Literate programming source code processor with a command line app and library interface

npm install dotlit
22 downloads in the last month


A literate programming source code processor with a command line app and library interface.

What is dotlit?

dotlit is a transparent extension to Markdown that adds semantics to Markdown code blocks. It allows you to easily:

  1. Use anonymous code blocks to document a single source file using Markdown and then filter out the Markdown to recover the original source.
  2. Use named code blocks to embed any number of source code files from any programming language into a Markdown document and then easily extract them out later.
  3. Combine named and anonymous code blocks to include dependent files inside a main source file.

Since it is a transparent extension any existing Markdown processor can process any dotlit file. This documentation is a dotlit file itself and you can see how GitHub renders it here.

Uses for dotlit?

  1. Adding meta documentation that you don't want to be shipped with the source code. These can be things like design decisions you made, checklists of requirements, references to online documentation or comments to files that don't provide for a comments, like json files. This literate programming style has recently had a reawakening due to Jeremy Ashkenas' Literate CoffeeScript (which inpart inspired dotlit's creation).
  2. Gathering and documenting a set of files. A good example is a chef recipe that with five customized files. Instead of having separate (or no) documenation on what was changed in the files you can have a single dotlit file that contains all the documenation and the files with their full paths preserved.
  3. Bundling the tests for a source file within the actual source file. Imagine being able to write the test code for the function right next to the actual function instead of maintaining seperate files. Using a combination of anonymous code blocks in the source file with named test file blocks you can easily embed the test code and then extract it out for testing.
  4. Developing code using a single file but compiling/serving it using multiple files. In web development combining html, css and Javascript in a single file is not a good idea because it limits testability and often violates the DRY principle. However, keeping all the related code in one file does often provide for a quicker development cycle because you are not constantly switching between files. With dotlit you can combine the files during development and easily extract them for testing/compiling/serving and get a meta documenation facility as a bonus.
  5. Writing tutorials or ebooks containing code you can validate. How many tutorials or ebooks have you read that have syntax errors in the code samples because they were just pasted in after the fact? Because dotlit is simply Markdown you can use any of the wide variety of tools to render the dotlit file as HTML, PDF or the various eBook formats and still be able to extract and verify the code with the bonus of being able to also deliver all of the source files separately. This was the original inspiration for dotlit (after talking to (Paul Bissex)[] at a developer's group about how he used Markdown to co-write his Django book).

dotlit Markup

Anonymous Code Blocks

Anonymous code blocks are simply Markdown code blocks.

#include <stdio.h>

int main() {
    puts("Hello, world!");
    return 0;

If this anonymous code block was inside a file named you could extract the code into hello.c using this command:

$ dotlit --extract

The placement of the .lit extension in the filename sets the end of the extracted filename in a document with anonymous code blocks. The following three files with anonymous code blocks would extract out to test.js.

  3. test.js.lit

Named Code Blocks

Named code blocks allow you embed a named file in a Markdown document. The $ character (when it is the first character in a code block and is followed by at least one whitespace character) denotes a dotlit file operation.

$ hello.c
#include <stdio.h>

int main() {
    puts("Hello, world!");
    return 0;

If this named code block was in a file named you could extract the code into hello.c using this command:

$ dotlit --extract hello.c

Simple Named Code Blocks Example

Imagine you are writing a simple tutorial about creating the "Hello, world!" code shown above. If you do it as a dotlit file you can extract your code from your tutorial as you write it to make sure it compiles and your reader could extract it later to get the code without copying and pasting. You could also render the dotlit file as HTML and deliver it along with the extracted files.

Tutorial: Creating "Hello, world!"

First create a file named hello.c using your favorite editor. The first thing you will add is preprocessor command to tell the C compiler to load the standard io libary.

$ hello.c
#include <stdio.h>

Then add the following lines to hello.c to define the main entry point into the program.

$ hello.c (+)

int main() {

Finally let's make the program print out "Hello, world!" and exit with a status code of 0. Add the following between the two braces:

$ hello.c (3+)
    puts("Hello, world!");
    return 0;

Compile hello.c with your favorite compiler and you are done!

dotlit File Operations

The previous simple example shows the three of the file operations that dotlit adds. Here are all operations:

$ test.txt
This creates a new file called test.txt since there is no operator after the filename.

$ test.txt (+)
This line gets appended to test.txt at line 2. The plain "+" means insert at the end of the file.

$ test.txt (2+)
These twos lines are inserted at line 2, pushing "This line gets appended to test.txt" to line 4.
The "2+" means insert the text at line 2.  You can have any number of lines here.

$ test.txt (1-2)
This deletes the first two lines of test.txt.  The "1-2" means delete two lines starting at line 1.
The content of this code block is ignored but will be flagged if the dotlit command renders it to HTML.

$ test.txt (1)
This is the new first line of the file.  The bare "1" means replace line 1 in test.txt
This second line is ignored since we only specified a single line replacement.  
This code block will be flagged if the dotlit command renders it to HTML since it has more than one line.

$ test.txt (1:2)
This is the new first line of the file.  The "1:2" means replace two lines starting at line 1 with these 
three lines.  The number of new lines in the code block does not need to match the number of lines
that are replaced, so this line would also be added to the file.

$ (1-2)
This is an anonymous code block that allows you to use any of the parenthetical operators mentioned above.
This example would delete two lines starting at line 1.

File Operation Format

  • (+) append the contents of the code block to end of file
  • (L+) append the contents of the code block starting at line L
  • (L-C) delete C lines starting at line L and ignore the contents of the code block
  • (L) replace line L with the first line of the code block
  • (L:C) replace C lines starting at line L with the entire contents of the code block

Getting Started

dotlit is a node.js module that has both a command line component and a library that you can use in your own node.js projects. You don't have to be a node.js programmer to use the command line interface but you do need to have node.js installed.

Once you install node.js install the module with: npm install -g dotlit

Command Line Interface

$ dotlit --help
dolit version 0.2.0

Usage: dotlit <file> [options]

  --list            Lists the embedded files (default if no option is given)
  --extract <file>  Extracts the specified embedded file
  --extract-all     Extracts all the embedded files
  --print <file>    Prints the specified embedded file to stdout
  --print-all       Prints all the embedded files to stdout
  --render <file>   Renders the dotlit file as html
  --force           Skips confirmation when extracting files outside the current directory
  --verbose         Shows messages while processing
  --help            Shows this information

Using dotlit files with named code blocks

List all the embedded files

$ dotlit --list 

Print out the contents of an embedded file

$ dotlit --print hello.c
#include <stdio.h>

int main() {
    puts("Hello, world!");
    return 0;

Extract all the files. You will need to give confirmation for any files extracted when they are outside of the current directory or its subdirectories. You use the --force parameter to disable these checks.

$ dotlit --extract-all --verbose
Extract /etc/nginx/sites-available/ (outside of tree) [y/N] y
[INFO] Extracted assets/js/api.js
[INFO] Extracted index.html
[INFO] Extracted assets/js/app.js
[INFO] Extracted assets/css/app.css
[INFO] Extracted /etc/nginx/sites-available/

Extract the foo.js file

$ dotlit --extract assets/js/foo.js

Render a dotlit file as HTML

$ dotlit --render named-blocks.html

Using dotlit files with only anonymous code blocks

List all the files

$ dotlit --list 

Extract the code into anonymous-blocks.js

$ dotlit --extract

Using dotlit files with a mix of named and anonymous code blocks

$ dotlit --list 

Extract all the files

$ dotlit --extract-all

Extract the foo.js file

$ dotlit --extract foo.js

Extract the mixed-blocks.js file

$ dotlit --extract

Library Interfaces

dotlit Interface

  • load(filename, callback(err, LitFile)):void - async loader, uses callback to return LitFile
  • loadSync(filename):LitFile - returns LitFile or throws an error if file can't be loaded
  • create(filename, text):LitFile - returns LitFile created from the passed text

LitFile Interface

  • filename:String - the filename (read-only)
  • extractFilename:String - the filename used when extracting anonymous code blocks (read-only)
  • text:String - text of the dotlit file (read-only)
  • files:Array of EmbeddedFile - array of EmbeddedFiles in the order they appear in the dotlit file (read-only)
  • filenames:Array of String - array of filenames in the order they appear in the dotlit file (read-only)
  • fileMap:Object (String to EmbeddedFile) - mapping of embedded filenames to EmbeddedFile instances in the dotlit file (read-only)
  • html:String - the contents of the dotlit file as HTML
  • extract(filename):EmbeddedFile returns the embedded file or null if file is not in the LitFile

EmbeddedFile Interface

  • filename:Sting - the filename (read-only)
  • contents:String - the contents of the file (read-only)


$ examples.js
var dotlit = require('dotlit');

// load a lit file asynchronously and print number of files inside of it
dotlit.load('', function (err, litFile) {
  if (!err) {

// load a lit file synchronously and print the names of all the files inside of it
var litFile = dotlit.loadSync('');
if (litFile) {
  litFile.filenames.forEach(function (filename) {

// load a lit file asynchronously, extract a file from it and then display the contents of the embedded file
dotlit.load('', function (err, litFile) {
  if (!err) {
    var file = litFile.extract('foo.js');
    if (file) {

// load a lit file asynchronously and render an HTML view
dotlit.load('', function (err, litFile) {
  if (!err) {

// create a lit file from a buffer (the filename is needed to resolve anonymous code blocks) and then render it as HTML
var litFile = dotlit.create('', 'Here is the main function:\n    #include <stdio.h>\n\n    int main() {\n        puts("Hello, world!");\n        return 0;\n    }\n');


Please maintain the existing coding style. Add unit tests for any new or changed functionality. Lint and test your code using the existing Gruntfile.

Release History

  • 0.1.0 Initial Release. README driven development phase.
  • 0.2.0 Initial command line options and library complete.
  • 0.2.1 Changed file operation to be $ as the first character with at least one space after it
  • 0.2.2 Added dependency to mkdirp to ensure directories exist when extracting
  • 0.2.3 Added dotlit export function to support grunt task
  • 0.2.4 Added check if file content changed before writing out exported file


  • 0.3.0 Expand tests to provide full code coverage.
  • Add external git repo with sample projects
  • Create free online service to extract files from raw or rendered dotlit files passed as urls


Copyright (c) 2013 by Doug Martin, Zoopdoop, LLC.
Licensed under the MIT license.

npm loves you