When doing codegen instead of transforming directly into C code instead better to create C AST and codegen into C AST. This will allow later to have dynamic break and continue statements. Which should know jump locations later, after codegen of current block is finished.
Right now it requires 2 runs for each async block which is not good for performance.
When doing codegen instead of transforming directly into C code instead better to create C AST and codegen into C AST. This will allow later to have dynamic break and continue statements. Which should know jump locations later, after codegen of current block is finished. Right now it requires 2 runs for each async block which is not good for performance.