bos@559: <!-- vim: set filetype=docbkxml shiftwidth=2 autoindent expandtab tw=77 : -->
bos@559:
bos@559: <chapter id="cha:collab">
bos@572: <?dbhtml filename="collaborating-with-other-people.html"?>
bos@559: <title>Collaborating with other people</title>
bos@559:
bos@584: <para id="x_44a">As a completely decentralised tool, Mercurial doesn't impose
bos@559: any policy on how people ought to work with each other. However,
bos@559: if you're new to distributed revision control, it helps to have
bos@559: some tools and examples in mind when you're thinking about
bos@559: possible workflow models.</para>
bos@559:
bos@559: <sect1>
bos@559: <title>Mercurial's web interface</title>
bos@559:
bos@584: <para id="x_44b">Mercurial has a powerful web interface that provides several
bos@559: useful capabilities.</para>
bos@559:
bos@584: <para id="x_44c">For interactive use, the web interface lets you browse a
bos@559: single repository or a collection of repositories. You can view
bos@559: the history of a repository, examine each change (comments and
bos@675: diffs), and view the contents of each directory and file. You
bos@675: can even get a view of history that gives a graphical view of
bos@675: the relationships between individual changes and merges.</para>
bos@675:
bos@675: <para id="x_44d">Also for human consumption, the web interface provides
bos@675: Atom and RSS feeds of the changes in a repository. This lets you
bos@674: <quote>subscribe</quote> to a repository using your favorite
bos@559: feed reader, and be automatically notified of activity in that
bos@559: repository as soon as it happens. I find this capability much
bos@559: more convenient than the model of subscribing to a mailing list
bos@559: to which notifications are sent, as it requires no additional
bos@559: configuration on the part of whoever is serving the
bos@559: repository.</para>
bos@559:
bos@584: <para id="x_44e">The web interface also lets remote users clone a repository,
bos@559: pull changes from it, and (when the server is configured to
bos@559: permit it) push changes back to it. Mercurial's HTTP tunneling
bos@559: protocol aggressively compresses data, so that it works
bos@559: efficiently even over low-bandwidth network connections.</para>
bos@559:
bos@584: <para id="x_44f">The easiest way to get started with the web interface is to
bos@559: use your web browser to visit an existing repository, such as
bos@559: the master Mercurial repository at <ulink
bos@675: url="http://www.selenic.com/repo/hg">http://www.selenic.com/repo/hg</ulink>.</para>
bos@559:
bos@592: <para id="x_450">If you're interested in providing a web interface
bos@675: to your own repositories, there are several good ways to do
bos@675: this.</para>
bos@675:
bos@676: <para id="x_69d">The easiest and fastest way to get started in an informal
bos@675: environment is to use the <command role="hg-cmd">hg
bos@592: serve</command> command, which is best suited to short-term
bos@592: <quote>lightweight</quote> serving. See <xref
bos@559: linkend="sec:collab:serve"/> below for details of how to use
bos@675: this command.</para>
bos@675:
bos@701: <para id="x_69e">For longer-lived repositories that you'd like to
bos@701: have permanently available, there are several public hosting
bos@701: services available. Some are free to open source projects,
bos@701: while others offer paid commercial hosting. An up-to-date list
bos@701: is available at <ulink
bos@701: url="http://www.selenic.com/mercurial/wiki/index.cgi/MercurialHosting">http://www.selenic.com/mercurial/wiki/index.cgi/MercurialHosting</ulink>.</para>
bos@675:
bos@676: <para id="x_6a0">If you would prefer to host your own repositories, Mercurial
bos@675: has built-in support for several popular hosting technologies,
bos@675: most notably CGI (Common Gateway Interface), and WSGI (Web
bos@675: Services Gateway Interface). See <xref
bos@675: linkend="sec:collab:cgi"/> for details of CGI and WSGI
bos@559: configuration.</para>
bos@559: </sect1>
bos@675:
bos@559: <sect1>
bos@559: <title>Collaboration models</title>
bos@559:
bos@584: <para id="x_451">With a suitably flexible tool, making decisions about
bos@559: workflow is much more of a social engineering challenge than a
bos@559: technical one. Mercurial imposes few limitations on how you can
bos@559: structure the flow of work in a project, so it's up to you and
bos@559: your group to set up and live with a model that matches your own
bos@559: particular needs.</para>
bos@559:
bos@559: <sect2>
bos@559: <title>Factors to keep in mind</title>
bos@559:
bos@584: <para id="x_452">The most important aspect of any model that you must keep
bos@559: in mind is how well it matches the needs and capabilities of
bos@559: the people who will be using it. This might seem
bos@559: self-evident; even so, you still can't afford to forget it for
bos@559: a moment.</para>
bos@559:
bos@584: <para id="x_453">I once put together a workflow model that seemed to make
bos@559: perfect sense to me, but that caused a considerable amount of
bos@559: consternation and strife within my development team. In spite
bos@559: of my attempts to explain why we needed a complex set of
bos@559: branches, and how changes ought to flow between them, a few
bos@559: team members revolted. Even though they were smart people,
bos@559: they didn't want to pay attention to the constraints we were
bos@559: operating under, or face the consequences of those constraints
bos@559: in the details of the model that I was advocating.</para>
bos@559:
bos@584: <para id="x_454">Don't sweep foreseeable social or technical problems under
bos@559: the rug. Whatever scheme you put into effect, you should plan
bos@559: for mistakes and problem scenarios. Consider adding automated
bos@559: machinery to prevent, or quickly recover from, trouble that
bos@559: you can anticipate. As an example, if you intend to have a
bos@559: branch with not-for-release changes in it, you'd do well to
bos@559: think early about the possibility that someone might
bos@559: accidentally merge those changes into a release branch. You
bos@559: could avoid this particular problem by writing a hook that
bos@559: prevents changes from being merged from an inappropriate
bos@559: branch.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Informal anarchy</title>
bos@559:
bos@584: <para id="x_455">I wouldn't suggest an <quote>anything goes</quote>
bos@559: approach as something sustainable, but it's a model that's
bos@559: easy to grasp, and it works perfectly well in a few unusual
bos@559: situations.</para>
bos@559:
bos@584: <para id="x_456">As one example, many projects have a loose-knit group of
bos@559: collaborators who rarely physically meet each other. Some
bos@559: groups like to overcome the isolation of working at a distance
bos@674: by organizing occasional <quote>sprints</quote>. In a sprint,
bos@559: a number of people get together in a single location (a
bos@559: company's conference room, a hotel meeting room, that kind of
bos@559: place) and spend several days more or less locked in there,
bos@559: hacking intensely on a handful of projects.</para>
bos@559:
bos@675: <para id="x_457">A sprint or a hacking session in a coffee shop are the perfect places to use the
bos@592: <command role="hg-cmd">hg serve</command> command, since
bos@592: <command role="hg-cmd">hg serve</command> does not require any
bos@592: fancy server infrastructure. You can get started with
bos@592: <command role="hg-cmd">hg serve</command> in moments, by
bos@592: reading <xref linkend="sec:collab:serve"/> below. Then simply
bos@592: tell the person next to you that you're running a server, send
bos@592: the URL to them in an instant message, and you immediately
bos@592: have a quick-turnaround way to work together. They can type
bos@592: your URL into their web browser and quickly review your
bos@592: changes; or they can pull a bugfix from you and verify it; or
bos@592: they can clone a branch containing a new feature and try it
bos@592: out.</para>
bos@559:
bos@675: <para id="x_458">The charm, and the problem, with doing things
bos@675: in an ad hoc fashion like this is that only people who know
bos@675: about your changes, and where they are, can see them. Such an
bos@675: informal approach simply doesn't scale beyond a handful
bos@675: people, because each individual needs to know about
bos@675: <emphasis>n</emphasis> different repositories to pull
bos@675: from.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>A single central repository</title>
bos@559:
bos@584: <para id="x_459">For smaller projects migrating from a centralised revision
bos@559: control tool, perhaps the easiest way to get started is to
bos@559: have changes flow through a single shared central repository.
bos@559: This is also the most common <quote>building block</quote> for
bos@559: more ambitious workflow schemes.</para>
bos@559:
bos@584: <para id="x_45a">Contributors start by cloning a copy of this repository.
bos@559: They can pull changes from it whenever they need to, and some
bos@559: (perhaps all) developers have permission to push a change back
bos@559: when they're ready for other people to see it.</para>
bos@559:
bos@584: <para id="x_45b">Under this model, it can still often make sense for people
bos@559: to pull changes directly from each other, without going
bos@559: through the central repository. Consider a case in which I
bos@559: have a tentative bug fix, but I am worried that if I were to
bos@559: publish it to the central repository, it might subsequently
bos@559: break everyone else's trees as they pull it. To reduce the
bos@559: potential for damage, I can ask you to clone my repository
bos@559: into a temporary repository of your own and test it. This
bos@559: lets us put off publishing the potentially unsafe change until
bos@559: it has had a little testing.</para>
bos@559:
bos@675: <para id="x_45c">If a team is hosting its own repository in this
bos@675: kind of scenario, people will usually use the
bos@675: <command>ssh</command> protocol to securely push changes to
bos@675: the central repository, as documented in <xref
bos@592: linkend="sec:collab:ssh"/>. It's also usual to publish a
bos@675: read-only copy of the repository over HTTP, as in
bos@675: <xref linkend="sec:collab:cgi"/>. Publishing over HTTP
bos@675: satisfies the needs of people who don't have push access, and
bos@675: those who want to use web browsers to browse the repository's
bos@675: history.</para>
bos@675: </sect2>
bos@675:
bos@675: <sect2>
bos@675: <title>A hosted central repository</title>
bos@675:
bos@676: <para id="x_6a1">A wonderful thing about public hosting services like
bos@675: <ulink url="http://bitbucket.org/">Bitbucket</ulink> is that
bos@675: not only do they handle the fiddly server configuration
bos@675: details, such as user accounts, authentication, and secure
bos@675: wire protocols, they provide additional infrastructure to make
bos@675: this model work well.</para>
bos@675:
bos@676: <para id="x_6a2">For instance, a well-engineered hosting service will let
bos@675: people clone their own copies of a repository with a single
bos@675: click. This lets people work in separate spaces and share
bos@675: their changes when they're ready.</para>
bos@675:
bos@676: <para id="x_6a3">In addition, a good hosting service will let people
bos@675: communicate with each other, for instance to say <quote>there
bos@675: are changes ready for you to review in this
bos@675: tree</quote>.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Working with multiple branches</title>
bos@559:
bos@584: <para id="x_45d">Projects of any significant size naturally tend to make
bos@559: progress on several fronts simultaneously. In the case of
bos@559: software, it's common for a project to go through periodic
bos@559: official releases. A release might then go into
bos@559: <quote>maintenance mode</quote> for a while after its first
bos@559: publication; maintenance releases tend to contain only bug
bos@559: fixes, not new features. In parallel with these maintenance
bos@559: releases, one or more future releases may be under
bos@559: development. People normally use the word
bos@559: <quote>branch</quote> to refer to one of these many slightly
bos@559: different directions in which development is
bos@559: proceeding.</para>
bos@559:
bos@584: <para id="x_45e">Mercurial is particularly well suited to managing a number
bos@559: of simultaneous, but not identical, branches. Each
bos@559: <quote>development direction</quote> can live in its own
bos@559: central repository, and you can merge changes from one to
bos@559: another as the need arises. Because repositories are
bos@559: independent of each other, unstable changes in a development
bos@559: branch will never affect a stable branch unless someone
bos@675: explicitly merges those changes into the stable branch.</para>
bos@559:
bos@584: <para id="x_45f">Here's an example of how this can work in practice. Let's
bos@559: say you have one <quote>main branch</quote> on a central
bos@567: server.</para>
bos@567:
bos@567: &interaction.branching.init;
bos@567:
bos@584: <para id="x_460">People clone it, make changes locally, test them, and push
bos@567: them back.</para>
bos@559:
bos@584: <para id="x_461">Once the main branch reaches a release milestone, you can
bos@559: use the <command role="hg-cmd">hg tag</command> command to
bos@567: give a permanent name to the milestone revision.</para>
bos@567:
bos@567: &interaction.branching.tag;
bos@567:
bos@584: <para id="x_462">Let's say some ongoing
bos@567: development occurs on the main branch.</para>
bos@567:
bos@567: &interaction.branching.main;
bos@567:
bos@584: <para id="x_463">Using the tag that was recorded at the milestone, people
bos@567: who clone that repository at any time in the future can use
bos@567: <command role="hg-cmd">hg update</command> to get a copy of
bos@567: the working directory exactly as it was when that tagged
bos@567: revision was committed.</para>
bos@567:
bos@567: &interaction.branching.update;
bos@559:
bos@584: <para id="x_464">In addition, immediately after the main branch is tagged,
bos@675: we can then clone the main branch on the server to a new
bos@567: <quote>stable</quote> branch, also on the server.</para>
bos@567:
bos@567: &interaction.branching.clone;
bos@559:
bos@675: <para id="x_465">If we need to make a change to the stable
bos@675: branch, we can then clone <emphasis>that</emphasis>
bos@675: repository, make our changes, commit, and push our changes
bos@675: back there.</para>
bos@567:
bos@567: &interaction.branching.stable;
bos@567:
bos@584: <para id="x_466">Because Mercurial repositories are independent, and
bos@567: Mercurial doesn't move changes around automatically, the
bos@567: stable and main branches are <emphasis>isolated</emphasis>
bos@675: from each other. The changes that we made on the main branch
bos@567: don't <quote>leak</quote> to the stable branch, and vice
bos@567: versa.</para>
bos@559:
bos@675: <para id="x_467">We'll often want all of our bugfixes on the stable
bos@559: branch to show up on the main branch, too. Rather than
bos@675: rewrite a bugfix on the main branch, we can simply pull and
bos@559: merge changes from the stable to the main branch, and
bos@675: Mercurial will bring those bugfixes in for us.</para>
bos@675:
bos@675: &interaction.branching.merge;
bos@675:
bos@675: <para id="x_468">The main branch will still contain changes that
bos@675: are not on the stable branch, but it will also contain all of
bos@675: the bugfixes from the stable branch. The stable branch
bos@675: remains unaffected by these changes, since changes are only
bos@675: flowing from the stable to the main branch, and not the other
bos@675: way.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Feature branches</title>
bos@559:
bos@584: <para id="x_469">For larger projects, an effective way to manage change is
bos@559: to break up a team into smaller groups. Each group has a
bos@559: shared branch of its own, cloned from a single
bos@559: <quote>master</quote> branch used by the entire project.
bos@559: People working on an individual branch are typically quite
bos@559: isolated from developments on other branches.</para>
bos@559:
bos@591: <figure id="fig:collab:feature-branches">
bos@591: <title>Feature branches</title>
bos@591: <mediaobject>
dongsheng@655: <imageobject><imagedata width="100%" fileref="figs/feature-branches.png"/></imageobject>
bos@591: <textobject><phrase>XXX add text</phrase></textobject>
bos@591: </mediaobject>
bos@591: </figure>
bos@559:
bos@584: <para id="x_46b">When a particular feature is deemed to be in suitable
bos@559: shape, someone on that feature team pulls and merges from the
bos@559: master branch into the feature branch, then pushes back up to
bos@559: the master branch.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>The release train</title>
bos@559:
bos@674: <para id="x_46c">Some projects are organized on a <quote>train</quote>
bos@559: basis: a release is scheduled to happen every few months, and
bos@559: whatever features are ready when the <quote>train</quote> is
bos@559: ready to leave are allowed in.</para>
bos@559:
bos@584: <para id="x_46d">This model resembles working with feature branches. The
bos@559: difference is that when a feature branch misses a train,
bos@559: someone on the feature team pulls and merges the changes that
bos@559: went out on that train release into the feature branch, and
bos@559: the team continues its work on top of that release so that
bos@559: their feature can make the next release.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>The Linux kernel model</title>
bos@559:
bos@584: <para id="x_46e">The development of the Linux kernel has a shallow
bos@559: hierarchical structure, surrounded by a cloud of apparent
bos@559: chaos. Because most Linux developers use
bos@559: <command>git</command>, a distributed revision control tool
bos@559: with capabilities similar to Mercurial, it's useful to
bos@559: describe the way work flows in that environment; if you like
bos@559: the ideas, the approach translates well across tools.</para>
bos@559:
bos@584: <para id="x_46f">At the center of the community sits Linus Torvalds, the
bos@559: creator of Linux. He publishes a single source repository
bos@559: that is considered the <quote>authoritative</quote> current
bos@559: tree by the entire developer community. Anyone can clone
bos@559: Linus's tree, but he is very choosy about whose trees he pulls
bos@559: from.</para>
bos@559:
bos@584: <para id="x_470">Linus has a number of <quote>trusted lieutenants</quote>.
bos@559: As a general rule, he pulls whatever changes they publish, in
bos@559: most cases without even reviewing those changes. Some of
bos@559: those lieutenants are generally agreed to be
bos@559: <quote>maintainers</quote>, responsible for specific
bos@559: subsystems within the kernel. If a random kernel hacker wants
bos@559: to make a change to a subsystem that they want to end up in
bos@559: Linus's tree, they must find out who the subsystem's
bos@559: maintainer is, and ask that maintainer to take their change.
bos@559: If the maintainer reviews their changes and agrees to take
bos@559: them, they'll pass them along to Linus in due course.</para>
bos@559:
bos@584: <para id="x_471">Individual lieutenants have their own approaches to
bos@559: reviewing, accepting, and publishing changes; and for deciding
bos@559: when to feed them to Linus. In addition, there are several
bos@559: well known branches that people use for different purposes.
bos@559: For example, a few people maintain <quote>stable</quote>
bos@559: repositories of older versions of the kernel, to which they
bos@559: apply critical fixes as needed. Some maintainers publish
bos@559: multiple trees: one for experimental changes; one for changes
bos@559: that they are about to feed upstream; and so on. Others just
bos@559: publish a single tree.</para>
bos@559:
bos@584: <para id="x_472">This model has two notable features. The first is that
bos@559: it's <quote>pull only</quote>. You have to ask, convince, or
bos@559: beg another developer to take a change from you, because there
bos@559: are almost no trees to which more than one person can push,
bos@559: and there's no way to push changes into a tree that someone
bos@559: else controls.</para>
bos@559:
bos@584: <para id="x_473">The second is that it's based on reputation and acclaim.
bos@559: If you're an unknown, Linus will probably ignore changes from
bos@559: you without even responding. But a subsystem maintainer will
bos@559: probably review them, and will likely take them if they pass
bos@559: their criteria for suitability. The more <quote>good</quote>
bos@559: changes you contribute to a maintainer, the more likely they
bos@559: are to trust your judgment and accept your changes. If you're
bos@559: well-known and maintain a long-lived branch for something
bos@559: Linus hasn't yet accepted, people with similar interests may
bos@559: pull your changes regularly to keep up with your work.</para>
bos@559:
bos@584: <para id="x_474">Reputation and acclaim don't necessarily cross subsystem
bos@559: or <quote>people</quote> boundaries. If you're a respected
bos@559: but specialised storage hacker, and you try to fix a
bos@559: networking bug, that change will receive a level of scrutiny
bos@559: from a network maintainer comparable to a change from a
bos@559: complete stranger.</para>
bos@559:
bos@584: <para id="x_475">To people who come from more orderly project backgrounds,
bos@559: the comparatively chaotic Linux kernel development process
bos@559: often seems completely insane. It's subject to the whims of
bos@559: individuals; people make sweeping changes whenever they deem
bos@559: it appropriate; and the pace of development is astounding.
bos@559: And yet Linux is a highly successful, well-regarded piece of
bos@559: software.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Pull-only versus shared-push collaboration</title>
bos@559:
bos@584: <para id="x_476">A perpetual source of heat in the open source community is
bos@559: whether a development model in which people only ever pull
bos@559: changes from others is <quote>better than</quote> one in which
bos@559: multiple people can push changes to a shared
bos@559: repository.</para>
bos@559:
bos@584: <para id="x_477">Typically, the backers of the shared-push model use tools
bos@559: that actively enforce this approach. If you're using a
bos@559: centralised revision control tool such as Subversion, there's
bos@559: no way to make a choice over which model you'll use: the tool
bos@559: gives you shared-push, and if you want to do anything else,
bos@559: you'll have to roll your own approach on top (such as applying
bos@559: a patch by hand).</para>
bos@559:
bos@675: <para id="x_478">A good distributed revision control tool will
bos@675: support both models. You and your collaborators can then
bos@675: structure how you work together based on your own needs and
bos@675: preferences, not on what contortions your tools force you
bos@675: into.</para>
bos@559: </sect2>
bos@559: <sect2>
bos@559: <title>Where collaboration meets branch management</title>
bos@559:
bos@592: <para id="x_479">Once you and your team set up some shared
bos@592: repositories and start propagating changes back and forth
bos@592: between local and shared repos, you begin to face a related,
bos@592: but slightly different challenge: that of managing the
bos@592: multiple directions in which your team may be moving at once.
bos@592: Even though this subject is intimately related to how your
bos@592: team collaborates, it's dense enough to merit treatment of its
bos@592: own, in <xref linkend="chap:branch"/>.</para>
bos@559: </sect2>
bos@559: </sect1>
bos@675:
bos@559: <sect1>
bos@559: <title>The technical side of sharing</title>
bos@559:
bos@584: <para id="x_47a">The remainder of this chapter is devoted to the question of
bos@675: sharing changes with your collaborators.</para>
bos@559: </sect1>
bos@675:
bos@559: <sect1 id="sec:collab:serve">
bos@559: <title>Informal sharing with <command role="hg-cmd">hg
bos@559: serve</command></title>
bos@559:
bos@584: <para id="x_47b">Mercurial's <command role="hg-cmd">hg serve</command>
bos@559: command is wonderfully suited to small, tight-knit, and
bos@559: fast-paced group environments. It also provides a great way to
bos@559: get a feel for using Mercurial commands over a network.</para>
bos@559:
bos@584: <para id="x_47c">Run <command role="hg-cmd">hg serve</command> inside a
bos@559: repository, and in under a second it will bring up a specialised
bos@559: HTTP server; this will accept connections from any client, and
bos@559: serve up data for that repository until you terminate it.
bos@559: Anyone who knows the URL of the server you just started, and can
bos@559: talk to your computer over the network, can then use a web
bos@559: browser or Mercurial to read data from that repository. A URL
bos@559: for a <command role="hg-cmd">hg serve</command> instance running
bos@559: on a laptop is likely to look something like
bos@559: <literal>http://my-laptop.local:8000/</literal>.</para>
bos@559:
bos@584: <para id="x_47d">The <command role="hg-cmd">hg serve</command> command is
bos@559: <emphasis>not</emphasis> a general-purpose web server. It can do
bos@559: only two things:</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_47e">Allow people to browse the history of the
bos@559: repository it's serving, from their normal web
bos@559: browsers.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_47f">Speak Mercurial's wire protocol, so that people
bos@559: can <command role="hg-cmd">hg clone</command> or <command
bos@559: role="hg-cmd">hg pull</command> changes from that
bos@559: repository.</para>
bos@559: </listitem></itemizedlist>
bos@584: <para id="x_480">In particular, <command role="hg-cmd">hg serve</command>
bos@559: won't allow remote users to <emphasis>modify</emphasis> your
bos@559: repository. It's intended for read-only use.</para>
bos@559:
bos@584: <para id="x_481">If you're getting started with Mercurial, there's nothing to
bos@559: prevent you from using <command role="hg-cmd">hg serve</command>
bos@559: to serve up a repository on your own computer, then use commands
bos@559: like <command role="hg-cmd">hg clone</command>, <command
bos@559: role="hg-cmd">hg incoming</command>, and so on to talk to that
bos@559: server as if the repository was hosted remotely. This can help
bos@559: you to quickly get acquainted with using commands on
bos@559: network-hosted repositories.</para>
bos@559:
bos@559: <sect2>
bos@559: <title>A few things to keep in mind</title>
bos@559:
bos@584: <para id="x_482">Because it provides unauthenticated read access to all
bos@559: clients, you should only use <command role="hg-cmd">hg
bos@559: serve</command> in an environment where you either don't
bos@559: care, or have complete control over, who can access your
bos@559: network and pull data from your repository.</para>
bos@559:
bos@584: <para id="x_483">The <command role="hg-cmd">hg serve</command> command
bos@559: knows nothing about any firewall software you might have
bos@559: installed on your system or network. It cannot detect or
bos@559: control your firewall software. If other people are unable to
bos@559: talk to a running <command role="hg-cmd">hg serve</command>
bos@559: instance, the second thing you should do
bos@559: (<emphasis>after</emphasis> you make sure that they're using
bos@559: the correct URL) is check your firewall configuration.</para>
bos@559:
bos@584: <para id="x_484">By default, <command role="hg-cmd">hg serve</command>
bos@559: listens for incoming connections on port 8000. If another
bos@559: process is already listening on the port you want to use, you
bos@559: can specify a different port to listen on using the <option
bos@559: role="hg-opt-serve">-p</option> option.</para>
bos@559:
bos@584: <para id="x_485">Normally, when <command role="hg-cmd">hg serve</command>
bos@559: starts, it prints no output, which can be a bit unnerving. If
bos@559: you'd like to confirm that it is indeed running correctly, and
bos@559: find out what URL you should send to your collaborators, start
bos@559: it with the <option role="hg-opt-global">-v</option>
bos@559: option.</para>
bos@559: </sect2>
bos@559: </sect1>
bos@675:
bos@559: <sect1 id="sec:collab:ssh">
bos@559: <title>Using the Secure Shell (ssh) protocol</title>
bos@559:
bos@584: <para id="x_486">You can pull and push changes securely over a network
bos@559: connection using the Secure Shell (<literal>ssh</literal>)
bos@559: protocol. To use this successfully, you may have to do a little
bos@559: bit of configuration on the client or server sides.</para>
bos@559:
bos@675: <para id="x_487">If you're not familiar with ssh, it's the name of
bos@675: both a command and a network protocol that let you securely
bos@675: communicate with another computer. To use it with Mercurial,
bos@675: you'll be setting up one or more user accounts on a server so
bos@675: that remote users can log in and execute commands.</para>
bos@559:
bos@584: <para id="x_488">(If you <emphasis>are</emphasis> familiar with ssh, you'll
bos@559: probably find some of the material that follows to be elementary
bos@559: in nature.)</para>
bos@559:
bos@559: <sect2>
bos@559: <title>How to read and write ssh URLs</title>
bos@559:
bos@584: <para id="x_489">An ssh URL tends to look like this:</para>
bos@559: <programlisting>ssh://bos@hg.serpentine.com:22/hg/hgbook</programlisting>
bos@559: <orderedlist>
bos@584: <listitem><para id="x_48a">The <quote><literal>ssh://</literal></quote>
bos@559: part tells Mercurial to use the ssh protocol.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_48b">The <quote><literal>bos@</literal></quote>
bos@559: component indicates what username to log into the server
bos@559: as. You can leave this out if the remote username is the
bos@559: same as your local username.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_48c">The
bos@559: <quote><literal>hg.serpentine.com</literal></quote> gives
bos@559: the hostname of the server to log into.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_48d">The <quote>:22</quote> identifies the port
bos@559: number to connect to the server on. The default port is
bos@579: 22, so you only need to specify a colon and port number if
bos@579: you're <emphasis>not</emphasis> using port 22.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_48e">The remainder of the URL is the local path to
bos@559: the repository on the server.</para>
bos@559: </listitem></orderedlist>
bos@559:
bos@584: <para id="x_48f">There's plenty of scope for confusion with the path
bos@559: component of ssh URLs, as there is no standard way for tools
bos@559: to interpret it. Some programs behave differently than others
bos@559: when dealing with these paths. This isn't an ideal situation,
bos@559: but it's unlikely to change. Please read the following
bos@559: paragraphs carefully.</para>
bos@559:
bos@584: <para id="x_490">Mercurial treats the path to a repository on the server as
bos@559: relative to the remote user's home directory. For example, if
bos@559: user <literal>foo</literal> on the server has a home directory
bos@559: of <filename class="directory">/home/foo</filename>, then an
bos@559: ssh URL that contains a path component of <filename
bos@559: class="directory">bar</filename> <emphasis>really</emphasis>
bos@559: refers to the directory <filename
bos@559: class="directory">/home/foo/bar</filename>.</para>
bos@559:
bos@584: <para id="x_491">If you want to specify a path relative to another user's
bos@559: home directory, you can use a path that starts with a tilde
bos@559: character followed by the user's name (let's call them
bos@559: <literal>otheruser</literal>), like this.</para>
bos@559: <programlisting>ssh://server/~otheruser/hg/repo</programlisting>
bos@559:
bos@584: <para id="x_492">And if you really want to specify an
bos@559: <emphasis>absolute</emphasis> path on the server, begin the
bos@559: path component with two slashes, as in this example.</para>
bos@559: <programlisting>ssh://server//absolute/path</programlisting>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Finding an ssh client for your system</title>
bos@559:
bos@584: <para id="x_493">Almost every Unix-like system comes with OpenSSH
bos@559: preinstalled. If you're using such a system, run
bos@559: <literal>which ssh</literal> to find out if the
bos@559: <command>ssh</command> command is installed (it's usually in
bos@559: <filename class="directory">/usr/bin</filename>). In the
bos@559: unlikely event that it isn't present, take a look at your
bos@559: system documentation to figure out how to install it.</para>
bos@559:
bos@675: <para id="x_494">On Windows, the TortoiseHg package is bundled
bos@675: with a version of Simon Tatham's excellent
bos@675: <command>plink</command> command, and you should not need to
bos@675: do any further configuration.</para>
bos@675: </sect2>
bos@675:
bos@675: <sect2>
bos@675: <title>Generating a key pair</title>
bos@675:
bos@675: <para id="x_499">To avoid the need to repetitively type a
bos@675: password every time you need to use your ssh client, I
bos@675: recommend generating a key pair.</para>
bos@675:
bos@675: <tip>
bos@675: <title>Key pairs are not mandatory</title>
bos@675:
bos@676: <para id="x_6a4">Mercurial knows nothing about ssh authentication or key
bos@675: pairs. You can, if you like, safely ignore this section and
bos@675: the one that follows until you grow tired of repeatedly
bos@675: typing ssh passwords.</para>
bos@675: </tip>
bos@675:
bos@559: <itemizedlist>
bos@675: <listitem>
bos@676: <para id="x_6a5">On a Unix-like system, the
bos@675: <command>ssh-keygen</command> command will do the
bos@675: trick.</para>
bos@676: <para id="x_6a6">On Windows, if you're using TortoiseHg, you may need
bos@675: to download a command named <command>puttygen</command>
bos@675: from <ulink
bos@675: url="http://www.chiark.greenend.org.uk/~sgtatham/putty">the
bos@675: PuTTY web site</ulink> to generate a key pair. See
bos@675: <ulink
bos@675: url="http://the.earth.li/~sgtatham/putty/0.60/htmldoc/Chapter8.html#pubkey-puttygen">the
bos@675: <command>puttygen</command> documentation</ulink> for
bos@675: details of how use the command.</para>
bos@675: </listitem>
bos@675: </itemizedlist>
bos@559:
bos@584: <para id="x_49a">When you generate a key pair, it's usually
bos@559: <emphasis>highly</emphasis> advisable to protect it with a
bos@559: passphrase. (The only time that you might not want to do this
bos@559: is when you're using the ssh protocol for automated tasks on a
bos@559: secure network.)</para>
bos@559:
bos@584: <para id="x_49b">Simply generating a key pair isn't enough, however.
bos@559: You'll need to add the public key to the set of authorised
bos@559: keys for whatever user you're logging in remotely as. For
bos@559: servers using OpenSSH (the vast majority), this will mean
bos@559: adding the public key to a list in a file called <filename
bos@559: role="special">authorized_keys</filename> in their <filename
bos@559: role="special" class="directory">.ssh</filename>
bos@559: directory.</para>
bos@559:
bos@584: <para id="x_49c">On a Unix-like system, your public key will have a
bos@559: <filename>.pub</filename> extension. If you're using
bos@559: <command>puttygen</command> on Windows, you can save the
bos@559: public key to a file of your choosing, or paste it from the
bos@559: window it's displayed in straight into the <filename
bos@559: role="special">authorized_keys</filename> file.</para>
bos@559: </sect2>
bos@559: <sect2>
bos@559: <title>Using an authentication agent</title>
bos@559:
bos@584: <para id="x_49d">An authentication agent is a daemon that stores
bos@559: passphrases in memory (so it will forget passphrases if you
bos@559: log out and log back in again). An ssh client will notice if
bos@559: it's running, and query it for a passphrase. If there's no
bos@559: authentication agent running, or the agent doesn't store the
bos@559: necessary passphrase, you'll have to type your passphrase
bos@559: every time Mercurial tries to communicate with a server on
bos@559: your behalf (e.g. whenever you pull or push changes).</para>
bos@559:
bos@584: <para id="x_49e">The downside of storing passphrases in an agent is that
bos@559: it's possible for a well-prepared attacker to recover the
bos@559: plain text of your passphrases, in some cases even if your
bos@559: system has been power-cycled. You should make your own
bos@559: judgment as to whether this is an acceptable risk. It
bos@559: certainly saves a lot of repeated typing.</para>
bos@559:
bos@675: <itemizedlist>
bos@675: <listitem>
bos@675: <para id="x_49f">On Unix-like systems, the agent is called
bos@675: <command>ssh-agent</command>, and it's often run
bos@675: automatically for you when you log in. You'll need to use
bos@675: the <command>ssh-add</command> command to add passphrases
bos@675: to the agent's store.</para>
bos@675: </listitem>
bos@675: <listitem>
bos@676: <para id="x_6a7">On Windows, if you're using TortoiseHg, the
bos@675: <command>pageant</command> command acts as the agent. As
bos@675: with <command>puttygen</command>, you'll need to <ulink
bos@675: url="http://www.chiark.greenend.org.uk/%7Esgtatham/putty/download.html">download
bos@675: <command>pageant</command></ulink> from the PuTTY web
bos@675: site and read <ulink
bos@675: url="http://the.earth.li/~sgtatham/putty/0.60/htmldoc/Chapter9.html#pageant">its
bos@675: documentation</ulink>. The <command>pageant</command>
bos@675: command adds an icon to your system tray that will let you
bos@675: manage stored passphrases.</para>
bos@675: </listitem>
bos@675: </itemizedlist>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Configuring the server side properly</title>
bos@559:
bos@584: <para id="x_4a0">Because ssh can be fiddly to set up if you're new to it,
bos@675: a variety of things can go wrong. Add Mercurial
bos@559: on top, and there's plenty more scope for head-scratching.
bos@559: Most of these potential problems occur on the server side, not
bos@559: the client side. The good news is that once you've gotten a
bos@559: configuration working, it will usually continue to work
bos@559: indefinitely.</para>
bos@559:
bos@584: <para id="x_4a1">Before you try using Mercurial to talk to an ssh server,
bos@559: it's best to make sure that you can use the normal
bos@559: <command>ssh</command> or <command>putty</command> command to
bos@559: talk to the server first. If you run into problems with using
bos@559: these commands directly, Mercurial surely won't work. Worse,
bos@559: it will obscure the underlying problem. Any time you want to
bos@559: debug ssh-related Mercurial problems, you should drop back to
bos@559: making sure that plain ssh client commands work first,
bos@559: <emphasis>before</emphasis> you worry about whether there's a
bos@559: problem with Mercurial.</para>
bos@559:
bos@584: <para id="x_4a2">The first thing to be sure of on the server side is that
bos@559: you can actually log in from another machine at all. If you
bos@559: can't use <command>ssh</command> or <command>putty</command>
bos@559: to log in, the error message you get may give you a few hints
bos@559: as to what's wrong. The most common problems are as
bos@559: follows.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4a3">If you get a <quote>connection refused</quote>
bos@559: error, either there isn't an SSH daemon running on the
bos@559: server at all, or it's inaccessible due to firewall
bos@559: configuration.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4a4">If you get a <quote>no route to host</quote>
bos@559: error, you either have an incorrect address for the server
bos@559: or a seriously locked down firewall that won't admit its
bos@559: existence at all.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4a5">If you get a <quote>permission denied</quote>
bos@559: error, you may have mistyped the username on the server,
bos@559: or you could have mistyped your key's passphrase or the
bos@559: remote user's password.</para>
bos@559: </listitem></itemizedlist>
bos@584: <para id="x_4a6">In summary, if you're having trouble talking to the
bos@559: server's ssh daemon, first make sure that one is running at
bos@559: all. On many systems it will be installed, but disabled, by
bos@559: default. Once you're done with this step, you should then
bos@559: check that the server's firewall is configured to allow
bos@559: incoming connections on the port the ssh daemon is listening
bos@559: on (usually 22). Don't worry about more exotic possibilities
bos@559: for misconfiguration until you've checked these two
bos@559: first.</para>
bos@559:
bos@584: <para id="x_4a7">If you're using an authentication agent on the client side
bos@559: to store passphrases for your keys, you ought to be able to
bos@559: log into the server without being prompted for a passphrase or
bos@559: a password. If you're prompted for a passphrase, there are a
bos@559: few possible culprits.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4a8">You might have forgotten to use
bos@559: <command>ssh-add</command> or <command>pageant</command>
bos@559: to store the passphrase.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4a9">You might have stored the passphrase for the
bos@559: wrong key.</para>
bos@559: </listitem></itemizedlist>
bos@584: <para id="x_4aa">If you're being prompted for the remote user's password,
bos@559: there are another few possible problems to check.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4ab">Either the user's home directory or their
bos@559: <filename role="special" class="directory">.ssh</filename>
bos@559: directory might have excessively liberal permissions. As
bos@559: a result, the ssh daemon will not trust or read their
bos@559: <filename role="special">authorized_keys</filename> file.
bos@559: For example, a group-writable home or <filename
bos@559: role="special" class="directory">.ssh</filename>
bos@559: directory will often cause this symptom.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4ac">The user's <filename
bos@559: role="special">authorized_keys</filename> file may have
bos@559: a problem. If anyone other than the user owns or can write
bos@559: to that file, the ssh daemon will not trust or read
bos@559: it.</para>
bos@559: </listitem></itemizedlist>
bos@559:
bos@584: <para id="x_4ad">In the ideal world, you should be able to run the
bos@559: following command successfully, and it should print exactly
bos@559: one line of output, the current date and time.</para>
bos@559: <programlisting>ssh myserver date</programlisting>
bos@559:
bos@584: <para id="x_4ae">If, on your server, you have login scripts that print
bos@559: banners or other junk even when running non-interactive
bos@559: commands like this, you should fix them before you continue,
bos@559: so that they only print output if they're run interactively.
bos@559: Otherwise these banners will at least clutter up Mercurial's
bos@559: output. Worse, they could potentially cause problems with
bos@701: running Mercurial commands remotely. Mercurial tries to
bos@559: detect and ignore banners in non-interactive
bos@559: <command>ssh</command> sessions, but it is not foolproof. (If
bos@559: you're editing your login scripts on your server, the usual
bos@559: way to see if a login script is running in an interactive
bos@559: shell is to check the return code from the command
bos@559: <literal>tty -s</literal>.)</para>
bos@559:
bos@584: <para id="x_4af">Once you've verified that plain old ssh is working with
bos@559: your server, the next step is to ensure that Mercurial runs on
bos@559: the server. The following command should run
bos@559: successfully:</para>
bos@580:
bos@559: <programlisting>ssh myserver hg version</programlisting>
bos@580:
bos@584: <para id="x_4b0">If you see an error message instead of normal <command
bos@559: role="hg-cmd">hg version</command> output, this is usually
bos@559: because you haven't installed Mercurial to <filename
bos@559: class="directory">/usr/bin</filename>. Don't worry if this
bos@559: is the case; you don't need to do that. But you should check
bos@559: for a few possible problems.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4b1">Is Mercurial really installed on the server at
bos@559: all? I know this sounds trivial, but it's worth
bos@559: checking!</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4b2">Maybe your shell's search path (usually set
bos@559: via the <envar>PATH</envar> environment variable) is
bos@559: simply misconfigured.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4b3">Perhaps your <envar>PATH</envar> environment
bos@559: variable is only being set to point to the location of the
bos@559: <command>hg</command> executable if the login session is
bos@559: interactive. This can happen if you're setting the path
bos@559: in the wrong shell login script. See your shell's
bos@559: documentation for details.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4b4">The <envar>PYTHONPATH</envar> environment
bos@559: variable may need to contain the path to the Mercurial
bos@559: Python modules. It might not be set at all; it could be
bos@559: incorrect; or it may be set only if the login is
bos@559: interactive.</para>
bos@559: </listitem></itemizedlist>
bos@559:
bos@584: <para id="x_4b5">If you can run <command role="hg-cmd">hg version</command>
bos@559: over an ssh connection, well done! You've got the server and
bos@559: client sorted out. You should now be able to use Mercurial to
bos@559: access repositories hosted by that username on that server.
bos@559: If you run into problems with Mercurial and ssh at this point,
bos@559: try using the <option role="hg-opt-global">--debug</option>
bos@559: option to get a clearer picture of what's going on.</para>
bos@559: </sect2>
bos@559: <sect2>
bos@559: <title>Using compression with ssh</title>
bos@559:
bos@584: <para id="x_4b6">Mercurial does not compress data when it uses the ssh
bos@559: protocol, because the ssh protocol can transparently compress
bos@672: data. However, the default behavior of ssh clients is
bos@559: <emphasis>not</emphasis> to request compression.</para>
bos@559:
bos@584: <para id="x_4b7">Over any network other than a fast LAN (even a wireless
bos@559: network), using compression is likely to significantly speed
bos@559: up Mercurial's network operations. For example, over a WAN,
bos@559: someone measured compression as reducing the amount of time
bos@559: required to clone a particularly large repository from 51
bos@559: minutes to 17 minutes.</para>
bos@559:
bos@584: <para id="x_4b8">Both <command>ssh</command> and <command>plink</command>
bos@559: accept a <option role="cmd-opt-ssh">-C</option> option which
bos@559: turns on compression. You can easily edit your <filename
bos@580: role="special">~/.hgrc</filename> to enable compression for
bos@675: all of Mercurial's uses of the ssh protocol. Here is how to
bos@675: do so for regular <command>ssh</command> on Unix-like systems,
bos@675: for example.</para>
bos@579: <programlisting>[ui]
bos@579: ssh = ssh -C</programlisting>
bos@559:
bos@675: <para id="x_4b9">If you use <command>ssh</command> on a
bos@675: Unix-like system, you can configure it to always use
bos@675: compression when talking to your server. To do this, edit
bos@675: your <filename role="special">.ssh/config</filename> file
bos@675: (which may not yet exist), as follows.</para>
bos@675:
bos@579: <programlisting>Host hg
bos@579: Compression yes
bos@579: HostName hg.example.com</programlisting>
bos@675:
bos@675: <para id="x_4ba">This defines a hostname alias,
bos@675: <literal>hg</literal>. When you use that hostname on the
bos@675: <command>ssh</command> command line or in a Mercurial
bos@675: <literal>ssh</literal>-protocol URL, it will cause
bos@559: <command>ssh</command> to connect to
bos@559: <literal>hg.example.com</literal> and use compression. This
bos@559: gives you both a shorter name to type and compression, each of
bos@559: which is a good thing in its own right.</para>
bos@559: </sect2>
bos@559: </sect1>
bos@675:
bos@559: <sect1 id="sec:collab:cgi">
bos@559: <title>Serving over HTTP using CGI</title>
bos@559:
bos@676: <para id="x_6a8">The simplest way to host one or more repositories in a
bos@675: permanent way is to use a web server and Mercurial's CGI
bos@675: support.</para>
bos@675:
bos@584: <para id="x_4bb">Depending on how ambitious you are, configuring Mercurial's
bos@559: CGI interface can take anything from a few moments to several
bos@559: hours.</para>
bos@559:
bos@584: <para id="x_4bc">We'll begin with the simplest of examples, and work our way
bos@559: towards a more complex configuration. Even for the most basic
bos@559: case, you're almost certainly going to need to read and modify
bos@559: your web server's configuration.</para>
bos@559:
bos@559: <note>
bos@675: <title>High pain tolerance required</title>
bos@675:
bos@675: <para id="x_4bd">Configuring a web server is a complex, fiddly,
bos@675: and highly system-dependent activity. I can't possibly give
bos@675: you instructions that will cover anything like all of the
bos@675: cases you will encounter. Please use your discretion and
bos@675: judgment in following the sections below. Be prepared to make
bos@675: plenty of mistakes, and to spend a lot of time reading your
bos@675: server's error logs.</para>
bos@675:
bos@676: <para id="x_6a9">If you don't have a strong stomach for tweaking
bos@675: configurations over and over, or a compelling need to host
bos@675: your own services, you might want to try one of the public
bos@675: hosting services that I mentioned earlier.</para>
bos@559: </note>
bos@559:
bos@559: <sect2>
bos@559: <title>Web server configuration checklist</title>
bos@559:
bos@584: <para id="x_4be">Before you continue, do take a few moments to check a few
bos@559: aspects of your system's setup.</para>
bos@559:
bos@559: <orderedlist>
bos@675: <listitem><para id="x_4bf">Do you have a web server installed
bos@675: at all? Mac OS X and some Linux distributions ship with
bos@675: Apache, but many other systems may not have a web server
bos@675: installed.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4c0">If you have a web server installed, is it
bos@559: actually running? On most systems, even if one is
bos@559: present, it will be disabled by default.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4c1">Is your server configured to allow you to run
bos@559: CGI programs in the directory where you plan to do so?
bos@559: Most servers default to explicitly disabling the ability
bos@559: to run CGI programs.</para>
bos@559: </listitem></orderedlist>
bos@559:
bos@584: <para id="x_4c2">If you don't have a web server installed, and don't have
bos@559: substantial experience configuring Apache, you should consider
bos@559: using the <literal>lighttpd</literal> web server instead of
bos@559: Apache. Apache has a well-deserved reputation for baroque and
bos@559: confusing configuration. While <literal>lighttpd</literal> is
bos@559: less capable in some ways than Apache, most of these
bos@559: capabilities are not relevant to serving Mercurial
bos@559: repositories. And <literal>lighttpd</literal> is undeniably
bos@559: <emphasis>much</emphasis> easier to get started with than
bos@559: Apache.</para>
bos@675: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Basic CGI configuration</title>
bos@559:
bos@584: <para id="x_4c3">On Unix-like systems, it's common for users to have a
bos@559: subdirectory named something like <filename
bos@559: class="directory">public_html</filename> in their home
bos@559: directory, from which they can serve up web pages. A file
bos@559: named <filename>foo</filename> in this directory will be
bos@559: accessible at a URL of the form
bos@580: <literal>http://www.example.com/username/foo</literal>.</para>
bos@559:
bos@584: <para id="x_4c4">To get started, find the <filename
bos@559: role="special">hgweb.cgi</filename> script that should be
bos@559: present in your Mercurial installation. If you can't quickly
bos@559: find a local copy on your system, simply download one from the
bos@559: master Mercurial repository at <ulink
bos@559: url="http://www.selenic.com/repo/hg/raw-file/tip/hgweb.cgi">http://www.selenic.com/repo/hg/raw-file/tip/hgweb.cgi</ulink>.</para>
bos@559:
bos@584: <para id="x_4c5">You'll need to copy this script into your <filename
bos@559: class="directory">public_html</filename> directory, and
bos@559: ensure that it's executable.</para>
bos@579: <programlisting>cp .../hgweb.cgi ~/public_html
bos@579: chmod 755 ~/public_html/hgweb.cgi</programlisting>
bos@584: <para id="x_4c6">The <literal>755</literal> argument to
bos@559: <command>chmod</command> is a little more general than just
bos@559: making the script executable: it ensures that the script is
bos@559: executable by anyone, and that <quote>group</quote> and
bos@559: <quote>other</quote> write permissions are
bos@559: <emphasis>not</emphasis> set. If you were to leave those
bos@559: write permissions enabled, Apache's <literal>suexec</literal>
bos@559: subsystem would likely refuse to execute the script. In fact,
bos@559: <literal>suexec</literal> also insists that the
bos@559: <emphasis>directory</emphasis> in which the script resides
bos@559: must not be writable by others.</para>
bos@559: <programlisting>chmod 755 ~/public_html</programlisting>
bos@559:
bos@559: <sect3 id="sec:collab:wtf">
bos@559: <title>What could <emphasis>possibly</emphasis> go
bos@559: wrong?</title>
bos@559:
hg@677: <para id="x_4c7">Once you've copied the CGI script into place,
bos@679: go into a web browser, and try to open the URL
bos@679: <literal>http://myhostname/~myuser/hgweb.cgi</literal>,
bos@679: <emphasis>but</emphasis> brace yourself for instant failure.
bos@679: There's a high probability that trying to visit this URL
bos@679: will fail, and there are many possible reasons for this. In
bos@679: fact, you're likely to stumble over almost every one of the
bos@679: possible errors below, so please read carefully. The
bos@679: following are all of the problems I ran into on a system
bos@679: running Fedora 7, with a fresh installation of Apache, and a
bos@679: user account that I created specially to perform this
bos@679: exercise.</para>
bos@559:
bos@584: <para id="x_4c8">Your web server may have per-user directories disabled.
bos@559: If you're using Apache, search your config file for a
bos@559: <literal>UserDir</literal> directive. If there's none
bos@559: present, per-user directories will be disabled. If one
bos@559: exists, but its value is <literal>disabled</literal>, then
bos@559: per-user directories will be disabled. Otherwise, the
bos@559: string after <literal>UserDir</literal> gives the name of
bos@559: the subdirectory that Apache will look in under your home
bos@559: directory, for example <filename
bos@559: class="directory">public_html</filename>.</para>
bos@559:
bos@584: <para id="x_4c9">Your file access permissions may be too restrictive.
bos@559: The web server must be able to traverse your home directory
bos@559: and directories under your <filename
bos@559: class="directory">public_html</filename> directory, and
bos@559: read files under the latter too. Here's a quick recipe to
bos@559: help you to make your permissions more appropriate.</para>
bos@579: <programlisting>chmod 755 ~
bos@579: find ~/public_html -type d -print0 | xargs -0r chmod 755
bos@579: find ~/public_html -type f -print0 | xargs -0r chmod 644</programlisting>
bos@559:
bos@584: <para id="x_4ca">The other possibility with permissions is that you might
bos@559: get a completely empty window when you try to load the
bos@559: script. In this case, it's likely that your access
ori@561: permissions are <emphasis>too permissive</emphasis>. Apache's
bos@559: <literal>suexec</literal> subsystem won't execute a script
bos@559: that's group- or world-writable, for example.</para>
bos@559:
bos@584: <para id="x_4cb">Your web server may be configured to disallow execution
bos@559: of CGI programs in your per-user web directory. Here's
bos@559: Apache's default per-user configuration from my Fedora
bos@559: system.</para>
bos@579:
bos@579: &ch06-apache-config.lst;
bos@579:
bos@584: <para id="x_4cc">If you find a similar-looking
bos@559: <literal>Directory</literal> group in your Apache
bos@559: configuration, the directive to look at inside it is
bos@559: <literal>Options</literal>. Add <literal>ExecCGI</literal>
bos@559: to the end of this list if it's missing, and restart the web
bos@559: server.</para>
bos@559:
bos@584: <para id="x_4cd">If you find that Apache serves you the text of the CGI
bos@559: script instead of executing it, you may need to either
bos@559: uncomment (if already present) or add a directive like
bos@559: this.</para>
bos@559: <programlisting>AddHandler cgi-script .cgi</programlisting>
bos@559:
bos@584: <para id="x_4ce">The next possibility is that you might be served with a
bos@559: colourful Python backtrace claiming that it can't import a
bos@559: <literal>mercurial</literal>-related module. This is
bos@559: actually progress! The server is now capable of executing
bos@559: your CGI script. This error is only likely to occur if
bos@559: you're running a private installation of Mercurial, instead
bos@559: of a system-wide version. Remember that the web server runs
bos@559: the CGI program without any of the environment variables
bos@559: that you take for granted in an interactive session. If
bos@559: this error happens to you, edit your copy of <filename
bos@559: role="special">hgweb.cgi</filename> and follow the
bos@559: directions inside it to correctly set your
bos@559: <envar>PYTHONPATH</envar> environment variable.</para>
bos@559:
bos@701: <para id="x_4cf">Finally, you are <emphasis>certain</emphasis> to be
bos@559: served with another colourful Python backtrace: this one
bos@559: will complain that it can't find <filename
bos@559: class="directory">/path/to/repository</filename>. Edit
bos@559: your <filename role="special">hgweb.cgi</filename> script
bos@559: and replace the <filename
bos@559: class="directory">/path/to/repository</filename> string
bos@559: with the complete path to the repository you want to serve
bos@559: up.</para>
bos@559:
bos@584: <para id="x_4d0">At this point, when you try to reload the page, you
bos@559: should be presented with a nice HTML view of your
bos@559: repository's history. Whew!</para>
bos@559: </sect3>
bos@675:
bos@559: <sect3>
bos@559: <title>Configuring lighttpd</title>
bos@559:
bos@584: <para id="x_4d1">To be exhaustive in my experiments, I tried configuring
bos@559: the increasingly popular <literal>lighttpd</literal> web
bos@559: server to serve the same repository as I described with
bos@559: Apache above. I had already overcome all of the problems I
bos@559: outlined with Apache, many of which are not server-specific.
bos@559: As a result, I was fairly sure that my file and directory
bos@559: permissions were good, and that my <filename
bos@559: role="special">hgweb.cgi</filename> script was properly
bos@559: edited.</para>
bos@559:
bos@584: <para id="x_4d2">Once I had Apache running, getting
bos@559: <literal>lighttpd</literal> to serve the repository was a
bos@559: snap (in other words, even if you're trying to use
bos@559: <literal>lighttpd</literal>, you should read the Apache
bos@559: section). I first had to edit the
bos@559: <literal>mod_access</literal> section of its config file to
bos@559: enable <literal>mod_cgi</literal> and
bos@559: <literal>mod_userdir</literal>, both of which were disabled
bos@559: by default on my system. I then added a few lines to the
bos@559: end of the config file, to configure these modules.</para>
bos@580: <programlisting>userdir.path = "public_html"
bos@580: cgi.assign = (".cgi" => "" )</programlisting>
bos@584: <para id="x_4d3">With this done, <literal>lighttpd</literal> ran
bos@559: immediately for me. If I had configured
bos@559: <literal>lighttpd</literal> before Apache, I'd almost
bos@559: certainly have run into many of the same system-level
bos@559: configuration problems as I did with Apache. However, I
bos@559: found <literal>lighttpd</literal> to be noticeably easier to
bos@559: configure than Apache, even though I've used Apache for over
bos@559: a decade, and this was my first exposure to
bos@559: <literal>lighttpd</literal>.</para>
bos@559: </sect3>
bos@559: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Sharing multiple repositories with one CGI script</title>
bos@559:
bos@584: <para id="x_4d4">The <filename role="special">hgweb.cgi</filename> script
bos@559: only lets you publish a single repository, which is an
bos@559: annoying restriction. If you want to publish more than one
bos@559: without wracking yourself with multiple copies of the same
bos@559: script, each with different names, a better choice is to use
bos@559: the <filename role="special">hgwebdir.cgi</filename>
bos@559: script.</para>
bos@559:
bos@584: <para id="x_4d5">The procedure to configure <filename
bos@559: role="special">hgwebdir.cgi</filename> is only a little more
bos@559: involved than for <filename
bos@559: role="special">hgweb.cgi</filename>. First, you must obtain
bos@559: a copy of the script. If you don't have one handy, you can
bos@559: download a copy from the master Mercurial repository at <ulink
bos@559: url="http://www.selenic.com/repo/hg/raw-file/tip/hgwebdir.cgi">http://www.selenic.com/repo/hg/raw-file/tip/hgwebdir.cgi</ulink>.</para>
bos@559:
bos@584: <para id="x_4d6">You'll need to copy this script into your <filename
bos@559: class="directory">public_html</filename> directory, and
bos@559: ensure that it's executable.</para>
bos@592:
bos@580: <programlisting>cp .../hgwebdir.cgi ~/public_html
bos@580: chmod 755 ~/public_html ~/public_html/hgwebdir.cgi</programlisting>
bos@592:
bos@592: <para id="x_4d7">With basic configuration out of the way, try to
bos@679: visit <literal>http://myhostname/~myuser/hgwebdir.cgi</literal>
hg@677: in your browser. It should
bos@559: display an empty list of repositories. If you get a blank
bos@559: window or error message, try walking through the list of
bos@592: potential problems in <xref
bos@559: linkend="sec:collab:wtf"/>.</para>
bos@559:
bos@584: <para id="x_4d8">The <filename role="special">hgwebdir.cgi</filename>
bos@559: script relies on an external configuration file. By default,
bos@559: it searches for a file named <filename
bos@559: role="special">hgweb.config</filename> in the same directory
bos@559: as itself. You'll need to create this file, and make it
bos@559: world-readable. The format of the file is similar to a
bos@559: Windows <quote>ini</quote> file, as understood by Python's
bos@559: <literal>ConfigParser</literal>
bos@559: <citation>web:configparser</citation> module.</para>
bos@559:
bos@584: <para id="x_4d9">The easiest way to configure <filename
bos@559: role="special">hgwebdir.cgi</filename> is with a section
bos@559: named <literal>collections</literal>. This will automatically
bos@559: publish <emphasis>every</emphasis> repository under the
bos@559: directories you name. The section should look like
bos@559: this:</para>
bos@580: <programlisting>[collections]
bos@580: /my/root = /my/root</programlisting>
bos@584: <para id="x_4da">Mercurial interprets this by looking at the directory name
bos@559: on the <emphasis>right</emphasis> hand side of the
bos@559: <quote><literal>=</literal></quote> sign; finding repositories
bos@559: in that directory hierarchy; and using the text on the
bos@559: <emphasis>left</emphasis> to strip off matching text from the
bos@559: names it will actually list in the web interface. The
bos@559: remaining component of a path after this stripping has
bos@559: occurred is called a <quote>virtual path</quote>.</para>
bos@559:
bos@679: <para id="x_4db">Given the example above, if we have a
bos@679: repository whose local path is <filename
bos@559: class="directory">/my/root/this/repo</filename>, the CGI
bos@559: script will strip the leading <filename
bos@559: class="directory">/my/root</filename> from the name, and
bos@559: publish the repository with a virtual path of <filename
bos@559: class="directory">this/repo</filename>. If the base URL for
bos@679: our CGI script is
bos@679: <literal>http://myhostname/~myuser/hgwebdir.cgi</literal>, the
bos@679: complete URL for that repository will be
bos@679: <literal>http://myhostname/~myuser/hgwebdir.cgi/this/repo</literal>.</para>
bos@559:
bos@584: <para id="x_4dc">If we replace <filename
bos@559: class="directory">/my/root</filename> on the left hand side
bos@559: of this example with <filename
bos@559: class="directory">/my</filename>, then <filename
bos@559: role="special">hgwebdir.cgi</filename> will only strip off
bos@559: <filename class="directory">/my</filename> from the repository
bos@559: name, and will give us a virtual path of <filename
bos@559: class="directory">root/this/repo</filename> instead of
bos@559: <filename class="directory">this/repo</filename>.</para>
bos@559:
bos@584: <para id="x_4dd">The <filename role="special">hgwebdir.cgi</filename>
bos@559: script will recursively search each directory listed in the
bos@559: <literal>collections</literal> section of its configuration
bos@559: file, but it will <literal>not</literal> recurse into the
bos@559: repositories it finds.</para>
bos@559:
bos@584: <para id="x_4de">The <literal>collections</literal> mechanism makes it easy
bos@559: to publish many repositories in a <quote>fire and
bos@559: forget</quote> manner. You only need to set up the CGI
bos@559: script and configuration file one time. Afterwards, you can
bos@559: publish or unpublish a repository at any time by simply moving
bos@559: it into, or out of, the directory hierarchy in which you've
bos@559: configured <filename role="special">hgwebdir.cgi</filename> to
bos@559: look.</para>
bos@559:
bos@559: <sect3>
bos@559: <title>Explicitly specifying which repositories to
bos@559: publish</title>
bos@559:
bos@584: <para id="x_4df">In addition to the <literal>collections</literal>
bos@559: mechanism, the <filename
bos@559: role="special">hgwebdir.cgi</filename> script allows you
bos@559: to publish a specific list of repositories. To do so,
bos@559: create a <literal>paths</literal> section, with contents of
bos@559: the following form.</para>
bos@580: <programlisting>[paths]
bos@580: repo1 = /my/path/to/some/repo
bos@580: repo2 = /some/path/to/another</programlisting>
bos@584: <para id="x_4e0">In this case, the virtual path (the component that will
bos@559: appear in a URL) is on the left hand side of each
bos@559: definition, while the path to the repository is on the
bos@559: right. Notice that there does not need to be any
bos@559: relationship between the virtual path you choose and the
bos@559: location of a repository in your filesystem.</para>
bos@559:
bos@584: <para id="x_4e1">If you wish, you can use both the
bos@559: <literal>collections</literal> and <literal>paths</literal>
bos@559: mechanisms simultaneously in a single configuration
bos@559: file.</para>
bos@559:
bos@559: <note>
bos@675: <title>Beware duplicate virtual paths</title>
bos@675:
bos@675: <para id="x_4e2"> If several repositories have the same
bos@675: virtual path, <filename
bos@675: role="special">hgwebdir.cgi</filename> will not report
bos@675: an error. Instead, it will behave unpredictably.</para>
bos@559: </note>
bos@559: </sect3>
bos@559: </sect2>
bos@675:
bos@559: <sect2>
bos@559: <title>Downloading source archives</title>
bos@559:
bos@584: <para id="x_4e3">Mercurial's web interface lets users download an archive
bos@559: of any revision. This archive will contain a snapshot of the
bos@559: working directory as of that revision, but it will not contain
bos@559: a copy of the repository data.</para>
bos@559:
bos@584: <para id="x_4e4">By default, this feature is not enabled. To enable it,
bos@559: you'll need to add an <envar
bos@559: role="rc-item-web">allow_archive</envar> item to the
bos@559: <literal role="rc-web">web</literal> section of your <filename
bos@675: role="special">~/.hgrc</filename>; see below for details.</para>
bos@559: </sect2>
bos@559: <sect2>
bos@559: <title>Web configuration options</title>
bos@559:
bos@584: <para id="x_4e5">Mercurial's web interfaces (the <command role="hg-cmd">hg
bos@559: serve</command> command, and the <filename
bos@559: role="special">hgweb.cgi</filename> and <filename
bos@559: role="special">hgwebdir.cgi</filename> scripts) have a
bos@559: number of configuration options that you can set. These
bos@559: belong in a section named <literal
bos@559: role="rc-web">web</literal>.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4e6"><envar
bos@559: role="rc-item-web">allow_archive</envar>: Determines
bos@559: which (if any) archive download mechanisms Mercurial
bos@559: supports. If you enable this feature, users of the web
bos@559: interface will be able to download an archive of whatever
bos@559: revision of a repository they are viewing. To enable the
bos@559: archive feature, this item must take the form of a
bos@559: sequence of words drawn from the list below.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4e7"><literal>bz2</literal>: A
bos@559: <command>tar</command> archive, compressed using
bos@559: <literal>bzip2</literal> compression. This has the
bos@559: best compression ratio, but uses the most CPU time on
bos@559: the server.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4e8"><literal>gz</literal>: A
bos@559: <command>tar</command> archive, compressed using
bos@559: <literal>gzip</literal> compression.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4e9"><literal>zip</literal>: A
bos@559: <command>zip</command> archive, compressed using LZW
bos@559: compression. This format has the worst compression
bos@559: ratio, but is widely used in the Windows world.</para>
bos@559: </listitem>
bos@559: </itemizedlist>
bos@584: <para id="x_4ea"> If you provide an empty list, or don't have an
bos@559: <envar role="rc-item-web">allow_archive</envar> entry at
bos@559: all, this feature will be disabled. Here is an example of
bos@559: how to enable all three supported formats.</para>
bos@580: <programlisting>[web]
bos@580: allow_archive = bz2 gz zip</programlisting>
bos@559: </listitem>
bos@584: <listitem><para id="x_4eb"><envar role="rc-item-web">allowpull</envar>:
bos@559: Boolean. Determines whether the web interface allows
bos@559: remote users to <command role="hg-cmd">hg pull</command>
bos@559: and <command role="hg-cmd">hg clone</command> this
bos@559: repository over HTTP. If set to <literal>no</literal> or
bos@559: <literal>false</literal>, only the
bos@559: <quote>human-oriented</quote> portion of the web interface
bos@559: is available.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4ec"><envar role="rc-item-web">contact</envar>:
bos@559: String. A free-form (but preferably brief) string
bos@559: identifying the person or group in charge of the
bos@559: repository. This often contains the name and email
bos@559: address of a person or mailing list. It often makes sense
bos@559: to place this entry in a repository's own <filename
bos@559: role="special">.hg/hgrc</filename> file, but it can make
bos@580: sense to use in a global <filename
bos@580: role="special">~/.hgrc</filename> if every repository
bos@580: has a single maintainer.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4ed"><envar role="rc-item-web">maxchanges</envar>:
bos@559: Integer. The default maximum number of changesets to
bos@559: display in a single page of output.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4ee"><envar role="rc-item-web">maxfiles</envar>:
bos@559: Integer. The default maximum number of modified files to
bos@559: display in a single page of output.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4ef"><envar role="rc-item-web">stripes</envar>:
bos@559: Integer. If the web interface displays alternating
bos@559: <quote>stripes</quote> to make it easier to visually align
bos@559: rows when you are looking at a table, this number controls
bos@559: the number of rows in each stripe.</para>
bos@559: </listitem>
bos@592: <listitem><para id="x_4f0"><envar
bos@592: role="rc-item-web">style</envar>: Controls the template
bos@592: Mercurial uses to display the web interface. Mercurial
bos@675: ships with several web templates.</para>
bos@675: <itemizedlist>
bos@675: <listitem>
bos@676: <para id="x_6aa"><literal>coal</literal> is monochromatic.</para>
bos@675: </listitem>
bos@675: <listitem>
bos@676: <para id="x_6ab"><literal>gitweb</literal> emulates the visual
bos@675: style of git's web interface.</para>
bos@675: </listitem>
bos@675: <listitem>
bos@676: <para id="x_6ac"><literal>monoblue</literal> uses solid blues and
bos@675: greys.</para>
bos@675: </listitem>
bos@675: <listitem>
bos@676: <para id="x_6ad"><literal>paper</literal> is the default.</para>
bos@675: </listitem>
bos@675: <listitem>
bos@676: <para id="x_6ae"><literal>spartan</literal> was the default for a
bos@675: long time.</para>
bos@675: </listitem>
bos@675: </itemizedlist>
bos@676: <para id="x_6af">You can
bos@592: also specify a custom template of your own; see
bos@592: <xref linkend="chap:template"/> for details. Here, you can
bos@592: see how to enable the <literal>gitweb</literal>
bos@592: style.</para>
bos@580: <programlisting>[web]
bos@580: style = gitweb</programlisting>
bos@559: </listitem>
bos@584: <listitem><para id="x_4f1"><envar role="rc-item-web">templates</envar>:
bos@559: Path. The directory in which to search for template
bos@559: files. By default, Mercurial searches in the directory in
bos@559: which it was installed.</para>
bos@559: </listitem></itemizedlist>
bos@584: <para id="x_4f2">If you are using <filename
bos@559: role="special">hgwebdir.cgi</filename>, you can place a few
bos@559: configuration items in a <literal role="rc-web">web</literal>
bos@559: section of the <filename
bos@559: role="special">hgweb.config</filename> file instead of a
bos@580: <filename role="special">~/.hgrc</filename> file, for
bos@559: convenience. These items are <envar
bos@559: role="rc-item-web">motd</envar> and <envar
bos@559: role="rc-item-web">style</envar>.</para>
bos@559:
bos@559: <sect3>
bos@559: <title>Options specific to an individual repository</title>
bos@559:
bos@584: <para id="x_4f3">A few <literal role="rc-web">web</literal> configuration
bos@559: items ought to be placed in a repository's local <filename
bos@559: role="special">.hg/hgrc</filename>, rather than a user's
bos@580: or global <filename role="special">~/.hgrc</filename>.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4f4"><envar
bos@559: role="rc-item-web">description</envar>: String. A
bos@559: free-form (but preferably brief) string that describes
bos@559: the contents or purpose of the repository.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4f5"><envar role="rc-item-web">name</envar>:
bos@559: String. The name to use for the repository in the web
bos@559: interface. This overrides the default name, which is
bos@559: the last component of the repository's path.</para>
bos@559: </listitem></itemizedlist>
bos@559: </sect3>
bos@675:
bos@559: <sect3>
bos@559: <title>Options specific to the <command role="hg-cmd">hg
bos@559: serve</command> command</title>
bos@559:
bos@584: <para id="x_4f6">Some of the items in the <literal
bos@559: role="rc-web">web</literal> section of a <filename
bos@580: role="special">~/.hgrc</filename> file are only for use
bos@559: with the <command role="hg-cmd">hg serve</command>
bos@559: command.</para>
bos@559: <itemizedlist>
bos@584: <listitem><para id="x_4f7"><envar role="rc-item-web">accesslog</envar>:
bos@559: Path. The name of a file into which to write an access
bos@559: log. By default, the <command role="hg-cmd">hg
bos@559: serve</command> command writes this information to
bos@559: standard output, not to a file. Log entries are written
bos@559: in the standard <quote>combined</quote> file format used
bos@559: by almost all web servers.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4f8"><envar role="rc-item-web">address</envar>:
bos@559: String. The local address on which the server should
bos@559: listen for incoming connections. By default, the server
bos@559: listens on all addresses.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4f9"><envar role="rc-item-web">errorlog</envar>:
bos@559: Path. The name of a file into which to write an error
bos@559: log. By default, the <command role="hg-cmd">hg
bos@559: serve</command> command writes this information to
bos@559: standard error, not to a file.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4fa"><envar role="rc-item-web">ipv6</envar>:
bos@559: Boolean. Whether to use the IPv6 protocol. By default,
bos@559: IPv6 is not used.</para>
bos@559: </listitem>
bos@584: <listitem><para id="x_4fb"><envar role="rc-item-web">port</envar>:
bos@559: Integer. The TCP port number on which the server should
bos@559: listen. The default port number used is 8000.</para>
bos@559: </listitem></itemizedlist>
bos@559: </sect3>
bos@675:
bos@559: <sect3>
bos@580: <title>Choosing the right <filename
bos@580: role="special">~/.hgrc</filename> file to add <literal
bos@559: role="rc-web">web</literal> items to</title>
bos@559:
bos@584: <para id="x_4fc">It is important to remember that a web server like
bos@559: Apache or <literal>lighttpd</literal> will run under a user
bos@559: ID that is different to yours. CGI scripts run by your
bos@559: server, such as <filename
bos@559: role="special">hgweb.cgi</filename>, will usually also run
bos@559: under that user ID.</para>
bos@559:
bos@584: <para id="x_4fd">If you add <literal role="rc-web">web</literal> items to
bos@580: your own personal <filename role="special">~/.hgrc</filename> file, CGI scripts won't read that
bos@580: <filename role="special">~/.hgrc</filename> file. Those
bos@672: settings will thus only affect the behavior of the <command
bos@559: role="hg-cmd">hg serve</command> command when you run it.
bos@559: To cause CGI scripts to see your settings, either create a
bos@580: <filename role="special">~/.hgrc</filename> file in the
bos@559: home directory of the user ID that runs your web server, or
bos@559: add those settings to a system-wide <filename
bos@675: role="special">hgrc</filename> file.</para>
bos@559: </sect3>
bos@559: </sect2>
bos@559: </sect1>
bos@675:
bos@675: <sect1>
bos@675: <title>System-wide configuration</title>
bos@675:
bos@676: <para id="x_6b0">On Unix-like systems shared by multiple users (such as a
bos@675: server to which people publish changes), it often makes sense to
bos@675: set up some global default behaviors, such as what theme to use
bos@675: in web interfaces.</para>
bos@675:
bos@676: <para id="x_6b1">If a file named <filename>/etc/mercurial/hgrc</filename>
bos@675: exists, Mercurial will read it at startup time and apply any
bos@675: configuration settings it finds in that file. It will also look
bos@675: for files ending in a <literal>.rc</literal> extension in a
bos@675: directory named <filename>/etc/mercurial/hgrc.d</filename>, and
bos@675: apply any configuration settings it finds in each of those
bos@675: files.</para>
bos@675:
bos@675: <sect2>
bos@675: <title>Making Mercurial more trusting</title>
bos@675:
bos@676: <para id="x_6b2">One situation in which a global <filename>hgrc</filename>
bos@675: can be useful is if users are pulling changes owned by other
bos@675: users. By default, Mercurial will not trust most of the
bos@675: configuration items in a <filename>.hg/hgrc</filename> file
bos@675: inside a repository that is owned by a different user. If we
bos@675: clone or pull changes from such a repository, Mercurial will
bos@675: print a warning stating that it does not trust their
bos@675: <filename>.hg/hgrc</filename>.</para>
bos@675:
bos@676: <para id="x_6b3">If everyone in a particular Unix group is on the same team
bos@675: and <emphasis>should</emphasis> trust each other's
bos@675: configuration settings, or we want to trust particular users,
bos@675: we can override Mercurial's skeptical defaults by creating a
bos@675: system-wide <filename>hgrc</filename> file such as the
bos@675: following:</para>
bos@675:
bos@675: <programlisting># Save this as e.g. /etc/mercurial/hgrc.d/trust.rc
bos@675: [trusted]
bos@675: # Trust all entries in any hgrc file owned by the "editors" or
bos@675: # "www-data" groups.
bos@675: groups = editors, www-data
bos@675:
bos@675: # Trust entries in hgrc files owned by the following users.
bos@675: users = apache, bobo
bos@675: </programlisting>
bos@675: </sect2>
bos@675: </sect1>
bos@559: </chapter>
bos@559:
bos@559: <!--
bos@559: local variables:
bos@559: sgml-parent-document: ("00book.xml" "book" "chapter")
bos@559: end:
bos@559: -->