Merge pull request #164 from jasagredo/js/debug-auto-label

Cabal flag to control auto-labelling of threads

Author: simonmar

Control/Concurrent/Async/Internal.hs

@@ -57,6 +57,16 @@ import GHC.Exts
 import GHC.IO hiding (finally, onException)
 import GHC.Conc (ThreadId(..))
 
+#ifdef DEBUG_AUTO_LABEL
+import qualified GHC.Stack
+#endif
+
+#ifdef DEBUG_AUTO_LABEL
+#define CALLSTACK GHC.Stack.HasCallStack =>
+#else
+#define CALLSTACK
+#endif
+
 -- -----------------------------------------------------------------------------
 -- STM Async API
 
@@ -95,40 +105,53 @@ compareAsyncs (Async t1 _) (Async t2 _) = compare t1 t2
 -- (see module-level documentation for details).
 --
 -- __Use 'withAsync' style functions wherever you can instead!__
-async :: IO a -> IO (Async a)
+async ::
+  CALLSTACK
+  IO a -> IO (Async a)
 async = inline asyncUsing rawForkIO
 
 -- | Like 'async' but using 'forkOS' internally.
-asyncBound :: IO a -> IO (Async a)
+asyncBound ::
+  CALLSTACK
+  IO a -> IO (Async a)
 asyncBound = asyncUsing forkOS
 
 -- | Like 'async' but using 'forkOn' internally.
-asyncOn :: Int -> IO a -> IO (Async a)
+asyncOn ::
+  CALLSTACK
+  Int -> IO a -> IO (Async a)
 asyncOn = asyncUsing . rawForkOn
 
 -- | Like 'async' but using 'forkIOWithUnmask' internally.  The child
 -- thread is passed a function that can be used to unmask asynchronous
 -- exceptions.
-asyncWithUnmask :: ((forall b . IO b -> IO b) -> IO a) -> IO (Async a)
+asyncWithUnmask ::
+  CALLSTACK
+  ((forall b . IO b -> IO b) -> IO a) -> IO (Async a)
 asyncWithUnmask actionWith = asyncUsing rawForkIO (actionWith unsafeUnmask)
 
 -- | Like 'asyncOn' but using 'forkOnWithUnmask' internally.  The
 -- child thread is passed a function that can be used to unmask
 -- asynchronous exceptions.
-asyncOnWithUnmask :: Int -> ((forall b . IO b -> IO b) -> IO a) -> IO (Async a)
+asyncOnWithUnmask ::
+  CALLSTACK
+  Int -> ((forall b . IO b -> IO b) -> IO a) -> IO (Async a)
 asyncOnWithUnmask cpu actionWith =
   asyncUsing (rawForkOn cpu) (actionWith unsafeUnmask)
 
-asyncUsing :: (IO () -> IO ThreadId)
-           -> IO a -> IO (Async a)
+asyncUsing ::
+  CALLSTACK
+  (IO () -> IO ThreadId) -> IO a -> IO (Async a)
 asyncUsing doFork = \action -> do
    var <- newEmptyTMVarIO
+   let action_plus = debugLabelMe >> action
    -- t <- forkFinally action (\r -> atomically $ putTMVar var r)
    -- slightly faster:
    t <- mask $ \restore ->
-          doFork $ try (restore action) >>= atomically . putTMVar var
+          doFork $ try (restore action_plus) >>= atomically . putTMVar var
    return (Async t (readTMVar var))
 
+
 -- | Spawn an asynchronous action in a separate thread, and pass its
 -- @Async@ handle to the supplied function.  When the function returns
 -- or throws an exception, 'uninterruptibleCancel' is called on the @Async@.
@@ -144,41 +167,51 @@ asyncUsing doFork = \action -> do
 -- to `withAsync` returns, so nesting many `withAsync` calls requires
 -- linear memory.
 --
-withAsync :: IO a -> (Async a -> IO b) -> IO b
+withAsync ::
+  CALLSTACK
+  IO a -> (Async a -> IO b) -> IO b
 withAsync = inline withAsyncUsing rawForkIO
 
 -- | Like 'withAsync' but uses 'forkOS' internally.
-withAsyncBound :: IO a -> (Async a -> IO b) -> IO b
+withAsyncBound ::
+  CALLSTACK
+  IO a -> (Async a -> IO b) -> IO b
 withAsyncBound = withAsyncUsing forkOS
 
 -- | Like 'withAsync' but uses 'forkOn' internally.
-withAsyncOn :: Int -> IO a -> (Async a -> IO b) -> IO b
+withAsyncOn ::
+  CALLSTACK
+  Int -> IO a -> (Async a -> IO b) -> IO b
 withAsyncOn = withAsyncUsing . rawForkOn
 
 -- | Like 'withAsync' but uses 'forkIOWithUnmask' internally.  The
 -- child thread is passed a function that can be used to unmask
 -- asynchronous exceptions.
-withAsyncWithUnmask
-  :: ((forall c. IO c -> IO c) -> IO a) -> (Async a -> IO b) -> IO b
+withAsyncWithUnmask ::
+  CALLSTACK
+  ((forall c. IO c -> IO c) -> IO a) -> (Async a -> IO b) -> IO b
 withAsyncWithUnmask actionWith =
   withAsyncUsing rawForkIO (actionWith unsafeUnmask)
 
 -- | Like 'withAsyncOn' but uses 'forkOnWithUnmask' internally.  The
 -- child thread is passed a function that can be used to unmask
 -- asynchronous exceptions
-withAsyncOnWithUnmask
-  :: Int -> ((forall c. IO c -> IO c) -> IO a) -> (Async a -> IO b) -> IO b
+withAsyncOnWithUnmask ::
+  CALLSTACK
+  Int -> ((forall c. IO c -> IO c) -> IO a) -> (Async a -> IO b) -> IO b
 withAsyncOnWithUnmask cpu actionWith =
   withAsyncUsing (rawForkOn cpu) (actionWith unsafeUnmask)
 
-withAsyncUsing :: (IO () -> IO ThreadId)
-               -> IO a -> (Async a -> IO b) -> IO b
+withAsyncUsing ::
+  CALLSTACK
+  (IO () -> IO ThreadId) -> IO a -> (Async a -> IO b) -> IO b
 -- The bracket version works, but is slow.  We can do better by
 -- hand-coding it:
 withAsyncUsing doFork = \action inner -> do
   var <- newEmptyTMVarIO
   mask $ \restore -> do
-    t <- doFork $ try (restore action) >>= atomically . putTMVar var
+    let action_plus = debugLabelMe >> action
+    t <- doFork $ try (restore action_plus) >>= atomically . putTMVar var
     let a = Async t (readTMVar var)
     r <- restore (inner a) `catchAll` \e -> do
       uninterruptibleCancel a
@@ -554,11 +587,15 @@ isCancel e
 -- >   withAsync right $ \b ->
 -- >   waitEither a b
 --
-race :: IO a -> IO b -> IO (Either a b)
+race ::
+  CALLSTACK
+  IO a -> IO b -> IO (Either a b)
 
 -- | Like 'race', but the result is ignored.
 --
-race_ :: IO a -> IO b -> IO ()
+race_ ::
+  CALLSTACK
+  IO a -> IO b -> IO ()
 
 
 -- | Run two @IO@ actions concurrently, and return both results.  If
@@ -570,19 +607,25 @@ race_ :: IO a -> IO b -> IO ()
 -- >   withAsync left $ \a ->
 -- >   withAsync right $ \b ->
 -- >   waitBoth a b
-concurrently :: IO a -> IO b -> IO (a,b)
+concurrently ::
+  CALLSTACK
+  IO a -> IO b -> IO (a,b)
 
 
 -- | Run two @IO@ actions concurrently. If both of them end with @Right@,
 -- return both results.  If one of then ends with @Left@, interrupt the other
 -- action and return the @Left@.
 --
-concurrentlyE :: IO (Either e a) -> IO (Either e b) -> IO (Either e (a, b))
+concurrentlyE ::
+  CALLSTACK
+  IO (Either e a) -> IO (Either e b) -> IO (Either e (a, b))
 
 -- | 'concurrently', but ignore the result values
 --
 -- @since 2.1.1
-concurrently_ :: IO a -> IO b -> IO ()
+concurrently_ ::
+  CALLSTACK
+  IO a -> IO b -> IO ()
 
 #define USE_ASYNC_VERSIONS 0
 
@@ -643,9 +686,11 @@ concurrentlyE left right = concurrently' left right (collect [])
             Left ex -> throwIO ex
             Right r -> collect (r:xs) m
 
-concurrently' :: IO a -> IO b
-             -> (IO (Either SomeException (Either a b)) -> IO r)
-             -> IO r
+concurrently' ::
+  CALLSTACK
+  IO a -> IO b
+  -> (IO (Either SomeException (Either a b)) -> IO r)
+  -> IO r
 concurrently' left right collect = do
     done <- newEmptyMVar
     mask $ \restore -> do
@@ -721,37 +766,49 @@ concurrently_ left right = concurrently' left right (collect 0)
 -- for each element of the @Traversable@, so running this on large
 -- inputs without care may lead to resource exhaustion (of memory,
 -- file descriptors, or other limited resources).
-mapConcurrently :: Traversable t => (a -> IO b) -> t a -> IO (t b)
+mapConcurrently ::
+  CALLSTACK
+  Traversable t => (a -> IO b) -> t a -> IO (t b)
 mapConcurrently f = runConcurrently . traverse (Concurrently . f)
 
 -- | `forConcurrently` is `mapConcurrently` with its arguments flipped
 --
 -- > pages <- forConcurrently ["url1", "url2", "url3"] $ \url -> getURL url
 --
 -- @since 2.1.0
-forConcurrently :: Traversable t => t a -> (a -> IO b) -> IO (t b)
+forConcurrently ::
+  CALLSTACK
+  Traversable t => t a -> (a -> IO b) -> IO (t b)
 forConcurrently = flip mapConcurrently
 
 -- | `mapConcurrently_` is `mapConcurrently` with the return value discarded;
 -- a concurrent equivalent of 'mapM_'.
-mapConcurrently_ :: F.Foldable f => (a -> IO b) -> f a -> IO ()
+mapConcurrently_ ::
+  CALLSTACK
+  F.Foldable f => (a -> IO b) -> f a -> IO ()
 mapConcurrently_ f = runConcurrently . F.foldMap (Concurrently . void . f)
 
 -- | `forConcurrently_` is `forConcurrently` with the return value discarded;
 -- a concurrent equivalent of 'forM_'.
-forConcurrently_ :: F.Foldable f => f a -> (a -> IO b) -> IO ()
+forConcurrently_ ::
+  CALLSTACK
+  F.Foldable f => f a -> (a -> IO b) -> IO ()
 forConcurrently_ = flip mapConcurrently_
 
 -- | Perform the action in the given number of threads.
 --
 -- @since 2.1.1
-replicateConcurrently :: Int -> IO a -> IO [a]
+replicateConcurrently ::
+  CALLSTACK
+  Int -> IO a -> IO [a]
 replicateConcurrently cnt = runConcurrently . sequenceA . replicate cnt . Concurrently
 
 -- | Same as 'replicateConcurrently', but ignore the results.
 --
 -- @since 2.1.1
-replicateConcurrently_ :: Int -> IO a -> IO ()
+replicateConcurrently_ ::
+  CALLSTACK
+  Int -> IO a -> IO ()
 replicateConcurrently_ cnt = runConcurrently . F.fold . replicate cnt . Concurrently . void
 
 -- -----------------------------------------------------------------------------
@@ -845,14 +902,16 @@ instance (Semigroup a, Monoid a) => Monoid (ConcurrentlyE e a) where
 -- | Fork a thread that runs the supplied action, and if it raises an
 -- exception, re-runs the action.  The thread terminates only when the
 -- action runs to completion without raising an exception.
-forkRepeat :: IO a -> IO ThreadId
+forkRepeat ::
+  CALLSTACK
+  IO a -> IO ThreadId
 forkRepeat action =
   mask $ \restore ->
     let go = do r <- tryAll (restore action)
                 case r of
                   Left _ -> go
                   _      -> return ()
-    in forkIO go
+    in forkIO (debugLabelMe >> go)
 
 catchAll :: IO a -> (SomeException -> IO a) -> IO a
 catchAll = catch
@@ -864,11 +923,29 @@ tryAll = try
 -- handler: saves a bit of time when we will be installing our own
 -- exception handler.
 {-# INLINE rawForkIO #-}
-rawForkIO :: IO () -> IO ThreadId
-rawForkIO (IO action) = IO $ \ s ->
-   case (fork# action s) of (# s1, tid #) -> (# s1, ThreadId tid #)
+rawForkIO ::
+  CALLSTACK
+  IO () -> IO ThreadId
+rawForkIO action = IO $ \ s ->
+   case (fork# action_plus s) of (# s1, tid #) -> (# s1, ThreadId tid #)
+  where
+    (IO action_plus) = debugLabelMe >> action
 
 {-# INLINE rawForkOn #-}
-rawForkOn :: Int -> IO () -> IO ThreadId
-rawForkOn (I# cpu) (IO action) = IO $ \ s ->
-   case (forkOn# cpu action s) of (# s1, tid #) -> (# s1, ThreadId tid #)
+rawForkOn ::
+  CALLSTACK
+  Int -> IO () -> IO ThreadId
+rawForkOn (I# cpu) action = IO $ \ s ->
+   case (forkOn# cpu action_plus s) of (# s1, tid #) -> (# s1, ThreadId tid #)
+  where
+   (IO action_plus) = debugLabelMe >> action
+
+debugLabelMe ::
+  CALLSTACK
+  IO ()
+debugLabelMe =
+#ifdef DEBUG_AUTO_LABEL
+  myThreadId >>= flip labelThread (GHC.Stack.prettyCallStack callStack)
+#else
+  pure ()
+#endif

async.cabal

@@ -64,6 +64,16 @@ source-repository head
     type: git
     location: https://github.com/simonmar/async.git
 
+flag debug-auto-label
+   description:
+     Strictly for debugging as it might have a non-negligible overhead.
+
+     Enabling this flag will auto-label the threads spawned by @async@. Use it to
+     find where are unlabelled threads spawned in your program (be it your code or
+     dependency code).
+   default: False
+   manual: True
+
 library
     default-language:    Haskell2010
     other-extensions:    CPP, MagicHash, RankNTypes, UnboxedTuples
@@ -74,6 +84,8 @@ library
     build-depends:       base     >= 4.3     && < 4.22,
                          hashable >= 1.1.2.0 && < 1.6,
                          stm      >= 2.2     && < 2.6
+    if flag(debug-auto-label)
+      cpp-options: -DDEBUG_AUTO_LABEL
 
 test-suite test-async
     default-language: Haskell2010