I would file a bug on the GHC bug tracker: http://hackage.haskell.org/trac/ghc/newticket?type=bug You will need to use the guest login to trac, though (in grey text at the bottom of the page). I've also CCed the GHC-users list, as there are folks over there that might have knowledge of the concurrency libraries. Take care, Antoine On Sat, Jan 1, 2011 at 7:03 PM, Eric Stansifer wrote:

Hi,

I didn't get a response to my question on haskell-cafe, perhaps libraries is a more appropriate place to ask.

doc:  http://www.haskell.org/ghc/docs/7.0-latest/html/libraries/base-4.3.0.0/Contr... source:  http://www.haskell.org/ghc/docs/7.0-latest/html/libraries/base-4.3.0.0/src/C...

The isEmptySampleVar function,

isEmptySampleVar :: SampleVar a -> IO Bool isEmptySampleVar (SampleVar svar) = do   (readers, _) <- readMVar svar   return (readers == 0)

returns False whenever readers < 0.  However, readers < 0 occurs when there are threads waiting on an empty SampleVar.

The documentation on SampleVar is a bit sparse on explaining the intended behavior;  I wouldn't have expected this behavior if I hadn't read the source.  Can someone clarify the semantics of SampleVar?

For a contrived example, consider

do_something = threadDelay 100000     -- 100 ms

produce, consume :: SampleVar Int -> IO () produce svar = do    do_something    b <- isEmptySampleVar svar    if b then    randomIO >>= writeSampleVar svar    else   return ()    produce svar

consume svar = do    x <- readSampleVar svar    print x    consume svar

main = do    svar <- newEmptySampleVar    forkIO $ produce svar    forkIO $ consume svar    threadDelay 1000000       -- one second

This code deadlocks instead of printing random numbers.

Thanks, Eric

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries