Irregular Regular Expressions


Ire is a sort of scripting language thingy based around regular expressions.

I often find myself string together bits of grep and sed just because I want to do something like listing all the lines of a file and reformatting certain lines depending on conditions or perhaps only listing certain lines.

Ire is a scripting engine that adds structure to regular expressions so that you can have certain regexps applied only if others match etc.


Ire's only requirement is that you have node.js installed. Packages exist for most operating systems. Get it, it's great.


Ire is in npm and can be installed with:

npm install -g ire


You can run a script from a file with:

ire myscript.ire

Or you can pass a script in with the -e option:

ire -e '//Hello, world\n/p'

If you add the -p flag, ire will print the data line at the end of the script. This is useful when using ire in pipe such as:

cat /proc/cpuinfo | ire -p -e '/\s*:\s*/ = /'


Ire programs are made up of blocks of statements.

Statements can be indented after another statement to show dependency or grouping. If you prefer, you can group statements by using { and }.

While inside a group using { and }, you cannot use indentation to show grouping but you can use semicolons to separate statements as well as newlines.

A statement can be either:

Reference markers

    # some code to run
    # if myref is later imported


>mref { #some code to run ; # if myref is later imported }

To give a name to a block of code, use a reference marker. The code under the marker will not be executed immediately, but can be called by an include statement

Import statements


Import statements simply execute a block that has been marked with the reference marker of the same name.


A command can be in one of three formats:

SEPARATOR can be any charater except < or > (to avoid confusion with refs and imports)

regexp is a regular expression to match against the current data. If it matches, any indented block following this command will be executed.

replacement is a string to replace the matched portion of the data with. Use $1, $2, etc. to refer back to capture groups in regexp.

flags is a string of letters each of which modifies the behaviour of either the regexp or replacement or performs an action.

List of currently supported flags

Not yet implemented/tested flags


There's no better way to understand something that by looking at a few examples. Here goes...

Hello, world

Everyone's favourite program

# Print out "Hello, world"
//Hello, world/p

Programs start off with an empty data string. The simple program above works by matching nothing (this will match against any string) and replacing it will "Hello, world" and then printing the result.

If you were to include this as part of a larger program, you would want:

# Print out "Hello, world"
//Hello, world/tp

The t means that the change is not permanent and only applies for this line.

Note that the comment line (# print out "Hello, world") is actually also a command. It uses # as it's separator and then looks for something that matches Print out "Hello, world". Even if the data string did happen to match, it doesn't perform any action on it so it effectively gets ignored.