Ruby API for Mercurial DVCS. Powers Mercurial on


Please refer to YARD documentation here:

Github doesn't support some YARd-specific syntax so this README can look broken.


gem install mercurial-ruby


Tested with Mercurial versions 1.9, 1.9.1, 2.1, 2.5.2; Ruby versions 1.8.7, 1.9.3 and 2.0.0.


There are several settings you can configure:

Mercurial.configure do |conf|
  conf.hg_binary_path = "/usr/bin/hg"

See Configuration class for details.


All actions are performed through Repository object. So before you can start doing anything you need to initialize one. Either by creating a new repository:

repository = Mercurial::Repository.create("/path/to/new_repository")

or opening an existing one:

repository ="/path/to/existing_repository")

Now feel free to invoke various methods to get changesets, branches, nodes, etc:

repository.commits.by_hash_ids('291a498f04e9', '63f70b2314ed')

See Features section below for a full list of entities and their methods.


Mercurial Entities

Custom Commands

You can use Shell class to execute custom shell commands that weren't added to the gem as first-class citizens yet.

Built-in Caching

There's a simple caching mechanism built into the gem. If you pass Rails CacheStore-compatible caching store to the Configuration block, mercurial-ruby will cache output of all hg commands it's executing. Then if you execute same method again and it will run the same command, the gem will return the output from cache.

Here's how you configure it:

Mercurial.configure do |conf|
  conf.cache_store = Rails.cache

The gem is using a single method of the CacheStore called fetch. Cache expires automatically when repository's mtime changes, and it's your job to update it.

Built-in Timeouts

You can provide a timeout for pretty much any command you are running. Do it like this:

repository.commits.all(:timeout => 5)


Copyright © 2013 Ilya Sabanin. See LICENSE.txt for further details.