Fix Geometry in Firefox

[bagnell]

Change Enumeration reference equality checks to perform a value comparison. Work around transferable object bug in Firefox:

https://bugzilla.mozilla.org/show_bug.cgi?id=841904

This fixes #1039.

[mramato]

This is more of a question for @shunter. Do we expect to always have to do value comparisons for enums in webworkers? It kind of negates the point of having Enumeration be an object rather than just a constant value, doesn't it? It's also really easy to write code that appears to run fine (and works under Chrome) but then fails in Firefox. If this is temporary, fine; but if now, maybe we should rethink our enum strategy? Just a thought.

[bagnell]

We always have to do value comparisons for enums transferred to or from a
worker. Its the same in both Firefox and Chrome. The reason it was failing
in Firefox is because the extension for unsigned int indices was
unavailable. The function to fit the indices to unsigned shorts was being
called in Firefox and not Chrome.
On Aug 19, 2013 10:34 PM, "Matthew Amato" notifications@github.com wrote:

This is more of a question for @shunter https://github.com/shunter. Do
we expect to always have to do value comparisons for enums in webworkers?
It kind of negates the point of having Enumeration be an object rather
than just a constant value, doesn't it? It's also really easy to write code
that appears to run fine (and works under Chrome) but then fails in
Firefox. If this is temporary, fine; but if now, maybe we should rethink
our enum strategy? Just a thought.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/1042#issuecomment-22918875
.

[mramato]

I can confirm that the tests all pass now and Firefox Sandcastle examples work as expected. As an aside, is there any sort of event that gets fired as geometry is processed? It would be useful to put up some sort of loading indicator for the G&A Sandcastle example, it's really a dog on Firefox.

I'll let @shunter do the actual merge.

[pjcozzi]

maybe we should rethink our enum strategy?

I have agreed with this for a while, and in #1023, @shunter is also starting to consider it. Propose something on the forum or just open an issue.

As an aside, is there any sort of event that gets fired as geometry is processed?

Good idea. We'll put it on the roadmap, but it may be a bit before we get to it.

[shunter]

You need to convert the typed arrays even if the ArrayBuffer is already in transferableObjects.

[bagnell]

@shunter This is ready for another review.