hover / hoverOff never work for L.glify.lines()

[utilmind]

Hi! I investigated the issue why hover and hoverOff methods are never triggered for Glify lines...

Upon debugging static method tryHover in lines.ts I noticed that when it checks whether the hover point in inBounds(e.latlng, bounds) it accepts invalid (impossible) bounds, like the following ones:

{
    "_southWest": {
        "lat": -185.939713,
        "lng": 24.563612
    },
    "_northEast": {
        "lat": -66.970467,
        "lng": 71.27343
    }
}

It looks like lat and lng are mixed up.

After further investigation I also noticed that bounds constant in the TryHover method receiving incorrect input data, array of data.features instead of the featureCollection object,

const bounds = geoJSON(data.features).getBounds();

instead of

const bounds = geoJSON(data).getBounds();

I'm still testing it, then will prepare a patch / pull request if everything good...

UPD. data.features are not the issue, Leaflet's geoJSON() accept both array or object, so it's just micro optimization. But the real issue is that Leaflet's getBounds() returns bounds (lat/lng) in different sequence than expected for WebGL. lat/lng should be just swapped. This seems easier than swap lat/lng in all features of input data.

Or alternatively, something different should be used instead of Leaflet's getJSON().

[utilmind]

Updated PR #179 .

I removed incompatible Leaflet's geoJSON() function and added own getBounds() method for the Lines class. It's optimized for performance, bounding rectangle for all features is cached to avoid recalculation if the number of features is not changed.

So the main fix is removing the line with Leaflet's geoJSON() and checking rectangle with new getBounds()’ (actually incBounds(), which is getBounds()+ little padding specified insensitivityHover` option).

However, in general, that inBounds() check can be skipped completely. Its only purpose is a small speed optimization to avoid checking if the mouse cursor is outside the area with the features.

[utilmind]

Please close this issue after approval of Pull Request #179 .