Tom Lauwaerts

Fixes #78

I agree. The callback system should definitely be reset completely when a new module gets loaded. Except for the event queue imo. For the persistant module. When that feature is...

Version [0.5.0](https://github.com/TOPLLab/WARDuino/releases/tag/v0.5.0) just released with this fix.

One issue that also needs further investigation and understanding, is how the code even works now. Since the proxy handler should only return the result once it encounters the guard...

Guard blocks are implemented incorrectly for callbacks. Currently pushed before the callback frames, and popped after the first callback has executed. Fix: guard pushed before callback frames and the return...

https://github.com/TOPLLab/WARDuino/blob/e7613947f4f662b1519d0509d24cd49e0efcb620/tests/integration/wasm_tests.cpp#L67-L70 > This would indeed be a convenient addition to the WARDuino class _Originally posted by @beardhatcode in https://github.com/TOPLLab/WARDuino/pull/26#discussion_r775922883_

> @carllocos , I have been digging a bit into emcc and the following works for me with EMCC 2.0 > > ``` > emcc FILE \ > -s STANDALONE_WASM...

Rerunning the benchmarks with the same setup (`arduino-cli` and arduino-esp32 `v2.0.3`) gives slightly slower results as before (compared to SPE results). | benchmark | SPE results (v0.1.0) | v0.2.0 |...

Executing the `catalan` benchmark with the emulator gives the following performance profile:  How representative this is, is in doubt.

IO fflushes were the culprits. Removed in ac669055e1e49e043b3d568a269bda1a776accae.