|Deletions are marked like this.||Additions are marked like this.|
|Line 19:||Line 19:|
[[http://www.njoyz.com/anal-toys.html|anal sex toys]]
This page describes how to create repositories accessible via a single shared SSH account without needing to give full shell access to other people. This is just one of many ways to make your repository available to multiple committers, and not necessarily the most common. See PublishingRepositories for a good overview of many ways to allow others to interact with your repository.
hg-ssh is a python script available in contrib/hg-ssh and was probably installed along with your Mercurial software. Allowed repositories are managed directly in the authorized_keys file.
Look at the start of the script for usage instructions. When possible use the version that matches your installed version of Mercurial.
Despite its name, this is not a Mercurial server. It offers an improved management interface for the shared ssh mechanism like that provided by hg-ssh.
mercurial-server provides the most complete and easiest-to-use solution to this problem for hosting a collection of repositories on Unix systems. Installing mercurial-server creates a new user, hg, which will own all the repositories to be shared. Giving access to a new user is as simple as adding their SSH key to a special repository and pushing the changes. mercurial-server can enforce fine-grained permissions and logs all events.
mercurial-server is descended from hg-ssh.
How these work
When accessing a remote repository via Mercurial's ssh repository type, hg basically does the following:
$ ssh hg.example.com hg -R /path/to/repos serve --stdio
It relies on ssh for authentication and tunneling. When using public key authentication, ssh allows limiting the user to one specific command (as described in the sshd manual page in the section concerning the authorized_keys file format). Such a command, provided by the solutions listed above, can do the necessary sanity checking around the requested operation, and can then call hg just like ssh would do in the example above. Since every user gets his own private key and his own entry in authorized_keys, the solutions presented here are able to distinguish between different users and thus enforce things like access control, even though a single system account (or system user) may be providing the underlying services. Moreover, since a designated command must be executed when those accessing the repository authenticate themselves, it should not be possible for users to start a normal shell and bypass access controls implemented by the designated command (although this does depend on the implementation and proper functioning of the command itself).