process-extras: Process extras

[ library, mit, system ] [ Propose Tags ]

Extends Read process input and output as ByteStrings or Text, or write your own ProcessOutput instance. Lazy process input and output. ProcessMaker class for more flexibility in the process creation API.

[Skip to Readme]


Maintainer's Corner

Package maintainers

For package maintainers and hackage trustees


Versions [RSS] 0.1.0, 0.1.1, 0.1.2, 0.1.3, 0.2.0, 0.3.0,, 0.3.1,,,,,,,,, 0.3.4, 0.4, 0.4.1,,,, 0.7, 0.7.1, 0.7.2, 0.7.3, 0.7.4 (info)
Change log changelog
Dependencies base (>=4.5 && <5), bytestring, data-default, deepseq (>=1.1), generic-deriving (>=1.10), ListLike (>=4.4), mtl, process, text [details]
License MIT
Author David Lazar, Bas van Dijk, David Fox
Maintainer David Fox <>
Category System
Home page
Source repo head: git clone
Uploaded by DavidFox at 2018-02-03T16:53:15Z
Distributions Arch:0.7.4, Debian:0.7.4, LTSHaskell:0.7.4, NixOS:0.7.4, Stackage:0.7.4
Reverse Dependencies 30 direct, 36 indirect [details]
Downloads 52084 total (198 in the last 30 days)
Rating (no votes yet) [estimated by Bayesian average]
Your Rating
  • λ
  • λ
  • λ
Status Docs available [build log]
Last success reported on 2018-02-03 [all 1 reports]

Readme for process-extras-0.7.4

[back to package description]

What is process-extras

You might want to use this package if

  • You want to read and write ByteStrings or Text to a process rather than just Strings
  • You want to read output from a non-terminating process (e.g. yes(1))
  • You want more flexibility in process creation
  • You want various types of progress output from the process to the console before capturing its output - indented and prefixed output, reporting of the command that started the process, dots to reflect the size of the process output, etc.


This project is available on GitHub. You may contribute changes there.

Please report bugs and feature requests using the GitHub issue tracker.


The output type of the raw system process functions is ByteString. Instances of ListLikeProcessIO are provided to read as type String, Text, Lazy Text, ByteString, or Lazy ByteString. Select by casting the result, or by specifying the module containing the specialized function:

> :m +System.Process.ListLike Data.ByteString Data.Text.Lazy
> readCreateProcess (shell "echo 'λ'") mempty :: IO (ExitCode, ByteString, ByteString)
> readCreateProcess (shell "echo 'λ'") mempty :: IO (ExitCode, Text, Text)
> readCreateProcess (shell "echo 'λ'") mempty :: IO (ExitCode, String, String)
> readCreateProcess (shell "yes | head -10") mempty :: IO (ExitCode, Text, Text)

Although the output type can be lazy, normal process functions still need to read until EOF on the process output before returing anything. If you have a process whose output never ends you can use the readCreateProcessLazy function to read it. Functions like readProcess would block waiting for EOF on the process output:

> (Prelude.take 4 <$> readCreateProcessLazy (proc "yes" []) mempty :: IO [Chunk Text]) >>= mapM_ (putStrLn . show)
ProcessHandle <process>
Stdout "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny ..."

The output type can be any instance of ProcessOutput, instances for types (ExitCode, a, a), [Chunk a], and (ExitCode, [Chunk a]) are provided. [Chunk a] can be converted to any other instance of ProcessOutput using collectOutput

> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk ByteString]) >>= mapM_ (Prelude.putStrLn . show)
Stdout "\US\139\b\NUL\237\136\&7W\NUL\ETX345\183\&403\215\&31Q04267\177\&0\177\212\&33\225\STX\NUL_\169\142\178\ETB\NUL\NUL\NUL"
Stderr "gzip: stdin: file size changed while zipping\n -8.7%\n"
Result ExitSuccess
> (readCreateProcess (shell "uptime") mempty :: IO [Chunk ByteString]) >>= writeOutput
 14:00:34 up 18 days,  7:16,  6 users,  load average: 0.04, 0.10, 0.08
> collectOutput <$> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk ByteString]) :: IO (ExitCode, ByteString, ByteString)
(ExitSuccess,"\US\139\b\NUL\185\137\&7W\NUL\ETX345\183\&427\212\&33W0426731\177\208\&35\225\STX\NUL\237\192\CAN\224\ETB\NUL\NUL\NUL","gzip: stdin: file size changed while zipping\n -8.7%\n")
> collectOutput <$> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk ByteString]) :: IO (ExitCode, ByteString, ByteString)
(ExitSuccess,"\US\139\b\NUL\185\137\&7W\NUL\ETX345\183\&427\212\&33W0426731\177\208\&35\225\STX\NUL\237\192\CAN\224\ETB\NUL\NUL\NUL","gzip: stdin: file size changed while zipping\n -8.7%\n")
> (collectOutput . Prelude.filter (\x -> case x of Stderr _ -> False; _ -> True)) <$> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk ByteString]) :: IO (ExitCode, ByteString, ByteString)

Some cases that need investigation:

> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk String]) >>= mapM_ (putStrLn . show)
*** Exception: fd:13: hGetContents: invalid argument (invalid byte sequence)
> (readCreateProcess (shell "gzip -v < /proc/uptime") mempty :: IO [Chunk Text]) >>= mapM_ (putStrLn . show)
*** Exception: fd:13: hClose: invalid argument (Bad file descriptor)