1413920 - nsMultiplexInputStream should call OnInputStreamReady on close

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, enhancement)

Description

[Andrea Marchesini [:baku]]

+++ This bug was initially created as a clone of Bug #1413863 +++

We don't have valid tests for the asyncWait call. This bug is about implementing a test.

Comment 1

[Andrea Marchesini [:baku]]

Attachment: multiplexTest.patch

Note that if AsyncWait() is called without an EventTarget, nsMultiplexInputStream uses the main-thread.

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 8924546 [details] [diff] [review]
multiplexTest.patch

This is changing some non-testing only code. Automatic r-.
Re-purpose this bug and change its title or something, but please no functionality changes in patches which are supposed to be about testing.

(bug 1404910 was weird enough)

Comment 3

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 8924546 [details] [diff] [review]
multiplexTest.patch

>+// AsyncWait - sync - closureOnly
>+TEST(TestMultiplexInputStream, AsyncWait_sync_closureOnly) {
>+  nsCOMPtr<nsIInputStream> is = CreateStreamHelper();
>+
>+  nsCOMPtr<nsIAsyncInputStream> ais = do_QueryInterface(is);
>+  ASSERT_TRUE(!!ais);
>+
>+  RefPtr<testing::InputStreamCallback> cb =
>+    new testing::InputStreamCallback();
>+
>+  ASSERT_EQ(NS_OK, ais->AsyncWait(cb, nsIAsyncInputStream::WAIT_CLOSURE_ONLY,
>+                                  0, nullptr));
>+  ASSERT_FALSE(cb->Called());
>+
>+  ais->CloseWithStatus(NS_ERROR_FAILURE);
>+  ASSERT_FALSE(cb->Called());
>+
>+  // Eventually it is called.
>+  MOZ_ALWAYS_TRUE(mozilla::SpinEventLoopUntil([&]() { return cb->Called(); }));
>+  ASSERT_TRUE(cb->Called());
>+}
>+
>+// AsyncWait - async
>+TEST(TestMultiplexInputStream, AsyncWait_async_closureOnly) {
>+  nsCOMPtr<nsIInputStream> is = CreateStreamHelper();
>+
>+  nsCOMPtr<nsIAsyncInputStream> ais = do_QueryInterface(is);
>+  ASSERT_TRUE(!!ais);
>+
>+  RefPtr<testing::InputStreamCallback> cb =
>+    new testing::InputStreamCallback();
>+  nsCOMPtr<nsIThread> thread = do_GetCurrentThread();
>+
>+  ASSERT_EQ(NS_OK, ais->AsyncWait(cb, nsIAsyncInputStream::WAIT_CLOSURE_ONLY,
>+                                  0, thread));
>+
>+  ASSERT_FALSE(cb->Called());
>+  ais->CloseWithStatus(NS_ERROR_FAILURE);
>+  ASSERT_FALSE(cb->Called());
>+
>+  // Eventually it is called.
>+  MOZ_ALWAYS_TRUE(mozilla::SpinEventLoopUntil([&]() { return cb->Called(); }));
>+  ASSERT_TRUE(cb->Called());
>+}
I don't undersatnd the difference here. What is sync in the sync case (the first case).
The only difference is whether thread is passed, but what is then sync and what is async.
If nothing else, add some comment what these are testing. It isn't clear now.

Comment 4

[Andrea Marchesini [:baku]]

Attachment: multiplexTest.patch

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/7c79d729d27e
nsMultiplexInputStream should call OnInputStreamReady on close, r=smaug

Comment 6

[Cristina Coroiu [:ccoroiu]]

https://hg.mozilla.org/mozilla-central/rev/7c79d729d27e