Fix: Abort in WASIX thread now returns non-zero exit code

lib/wasix/src/syscalls/wasix/thread_spawn.rs

@@ -213,6 +213,17 @@ fn call_module_internal<M: MemorySize>(
 
     let exit_code = handle_thread_result(env, store, thread_result)?;
 
+    // Fix for issue #5501: When a thread exits with a non-zero exit code (e.g., from abort()),
+    // we must set the thread's status BEFORE calling on_exit(). This is because on_exit()
+    // eventually calls process.terminate() which propagates the exit code to all threads.
+    // Without setting the status first, there's a race where the thread handle cleanup
+    // (WasiThreadHandleProtected::drop) could set the status to Success(0) before the
+    // correct exit code is recorded. The existing test `test_snapshot_worker_panicking`
+    // validates this behavior by ensuring a trap in a worker thread results in exit code 173.
+    if let Some(code) = exit_code {
+        env.data(store).thread.set_status_finished(Ok(code));
+    }
+
     // Clean up the environment on exit
     env.on_exit(store, exit_code);
     Ok(())