Colin White

It's likely related to this issue: https://github.com/orfjackal/retrolambda/issues/121

This will be fixed in the next alpha release via https://github.com/coil-kt/coil/pull/2398. Coil now places a global cap on an allocated bitmap's size (default to 4096x4096). In practice this prevents allocating...

I think I'd want to avoid adding a separate `maxHardwareBitmapSize` API since it's bit niche. That said, this should be easy to enforce with a custom `Interceptor`. You can do...

@Foolbar Ah gotcha - sorry you're right `maxBitmapSize` fulfill a different use case. `maxBitmapSize` guards against allocating a bitmap that could OOM or exceed the max hardware bitmap size (especially...

Thanks for taking a look at this! Though I'm a bit confused - won't `async(Dispatchers.Main.immediate)` always run sequentially (even if wrapped with `withContext(Dispatchers.Default)`) since `Dispatchers.Main.immediate` is a single threaded dispatcher?...

Interesting thanks for digging in! Looks like there are two main avenues to improve perf like you called out: - **Speed up `scope.launch` in `AsyncImagePainter.onRemembered`**: I'm not sure what we...

@mlykotom This is awesome, thanks! Do you want to merge this in with the scroll perf benchmark? To improve scroll perf even more, I think if we offer an opt-in...

@mlykotom Thanks for all the work on this! I'll take a look through the other optimizations this weekend. Unfortunately we can't make any binary incompatible changes to the 2.x branch,...

On second thought this might not be the best idea since it locks us into a max ~2GB memory cache size as mentioned previously. This could be a real limitation...

It looks like `androidx.collection.LruCache` doesn't publish versions for JS and wasmJs either unfortunately. Given these limitations I think it makes sense to keep Coil's custom `LruCache`. It's small (