Rename SymbolBucket::addFeatures and SymbolBucket::placeFeatures

[jfirebaugh]

From #2873 (comment):

There are actually three steps:

• parseFeatures: determines the text we need to display by replacing tokens and applying text transformations, and the glyphs we need
• addFeatures: determines the potential position of labels
• placeFeatures: actually places them and adds them to a buffer

placeFeatures is currently called as the last step in addFeatures, but there's no good reason to; we could separate that out and call the addFeatures without placing as soon as we have the symbols. Then we would have less work to do later. However, it means we'd have to introduce a third step beyond pending.

[jfirebaugh]

Also @kkaefer says "addFeatures is poorly named" -- we should rename it to something better.

[lucaswoj]

I agree that we should develop some better terminology around this API. How about

• resolveText
• calculateTextPlacement
• placeFeatures

[kkaefer]

text => symbol, but otherwise good

[ansis]

they were separated in #3543 but not renamed