[Haskell-cafe] Re: Waiting for thread to finish

[ChrisK]

A safer gimmick...

Ben Franksen wrote:
> 
> tickWhileDoing :: String -> IO a -> IO a
> tickWhileDoing msg act = do
>   hPutStr stderr msg >> hPutChar stderr ' ' >> hFlush stderr
>   start_time <- getCPUTime
>   tickerId <- forkIO ticker
... an async exception here will leave the ticker runnning....
>   res <- act `finally` killThread tickerId

The best way to make this safe that I know of is:

>   res <- block $ do
>     tickerId <- forkIO ticker
>     unblock act `finally` killThread tickerId


>   stop_time <- getCPUTime
>   let time_diff = realToFrac (stop_time - start_time) / 1e12
>   hPutStrLn stderr $ " done (took us " ++ show time_diff ++ " seconds)"
>   return res
>   where
>     ticker = do
>       hPutChar stderr '.' >> hFlush stderr
>       threadDelay 100000 {-microsec-}
>       ticker
> 
> I think nobody in his right mind would even try to do something like that in
> C or Perl or whatever, at least not if it wasn't strictly a requirement and
> correct operation is important (the script gets executed as part of our
> build process and a subtle concurrency bug could lead to a wrong
> configuration for the target control system). In Haskell it was so easy to
> do that I just couldn't resist.
> 
> Cheers
> Ben
> 
> PS (completely off-topic, sorry): I've been using the collections library
> throughout the project & I must say it is a lot nicer to work with than the
> base library mumble-jumble of duplicate interfaces, qualified imports and
> whatnot. The only disadvantages are that the API is not yet as complete as
> e.g. Data.Map, and that I have to manually hide name-clashing Prelude
> functions in almost every module. Many thanks to Jean-Philippe Bernardy for
> pioneering this work.