Fix for marker content update bug

[tst-dhudlow]

See #286

[ebidel]

Please add a test.

[ebidel]

ping

[tst-dhudlow]

When I run the current tests with polymer test I get a bunch of failures. (Timeouts and expected { Object (__gm, gm_accessors_, ...) } to equal undefined...). Can you point me to any docs on how to get these running so I can add?

[ebidel]

Running them myself, it looks like there are errors. I can help get that working again.

Mind just adding a test for yourself?

[tst-dhudlow]

Sure thing, I'll give it a shot. I just figured I was probably not configured correctly.

[tst-dhudlow]

FYI, still working on reproducing this during tests. Having a bit of a time with it, but making progress.

[tst-dhudlow]

Test created.

Note that this test will also pass with only changes to google-map-marker.html from #294

I still believe this change is logically correct though it may not technically be necessary after we hammer out #294.

[ebidel]

TBH, I'm not sure about this change. Generally, hacking around timing issues like this means there's an underlying problem that needs to be addressed. In this case, #286 is using google-map-marker outside of google-map. That's not intended behavior.

[tst-dhudlow]

Well, it's possible you're right that issues related to #294 are the real problem. I could try to verify this by tweaking the test, but it's my understanding that the issues is caused because when the marker is detached, its content is updated, and then it's re-attached, it doesn't update its content. Templates detach stuff from the DOM so they can be re-used and that's why this came up here.

The question is, do you want content to be updated if it is added/modified while the marker is in a detached state?