This seems low-priority, but test/async/cancellable.wast in the CM repo main branch has a test case that currently fails on dev Wasmtime that I believe should pass. It tests whether thread.yield with cancellable set is directly resumed by subtask.cancel (allowing it to eagerly cancel itself, rather saving a "pending cancellation" that gets delivered some time later). Currently, subtask.cancel returns BLOCKED (-1), which suggests that the cancellable is not being considered.
This seems low-priority, but
test/async/cancellable.wastin the CM repomainbranch has a test case that currently fails ondevWasmtime that I believe should pass. It tests whetherthread.yieldwithcancellableset is directly resumed bysubtask.cancel(allowing it to eagerly cancel itself, rather saving a "pending cancellation" that gets delivered some time later). Currently,subtask.cancelreturnsBLOCKED(-1), which suggests that thecancellableis not being considered.