separate function makes the logic a little weaker, because now the guarantee that the load event is sent exactly once is pretty subtle. If anyone called callReadyStateChangeListener an extra time when the state was already Completed, we'd fire a separate load event.

The fix looks good, but I think it can be done in a way that's slightly clearer.

[Darin Adler]

Sorry, I see now that you already went one round with John Sullivan on trying to make this clearer. However, even given that I still think there's room for improvement.

[Eric Seidel (no email)]

Created attachment 6628 [details]
Addressing Darin's concerns

[Eric Seidel (no email)]

This is also in Radar:
<rdar://problem/4449442> GMAIL: XMLHttpRequest does not correctly report "Interactive" state on receipt of load data

[Alexey Proskuryakov]

Just for the record: the current version of Web Applications spec says "A readystatechange event shall immediately be dispatched at the object whenever the readyState attribute changes value. The readystatechange event must never be dispatched by the UA if the readyState attribute did not change."

But I agree that such behavior doesn't sound right.

[Darin Adler]

Comment on attachment 6628 [details]
Addressing Darin's concerns

Typo, "recieves" instead of "receives".

Looks fine, r=me.

[Alexey Proskuryakov]

A test for this is http/tests/xmlhttprequest/interactive-state.html

[Lucas Forschler]

Mass moving XML DOM bugs to the "DOM" Component.