r3bl-open-core: Async TUI Framework in Rust
ToolingComments
does an async pipeline actually help if the bottleneck is the emulator's render loop?
What if the trend toward web-style wrappers is a pragmatic response to the complexity of modern terminal capabilities? If developers prefer the ergonomics of something like iocraft, a purely async pipeline might be solving a performance bottleneck that the market is already comfortable trading for ease of styling.
We saw this with early reactive UI attempts in C++ where styling ease served as a mask for massive memory leaks. The overhead of a browser engine in a TUI is a technical debt that always comes due.
The async architecture could actually simplify the multi-process orchestration mentioned in the post. This could enable TUIs that handle high-frequency telemetry without locking the main thread, which would be a significant win for observability tools.
If the trade-off is performance versus ergonomics, where are the benchmarks? I would like to see the specific memory delta between this and an Ink-based wrapper before concluding the overhead is the primary constraint.