Dane Springmeyer

> I wonder if closest_point_impl can break out of the loop if it gets a hit? Yes, that could work - @artemp thoughts? Something like this perhaps? ```diff diff --git...

Thanks for writing this up and providing a view into the state of things. non-deterministic builds feels like a major drawback. > If we can demonstrate a significant benefit for...

> It seems to me that the only reliable way to support this is having in the OSM database itself the information about the language that was used in the...

You may also be interested to check out https://github.com/mapbox/vtshaver. It is the library that powers the Mapbox "style-optimized" API which you reference above. As context, there are a few ways...

> I noticed that adding the dependency on the upstream arc package did slow down the benchmark a little bit. @mfedderly thanks for calling that out. The Arc.js package (both...

I've taken a look at this and I think the problem is that icons (as well as text) have a `perspective_ratio` applied to their size on the GPU (in the...

One potential solution for this would be to detect when icons are being offset and skip applying the perspective_ratio. This would then respect their exact size and placement without modification....

> Interesting, how does this solution affect performance? My solution currently sends another uniform type into the shaders, so there will be some overhead there. But in general it avoids...

These all passed on my fork: They all run in seconds, except for the python 3.14 job which takes over 10 minutes do to needing to source compile grpcio (but...

@mc-nv is this something that would be useful? I can imagine you have similar testing internally, but having this external on the open source repo could be useful for ensuring...