Mercurial > hgbook

view en/appA-svn.xml @ 825:d7d09cda83d2

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add paragraph IDs
author Bryan O'Sullivan <bos@serpentine.com>
date Sun, 03 May 2009 19:23:31 -0700
parents fd2e83ffb165
children
line wrap: on
line source

<!-- vim: set filetype=docbkxml shiftwidth=2 autoindent expandtab tw=77 : -->

<appendix id="svn">
  <?dbhtml filename="migrating-to-mercurial.html"?>
<title>Migrating to Mercurial</title>

  <para id="x_6e1">A common way to test the waters with a new revision control
    tool is to experiment with switching an existing project, rather
    than starting a new project from scratch.</para>

  <para id="x_6e2">In this appendix, we discuss how to import a project's history
    into Mercurial, and what to look out for if you are used to a
    different revision control system.</para>

  <sect1>
    <title>Importing history from another system</title>

    <para id="x_6e3">Mercurial ships with an extension named
      <literal>convert</literal>, which can import project history
      from most popular revision control systems.  At the time this
      book was written, it could import history from the following
      systems:</para>
    <itemizedlist>
      <listitem>
	<para id="x_6e4">Subversion</para>
      </listitem>
      <listitem>
	<para id="x_6e5">CVS</para>
      </listitem>
      <listitem>
	<para id="x_6e6">git</para>
      </listitem>
      <listitem>
	<para id="x_6e7">Darcs</para>
      </listitem>
      <listitem>
	<para id="x_6e8">Bazaar</para>
      </listitem>
      <listitem>
	<para id="x_6e9">Monotone</para>
      </listitem>
      <listitem>
	<para id="x_6ea">GNU Arch</para>
      </listitem>
      <listitem>
	<para id="x_6eb">Mercurial</para>
      </listitem>
    </itemizedlist>

    <para id="x_6ec">(To see why Mercurial itself is supported as a source, see
      <xref linkend="svn.filemap"/>.)</para>

    <para id="x_6ed">You can enable the extension in the usual way, by editing
      your <filename>~/.hgrc</filename> file.</para>

    <programlisting>[extensions]
convert =</programlisting>

    <para id="x_6ee">This will make a <command>hg convert</command> command
      available.  The command is easy to use.  For instance, this
      command will import the Subversion history for the Nose unit
      testing framework into Mercurial.</para>

    <screen><prompt>$</prompt> <userinput>hg convert http://python-nose.googlecode.com/svn/trunk</userinput></screen>

    <para id="x_6ef">The <literal>convert</literal> extension operates
      incrementally.  In other words, after you have run <command>hg
	convert</command> once, running it again will import any new
      revisions committed after the first run began.  Incremental
      conversion will only work if you run <command>hg
	convert</command> in the same Mercurial repository that you
      originally used, because the <literal>convert</literal>
      extension saves some private metadata in a
      non-revision-controlled file named
      <filename>.hg/shamap</filename> inside the target
      repository.</para>

    <para id="x_707">When you want to start making changes using Mercurial, it's
      best to clone the tree in which you are doing your conversions,
      and leave the original tree for future incremental conversions.
      This is the safest way to let you pull and merge future commits
      from the source revision control system into your newly active
      Mercurial project.</para>

    <sect2>
      <title>Converting multiple branches</title>

      <para id="x_708">The <command>hg convert</command> command given above
	converts only the history of the <literal>trunk</literal>
	branch of the Subversion repository.  If we instead use the
	URL <literal>http://python-nose.googlecode.com/svn</literal>,
	Mercurial will automatically detect the
	<literal>trunk</literal>, <literal>tags</literal> and
	<literal>branches</literal> layout that Subversion projects
	usually use, and it will import each as a separate Mercurial
	branch.</para>

      <para id="x_709">By default, each Subversion branch imported into Mercurial
	is given a branch name.  After the conversion completes, you
	can get a list of the active branch names in the Mercurial
	repository using <command>hg branches -a</command>. If you
	would prefer to import the Subversion branches without names,
	pass the <option>--config
	  convert.hg.usebranchnames=false</option> option to
	<command>hg convert</command>.</para>

      <para id="x_70a">Once you have converted your tree, if you want to follow
	the usual Mercurial practice of working in a tree that
	contains a single branch, you can clone that single branch
	using <command>hg clone -r mybranchname</command>.</para>
    </sect2>

    <sect2>
      <title>Mapping user names</title>

      <para id="x_6f0">Some revision control tools save only short usernames with
	commits, and these can be difficult to interpret.  The norm
	with Mercurial is to save a committer's name and email
	address, which is much more useful for talking to them after
	the fact.</para>

      <para id="x_6f1">If you are converting a tree from a revision control
	system that uses short names, you can map those names to
	longer equivalents by passing a <option>--authors</option>
	option to <command>hg convert</command>.  This option accepts
	a file name that should contain entries of the following
	form.</para>

      <programlisting>arist = Aristotle &lt;aristotle@phil.example.gr&gt;
soc = Socrates &lt;socrates@phil.example.gr&gt;</programlisting>

      <para id="x_6f2">Whenever <literal>convert</literal> encounters a commit
	with the username <literal>arist</literal> in the source
	repository, it will use the name <literal>Aristotle
	  &lt;aristotle@phil.example.gr&gt;</literal> in the converted
	Mercurial revision.  If no match is found for a name, it is
	used verbatim.</para>
    </sect2>

    <sect2 id="svn.filemap">
      <title>Tidying up the tree</title>

      <para id="x_6f3">Not all projects have pristine history.  There may be a
	directory that should never have been checked in, a file that
	is too big, or a whole hierarchy that needs to be
	refactored.</para>

      <para id="x_6f4">The <literal>convert</literal> extension supports the idea
	of a <quote>file map</quote> that can reorganize the files and
	directories in a project as it imports the project's history.
	This is useful not only when importing history from other
	revision control systems, but also to prune or refactor a
	Mercurial tree.</para>

      <para id="x_6f5">To specify a file map, use the <option>--filemap</option>
	option and supply a file name.  A file map contains lines of the
	following forms.</para>

      <programlisting># This is a comment.
# Empty lines are ignored.	

include path/to/file

exclude path/to/file

rename from/some/path to/some/other/place
</programlisting>

      <para id="x_6f6">The <literal>include</literal> directive causes a file, or
	all files under a directory, to be included in the destination
	repository.  This also excludes all other files and dirs not
	explicitely included.  The <literal>exclude</literal>
	directive causes files or directories to be omitted, and
	others not explicitly mentioned to be included.</para>

      <para id="x_6f7">To move a file or directory from one location to another,
	use the <literal>rename</literal> directive.  If you need to
	move a file or directory from a subdirectory into the root of
	the repository, use <literal>.</literal> as the second
	argument to the <literal>rename</literal> directive.</para>
    </sect2>

    <sect2>
      <title>Improving Subversion conversion performance</title>

      <para id="x_70b">You will often need several attempts before you hit the
	perfect combination of user map, file map, and other
	conversion parameters.  Converting a Subversion repository
	over an access protocol like <literal>ssh</literal> or
	<literal>http</literal> can proceed thousands of times more
	slowly than Mercurial is capable of actually operating, due to
	network delays.  This can make tuning that perfect conversion
	recipe very painful.</para>

      <para id="x_70c">The <ulink
	  url="http://svn.collab.net/repos/svn/trunk/notes/svnsync.txt"><command>svnsync</command></ulink> 
	command can greatly speed up the conversion of a Subversion
	repository.  It is a read-only mirroring program for
	Subversion repositories.  The idea is that you create a local
	mirror of your Subversion tree, then convert the mirror into a
	Mercurial repository.</para>

      <para id="x_70d">Suppose we want to convert the Subversion repository for
	the popular Memcached project into a Mercurial tree.  First,
	we create a local Subversion repository.</para>

      <screen><prompt>$</prompt> <userinput>svnadmin create memcached-mirror</userinput></screen>

      <para id="x_70e">Next, we set up a Subversion hook that
	<command>svnsync</command> needs.</para>

      <screen><prompt>$</prompt> <userinput>echo '#!/bin/sh' > memcached-mirror/hooks/pre-revprop-change</userinput>
<prompt>$</prompt> <userinput>chmod +x memcached-mirror/hooks/pre-revprop-change</userinput></screen>

      <para id="x_70f">We then initialize <command>svnsync</command> in this
	repository.</para>

      <screen><prompt>$</prompt> <userinput>svnsync --init file://`pwd`/memcached-mirror \
  http://code.sixapart.com/svn/memcached</userinput></screen>

      <para id="x_710">Our next step is to begin the <command>svnsync</command>
	mirroring process.</para>

      <screen><prompt>$</prompt> <userinput>svnsync sync file://`pwd`/memcached-mirror</userinput></screen>

      <para id="x_711">Finally, we import the history of our local Subversion
	mirror into Mercurial.</para>

      <screen><prompt>$</prompt> <userinput>hg convert memcached-mirror</userinput></screen>
      
      <para id="x_712">We can use this process incrementally if the Subversion
	repository is still in use.  We run <command>svnsync</command>
	to pull new changes into our mirror, then <command>hg
	  convert</command> to import them into our Mercurial
	tree.</para>

      <para id="x_713">There are two advantages to doing a two-stage import with
	<command>svnsync</command>.  The first is that it uses more
	efficient Subversion network syncing code than <command>hg
	  convert</command>, so it transfers less data over the
	network.  The second is that the import from a local
	Subversion tree is so fast that you can tweak your conversion
	setup repeatedly without having to sit through a painfully
	slow network-based conversion process each time.</para>
    </sect2>
  </sect1>

  <sect1>
    <title>Migrating from Subversion</title>

    <para id="x_6f8">Subversion is currently the most popular open source
      revision control system. Although there are many differences
      between Mercurial and Subversion, making the transition from
      Subversion to Mercurial is not particularly difficult.  The two
      have similar command sets and generally uniform
      interfaces.</para>

    <sect2>
      <title>Philosophical differences</title>

      <para id="x_6f9">The fundamental difference between Subversion and
	Mercurial is of course that Subversion is centralized, while
	Mercurial is distributed.  Since Mercurial stores all of a
	project's history on your local drive, it only needs to
	perform a network access when you want to explicitly
	communicate with another repository. In contrast, Subversion
	stores very little information locally, and the client must
	thus contact its server for many common operations.</para>

      <para id="x_6fa">Subversion more or less gets away without a well-defined
	notion of a branch: which portion of a server's namespace
	qualifies as a branch is a matter of convention, with the
	software providing no enforcement.  Mercurial treats a
	repository as the unit of branch management.</para>

      <sect3>
	<title>Scope of commands</title>

	<para id="x_6fb">Since Subversion doesn't know what parts of its
	  namespace are really branches, it treats most commands as
	  requests to operate at and below whatever directory you are
	  currently visiting.  For instance, if you run <command>svn
	    log</command>, you'll get the history of whatever part of
	  the tree you're looking at, not the tree as a whole.</para>

	<para id="x_6fc">Mercurial's commands behave differently, by defaulting
	  to operating over an entire repository.  Run <command>hg
	    log</command> and it will tell you the history of the
	  entire tree, no matter what part of the working directory
	  you're visiting at the time.  If you want the history of
	  just a particular file or directory, simply supply it by
	  name, e.g. <command>hg log src</command>.</para>

	<para id="x_6fd">From my own experience, this difference in default
	  behaviors is probably the most likely to trip you up if you
	  have to switch back and forth frequently between the two
	  tools.</para>
      </sect3>

      <sect3>
	<title>Multi-user operation and safety</title>

	<para id="x_6fe">With Subversion, it is normal (though slightly frowned
	  upon) for multiple people to collaborate in a single branch.
	  If Alice and Bob are working together, and Alice commits
	  some changes to their shared branch, Bob must update his
	  client's view of the branch before he can commit.  Since at
	  this time he has no permanent record of the changes he has
	  made, he can corrupt or lose his modifications during and
	  after his update.</para>

	<para id="x_6ff">Mercurial encourages a commit-then-merge model instead.
	  Bob commits his changes locally before pulling changes from,
	  or pushing them to, the server that he shares with Alice.
	  If Alice pushed her changes before Bob tries to push his, he
	  will not be able to push his changes until he pulls hers,
	  merges with them, and commits the result of the merge.  If
	  he makes a mistake during the merge, he still has the option
	  of reverting to the commit that recorded his changes.</para>

	<para id="x_700">It is worth emphasizing that these are the common ways
	  of working with these tools. Subversion supports a safer
	  work-in-your-own-branch model, but it is cumbersome enough
	  in practice to not be widely used.  Mercurial can support
	  the less safe mode of allowing changes to be pulled in and
	  merged on top of uncommitted edits, but this is considered
	  highly unusual.</para>
      </sect3>

      <sect3>
	<title>Published vs local changes</title>

	<para id="x_701">A Subversion <command>svn commit</command> command
	  immediately publishes changes to a server, where they can be
	  seen by everyone who has read access.</para>

	<para id="x_702">With Mercurial, commits are always local, and must be
	  published via a <command>hg push</command> command
	  afterwards.</para>

	<para id="x_703">Each approach has its advantages and disadvantages.  The
	  Subversion model means that changes are published, and hence
	  reviewable and usable, immediately.  On the other hand, this
	  means that a user must have commit access to a repository in
	  order to use the software in a normal way, and commit access
	  is not lightly given out by most open source
	  projects.</para>

	<para id="x_704">The Mercurial approach allows anyone who can clone a
	  repository to commit changes without the need for someone
	  else's permission, and they can then publish their changes
	  and continue to participate however they see fit.  The
	  distinction between committing and pushing does open up the
	  possibility of someone committing changes to their laptop
	  and walking away for a few days having forgotten to push
	  them, which in rare cases might leave collaborators
	  temporarily stuck.</para>
      </sect3>
    </sect2>

    <sect2>
      <title>Quick reference</title>

      <table>
	<title>Subversion commands and Mercurial equivalents</title>
	<tgroup cols="3">
	  <thead>
	    <row>
	      <entry>Subversion</entry>
	      <entry>Mercurial</entry>
	      <entry>Notes</entry>
	    </row>
	  </thead>
	  <tbody>
	    <row>
	      <entry><command>svn add</command></entry>
	      <entry><command>hg add</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn blame</command></entry>
	      <entry><command>hg annotate</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn cat</command></entry>
	      <entry><command>hg cat</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn checkout</command></entry>
	      <entry><command>hg clone</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn cleanup</command></entry>
	      <entry>n/a</entry>
	      <entry>No cleanup needed</entry>
	    </row>
	    <row>
	      <entry><command>svn commit</command></entry>
	      <entry><command>hg commit</command>; <command>hg
		  push</command></entry>
	      <entry><command>hg push</command> publishes after
		commit</entry>
	    </row>
	    <row>
	      <entry><command>svn copy</command></entry>
	      <entry><command>hg clone</command></entry>
	      <entry>To create a new branch</entry>
	    </row>
	    <row>
	      <entry><command>svn copy</command></entry>
	      <entry><command>hg copy</command></entry>
	      <entry>To copy files or directories</entry>
	    </row>
	    <row>
	      <entry><command>svn delete</command> (<command>svn
		  remove</command>)</entry>
	      <entry><command>hg remove</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn diff</command></entry>
	      <entry><command>hg diff</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn export</command></entry>
	      <entry><command>hg archive</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn help</command></entry>
	      <entry><command>hg help</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn import</command></entry>
	      <entry><command>hg addremove</command>; <command>hg
		  commit</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn info</command></entry>
	      <entry><command>hg parents</command></entry>
	      <entry>Shows what revision is checked out</entry>
	    </row>
	    <row>
	      <entry><command>svn info</command></entry>
	      <entry><command>hg showconfig
		  paths.parent</command></entry>
	      <entry>Shows what URL is checked out</entry>
	    </row>
	    <row>
	      <entry><command>svn list</command></entry>
	      <entry><command>hg manifest</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn log</command></entry>
	      <entry><command>hg log</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn merge</command></entry>
	      <entry><command>hg merge</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn mkdir</command></entry>
	      <entry>n/a</entry>
	      <entry>Mercurial does not track directories</entry>
	    </row>
	    <row>
	      <entry><command>svn move</command> (<command>svn
		  rename</command>)</entry>
	      <entry><command>hg rename</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn resolved</command></entry>
	      <entry><command>hg resolve -m</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn revert</command></entry>
	      <entry><command>hg revert</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn status</command></entry>
	      <entry><command>hg status</command></entry>
	      <entry></entry>
	    </row>
	    <row>
	      <entry><command>svn update</command></entry>
	      <entry><command>hg pull -u</command></entry>
	      <entry></entry>
	    </row>
	  </tbody>
	</tgroup>
      </table>
    </sect2>
  </sect1>

  <sect1>
    <title>Useful tips for newcomers</title>

    <para id="x_705">Under some revision control systems, printing a diff for a
      single committed revision can be painful. For instance, with
      Subversion, to see what changed in revision 104654, you must
      type <command>svn diff -r104653:104654</command>. Mercurial
      eliminates the need to type the revision ID twice in this common
      case. For a plain diff, <command>hg export 104654</command>. For
      a log message followed by a diff, <command>hg log -r104654
	-p</command>.</para>

    <para id="x_706">When you run <command>hg status</command> without any
      arguments, it prints the status of the entire tree, with paths
      relative to the root of the repository.  This makes it tricky to
      copy a file name from the output of <command>hg status</command>
      into the command line.  If you supply a file or directory name
      to <command>hg status</command>, it will print paths relative to
      your current location instead.  So to get tree-wide status from
      <command>hg status</command>, with paths that are relative to
      your current directory and not the root of the repository, feed
      the output of <command>hg root</command> into <command>hg
	status</command>.  You can easily do this as follows on a
      Unix-like system:</para>

    <screen><prompt>$</prompt> <userinput>hg status `hg root`</userinput></screen>
  </sect1>
</appendix>

<!--
local variables: 
sgml-parent-document: ("00book.xml" "book" "appendix")
end:
-->