RIO.Process

Context in which to run processes.

Get the ProcessContext from the environment.

The environment variable map

Create a new ProcessContext from the given environment variable map.

Same as mkProcessContext but uses the system environment (from getEnvironment).

Modify the environment variables of a ProcessContext. This will not change the working directory.

Note that this requires MonadIO, as it will create a new IORef for the cache.

Use modifyEnvVars to create a new ProcessContext, and then use it in the provided action.

Look into the ProcessContext and return the specified environmet variable if one is available.

Set the working directory to be used by child processes.

Provide a ProcessConfig based on the ProcessContext in scope. Deals with resolving the full path, setting the child process's environment variables, setting the working directory, and wrapping the call with withProcessTimeLog for debugging output.

This is intended to be analogous to the proc function provided by the System.Process.Typed module, but has a different type signature to (1) allow it to perform IO actions for looking up paths, and (2) allow logging and timing of the running action.

Same as withProcess, but generalized to MonadUnliftIO.

Same as withProcess_, but generalized to MonadUnliftIO.

Same as withProcessWait, but generalized to MonadUnliftIO.

Same as withProcessWait_, but generalized to MonadUnliftIO.

Same as withProcessTerm, but generalized to MonadUnliftIO.

Same as withProcessTerm_, but generalized to MonadUnliftIO.

Execute a process within the configured environment.

Execution will not return, because either:

1) On non-windows, execution is taken over by execv of the sub-process. This allows signals to be propagated (#527)

2) On windows, an ExitCode exception will be thrown.

Like exec, but does not use execv on non-windows. This way, there is a sub-process, which is helpful in some cases (https://github.com/commercialhaskell/stack/issues/1306).

This function only exits by throwing ExitCode.

A convenience environment combining a LogFunc and a ProcessContext

Run an action using a LoggedProcessContext with default settings and no logging.

Exception type which may be generated in this module.

NOTE Other exceptions may be thrown by underlying libraries!

Check if the given executable exists on the given PATH.

Find the complete path for the given executable name.

On POSIX systems, filenames that match but are not exectuables are excluded.

On Windows systems, the executable names tried, in turn, are the supplied name (only if it has an extension) and that name extended by each of the exeExtensions. Also, this function may behave differently from findExecutable. The latter excludes as executables filenames without a .bat, .cmd, .com or .exe extension (case-insensitive).

Get the filename extensions for executable files, including the dot (if any).

On POSIX systems, this is [""].

On Windows systems, the list is determined by the value of the PATHEXT environment variable, if it present in the environment. If the variable is absent, this is its default value on a Windows system. This function may, therefore, behave differently from exeExtension, which returns only ".exe".

Augment the given value (assumed to be that of an environment variable that lists paths, such as PATH; this is not checked) with the given extra paths. Those paths are prepended (as in: they take precedence).

Apply augmentPath on the value of the PATH environment variable in the given EnvVars.

Apply augmentPath on the value of the given environment variable in the given EnvVars.

Show a process arg including speechmarks when necessary. Just for debugging purposes, not functionally important.

An abstract configuration for a process, which can then be launched into an actual running Process. Takes three type parameters, providing the types of standard input, standard output, and standard error, respectively.

There are three ways to construct a value of this type:

• With the proc smart constructor, which takes a command name and a list of arguments.

• With the shell smart constructor, which takes a shell string

• With the IsString instance via OverloadedStrings. If you provide it a string with no spaces (e.g., "date"), it will treat it as a raw command with no arguments (e.g., proc "date" []). If it has spaces, it will use shell.

In all cases, the default for all three streams is to inherit the streams from the parent process. For other settings, see the setters below for default values.

Once you have a ProcessConfig you can launch a process from it using the functions in the section Launch a process.

A specification for how to create one of the three standard child streams, stdin, stdout and stderr. A StreamSpec can be thought of as containing

1. A type safe version of StdStream from System.Process. This determines whether the stream should be inherited from the parent process, piped to or from a Handle, etc.

2. A means of accessing the stream as a value of type a

3. A cleanup action which will be run on the stream once the process terminates

To create a StreamSpec see the section Stream specs.

Whether a stream is an input stream or output stream. Note that this is from the perspective of the child process, so that a child's standard input stream is an STInput, even though the parent process will be writing to it.

A running process. The three type parameters provide the type of the standard input, standard output, and standard error streams.

To interact with a Process use the functions from the section Interact with a process.

Set the child's standard input stream to the given StreamSpec.

Default: inherit

Set the child's standard output stream to the given StreamSpec.

Default: inherit

Set the child's standard error stream to the given StreamSpec.

Default: inherit

Should we close all file descriptors besides stdin, stdout, and stderr? See close_fds for more information.

Default: False

Should we create a new process group?

Default: False

Delegate handling of Ctrl-C to the child. For more information, see delegate_ctlc.

Default: False

Detach console on Windows, see detach_console.

Default: False

Create new console on Windows, see create_new_console.

Default: False

Set a new session with the POSIX setsid syscall, does nothing on non-POSIX. See new_session.

Default: False

Set the child process's group ID with the POSIX setgid syscall, does nothing on non-POSIX. See child_group.

Default: False

Set the child process's user ID with the POSIX setuid syscall, does nothing on non-POSIX. See child_user.

Default: False

Create a new StreamSpec from the given StdStream and a helper function. This function:

• Takes as input the raw Maybe Handle returned by the createProcess function. The handle will be Just Handle if the StdStream argument is CreatePipe and Nothing otherwise. See createProcess for more details.

• Returns the actual stream value a, as well as a cleanup function to be run when calling stopProcess.

If making a StreamSpec with CreatePipe, prefer mkPipeStreamSpec, which encodes the invariant that a Handle is created.

A stream spec which simply inherits the stream of the parent process.

A stream spec which will close the stream for the child process. You usually do not want to use this, as it will leave the corresponding file descriptor unassigned and hence available for re-use in the child process. Prefer nullStream unless you're certain you want this behavior.

An input stream spec which sets the input to the given ByteString. A separate thread will be forked to write the contents to the child process.

Capture the output of a process in a ByteString.

This function will fork a separate thread to consume all input from the process, and will only make the results available when the underlying Handle is closed. As this is provided as an STM action, you can either check if the result is available, or block until it's ready.

In the event of any exception occurring when reading from the Handle, the STM action will throw a ByteStringOutputException.

Create a new pipe between this process and the child, and return a Handle to communicate with the child.

Use the provided Handle for the child process, and when the process exits, do not close it. This is useful if, for example, you want to have multiple processes write to the same log file sequentially.

Use the provided Handle for the child process, and when the process exits, close it. If you have no reason to keep the Handle open, you should use this over useHandleOpen.

Launch a process based on the given ProcessConfig. You should ensure that you call stopProcess on the result. It's usually better to use one of the functions in this module which ensures stopProcess is called, such as withProcessWait.

Close a process and release any resources acquired. This will ensure terminateProcess is called, wait for the process to actually exit, and then close out resources allocated for the streams. In the event of any cleanup exceptions being thrown this will throw an exception.

Run a process, capture its standard output and error as a ByteString, wait for it to complete, and then return its exit code, output, and error.

Note that any previously used setStdout or setStderr will be overridden.

Same as readProcess, but instead of returning the ExitCode, checks it with checkExitCode.

Exceptions thrown by this function will include stdout and stderr.

Run the given process, wait for it to exit, and returns its ExitCode.

Same as runProcess, but instead of returning the ExitCode, checks it with checkExitCode.

Same as readProcess, but only read the stdout of the process. Original settings for stderr remain.

Same as readProcessStdout, but instead of returning the ExitCode, checks it with checkExitCode.

Exceptions thrown by this function will include stdout.

Same as readProcess, but only read the stderr of the process. Original settings for stdout remain.

Same as readProcessStderr, but instead of returning the ExitCode, checks it with checkExitCode.

Exceptions thrown by this function will include stderr.

Wait for the process to exit and then return its ExitCode.

Same as waitExitCode, but in STM.

Check if a process has exited and, if so, return its ExitCode.

Same as getExitCode, but in STM.

Wait for a process to exit, and ensure that it exited successfully. If not, throws an ExitCodeException.

Exceptions thrown by this function will not include stdout or stderr (This prevents unbounded memory usage from reading them into memory). However, some callers such as readProcess_ catch the exception, add the stdout and stderr, and rethrow.

Same as checkExitCode, but in STM.

Get the child's standard input stream value.

Get the child's standard output stream value.

Get the child's standard error stream value.

Exception thrown by checkExitCode in the event of a non-success exit code. Note that checkExitCode is called by other functions as well, like runProcess_ or readProcess_.

Note that several functions that throw an ExitCodeException intentionally do not populate eceStdout or eceStderr. This prevents unbounded memory usage for large stdout and stderrs.

Functions which do include eceStdout or eceStderr (like readProcess_) state so in their documentation.

Wrapper for when an exception is thrown when reading from a child process, used by byteStringOutput.

Take ProcessHandle out of the Process. This method is needed in cases one need to use low level functions from the process package. Use cases for this method are:

1. Send a special signal to the process.

2. Terminate the process group instead of terminating single process.

3. Use platform specific API on the underlying process.

This method is considered unsafe because the actions it performs on the underlying process may overlap with the functionality that typed-process provides. For example the user should not call waitForProcess on the process handle as either waitForProcess or stopProcess will lock. Additionally, even if process was terminated by the terminateProcess or by sending signal, stopProcess should be called either way in order to cleanup resources allocated by the typed-process.