Anand Eng-Thakker
                                            Anand Eng-Thakker
                                        
                                    I think the biggest open issue we need to resolve here is the one about cross-platform compatibility. Given @1ec5's point in https://github.com/mapbox/mapbox-gl-js/issues/4089#issuecomment-276799020 that ICU and ECMAScript regexp syntaxes mostly overlap,...
Thanks for this proposal @indus! Separate from the `copyLayers` part and the alternative nested styles proposal in #4000, I think that `copySources` (or something like it) might fill a current...
As I understand it, the `copySources` option basically _is_ developer-provided metadata, except perhaps for the word "copy" in its name. It is an option that dictates the interpretation of the...
> Having GL JS able to decide whether to preserve or replace a source without the user explicitly having to make this performance decision is even better. If we want...
Alternative proposal at #6701
> We may want to create some infrastructure to automatically plumb together constructor options and getters @lucaswoj might be worth considering doing this auto-plumbing in conjunction with https://github.com/mapbox/mapbox-gl-js/issues/2741
Thanks for describing this pain point @gmaclennan. You're right: `map.loaded()` answers _both_ whether the style has been initially loaded _and_ whether there have been source/layer changes that haven't yet been...
I think `style.load` is equivalent to the _first_ `styledata`: https://github.com/mapbox/mapbox-gl-js/blob/a5a8dfe5b90737dbb295eaca7f41c667ae4060a8/src/style/style.js#L243-L244 (Confusingly, `this.fire(new Event('data', {dataType: 'style'}))` results in a `styledata` event) `styledata` would also be fired subsequently, after other changes to...
The Rollup plugin docs [here](https://github.com/rollup/rollup/wiki/Plugins) suggest that the sourcemap should be `null` if the transform doesn't change the source, versus `{mappings: ''}` if "it doesn't make sense to generate a...
I'm mostly 👍 on a barebones nested-list approach to object types _if_ dx is going to have its own type system. Couple other questions about that prospect, though: - Would...