performance regression in 2.8.0

[veggiesaurus]

More of a question than a bug report. There is a significant performance regression from 2.7.3 -> 2.8.0 in situations with lots of data points. It seems to be related to _resolvePointOptions in updateElement.

Codepens demonstrating the performance difference:

2.7.3
2.8.0

(the y-value of the top trace is based on the time taken to update the graph).

A more general question: Is there a more performant route for displaying homogenous, data-heavy traces? (such as a line with 1000 points, all with the exact same properties, other than position)

edit: master is a bit better, but still slower than 2.7.3 in this case, and about 40% of the time appears to be spent in _resolveDataElementOptions.

[benmccann]

I've been bothered by this as well and agree it should be possible to improve performance. I had one idea here: #6574