Follow upstream WASAPI teardown without stderr diagnostics

This commit is contained in:
fufesou 2026-09-12 23:16:49 +08:00
parent 8aa24a369b
commit b8ad274267

View file

@ -4,7 +4,6 @@ use crate::{
BackendSpecificError, Data, InputCallbackInfo, OutputCallbackInfo, PauseStreamError,
PlayStreamError, SampleFormat, StreamError,
};
use std::io::{self, Write};
use std::mem;
use std::os::windows::io::{AsRawHandle, FromRawHandle, OwnedHandle};
use std::ptr;
@ -176,21 +175,10 @@ impl Stream {
impl Drop for Stream {
#[inline]
fn drop(&mut self) {
// eprintln! panics on stderr write errors, e.g. a closed pipe reporting
// ERROR_NO_DATA (232), which Rust maps to BrokenPipe. Ignore diagnostic
// I/O errors so worker joining and handle cleanup still run.
if let Err(error) = self.push_command(Command::Terminate) {
let _ = writeln!(
io::stderr().lock(),
"cpal: WASAPI termination notification failed: {error}"
);
}
let _ = self.push_command(Command::Terminate);
if let Some(thread) = self.thread.take() {
if thread.thread().id() != thread::current().id() && thread.join().is_err() {
let _ = writeln!(
io::stderr().lock(),
"cpal: WASAPI worker panicked during stream processing"
);
if thread.thread().id() != thread::current().id() {
let _ = thread.join();
}
}
}