Merge ESP32 arduino 3.0x patches from upstream

[willmmiles]

Import the Arduino 3.0.0 support patches from upstream. While WLED may not be looking at updating in the near future, we can keep our repo in sync.

[DRSDavidSoft]

I wish you could also consider incorporatng mathieucarbou's updates. It has significantly enhanced the original library:
https://github.com/mathieucarbou/ESPAsyncWebServer

This is the only fork with an excellent continuation of the original lib. If WLED's fork could be rebased on it, it would greatly improve the development experience moving forward!

[willmmiles]

I wish you could also consider incorporatng mathieucarbou's updates. It has significantly enhanced the original library: https://github.com/mathieucarbou/ESPAsyncWebServer

This is the only fork with an excellent continuation of the original lib. If WLED's fork could be rebased on it, it would greatly improve the development experience moving forward!

Is there a specific feature or feature set you're looking for? I haven't tested that fork specifically, but I did independently try doing some of the "custom structure to STL" replacements there, and saw a substantial increase in the code size on ESP32 -- enough that I ultimately backed out those changes and patched up the local abstractions instead. Unfortunately code size is the number one limited resource for the WLED project these days.

[DRSDavidSoft]

@willmmiles I understand, I was mostly looking for the optimizations such as mathieucarbou#55 regarding strings (you can read more about it at mathieucarbou#54), as well as some recent new features that is super useful when developing a custom interface for WLED or delivering it to the user in, when the bundle size is larger.

I didn't consider the increase in code size when I suggested this. With that being said, I believe the idealistic approach is to do the rebase anyway, and utilize flags to disable parts that aren't needed to decrease code size. Although this will take development costs and is not a priority. I also support taking WLED patches there if it can benefit the library as a whole, especially ones related to the handling of WebSocket.

For the time being, may I suggest opening the issues tab? That way if some important features are missing in this fork, we can simply create a new issue to coordinate adding it in a way that wouldn't cause substantial code size increase.

I hope in the future, the branches for this lib would be streamlined. There are some quality patches in both repos and it'd be exciting if they get mainlined.