Turtle.Prelude

Print exactly one line to stdout

To print more than one line see printf, which also supports formatted output

Print exactly one line to stderr

Read in a line from stdin

Returns Nothing if at end of input

Read in a file as Text

Write out a file as Text

Get command line arguments in a list

Set or modify an environment variable

Note: This will change the current environment for all of your program's threads since this modifies the global state of the process

Delete an environment variable

Look up an environment variable

Retrieve all environment variables

Change the current directory

Note: This will change the current directory for all of your program's threads since this modifies the global state of the process

Get the current directory

Get the home directory

Get the path pointed to by a symlink

Canonicalize a path

Move a file or directory

Works if the two paths are on the same filesystem. If not, mv will still work when dealing with a regular file, but the operation will not be atomic

Create a directory

Fails if the directory is present

Create a directory tree (equivalent to mkdir -p)

Does not fail if the directory is present

Copy a file

Copy a directory tree and preserve symbolic links

Copy a directory tree and dereference symbolic links

Create a symlink from one FilePath to another

Returns True if the given FilePath is not a symbolic link

This comes in handy in conjunction with lsif:

lsif isNotSymbolicLink

Remove a file

Remove a directory

Remove a directory tree (equivalent to rm -r)

Use at your own risk

Check if a file exists

Check if a directory exists

Check if a path exists

Get the current time

Get the time a file was last modified

Touch a file, updating the access and modification times to the current time

Creates an empty file if it does not exist

Time how long a command takes in monotonic wall clock time

Returns the duration alongside the return value

Get the system's host name

Show the full path of an executable file

Show all matching executables in PATH, not just the first

Sleep for the given duration

A numeric literal argument is interpreted as seconds. In other words, (sleep 2.0) will sleep for two seconds.

Exit with the given exit code

An exit code of 0 indicates success

Throw an exception using the provided Text message

Analogous to && in Bash

Runs the second command only if the first one returns ExitSuccess

Analogous to || in Bash

Run the second command only if the first one returns ExitFailure

Acquire a Managed read-only Handle from a FilePath

Acquire a Managed write-only Handle from a FilePath

Acquire a Managed append-only Handle from a FilePath

Create a temporary file underneath the given directory

Deletes the temporary file when done

Note that this provides the Handle of the file in order to avoid a potential race condition from the file being moved or deleted before you have a chance to open the file. The mktempfile function provides a simpler API if you don't need to worry about that possibility.

Create a temporary file underneath the given directory

Deletes the temporary file when done

Create a temporary directory underneath the given directory

Deletes the temporary directory when done

Fork a thread, acquiring an Async value

Wait for an Async action to complete

Change the current directory. Once the current Shell is done, it returns back to the original directory.

>>> :set -XOverloadedStrings>>> cd "/">>> view (pushd "/tmp" >> pwd)FilePath "/tmp">>> pwdFilePath "/"

Read lines of Text from standard input

Read lines of Text from a file

Read lines of Text from a Handle

Stream lines of Text to standard output

Stream lines of Text to a file

Stream lines of Text to a Handle

Stream lines of Text to append to a file

Stream lines of Text to standard error

Read in a stream's contents strictly

Stream all immediate children of the given directory, excluding "." and ".."

Stream all recursive descendents of the given directory

This skips any directories that fail the supplied predicate

lstree = lsif (\_ -> return True)

Stream all recursive descendents of the given directory

Stream the recursive descendents of a given directory between a given minimum and maximum depth

Combine the output of multiple Shells, in order

Keep all lines that match the given Pattern

Keep every Text element that matches the given Pattern

Replace all occurrences of a Pattern with its Text result

sed performs substitution on a line-by-line basis, meaning that substitutions may not span multiple lines. Additionally, substitutions may occur multiple times within the same line, like the behavior of s/.../.../g.

Warning: Do not use a Pattern that matches the empty string, since it will match an infinite number of times. sed tries to detect such Patterns and die with an error message if they occur, but this detection is necessarily incomplete.

Like sed, but the provided substitution must match the beginning of the line

Like sed, but the provided substitution must match the end of the line

Like sed, but the provided substitution must match the entire line

Make a `Shell Text -> Shell Text` function work on FilePaths instead. | Ignores any paths which cannot be decoded as valid Text.

Like sed, but operates in place on a FilePath (analogous to sed -i)

Like sedPrefix, but operates in place on a FilePath

Like sedSuffix, but operates in place on a FilePath

Like sedEntire, but operates in place on a FilePath

Update a file in place using a Shell transformation

For example, this is used to implement the inplace* family of utilities

Search a directory recursively for all files matching the given Pattern

Filter a shell of FilePaths according to a given pattern

A Stream of "y"s

Number each element of a Shell (starting at 0)

Merge two Shells together, element-wise

If one Shell is longer than the other, the excess elements are truncated

A Shell that endlessly emits ()

Limit a Shell to a fixed number of values

NOTE: This is not lazy and will still consume the entire input stream. There is no way to implement a lazy version of this utility.

Limit a Shell to values that satisfy the predicate

This terminates the stream on the first value that does not satisfy the predicate

Cache a Shell's output so that repeated runs of the script will reuse the result of previous runs. You must supply a FilePath where the cached result will be stored.

The stored result is only reused if the Shell successfully ran to completion without any exceptions. Note: on some platforms Ctrl-C will flush standard input and signal end of file before killing the program, which may trick the program into "successfully" completing.

Run a list of IO actions in parallel using fork and wait.

>>> view (parallel [(sleep 3) >> date, date, date])2016-12-01 17:22:10.83296 UTC2016-12-01 17:22:07.829876 UTC2016-12-01 17:22:07.829963 UTC

Returns the result of a Shell that outputs a single line. Note that if no lines / more than 1 line is produced by the Shell, this function will die with an error message.

main = do
  directory <- single (inshell "pwd" empty)
  print directory

Filter adjacent duplicate elements:

>>> view (uniq (select [1,1,2,1,3]))1213

Filter adjacent duplicates determined after applying the function to the element:

>>> view (uniqOn fst (select [(1,'a'),(1,'b'),(2,'c'),(1,'d'),(3,'e')]))(1,'a')(2,'c')(1,'d')(3,'e')

Filter adjacent duplicate elements determined via the given function:

>>> view (uniqBy (==) (select [1,1,2,1,3]))1213

Return a new Shell that discards duplicates from the input Shell:

>>> view (nub (select [1, 1, 2, 3, 3, 4, 3]))1234

Return a new Shell that discards duplicates determined via the given function from the input Shell:

>>> view (nubOn id (select [1, 1, 2, 3, 3, 4, 3]))1234

Return a list of the sorted elements of the given Shell, keeping duplicates:

>>> sort (select [1,4,2,3,3,7])[1,2,3,3,4,7]

Return a list of the elements of the given Shell, sorted after applying the given function and keeping duplicates:

>>> sortOn id (select [1,4,2,3,3,7])[1,2,3,3,4,7]

Return a list of the elements of the given Shell, sorted by the given function and keeping duplicates:

>>> sortBy (comparing fst) (select [(1,'a'),(4,'b'),(2,'c'),(3,'d'),(3,'e'),(7,'f')])[(1,'a'),(2,'c'),(3,'d'),(3,'e'),(4,'b'),(7,'f')]

Group an arbitrary stream of Text into newline-delimited Lines

>>> stdout (toLines ("ABC" <|> "DEF" <|> "GHI")ABCDEFGHI>>> stdout (toLines empty)  -- Note that this always emits at least 1 `Line`

>>> stdout (toLines ("ABC\nDEF" <|> "" <|> "GHI\nJKL"))ABCDEFGHIJKL

Count the number of characters in the stream (like wc -c)

This uses the convention that the elements of the stream are implicitly ended by newlines that are one character wide

Count the number of words in the stream (like wc -w)

Count the number of lines in the stream (like wc -l)

This uses the convention that each element of the stream represents one line

Split a line into chunks delimited by the given Pattern

Run a command using execvp, retrieving the exit code

The command inherits stdout and stderr for the current process

Run a command line using the shell, retrieving the exit code

This command is more powerful than proc, but highly vulnerable to code injection if you template the command line with untrusted input

The command inherits stdout and stderr for the current process

This function is identical to proc except this throws ProcFailed for non-zero exit codes

This function is identical to shell except this throws ShellFailed for non-zero exit codes

Run a command using execvp, streaming stdout as lines of Text

The command inherits stderr for the current process

Run a command line using the shell, streaming stdout as lines of Text

This command is more powerful than inproc, but highly vulnerable to code injection if you template the command line with untrusted input

The command inherits stderr for the current process

Throws an ExitCode exception if the command returns a non-zero exit code

Run a command using the shell, streaming stdout and stderr as lines of Text. Lines from stdout are wrapped in Right and lines from stderr are wrapped in Left.

Throws an ExitCode exception if the command returns a non-zero exit code

Run a command line using the shell, streaming stdout and stderr as lines of Text. Lines from stdout are wrapped in Right and lines from stderr are wrapped in Left.

This command is more powerful than inprocWithErr, but highly vulnerable to code injection if you template the command line with untrusted input

Throws an ExitCode exception if the command returns a non-zero exit code

Run a command using execvp, retrieving the exit code and stdout as a non-lazy blob of Text

The command inherits stderr for the current process

Run a command line using the shell, retrieving the exit code and stdout as a non-lazy blob of Text

This command is more powerful than proc, but highly vulnerable to code injection if you template the command line with untrusted input

The command inherits stderr for the current process

Run a command using execvp, retrieving the exit code, stdout, and stderr as a non-lazy blob of Text

Run a command line using the shell, retrieving the exit code, stdout, and stderr as a non-lazy blob of Text

This command is more powerful than proc, but highly vulnerable to code injection if you template the command line with untrusted input

system generalizes shell and proc by allowing you to supply your own custom CreateProcess. This is for advanced users who feel comfortable using the lower-level process API

stream generalizes inproc and inshell by allowing you to supply your own custom CreateProcess. This is for advanced users who feel comfortable using the lower-level process API

Throws an ExitCode exception if the command returns a non-zero exit code

streamWithErr generalizes inprocWithErr and inshellWithErr by allowing you to supply your own custom CreateProcess. This is for advanced users who feel comfortable using the lower-level process API

Throws an ExitCode exception if the command returns a non-zero exit code

systemStrict generalizes shellStrict and procStrict by allowing you to supply your own custom CreateProcess. This is for advanced users who feel comfortable using the lower-level process API

systemStrictWithErr generalizes shellStrictWithErr and procStrictWithErr by allowing you to supply your own custom CreateProcess. This is for advanced users who feel comfortable using the lower-level process API

This type is the same as System.Directory.Permissions type except combining the executable and searchable fields into a single executable field for consistency with the Unix chmod. This simplification is still entirely consistent with the behavior of System.Directory, which treats the two fields as interchangeable.

Update a file or directory's user permissions

chmod rwo         "foo.txt"  -- chmod u=rw foo.txt
chmod executable  "foo.txt"  -- chmod u+x foo.txt
chmod nonwritable "foo.txt"  -- chmod u-w foo.txt

The meaning of each permission is:

• readable (+r for short): For files, determines whether you can read from that file (such as with input). For directories, determines whether or not you can list the directory contents (such as with ls). Note: if a directory is not readable then ls will stream an empty list of contents

• writable (+w for short): For files, determines whether you can write to that file (such as with output). For directories, determines whether you can create a new file underneath that directory.

• executable (+x for short): For files, determines whether or not that file is executable (such as with proc). For directories, determines whether or not you can read or execute files underneath that directory (such as with input or proc)

Get a file or directory's user permissions

Set a file or directory's user permissions

Copy a file or directory's permissions (analogous to chmod --reference)

Get the size of a file or a directory

An abstract file size

Specify the units you want by using an accessor like kilobytes

The Num instance for Size interprets numeric literals as bytes

Construct a Size from an integer in bytes

>>> format sz (B 42)"42 B"

Construct a Size from an integer in kilobytes

>>> format sz (KB 42)"42.0 KB">>> let B n = KB 1 in n1000

Construct a Size from an integer in megabytes

>>> format sz (MB 42)"42.0 MB">>> let KB n = MB 1 in n1000

Construct a Size from an integer in gigabytes

>>> format sz (GB 42)"42.0 GB">>> let MB n = GB 1 in n1000

Construct a Size from an integer in terabytes

>>> format sz (TB 42)"42.0 TB">>> let GB n = TB 1 in n1000

Construct a Size from an integer in kibibytes

>>> format sz (KiB 42)"43.8 KB">>> let B n = KiB 1 in n1024

Construct a Size from an integer in mebibytes

>>> format sz (MiB 42)"44.40 MB">>> let KiB n = MiB 1 in n1024

Construct a Size from an integer in gibibytes

>>> format sz (GiB 42)"45.97 GB">>> let MiB n = GiB 1 in n1024

Construct a Size from an integer in tebibytes

>>> format sz (TiB 42)"46.179 TB">>> let GiB n = TiB 1 in n1024

Format a Size using a human readable representation

>>> format sz 42"42 B">>> format sz 2309"2.309 KB">>> format sz 949203"949.203 KB">>> format sz 1600000000"1.600 GB">>> format sz 999999999999999999"999999.999 TB"

Extract a size in bytes

POSIX defines operations to get information, such as owner, permissions, size and access times, about a file. This information is represented by the FileStatus type.

Note: see chmod.

Limitations: Support for high resolution timestamps is filesystem dependent:

• HFS+ volumes on OS X only support whole-second times.

Get the status of a file

Get the status of a file, but don't follow symbolic links

Size of the file in bytes. Does not follow symlinks

Time of last access

Time of last modification

Time of last status change (i.e. owner, group, link count, mode, etc.)

Checks if this file is a block device.

Checks if this file is a character device.

Checks if this file is a named pipe device.

Checks if this file is a regular file device.

Checks if this file is a directory device.

Checks if this file is a symbolic link device.

Checks if this file is a socket device.

Check if a file was last modified after a given timestamp

Check if a file was last modified before a given timestamp