Is it possible to call a dart function synchronously?

[EugenePisotsky]

I'm trying to integrate taffy library and it needs a function that measures the size of the text (something like this). I want to pass a dart function, but it needs to be synchronous because otherwise I guess I won't be able to use it with taffy. Is it impossible right now?

[fzyzcjy]

Hi, this is not yet possible, because rust->dart uses Dart's "port" mechanism which is async. However, if the Rust code runs on the main thread that dart is on, then theoretically it may be possible to call dart synchronously (not yet supported though). Feel free to PR for it! Alternatively, I may work on it but may not be soon, since this requires some amount of work.

Btw, curious the use case of taffy in Dart - it seems Flutter already has layout system

[EugenePisotsky]

Hi. Thanks for the quick response!

Btw, curious the use case of taffy in Dart - it seems Flutter already has layout system

This isn’t for the main app. I need it for small extensions written in JS that can provide custom UI blocks. The QuickJS runtime executes the code, Taffy calculates the layout, and everything is rendered in Flutter. While it would be possible to map directly to Flutter’s layout system, I have other reasons for using Taffy in this case :)

Feel free to PR for it!

I’ll consider it.

P.S. Thanks for your work on this library, it's a gem!

[fzyzcjy]

Hi, you are welcome!

This isn’t for the main app. I need it for small extensions written in JS that can provide custom UI blocks. The QuickJS runtime executes the code, Taffy calculates the layout, and everything is rendered in Flutter. While it would be possible to map directly to Flutter’s layout system, I have other reasons for using Taffy in this case :)

I see :)

I’ll consider it.

Feel free to ping me if any hints are needed! Roughly speaking, Dart already has the ability to allow raw C (thus Rust) code to call Dart synchronously. We can simply utilize the SSE codec (one codec in frb that converts arbitrary types into a single Vec<u8>), then we only need to implement the logic to transfer such a byte array and everything is done.