view en/collab.tex @ 576:1f9ef8b3ab79

translated a bit of the hook reference

\chapter{Collaborating with other people}
\label{cha:collab}

As a completely decentralised tool, Mercurial doesn't impose any
policy on how people ought to work with each other.  However, if
you're new to distributed revision control, it helps to have some
tools and examples in mind when you're thinking about possible
workflow models.

\section{Mercurial's web interface}

Mercurial has a powerful web interface that provides several 
useful capabilities.

For interactive use, the web interface lets you browse a single
repository or a collection of repositories.  You can view the history
of a repository, examine each change (comments and diffs), and view
the contents of each directory and file.

Also for human consumption, the web interface provides an RSS feed of
the changes in a repository.  This lets you ``subscribe'' to a
repository using your favourite feed reader, and be automatically
notified of activity in that repository as soon as it happens.  I find
this capability much more convenient than the model of subscribing to
a mailing list to which notifications are sent, as it requires no
additional configuration on the part of whoever is serving the
repository.

The web interface also lets remote users clone a repository, pull
changes from it, and (when the server is configured to permit it) push
changes back to it.  Mercurial's HTTP tunneling protocol aggressively
compresses data, so that it works efficiently even over low-bandwidth
network connections.

The easiest way to get started with the web interface is to use your
web browser to visit an existing repository, such as the master
Mercurial repository at
\url{http://www.selenic.com/repo/hg?style=gitweb}.

If you're interested in providing a web interface to your own
repositories, Mercurial provides two ways to do this.  The first is
using the \hgcmd{serve} command, which is best suited to short-term
``lightweight'' serving.  See section~\ref{sec:collab:serve} below for
details of how to use this command.  If you have a long-lived
repository that you'd like to make permanently available, Mercurial
has built-in support for the CGI (Common Gateway Interface) standard,
which all common web servers support.  See
section~\ref{sec:collab:cgi} for details of CGI configuration.

\section{Collaboration models}

With a suitably flexible tool, making decisions about workflow is much
more of a social engineering challenge than a technical one.
Mercurial imposes few limitations on how you can structure the flow of
work in a project, so it's up to you and your group to set up and live
with a model that matches your own particular needs.

\subsection{Factors to keep in mind}

The most important aspect of any model that you must keep in mind is
how well it matches the needs and capabilities of the people who will
be using it.  This might seem self-evident; even so, you still can't
afford to forget it for a moment.

I once put together a workflow model that seemed to make perfect sense
to me, but that caused a considerable amount of consternation and
strife within my development team.  In spite of my attempts to explain
why we needed a complex set of branches, and how changes ought to flow
between them, a few team members revolted.  Even though they were
smart people, they didn't want to pay attention to the constraints we
were operating under, or face the consequences of those constraints in
the details of the model that I was advocating.

Don't sweep foreseeable social or technical problems under the rug.
Whatever scheme you put into effect, you should plan for mistakes and
problem scenarios.  Consider adding automated machinery to prevent, or
quickly recover from, trouble that you can anticipate.  As an example,
if you intend to have a branch with not-for-release changes in it,
you'd do well to think early about the possibility that someone might
accidentally merge those changes into a release branch.  You could
avoid this particular problem by writing a hook that prevents changes
from being merged from an inappropriate branch.

\subsection{Informal anarchy}

I wouldn't suggest an ``anything goes'' approach as something
sustainable, but it's a model that's easy to grasp, and it works
perfectly well in a few unusual situations.

As one example, many projects have a loose-knit group of collaborators
who rarely physically meet each other.  Some groups like to overcome
the isolation of working at a distance by organising occasional
``sprints''.  In a sprint, a number of people get together in a single
location (a company's conference room, a hotel meeting room, that kind
of place) and spend several days more or less locked in there, hacking
intensely on a handful of projects.

A sprint is the perfect place to use the \hgcmd{serve} command, since
\hgcmd{serve} does not requires any fancy server infrastructure.  You
can get started with \hgcmd{serve} in moments, by reading
section~\ref{sec:collab:serve} below.  Then simply tell the person
next to you that you're running a server, send the URL to them in an
instant message, and you immediately have a quick-turnaround way to
work together.  They can type your URL into their web browser and
quickly review your changes; or they can pull a bugfix from you and
verify it; or they can clone a branch containing a new feature and try
it out.

The charm, and the problem, with doing things in an ad hoc fashion
like this is that only people who know about your changes, and where
they are, can see them.  Such an informal approach simply doesn't
scale beyond a handful people, because each individual needs to know
about $n$ different repositories to pull from.

\subsection{A single central repository}

For smaller projects migrating from a centralised revision control
tool, perhaps the easiest way to get started is to have changes flow
through a single shared central repository.  This is also the
most common ``building block'' for more ambitious workflow schemes.

Contributors start by cloning a copy of this repository.  They can
pull changes from it whenever they need to, and some (perhaps all)
developers have permission to push a change back when they're ready
for other people to see it.

Under this model, it can still often make sense for people to pull
changes directly from each other, without going through the central
repository.  Consider a case in which I have a tentative bug fix, but
I am worried that if I were to publish it to the central repository,
it might subsequently break everyone else's trees as they pull it.  To
reduce the potential for damage, I can ask you to clone my repository
into a temporary repository of your own and test it.  This lets us put
off publishing the potentially unsafe change until it has had a little
testing.

In this kind of scenario, people usually use the \command{ssh}
protocol to securely push changes to the central repository, as
documented in section~\ref{sec:collab:ssh}.  It's also usual to
publish a read-only copy of the repository over HTTP using CGI, as in
section~\ref{sec:collab:cgi}.  Publishing over HTTP satisfies the
needs of people who don't have push access, and those who want to use
web browsers to browse the repository's history.

\subsection{Working with multiple branches}

Projects of any significant size naturally tend to make progress on
several fronts simultaneously.  In the case of software, it's common
for a project to go through periodic official releases.  A release
might then go into ``maintenance mode'' for a while after its first
publication; maintenance releases tend to contain only bug fixes, not
new features.  In parallel with these maintenance releases, one or
more future releases may be under development.  People normally use
the word ``branch'' to refer to one of these many slightly different
directions in which development is proceeding.

Mercurial is particularly well suited to managing a number of
simultaneous, but not identical, branches.  Each ``development
direction'' can live in its own central repository, and you can merge
changes from one to another as the need arises.  Because repositories
are independent of each other, unstable changes in a development
branch will never affect a stable branch unless someone explicitly
merges those changes in.

Here's an example of how this can work in practice.  Let's say you
have one ``main branch'' on a central server.
\interaction{branching.init}
People clone it, make changes locally, test them, and push them back.

Once the main branch reaches a release milestone, you can use the
\hgcmd{tag} command to give a permanent name to the milestone
revision.
\interaction{branching.tag}
Let's say some ongoing development occurs on the main branch.
\interaction{branching.main}
Using the tag that was recorded at the milestone, people who clone
that repository at any time in the future can use \hgcmd{update} to
get a copy of the working directory exactly as it was when that tagged
revision was committed.  
\interaction{branching.update}

In addition, immediately after the main branch is tagged, someone can
then clone the main branch on the server to a new ``stable'' branch,
also on the server.
\interaction{branching.clone}

Someone who needs to make a change to the stable branch can then clone
\emph{that} repository, make their changes, commit, and push their
changes back there.
\interaction{branching.stable}
Because Mercurial repositories are independent, and Mercurial doesn't
move changes around automatically, the stable and main branches are
\emph{isolated} from each other.  The changes that you made on the
main branch don't ``leak'' to the stable branch, and vice versa.

You'll often want all of your bugfixes on the stable branch to show up
on the main branch, too.  Rather than rewrite a bugfix on the main
branch, you can simply pull and merge changes from the stable to the
main branch, and Mercurial will bring those bugfixes in for you.
\interaction{branching.merge}
The main branch will still contain changes that are not on the stable
branch, but it will also contain all of the bugfixes from the stable
branch.  The stable branch remains unaffected by these changes.

\subsection{Feature branches}

For larger projects, an effective way to manage change is to break up
a team into smaller groups.  Each group has a shared branch of its
own, cloned from a single ``master'' branch used by the entire
project.  People working on an individual branch are typically quite
isolated from developments on other branches.

\begin{figure}[ht]
  \centering
  \grafix{feature-branches}
  \caption{Feature branches}
  \label{fig:collab:feature-branches}
\end{figure}

When a particular feature is deemed to be in suitable shape, someone
on that feature team pulls and merges from the master branch into the
feature branch, then pushes back up to the master branch.

\subsection{The release train}

Some projects are organised on a ``train'' basis: a release is
scheduled to happen every few months, and whatever features are ready
when the ``train'' is ready to leave are allowed in.

This model resembles working with feature branches.  The difference is
that when a feature branch misses a train, someone on the feature team
pulls and merges the changes that went out on that train release into
the feature branch, and the team continues its work on top of that
release so that their feature can make the next release.

\subsection{The Linux kernel model}

The development of the Linux kernel has a shallow hierarchical
structure, surrounded by a cloud of apparent chaos.  Because most
Linux developers use \command{git}, a distributed revision control
tool with capabilities similar to Mercurial, it's useful to describe
the way work flows in that environment; if you like the ideas, the
approach translates well across tools.

At the center of the community sits Linus Torvalds, the creator of
Linux.  He publishes a single source repository that is considered the
``authoritative'' current tree by the entire developer community.
Anyone can clone Linus's tree, but he is very choosy about whose trees
he pulls from.

Linus has a number of ``trusted lieutenants''.  As a general rule, he
pulls whatever changes they publish, in most cases without even
reviewing those changes.  Some of those lieutenants are generally
agreed to be ``maintainers'', responsible for specific subsystems
within the kernel.  If a random kernel hacker wants to make a change
to a subsystem that they want to end up in Linus's tree, they must
find out who the subsystem's maintainer is, and ask that maintainer to
take their change.  If the maintainer reviews their changes and agrees
to take them, they'll pass them along to Linus in due course.

Individual lieutenants have their own approaches to reviewing,
accepting, and publishing changes; and for deciding when to feed them
to Linus.  In addition, there are several well known branches that
people use for different purposes.  For example, a few people maintain
``stable'' repositories of older versions of the kernel, to which they
apply critical fixes as needed.  Some maintainers publish multiple
trees: one for experimental changes; one for changes that they are
about to feed upstream; and so on.  Others just publish a single
tree.

This model has two notable features.  The first is that it's ``pull
only''.  You have to ask, convince, or beg another developer to take a
change from you, because there are almost no trees to which more than
one person can push, and there's no way to push changes into a tree
that someone else controls.

The second is that it's based on reputation and acclaim.  If you're an
unknown, Linus will probably ignore changes from you without even
responding.  But a subsystem maintainer will probably review them, and
will likely take them if they pass their criteria for suitability.
The more ``good'' changes you contribute to a maintainer, the more
likely they are to trust your judgment and accept your changes.  If
you're well-known and maintain a long-lived branch for something Linus
hasn't yet accepted, people with similar interests may pull your
changes regularly to keep up with your work.

Reputation and acclaim don't necessarily cross subsystem or ``people''
boundaries.  If you're a respected but specialised storage hacker, and
you try to fix a networking bug, that change will receive a level of
scrutiny from a network maintainer comparable to a change from a
complete stranger.

To people who come from more orderly project backgrounds, the
comparatively chaotic Linux kernel development process often seems
completely insane.  It's subject to the whims of individuals; people
make sweeping changes whenever they deem it appropriate; and the pace
of development is astounding.  And yet Linux is a highly successful,
well-regarded piece of software.

\subsection{Pull-only versus shared-push collaboration}

A perpetual source of heat in the open source community is whether a
development model in which people only ever pull changes from others
is ``better than'' one in which multiple people can push changes to a
shared repository.

Typically, the backers of the shared-push model use tools that
actively enforce this approach.  If you're using a centralised
revision control tool such as Subversion, there's no way to make a
choice over which model you'll use: the tool gives you shared-push,
and if you want to do anything else, you'll have to roll your own
approach on top (such as applying a patch by hand).

A good distributed revision control tool, such as Mercurial, will
support both models.  You and your collaborators can then structure
how you work together based on your own needs and preferences, not on
what contortions your tools force you into.

\subsection{Where collaboration meets branch management}

Once you and your team set up some shared repositories and start
propagating changes back and forth between local and shared repos, you
begin to face a related, but slightly different challenge: that of
managing the multiple directions in which your team may be moving at
once.  Even though this subject is intimately related to how your team
collaborates, it's dense enough to merit treatment of its own, in
chapter~\ref{chap:branch}.

\section{The technical side of sharing}

The remainder of this chapter is devoted to the question of serving
data to your collaborators.

\section{Informal sharing with \hgcmd{serve}}
\label{sec:collab:serve}

Mercurial's \hgcmd{serve} command is wonderfully suited to small,
tight-knit, and fast-paced group environments.  It also provides a
great way to get a feel for using Mercurial commands over a network.

Run \hgcmd{serve} inside a repository, and in under a second it will
bring up a specialised HTTP server; this will accept connections from
any client, and serve up data for that repository until you terminate
it.  Anyone who knows the URL of the server you just started, and can
talk to your computer over the network, can then use a web browser or
Mercurial to read data from that repository.  A URL for a
\hgcmd{serve} instance running on a laptop is likely to look something
like \Verb|http://my-laptop.local:8000/|.

The \hgcmd{serve} command is \emph{not} a general-purpose web server.
It can do only two things:
\begin{itemize}
\item Allow people to browse the history of the repository it's
  serving, from their normal web browsers.
\item Speak Mercurial's wire protocol, so that people can
  \hgcmd{clone} or \hgcmd{pull} changes from that repository.
\end{itemize}
In particular, \hgcmd{serve} won't allow remote users to \emph{modify}
your repository.  It's intended for read-only use.

If you're getting started with Mercurial, there's nothing to prevent
you from using \hgcmd{serve} to serve up a repository on your own
computer, then use commands like \hgcmd{clone}, \hgcmd{incoming}, and
so on to talk to that server as if the repository was hosted remotely.
This can help you to quickly get acquainted with using commands on
network-hosted repositories.

\subsection{A few things to keep in mind}

Because it provides unauthenticated read access to all clients, you
should only use \hgcmd{serve} in an environment where you either don't
care, or have complete control over, who can access your network and
pull data from your repository.

The \hgcmd{serve} command knows nothing about any firewall software
you might have installed on your system or network.  It cannot detect
or control your firewall software.  If other people are unable to talk
to a running \hgcmd{serve} instance, the second thing you should do
(\emph{after} you make sure that they're using the correct URL) is
check your firewall configuration.

By default, \hgcmd{serve} listens for incoming connections on
port~8000.  If another process is already listening on the port you
want to use, you can specify a different port to listen on using the
\hgopt{serve}{-p} option.

Normally, when \hgcmd{serve} starts, it prints no output, which can be
a bit unnerving.  If you'd like to confirm that it is indeed running
correctly, and find out what URL you should send to your
collaborators, start it with the \hggopt{-v} option.

\section{Using the Secure Shell (ssh) protocol}
\label{sec:collab:ssh}

You can pull and push changes securely over a network connection using
the Secure Shell (\texttt{ssh}) protocol.  To use this successfully,
you may have to do a little bit of configuration on the client or
server sides.

If you're not familiar with ssh, it's a network protocol that lets you
securely communicate with another computer.  To use it with Mercurial,
you'll be setting up one or more user accounts on a server so that
remote users can log in and execute commands.

(If you \emph{are} familiar with ssh, you'll probably find some of the
material that follows to be elementary in nature.)

\subsection{How to read and write ssh URLs}

An ssh URL tends to look like this:
\begin{codesample2}
  ssh://[e-mail]:22/hg/hgbook
\end{codesample2}
\begin{enumerate}
\item The ``\texttt{ssh://}'' part tells Mercurial to use the ssh
  protocol.
\item The ``\texttt{bos@}'' component indicates what username to log
  into the server as.  You can leave this out if the remote username
  is the same as your local username.
\item The ``\texttt{hg.serpentine.com}'' gives the hostname of the
  server to log into.
\item The ``:22'' identifies the port number to connect to the server
  on.  The default port is~22, so you only need to specify this part
  if you're \emph{not} using port~22.
\item The remainder of the URL is the local path to the repository on
  the server.
\end{enumerate}

There's plenty of scope for confusion with the path component of ssh
URLs, as there is no standard way for tools to interpret it.  Some
programs behave differently than others when dealing with these paths.
This isn't an ideal situation, but it's unlikely to change.  Please
read the following paragraphs carefully.

Mercurial treats the path to a repository on the server as relative to
the remote user's home directory.  For example, if user \texttt{foo}
on the server has a home directory of \dirname{/home/foo}, then an ssh
URL that contains a path component of \dirname{bar}
\emph{really} refers to the directory \dirname{/home/foo/bar}.

If you want to specify a path relative to another user's home
directory, you can use a path that starts with a tilde character
followed by the user's name (let's call them \texttt{otheruser}), like
this.
\begin{codesample2}
  ssh://server/~otheruser/hg/repo
\end{codesample2}

And if you really want to specify an \emph{absolute} path on the
server, begin the path component with two slashes, as in this example.
\begin{codesample2}
  ssh://server//absolute/path
\end{codesample2}

\subsection{Finding an ssh client for your system}

Almost every Unix-like system comes with OpenSSH preinstalled.  If
you're using such a system, run \Verb|which ssh| to find out if
the \command{ssh} command is installed (it's usually in
\dirname{/usr/bin}).  In the unlikely event that it isn't present,
take a look at your system documentation to figure out how to install
it.

On Windows, you'll first need to download a suitable ssh
client.  There are two alternatives.
\begin{itemize}
\item Simon Tatham's excellent PuTTY package~\cite{web:putty} provides
  a complete suite of ssh client commands.
\item If you have a high tolerance for pain, you can use the Cygwin
  port of OpenSSH.
\end{itemize}
In either case, you'll need to edit your \hgini\ file to tell
Mercurial where to find the actual client command.  For example, if
you're using PuTTY, you'll need to use the \command{plink} command as
a command-line ssh client.
\begin{codesample2}
  [ui]
  ssh = C:/path/to/plink.exe -ssh -i "C:/path/to/my/private/key"
\end{codesample2}

\begin{note}
  The path to \command{plink} shouldn't contain any whitespace
  characters, or Mercurial may not be able to run it correctly (so
  putting it in \dirname{C:\\Program Files} is probably not a good
  idea).
\end{note}

\subsection{Generating a key pair}

To avoid the need to repetitively type a password every time you need
to use your ssh client, I recommend generating a key pair.  On a
Unix-like system, the \command{ssh-keygen} command will do the trick.
On Windows, if you're using PuTTY, the \command{puttygen} command is
what you'll need.

When you generate a key pair, it's usually \emph{highly} advisable to
protect it with a passphrase.  (The only time that you might not want
to do this is when you're using the ssh protocol for automated tasks
on a secure network.)

Simply generating a key pair isn't enough, however.  You'll need to
add the public key to the set of authorised keys for whatever user
you're logging in remotely as.  For servers using OpenSSH (the vast
majority), this will mean adding the public key to a list in a file
called \sfilename{authorized\_keys} in their \sdirname{.ssh}
directory.

On a Unix-like system, your public key will have a \filename{.pub}
extension.  If you're using \command{puttygen} on Windows, you can
save the public key to a file of your choosing, or paste it from the
window it's displayed in straight into the
\sfilename{authorized\_keys} file.

\subsection{Using an authentication agent}

An authentication agent is a daemon that stores passphrases in memory
(so it will forget passphrases if you log out and log back in again).
An ssh client will notice if it's running, and query it for a
passphrase.  If there's no authentication agent running, or the agent
doesn't store the necessary passphrase, you'll have to type your
passphrase every time Mercurial tries to communicate with a server on
your behalf (e.g.~whenever you pull or push changes).

The downside of storing passphrases in an agent is that it's possible
for a well-prepared attacker to recover the plain text of your
passphrases, in some cases even if your system has been power-cycled.
You should make your own judgment as to whether this is an acceptable
risk.  It certainly saves a lot of repeated typing.

On Unix-like systems, the agent is called \command{ssh-agent}, and
it's often run automatically for you when you log in.  You'll need to
use the \command{ssh-add} command to add passphrases to the agent's
store.  On Windows, if you're using PuTTY, the \command{pageant}
command acts as the agent.  It adds an icon to your system tray that
will let you manage stored passphrases.

\subsection{Configuring the server side properly}

Because ssh can be fiddly to set up if you're new to it, there's a
variety of things that can go wrong.  Add Mercurial on top, and
there's plenty more scope for head-scratching.  Most of these
potential problems occur on the server side, not the client side.  The
good news is that once you've gotten a configuration working, it will
usually continue to work indefinitely.

Before you try using Mercurial to talk to an ssh server, it's best to
make sure that you can use the normal \command{ssh} or \command{putty}
command to talk to the server first.  If you run into problems with
using these commands directly, Mercurial surely won't work.  Worse, it
will obscure the underlying problem.  Any time you want to debug
ssh-related Mercurial problems, you should drop back to making sure
that plain ssh client commands work first, \emph{before} you worry
about whether there's a problem with Mercurial.

The first thing to be sure of on the server side is that you can
actually log in from another machine at all.  If you can't use
\command{ssh} or \command{putty} to log in, the error message you get
may give you a few hints as to what's wrong.  The most common problems
are as follows.
\begin{itemize}
\item If you get a ``connection refused'' error, either there isn't an
  SSH daemon running on the server at all, or it's inaccessible due to
  firewall configuration.
\item If you get a ``no route to host'' error, you either have an
  incorrect address for the server or a seriously locked down firewall
  that won't admit its existence at all.
\item If you get a ``permission denied'' error, you may have mistyped
  the username on the server, or you could have mistyped your key's
  passphrase or the remote user's password.
\end{itemize}
In summary, if you're having trouble talking to the server's ssh
daemon, first make sure that one is running at all.  On many systems
it will be installed, but disabled, by default.  Once you're done with
this step, you should then check that the server's firewall is
configured to allow incoming connections on the port the ssh daemon is
listening on (usually~22).  Don't worry about more exotic
possibilities for misconfiguration until you've checked these two
first.

If you're using an authentication agent on the client side to store
passphrases for your keys, you ought to be able to log into the server
without being prompted for a passphrase or a password.  If you're
prompted for a passphrase, there are a few possible culprits.
\begin{itemize}
\item You might have forgotten to use \command{ssh-add} or
  \command{pageant} to store the passphrase.
\item You might have stored the passphrase for the wrong key.
\end{itemize}
If you're being prompted for the remote user's password, there are
another few possible problems to check.
\begin{itemize}
\item Either the user's home directory or their \sdirname{.ssh}
  directory might have excessively liberal permissions.  As a result,
  the ssh daemon will not trust or read their
  \sfilename{authorized\_keys} file.  For example, a group-writable
  home or \sdirname{.ssh} directory will often cause this symptom.
\item The user's \sfilename{authorized\_keys} file may have a problem.
  If anyone other than the user owns or can write to that file, the
  ssh daemon will not trust or read it.
\end{itemize}

In the ideal world, you should be able to run the following command
successfully, and it should print exactly one line of output, the
current date and time.
\begin{codesample2}
  ssh myserver date
\end{codesample2}

If, on your server, you have login scripts that print banners or other
junk even when running non-interactive commands like this, you should
fix them before you continue, so that they only print output if
they're run interactively.  Otherwise these banners will at least
clutter up Mercurial's output.  Worse, they could potentially cause
problems with running Mercurial commands remotely.  Mercurial makes
tries to detect and ignore banners in non-interactive \command{ssh}
sessions, but it is not foolproof.  (If you're editing your login
scripts on your server, the usual way to see if a login script is
running in an interactive shell is to check the return code from the
command \Verb|tty -s|.)

Once you've verified that plain old ssh is working with your server,
the next step is to ensure that Mercurial runs on the server.  The
following command should run successfully:
\begin{codesample2}
  ssh myserver hg version
\end{codesample2}
If you see an error message instead of normal \hgcmd{version} output,
this is usually because you haven't installed Mercurial to
\dirname{/usr/bin}.  Don't worry if this is the case; you don't need
to do that.  But you should check for a few possible problems.
\begin{itemize}
\item Is Mercurial really installed on the server at all?  I know this
  sounds trivial, but it's worth checking!
\item Maybe your shell's search path (usually set via the \envar{PATH}
  environment variable) is simply misconfigured.
\item Perhaps your \envar{PATH} environment variable is only being set
  to point to the location of the \command{hg} executable if the login
  session is interactive.  This can happen if you're setting the path
  in the wrong shell login script.  See your shell's documentation for
  details.
\item The \envar{PYTHONPATH} environment variable may need to contain
  the path to the Mercurial Python modules.  It might not be set at
  all; it could be incorrect; or it may be set only if the login is
  interactive.
\end{itemize}

If you can run \hgcmd{version} over an ssh connection, well done!
You've got the server and client sorted out.  You should now be able
to use Mercurial to access repositories hosted by that username on
that server.  If you run into problems with Mercurial and ssh at this
point, try using the \hggopt{--debug} option to get a clearer picture
of what's going on.

\subsection{Using compression with ssh}

Mercurial does not compress data when it uses the ssh protocol,
because the ssh protocol can transparently compress data.  However,
the default behaviour of ssh clients is \emph{not} to request
compression.

Over any network other than a fast LAN (even a wireless network),
using compression is likely to significantly speed up Mercurial's
network operations.  For example, over a WAN, someone measured
compression as reducing the amount of time required to clone a
particularly large repository from~51 minutes to~17 minutes.

Both \command{ssh} and \command{plink} accept a \cmdopt{ssh}{-C}
option which turns on compression.  You can easily edit your \hgrc\ to
enable compression for all of Mercurial's uses of the ssh protocol.
\begin{codesample2}
  [ui]
  ssh = ssh -C
\end{codesample2}

If you use \command{ssh}, you can configure it to always use
compression when talking to your server.  To do this, edit your
\sfilename{.ssh/config} file (which may not yet exist), as follows.
\begin{codesample2}
  Host hg
    Compression yes
    HostName hg.example.com
\end{codesample2}
This defines an alias, \texttt{hg}.  When you use it on the
\command{ssh} command line or in a Mercurial \texttt{ssh}-protocol
URL, it will cause \command{ssh} to connect to \texttt{hg.example.com}
and use compression.  This gives you both a shorter name to type and
compression, each of which is a good thing in its own right.

\section{Serving over HTTP using CGI}
\label{sec:collab:cgi}

Depending on how ambitious you are, configuring Mercurial's CGI
interface can take anything from a few moments to several hours.

We'll begin with the simplest of examples, and work our way towards a
more complex configuration.  Even for the most basic case, you're
almost certainly going to need to read and modify your web server's
configuration.

\begin{note}
  Configuring a web server is a complex, fiddly, and highly
  system-dependent activity.  I can't possibly give you instructions
  that will cover anything like all of the cases you will encounter.
  Please use your discretion and judgment in following the sections
  below.  Be prepared to make plenty of mistakes, and to spend a lot
  of time reading your server's error logs.
\end{note}

\subsection{Web server configuration checklist}

Before you continue, do take a few moments to check a few aspects of
your system's setup.

\begin{enumerate}
\item Do you have a web server installed at all?  Mac OS X ships with
  Apache, but many other systems may not have a web server installed.
\item If you have a web server installed, is it actually running?  On
  most systems, even if one is present, it will be disabled by
  default.
\item Is your server configured to allow you to run CGI programs in
  the directory where you plan to do so?  Most servers default to
  explicitly disabling the ability to run CGI programs.
\end{enumerate}

If you don't have a web server installed, and don't have substantial
experience configuring Apache, you should consider using the
\texttt{lighttpd} web server instead of Apache.  Apache has a
well-deserved reputation for baroque and confusing configuration.
While \texttt{lighttpd} is less capable in some ways than Apache, most
of these capabilities are not relevant to serving Mercurial
repositories.  And \texttt{lighttpd} is undeniably \emph{much} easier
to get started with than Apache.

\subsection{Basic CGI configuration}

On Unix-like systems, it's common for users to have a subdirectory
named something like \dirname{public\_html} in their home directory,
from which they can serve up web pages.  A file named \filename{foo}
in this directory will be accessible at a URL of the form
\texttt{http://www.example.com/\~{}username/foo}.

To get started, find the \sfilename{hgweb.cgi} script that should be
present in your Mercurial installation.  If you can't quickly find a
local copy on your system, simply download one from the master
Mercurial repository at
\url{http://www.selenic.com/repo/hg/raw-file/tip/hgweb.cgi}.

You'll need to copy this script into your \dirname{public\_html}
directory, and ensure that it's executable.
\begin{codesample2}
  cp .../hgweb.cgi ~/public_html
  chmod 755 ~/public_html/hgweb.cgi
\end{codesample2}
The \texttt{755} argument to \command{chmod} is a little more general
than just making the script executable: it ensures that the script is
executable by anyone, and that ``group'' and ``other'' write
permissions are \emph{not} set.  If you were to leave those write
permissions enabled, Apache's \texttt{suexec} subsystem would likely
refuse to execute the script.  In fact, \texttt{suexec} also insists
that the \emph{directory} in which the script resides must not be
writable by others.
\begin{codesample2}
  chmod 755 ~/public_html
\end{codesample2}

\subsubsection{What could \emph{possibly} go wrong?}
\label{sec:collab:wtf}

Once you've copied the CGI script into place, go into a web browser,
and try to open the URL \url{http://myhostname/~myuser/hgweb.cgi},
\emph{but} brace yourself for instant failure.  There's a high
probability that trying to visit this URL will fail, and there are
many possible reasons for this.  In fact, you're likely to stumble
over almost every one of the possible errors below, so please read
carefully.  The following are all of the problems I ran into on a
system running Fedora~7, with a fresh installation of Apache, and a
user account that I created specially to perform this exercise.

Your web server may have per-user directories disabled.  If you're
using Apache, search your config file for a \texttt{UserDir}
directive.  If there's none present, per-user directories will be
disabled.  If one exists, but its value is \texttt{disabled}, then
per-user directories will be disabled.  Otherwise, the string after
\texttt{UserDir} gives the name of the subdirectory that Apache will
look in under your home directory, for example \dirname{public\_html}.

Your file access permissions may be too restrictive.  The web server
must be able to traverse your home directory and directories under
your \dirname{public\_html} directory, and read files under the latter
too.  Here's a quick recipe to help you to make your permissions more
appropriate.
\begin{codesample2}
  chmod 755 ~
  find ~/public_html -type d -print0 | xargs -0r chmod 755
  find ~/public_html -type f -print0 | xargs -0r chmod 644
\end{codesample2}

The other possibility with permissions is that you might get a
completely empty window when you try to load the script.  In this
case, it's likely that your access permissions are \emph{too
  permissive}.  Apache's \texttt{suexec} subsystem won't execute a
script that's group-~or world-writable, for example.

Your web server may be configured to disallow execution of CGI
programs in your per-user web directory.  Here's Apache's
default per-user configuration from my Fedora system.
\begin{codesample2}
  <Directory /home/*/public_html>
      AllowOverride FileInfo AuthConfig Limit
      Options MultiViews Indexes SymLinksIfOwnerMatch IncludesNoExec
      <Limit GET POST OPTIONS>
          Order allow,deny
          Allow from all
      </Limit>
      <LimitExcept GET POST OPTIONS>
          Order deny,allow
          Deny from all
      </LimitExcept>
  </Directory>
\end{codesample2}
If you find a similar-looking \texttt{Directory} group in your Apache
configuration, the directive to look at inside it is \texttt{Options}.
Add \texttt{ExecCGI} to the end of this list if it's missing, and
restart the web server.

If you find that Apache serves you the text of the CGI script instead
of executing it, you may need to either uncomment (if already present)
or add a directive like this.
\begin{codesample2}
  AddHandler cgi-script .cgi
\end{codesample2}

The next possibility is that you might be served with a colourful
Python backtrace claiming that it can't import a
\texttt{mercurial}-related module.  This is actually progress!  The
server is now capable of executing your CGI script.  This error is
only likely to occur if you're running a private installation of
Mercurial, instead of a system-wide version.  Remember that the web
server runs the CGI program without any of the environment variables
that you take for granted in an interactive session.  If this error
happens to you, edit your copy of \sfilename{hgweb.cgi} and follow the
directions inside it to correctly set your \envar{PYTHONPATH}
environment variable.

Finally, you are \emph{certain} to by served with another colourful
Python backtrace: this one will complain that it can't find
\dirname{/path/to/repository}.  Edit your \sfilename{hgweb.cgi} script
and replace the \dirname{/path/to/repository} string with the complete
path to the repository you want to serve up.

At this point, when you try to reload the page, you should be
presented with a nice HTML view of your repository's history.  Whew!

\subsubsection{Configuring lighttpd}

To be exhaustive in my experiments, I tried configuring the
increasingly popular \texttt{lighttpd} web server to serve the same
repository as I described with Apache above.  I had already overcome
all of the problems I outlined with Apache, many of which are not
server-specific.  As a result, I was fairly sure that my file and
directory permissions were good, and that my \sfilename{hgweb.cgi}
script was properly edited.

Once I had Apache running, getting \texttt{lighttpd} to serve the
repository was a snap (in other words, even if you're trying to use
\texttt{lighttpd}, you should read the Apache section).  I first had
to edit the \texttt{mod\_access} section of its config file to enable
\texttt{mod\_cgi} and \texttt{mod\_userdir}, both of which were
disabled by default on my system.  I then added a few lines to the end
of the config file, to configure these modules.
\begin{codesample2}
  userdir.path = "public_html"
  cgi.assign = ( ".cgi" => "" )
\end{codesample2}
With this done, \texttt{lighttpd} ran immediately for me.  If I had
configured \texttt{lighttpd} before Apache, I'd almost certainly have
run into many of the same system-level configuration problems as I did
with Apache.  However, I found \texttt{lighttpd} to be noticeably
easier to configure than Apache, even though I've used Apache for over
a decade, and this was my first exposure to \texttt{lighttpd}.

\subsection{Sharing multiple repositories with one CGI script}

The \sfilename{hgweb.cgi} script only lets you publish a single
repository, which is an annoying restriction.  If you want to publish
more than one without wracking yourself with multiple copies of the
same script, each with different names, a better choice is to use the
\sfilename{hgwebdir.cgi} script.

The procedure to configure \sfilename{hgwebdir.cgi} is only a little
more involved than for \sfilename{hgweb.cgi}.  First, you must obtain
a copy of the script.  If you don't have one handy, you can download a
copy from the master Mercurial repository at
\url{http://www.selenic.com/repo/hg/raw-file/tip/hgwebdir.cgi}.

You'll need to copy this script into your \dirname{public\_html}
directory, and ensure that it's executable.
\begin{codesample2}
  cp .../hgwebdir.cgi ~/public_html
  chmod 755 ~/public_html ~/public_html/hgwebdir.cgi
\end{codesample2}
With basic configuration out of the way, try to visit
\url{http://myhostname/~myuser/hgwebdir.cgi} in your browser.  It
should display an empty list of repositories.  If you get a blank
window or error message, try walking through the list of potential
problems in section~\ref{sec:collab:wtf}.

The \sfilename{hgwebdir.cgi} script relies on an external
configuration file.  By default, it searches for a file named
\sfilename{hgweb.config} in the same directory as itself.  You'll need
to create this file, and make it world-readable.  The format of the
file is similar to a Windows ``ini'' file, as understood by Python's
\texttt{ConfigParser}~\cite{web:configparser} module.

The easiest way to configure \sfilename{hgwebdir.cgi} is with a
section named \texttt{collections}.  This will automatically publish
\emph{every} repository under the directories you name.  The section
should look like this:
\begin{codesample2}
  [collections]
  /my/root = /my/root
\end{codesample2}
Mercurial interprets this by looking at the directory name on the
\emph{right} hand side of the ``\texttt{=}'' sign; finding
repositories in that directory hierarchy; and using the text on the
\emph{left} to strip off matching text from the names it will actually
list in the web interface.  The remaining component of a path after
this stripping has occurred is called a ``virtual path''.

Given the example above, if we have a repository whose local path is
\dirname{/my/root/this/repo}, the CGI script will strip the leading
\dirname{/my/root} from the name, and publish the repository with a
virtual path of \dirname{this/repo}.  If the base URL for our CGI
script is \url{http://myhostname/~myuser/hgwebdir.cgi}, the complete
URL for that repository will be
\url{http://myhostname/~myuser/hgwebdir.cgi/this/repo}.

If we replace \dirname{/my/root} on the left hand side of this example
with \dirname{/my}, then \sfilename{hgwebdir.cgi} will only strip off
\dirname{/my} from the repository name, and will give us a virtual
path of \dirname{root/this/repo} instead of \dirname{this/repo}.

The \sfilename{hgwebdir.cgi} script will recursively search each
directory listed in the \texttt{collections} section of its
configuration file, but it will \texttt{not} recurse into the
repositories it finds.

The \texttt{collections} mechanism makes it easy to publish many
repositories in a ``fire and forget'' manner.  You only need to set up
the CGI script and configuration file one time.  Afterwards, you can
publish or unpublish a repository at any time by simply moving it
into, or out of, the directory hierarchy in which you've configured
\sfilename{hgwebdir.cgi} to look.

\subsubsection{Explicitly specifying which repositories to publish}

In addition to the \texttt{collections} mechanism, the
\sfilename{hgwebdir.cgi} script allows you to publish a specific list
of repositories.  To do so, create a \texttt{paths} section, with
contents of the following form.
\begin{codesample2}
  [paths]
  repo1 = /my/path/to/some/repo
  repo2 = /some/path/to/another
\end{codesample2}
In this case, the virtual path (the component that will appear in a
URL) is on the left hand side of each definition, while the path to
the repository is on the right.  Notice that there does not need to be
any relationship between the virtual path you choose and the location
of a repository in your filesystem.

If you wish, you can use both the \texttt{collections} and
\texttt{paths} mechanisms simultaneously in a single configuration
file.

\begin{note}
  If multiple repositories have the same virtual path,
  \sfilename{hgwebdir.cgi} will not report an error.  Instead, it will
  behave unpredictably.
\end{note}

\subsection{Downloading source archives}

Mercurial's web interface lets users download an archive of any
revision.  This archive will contain a snapshot of the working
directory as of that revision, but it will not contain a copy of the
repository data.

By default, this feature is not enabled.  To enable it, you'll need to
add an \rcitem{web}{allow\_archive} item to the \rcsection{web}
section of your \hgrc.

\subsection{Web configuration options}

Mercurial's web interfaces (the \hgcmd{serve} command, and the
\sfilename{hgweb.cgi} and \sfilename{hgwebdir.cgi} scripts) have a
number of configuration options that you can set.  These belong in a
section named \rcsection{web}.
\begin{itemize}
\item[\rcitem{web}{allow\_archive}] Determines which (if any) archive
  download mechanisms Mercurial supports.  If you enable this
  feature, users of the web interface will be able to download an
  archive of whatever revision of a repository they are viewing.
  To enable the archive feature, this item must take the form of a
  sequence of words drawn from the list below.
  \begin{itemize}
  \item[\texttt{bz2}] A \command{tar} archive, compressed using
    \texttt{bzip2} compression.  This has the best compression ratio,
    but uses the most CPU time on the server.
  \item[\texttt{gz}] A \command{tar} archive, compressed using
    \texttt{gzip} compression.
  \item[\texttt{zip}] A \command{zip} archive, compressed using LZW
    compression.  This format has the worst compression ratio, but is
    widely used in the Windows world.
  \end{itemize}
  If you provide an empty list, or don't have an
  \rcitem{web}{allow\_archive} entry at all, this feature will be
  disabled.  Here is an example of how to enable all three supported
  formats.
  \begin{codesample4}
    [web]
    allow_archive = bz2 gz zip
  \end{codesample4}
\item[\rcitem{web}{allowpull}] Boolean.  Determines whether the web
  interface allows remote users to \hgcmd{pull} and \hgcmd{clone} this
  repository over~HTTP.  If set to \texttt{no} or \texttt{false}, only
  the ``human-oriented'' portion of the web interface is available.
\item[\rcitem{web}{contact}] String.  A free-form (but preferably
  brief) string identifying the person or group in charge of the
  repository.  This often contains the name and email address of a
  person or mailing list.  It often makes sense to place this entry in
  a repository's own \sfilename{.hg/hgrc} file, but it can make sense
  to use in a global \hgrc\ if every repository has a single
  maintainer.
\item[\rcitem{web}{maxchanges}] Integer.  The default maximum number
  of changesets to display in a single page of output.
\item[\rcitem{web}{maxfiles}] Integer.  The default maximum number
  of modified files to display in a single page of output.
\item[\rcitem{web}{stripes}] Integer.  If the web interface displays
  alternating ``stripes'' to make it easier to visually align rows
  when you are looking at a table, this number controls the number of
  rows in each stripe.
\item[\rcitem{web}{style}] Controls the template Mercurial uses to
  display the web interface.  Mercurial ships with two web templates,
  named \texttt{default} and \texttt{gitweb} (the latter is much more
  visually attractive).  You can also specify a custom template of
  your own; see chapter~\ref{chap:template} for details.  Here, you
  can see how to enable the \texttt{gitweb} style.
  \begin{codesample4}
    [web]
    style = gitweb
  \end{codesample4}
\item[\rcitem{web}{templates}] Path.  The directory in which to search
  for template files.  By default, Mercurial searches in the directory
  in which it was installed.
\end{itemize}
If you are using \sfilename{hgwebdir.cgi}, you can place a few
configuration items in a \rcsection{web} section of the
\sfilename{hgweb.config} file instead of a \hgrc\ file, for
convenience.  These items are \rcitem{web}{motd} and
\rcitem{web}{style}.

\subsubsection{Options specific to an individual repository}

A few \rcsection{web} configuration items ought to be placed in a
repository's local \sfilename{.hg/hgrc}, rather than a user's or
global \hgrc.
\begin{itemize}
\item[\rcitem{web}{description}] String.  A free-form (but preferably
  brief) string that describes the contents or purpose of the
  repository.
\item[\rcitem{web}{name}] String.  The name to use for the repository
  in the web interface.  This overrides the default name, which is the
  last component of the repository's path.
\end{itemize}

\subsubsection{Options specific to the \hgcmd{serve} command}

Some of the items in the \rcsection{web} section of a \hgrc\ file are
only for use with the \hgcmd{serve} command.
\begin{itemize}
\item[\rcitem{web}{accesslog}] Path.  The name of a file into which to
  write an access log.  By default, the \hgcmd{serve} command writes
  this information to standard output, not to a file.  Log entries are
  written in the standard ``combined'' file format used by almost all
  web servers.
\item[\rcitem{web}{address}] String.  The local address on which the
  server should listen for incoming connections.  By default, the
  server listens on all addresses.
\item[\rcitem{web}{errorlog}] Path.  The name of a file into which to
  write an error log.  By default, the \hgcmd{serve} command writes this
  information to standard error, not to a file.
\item[\rcitem{web}{ipv6}] Boolean.  Whether to use the IPv6 protocol.
  By default, IPv6 is not used. 
\item[\rcitem{web}{port}] Integer.  The TCP~port number on which the
  server should listen.  The default port number used is~8000.
\end{itemize}

\subsubsection{Choosing the right \hgrc\ file to add \rcsection{web}
  items to}

It is important to remember that a web server like Apache or
\texttt{lighttpd} will run under a user~ID that is different to yours.
CGI scripts run by your server, such as \sfilename{hgweb.cgi}, will
usually also run under that user~ID.

If you add \rcsection{web} items to your own personal \hgrc\ file, CGI
scripts won't read that \hgrc\ file.  Those settings will thus only
affect the behaviour of the \hgcmd{serve} command when you run it.  To
cause CGI scripts to see your settings, either create a \hgrc\ file in
the home directory of the user ID that runs your web server, or add
those settings to a system-wide \hgrc\ file.


%%% Local Variables: 
%%% mode: latex
%%% TeX-master: "00book"
%%% End: