Evaluate query results

[ryanhamley]

Launch Checklist

This is a potential fix for #8373 It's an alternate approach to #9198 but its performance seems to be about the same

• briefly describe the changes in this PR
• include before/after visuals or gifs if this PR includes visual changes
• write tests for all new functionality
• document any changes to public APIs
• post benchmark scores
• manually test the debug page
• tagged @mapbox/studio and/or @mapbox/map-design-team if this PR includes style spec or visual changes
• tagged @mapbox/gl-native if this PR includes shader changes or needs a native port
• apply changelog label ('bug', 'feature', 'docs', etc) or use the label 'skip changelog'
• add an entry inside this element for inclusion in the mapbox-gl-js changelog: <changelog></changelog>

[ryanhamley]

This shows a sample flame chart using the query_features.html debug page without a geometry passed to queryRenderedFeatures (e.g. queryRenderedFeatures(null, options))

This is for a random query with a geometry (queryRenderedFeatures(box, options)). In isolation, evaluateQueryResults seems to almost always be fast. This is confusing me since the QueryPoint benchmark shows a major performance regression with both approaches I've tried.

The benchmark results seem broadly similar to #9198

cc @mourner

[mourner]

This seems to be heading in the right direction — we see a major improvement in the QueryBox bench (the many-features use case which the original issue was about), although this stems from the serializedLayers caching which we could cut out and merge in a separate PR.

For the queryPoint bench, you could run the Chrome Profiler while the benchmark is running to see what's the major offender. My guess is listImages which could be a major hit considering previous issues like #9261.

[ryanhamley]

closing this in favor of #9198