Google Git
Sign in
apache / subversion / refs/heads/0.11.1 / . / www / webdav-usage.html
blob: da83a396779700294a953ec351a89a3b3a92fb10 [file] [log] [blame]
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML//EN">
<html>
<head>
<title>Use of WebDAV in Subversion</title>
<style type="text/css">
.comment {
color: #900000;
background: #ffffc0;
}
.method {
font-family: courier;
font-weight: bold;
}
.term {
font-style: italic;
}
.prop {
font-family: courier;
font-weight: bold;
}
.url {
font-family: courier;
}
<!-- ### this dt thing isn't working... need to fix -->
dt {
font-weight: bold;
}
</style>
</head>
<body bgcolor=white>
<h1>Use of WebDAV in Subversion</h1>
<p>
This document details how WebDAV is used within the
<a href="http://subversion.tigris.org/">Subversion
product</a>. Specifically, how the client side interfaces with
<a href="http://www.webdav.org/neon/">Neon</a> to generate
WebDAV requests over the wire, and what the
server must do to map incoming WebDAV requests into operations
against the Subversion repository. Note that the server side is
implemented as an
<a href="http://httpd.apache.org/">Apache 2.0</a> module,
operating as a back-end for its mod_dav functionality.
</p>
<p>
This document heavily refers to the
<a href="http://subversion.tigris.org/svn-design.html">Subversion
design document</a> and the
<a href="http://www.webdav.org/deltav/">latest Delta-V protocol
draft</a>. Details of those documents will <em>not</em> be
replicated here.
</p>
<!-- ### put some of this stuff into a stylesheet -->
<center>
<table width="70%" border="0" bgcolor="#ff0000" cellspacing="0"
cellpadding="10">
<tr>
<td>
<font color="white">
<strong>NOTE:</strong> Subversion uses DeltaV for its
communication, but the Subversion client is
<strong>not</strong> a general-purpose DeltaV client. In
fact, it expects some custom features from the
server. Further, the Subversion server is
<strong>not</strong> a general-purpose DeltaV server. It
implements a strict <strong>subset</strong> of the
DeltaV specification. A WebDAV or DeltaV client may very
well be able to interoperate with it, but only if that
client operates within the narrow confines of those
features the server has implemented.
</font>
</td>
</tr>
<tr>
<td>
<font color="white">
Version 2.0 of Subversion will address WebDAV
interoperability (Class 1, Class 2, and DeltaV
features). Each of the custom features expected by the
client actually has an alternate mechanism available in
DeltaV, but in a much less efficient form.
</font>
</td>
</tr>
<tr>
<td>
<font color="white">
It is expected that Version 1.0 will support read-only,
Class 1 WebDAV clients. Any "low-hanging fruit" to
increase DeltaV interoperability will be considered.
</font>
</td>
</tr>
</table>
</center>
<p></p>
<hr width="80&#37;">
<h2>Basic Concepts</h2>
<p>
Subversion uses a tree-based format to describe a change set
against the repository. This tree is constructed on the client
side (by "walking" over the working copy) to describe the
change. The tree is marshalled to the server as a linear
sequence of changes to be applied to the repository. The
repository can accept changes in a random-access manner, so the
mapping from a tree to a set of changes works very well for the
repository.
</p>
<p>
Subversion provides properties on files, directories, and even
the abstract concept of a revision. Each of the operations
involving properties are mapped directly to WebDAV properties,
which are manipulated with the <span
class="method">PROPFIND</span> and <span
class="method">PROPPATCH</span> HTTP methods. Revisions are
modeled as DeltaV <span class="term">baselines</span>, so
revision properties are available through a <span
class="method">PROPFIND</span> on the baseline.
</p>
<p>
The Subversion server can efficiently compute deltas between two
revisions (these deltas are complete <span class="term">tree
deltas</span>, not simple text deltas). DeltaV does not have a
direct analogue for the tree delta concept. A client could
discover changes by issuing a sequence of <span
class="method">PROPFIND</span> requests on the various WebDAV
resources, but this would be a time-consuming operation,
involving many requests. Instead, Subversion marshals this
concept as a custom WebDAV <span
class="term">report</span>. Using this report, the client learns
which items in the working copy are out of date and can issue
<span class="method">GET</span> and <span
class="method">PROPFIND</span> methods to fetch the new data.
</p>
<p>
Tags and branches are simple copies in Subversion, which are
handled with the WebDAV <span class="method">COPY</span>.
</p>
<blockquote class="comment">
need to talk about copies somewhere. need to discuss how copy
history is retained (svn does it automatically, but interop with
other servers may require us to set a custom property on those
servers.
</blockquote>
<hr width="80&#37;">
<h2>DeltaV Concepts Used by Subversion</h2>
<p>
Subversion uses many of the DeltaV concepts, as listed
below. Note that many of these concepts are not fully
implemented by Subversion; we have implemented enough to meet
our needs, but little more.
</p>
<dl>
<dt>Baseline</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Activity</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Version Resource</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Version-Controlled Configuration</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Baseline Collection</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Version-Controlled Resource</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Working Resource (Feature)</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Merge Feature</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Label Feature</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
<dt>Version-Controlled-Collection Feature</dt>
<dd>
<span class="comment">further info to come...</span><p></p>
</dd>
</dl>
<hr width="80&#37;">
<h2>Subversion Projects as URLs</h2>
<p>
The very first concept to define is how a project is exposed to
the client. Subversion will expose all projects as URLs on a
server. The files and subdirectories under this project will be
exposed through the URL namespace.
</p>
<p>
For example, let us assume that we have a project named
"example". And let us say that this project will be exposed at
the URL:
<span class="url">http://subversion.tigris.org/repos/example/</span>.
</p>
<p>
This mapping is set up through a set of configuration
parameters for the Apache HTTP Server (which is hosting the
Subversion code and the particular project in question). The
configuration could look like:
</p>
<blockquote>
<pre>&lt;Location /repos/example&gt;
DAV svn
SVNPath /home/svn-projects/example
&lt;/Location&gt;</pre>
</blockquote>
<p>
Files and directories within the project will be directly mapped
into the URL namespace. For example, if the project contains a
file "file.c" in a subdirectory "sub", then the URL for that
file will be
<span class="url">http://subversion.tigris.org/repos/example/sub/file.c</span>.
</p>
<hr width="80&#37;">
<h2>Initial Checkout</h2>
<p>
When the user performs the initial checkout of a Subversion
project, the client will issue a series of <span
class="method">PROPFIND</span> and <span
class="method">GET</span> requests. These requests will traverse
the repository, pick up some necessary metadata, and then fetch
the latest revision.
</p>
<p class="comment">
describe the OPTIONS request for fetching the activity
collection set. describe the sequence of PROPFINDs to reach the
baseline collection.
</p>
<p class="comment">
<i>(moved here from below; need to rewrite)</i><br>
When the initial checkout was performed, Subversion fetched
the <span class="prop">DAV:activity-collection-set</span> value
and stored it as a property on each directory in the working copy.
property for each
collection. This property lists all of the locations on the
server where an activity may be created. The first of these
locations will be stored on the client for use during the commit
process.
</p>
<p class="comment">
Should probably describe the metadata we fetch, and how a
checkout of "not the latest" (e.g. by date or revision) will
work.
</p>
<hr width="80&#37;">
<h2>Committing a Change</h2>
<p>
Subversion commits are modeled using the "activity" concept from
DeltaV. An activity can be viewed as a transaction for a set of
resources.
</p>
<h3>Creating the activity</h3>
<p>
At commit time, the Subversion client will retrieve the stored
<span class="prop">DAV:activity-collection-set</span> value to
know where it should create the activity. Next,
the client will generate a UUID (a unique value) to use for the
activity's location. Finally, the client will issue a
<span class="method">MKACTIVITY</span> method request, where the Request-URL is
composed from the activity location and the UUID. This request
will construct an activity to hold all of the changes for the
commit.
</p>
<p>Abbreviated summary:</p>
<dl>
<dt>At checkout time:</dt>
<dd>
Request: <span class="method">OPTIONS</span> for
<span class="prop">DAV:activity-collection-set</span>
<br>
Response: <span class="url">http://www.example.com/repos/foo/$svn/act/</span>
</dd>
<dt>At commit time:</dt>
<dd>
Request: <span class="method">MKACTIVITY</span>
<span class="url">http://www.example.com/repos/foo/$svn/act/01234567-89ab-cdef-0123-456789abcdef</span>
<br>
Response: 201 (Created)
</dd>
</dl>
<p>
The <span class="method">CHECKOUT</span> method can specify an
activity to use upon checkout. This feature is used to associate
all items with the newly-created activity.
</p>
<h3>Storing the commit message</h3>
<p class="comment">
talk about checking out the baseline and applying a PROPPATCH to
the working baseline.
</p>
<h3>Mapping changes to WebDAV</h3>
<p>
A change set in Subversion is specified with a "tree delta" (see
the SVN design for more details on the changes that can be
placed into a tree delta). The tree delta will be unravelled
into a set of requests. These requests will be one of the
following forms:
</p>
<dl>
<dt>Delete file or directory</dt>
<dd>
These changes are mapped onto a <span
class="method">DELETE</span> operation. The version resource
of the target's parent collection is checked out using the
<span class="method">CHECKOUT</span> method (into the current
activity). The target (name) is then deleted from the
resulting working collection using the <span
class="method">DELETE</span> method.
<p></p>
</dd>
<dt>Add file</dt>
<dd>
This is modeled by performing a <span
class="method">CHECKOUT</span> of the version resource of the
target's parent collection. The new file is created within the
resulting working collection using a <span
class="method">PUT</span> request. Properties are applied
using <span class="method">PROPPATCH</span>.
<p></p>
</dd>
<dt>Add directory</dt>
<dd>
This is modeled by performing a <span
class="method">CHECKOUT</span> on the version resource of the
target's parent collection. The new directory is created
within the resulting working collection with a <span
class="method">MKCOL</span> request. Properties are applied
using <span class="method">PROPPATCH</span>.
<p></p>
</dd>
<dt>Add file or directory, with previous ancestory (a copy)</dt>
<dd>
<span class="comment">need to fix this section</span>
<p>
A tree delta can specify that a file/directory originates as
a copy of another file/dir. This copy may be further
modified by additional elements the tree delta.
</p>
<p>
This change will be modeled by performing a
<code>CHECKOUT</code> on the version resource of the parent
collection which will contain the new resource. The
<code>VERSION-CONTROL</code> method will create a new
version-controlled resource (VCR) within the working
collection, with the VCR's <code>DAV:checked-in</code>
property referring to the ancestor's version resource.
</p>
<blockquote>
<strong>Note:</strong> it appears that we will use
<code>COPY</code> to copy the appropriate resource into the
working collection. This will create a new version history
which is then placed into the working collection. The
version history will use the <code>DAV:precursor-set</code>
property to specify the version resource of the ancestor.
<p>
Because a version resource does not specify the revision,
it will not be possible to <code>COPY</code> a version
resource into the working collection -- it will not tell
us what revision was copied. Instead, we will most likely
copy a version resource out of the appropriate
baseline. This implies the client must be able to map from
a URL/revision pair to a baselined version resource URL.
</p>
<p>
The second issue is whether/how we set the
<code>DAV:precursor-set</code> property of the version
history. Or, more precisely, how we synthesize the value
from information stored in the repository. This is still
under investigation.
</p>
</blockquote>
</dd>
<dt>Replace file/dir by another file/dir</dt>
<dd>
This change does not have a WebDAV modeling because tree
deltas model it as two, sequential operations: a
<em>delete</em>, followed by an <em>add</em>.
<p></p>
</dd>
<dt>Moving a file or directory</dt>
<dd>
This change does not have a WebDAV modeling because tree
deltas model it as two, distinct operations: a
<em>delete</em>, and an <em>add</em> with previous ancestry.
<p></p>
</dd>
<dt>Replace file</dt>
<dd>
This is modeled with a <span class="method">CHECKOUT</span> on
the target's version resource, followed by a <span
class="method">PUT</span> to the resulting working resource.
<p></p>
</dd>
<dt>Replace directory</dt>
<dd>
In Subversion terms, "replace directory" means that additions,
deletions, and other changes will occur <em>within</em> the
directory. Each of these changes are modeled individually, and
the change to the directory is performed
implicitly. Therefore, this "change" has no particular mapping
into WebDAV.
<p></p>
</dd>
<dt>Property delta</dt>
<dd>
A property delta (against a file or directory) maps directly
to a <span class="method">PROPPATCH</span> in WebDAV
terms. The target's version resource will be checked out using
<span class="method">CHECKOUT</span> and the <span
class="method">PROPPATCH</span> will be applied to the
resulting working resource.
</dd>
</dl>
<h3>Final Commit</h3>
<p>
The final action of the commit process is to issue a <span
class="method">MERGE</span> request to the Subversion server,
specifying that the activity (created earlier) be checked in and
the corresponding version-controlled resources be updated to
refer to the new version resources.
</p>
<p class="comment">
the comment below is not quite right. talk about the working
baseline, and how that is used to create a new baseline (with
the commit message on it)
</p>
<p>
The version-controlled resources are also baseline-controlled,
which means that updates to them will automatically create a new
baseline. In essence, the commit will create a new baseline
corresponding to the new Subversion revision.
</p>
<h3>Example</h3>
<p class="comment">
<strong>Warning:</strong> this section has not been updated to
reflect some recent changes to the SVN-to-DAV mapping. Consider
it out of date until this warning is removed.
</p>
<p>
Consider the following set of operations and its corresponding
tree delta (taken from the SVN design document):
</p>
<ol>
<li>rename <code>/dir1/dir2</code> to <code>/dir1/dir4</code>,</li>
<li>rename <code>/dir1/dir3</code> to <code>/dir1/dir2</code>, and</li>
<li>move <code>file3</code> from <var>/dir1/dir4</var> to <var>/dir1/dir2</var>.</li>
</ol>
<pre>&lt;tree-delta&gt;
&lt;replace name='dir1'&gt;
&lt;directory&gt;
&lt;tree-delta&gt;
&lt;replace name='dir2'&gt;
&lt;directory ancestor='/dir1/dir3'&gt; (1)
&lt;tree-delta&gt;
&lt;new name='file3'&gt; (2)
&lt;file ancestor='/dir1/dir2/file3'/&gt;
&lt;/new&gt;
&lt;/tree-delta&gt;
&lt;/directory&gt;
&lt;/replace&gt;
&lt;delete name='dir3'/&gt; (3)
&lt;new name='dir4'&gt; (4)
&lt;directory ancestor='/dir1/dir2'&gt;
&lt;tree-delta&gt;
&lt;delete name='file3'/&gt; (5)
&lt;/tree-delta&gt;
&lt;/directory&gt;
&lt;/new&gt;
&lt;/tree-delta&gt;
&lt;/directory&gt;
&lt;/replace&gt;
&lt;/tree-delta&gt;
</pre>
<p>
Walking through this delta, we map out the WebDAV requests
listed below. The numbers in the above delta roughly correspond
to the numbered entries below. The correspondence is not exact
because a specific, resulting behavior is typically based on a
combination of a few elements in the delta.
</p>
<ol>
<li>
The <code>&lt;directory ancestor="/dir1/dir3"&gt;</code>
specifies that we are overwriting <code>/dir1/dir2</code> with
<code>/dir1/dir3</code>.
<p>
<code>CHECKOUT&nbsp;/dir1/dir2/</code><br>
<i>(returns a working resource URL for the directory)</i>
</p>
<p>
<code>COPY&nbsp;/dir1/dir3/</code><br>
<code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../</code><br>
<code>Overwrite:&nbsp;T</code>
</p>
</li>
<li>
<code>/dir1/dir2/file3</code> is new (since we just overwrote
the original <code>dir2</code> directory), and originates from
<code>/dir1/dir2/file3</code>. Thus, we simply
<code>COPY</code> the file into the target directory's working
resource:
<p>
<code>COPY&nbsp;/dir1/dir2/file3</code><br>
<code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../file3</code>
</p>
</li>
<li>
<code>CHECKOUT&nbsp;/dir1/dir3/</code><br>
<i>(returns a working resource URL for the directory)</i>
<p>
<code>DELETE&nbsp;/$svn/wrk/.../</code>
</p>
</li>
<li>
We are going to creating a new subdirectory (<code>dir4</code>) in the
<code>/dir1</code> directory. Since we don't have
<code>/dir1</code> checked out yet, we do so:
<p>
<code>CHECKOUT&nbsp;/dir1/</code><br>
<i>(returns a working resource URL for the directory)</i>
</p>
<p>
And now we copy the right directory into the new working
resource:
</p>
<p>
<code>COPY&nbsp;/dir1/dir2/</code><br>
<code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../dir4/</code>
</p>
</li>
<li>
The <code>COPY</code> created a complete set of working
resources on the server, so we simply delete the part that we
don't want:
<p>
<code>DELETE:&nbsp;/$svn/wrk/.../dir4/file3</code>
</p>
</li>
</ol>
<hr width="80&#37;">
<h2>URL Layout</h2>
<p>
The Subversion server exposes repositories at user-defined
URLs. For example, the "foo" repository might be located at
<span
class="url">http://www.example.com/repos/foo/</span>. However,
the server also requires a number of other resources to be
exposed for proper operation. These additional resources will be
associated with each repository in a location under the main
repository URL. By default, this location is "<span
class="url">$svn</span>". It may be changed by using the
<code>SVNSpecialURI</code> directive:
</p>
<blockquote>
<pre>&lt;Location /repos/foo&gt;
DAV svn
SVNPath /home/svn-projects/foo
SVNSpecialURI .special
&lt;/Location&gt;</pre>
</blockquote>
<p>
Underneath the location specified by <code>SVNSpecialURI</code>,
we will expose several collections. Assuming we use the default
of "<span class="url">$svn</span>", the collections are:
</p>
<dl>
<dt><span class="url">$svn/act/</span></dt>
<dd>
This area is where activity resources are created. The client
will pick a unique name within this collection and issue a
<span class="method">MKACTIVITY</span> for that URL. The client will then use
the activity in further interactions.
<p>
No methods are allowed on the <span class="url">$svn/act/</span>
resource.
</p>
<blockquote>
Note: actually, we may want to allow a <code>PROPFIND</code>
with a <code>Depth:&nbsp;1</code> header to allow clients to
enumerate the current activities.
</blockquote>
<p>
Only a subset of methods are allowed on the activities
within the collection. They are: <span class="method">PROPFIND</span>,
<span class="method">MERGE</span> (commit the activity), and
<span class="method">DELETE</span> (abort the activity).
</p>
<p>
Per the Delta-V specification, all activity resources will
have a <span class="prop">DAV:resourcetype</span> of
<span class="prop">DAV:activity</span>.
</p>
</dd>
<dt><span class="url">$svn/his/</span></dt>
<dd>
<span class="comment">do something with this section; we
actually don't use version history resources. in the future,
they might be modeled like this</span>
<p>
This collection contains the version history resources for
files and directories in a project. Its internal layout is
completely server-defined. Clients will receive URLs into
this collection (or a subcollection) from various responses.
</p>
<p>
No methods are allowed on the <span class="url">$svn/his/</span>
resource.
</p>
<p>
Internally, the URL namespace is laid out with URLs of the
following form:
</p>
<blockquote class="url">
$svn/his/<var>node-id</var>
</blockquote>
<p>
The <var>node-id</var> is an internal value
that Subversion uses to reference individual files and
directories. This <var>node-id</var> is a single integer
defined by the Subversion repository. Note that this is a an
undotted node id, which is the base for the entire history
of a given node in the repository.
</p>
<p>
The <span class="prop">DAV:resourcetype</span> of the <var>node-id</var>
collection is <span class="prop">DAV:version-history</span>.
</p>
<blockquote class="comment">
<strong>Note:</strong> the above information is probably not
quite correct. The issue of linking one version history to
another is still open. Further, I think that node 73 and
node 73.4.1 are each version histories (where the latter is
linked to the former). 73.x and 73.4.1.x are the versions
within the version history.
</blockquote>
</dd>
<dt><span class="url">$svn/ver/</span></dt>
<dd>
This collection contains the version resources for the
project.
<p>
No methods are allowed on the <span class="url">$svn/ver/</span>
resource.
</p>
<p>
The layout of this collection is internal to the server. For
reference purposes here (and to describe the
implementation), it is laid out as:
</p>
<blockquote class="url">
$svn/ver/<var>node-id</var>/<var>path</var>
</blockquote>
<p>
Only read-only methods are allowed against these resources
(e.g. <span class="method">GET</span>, <span class="method">PROPFIND</span>,
<span class="method">REPORT</span>); all other methods are illegal.
</p>
<p>
The <span class="prop">DAV:resourcetype</span> of a version resource is
simply the value of the resource at checkin time
(e.g. <code>&lt;D:resourcetype/&gt;</code> or
<code>&lt;D:resourcetype&gt;&lt;D:collection/&gt;&lt;/D:resourcetype&gt;</code>).
</p>
</dd>
<dt><span class="url">$svn/wrk/</span></dt>
<dd>
This collection contains working resources for the resources
that have been checked out with the <span class="method">CHECKOUT</span>
method. The form and construction of this collection is
server-defined, but is also well-defined so that clients may
interact properly with collection versions that have been
checked out.
<p>
No methods are allowed on the <span class="method">$svn/wrk/</span>
resource.
</p>
<p>
For reference purposes, the working resource URLs are
constructed as:
</p>
<blockquote class="url">
$svn/wrk/<var>activity</var>/<var>path</var>
</blockquote>
<p>
Any method is allowed on the working resources, but no
methods are allowed on any of its parents.
</p>
<p>
The <span class="prop">DAV:resourcetype</span> of the working resources
follows normal resource typing:
<code>&lt;D:resourcetype/&gt;</code> for regular working
resources, and
<code>&lt;D:resourcetype&gt;&lt;D:collection/&gt;&lt;/D:resourcetype&gt;</code>
for working collections.
</p>
</dd>
<dt><span class="url">$svn/vcc/</span></dt>
<dd>
<span class="comment">This section is not yet complete.</span>
<p>
version-controlled configuration...
</p>
<p>
<code>$svn/vcc/root</code> as a singleton.
</p>
</dd>
<dt><code>$svn/bln/</code></dt>
<dd>
<span class="comment">This section is not yet complete.</span>
<p>
baselines...
</p>
<p>
<code>$svn/bln/<var>rev</var>/</code>
</p>
</dd>
<dt><code>$svn/wbl/</code></dt>
<dd>
<span class="comment">This section is not yet complete.</span>
<p>
working baseline...
</p>
</dd>
<dt><code>$svn/bc/</code></dt>
<dd>
<span class="comment">This section is not yet complete.</span>
<p>
baseline collection...
</p>
</dd>
</dl>
<hr width="80&#37;">
<h2>Property Management (and History/Log Reporting)</h2>
<p class="comment">
this section needs to be reworked. the properties occur on the
FS revisions (and exposed via baselines).
</p>
<p>
As mentioned before, Subversion properties map onto WebDAV
properties. For history/log reporting, the following WebDAV
properties will be applied to each baseline (a Subversion
revision) and to each version resource created by the
revision. Since these resources are all version resources, the
properties below are read-only.
</p>
<dl>
<dt><code>DAV:comment</code></dt>
<dd>
This is the standard (dead) property for specifying a checkin
comment.
<p></p>
</dd>
<dt><code>DAV:creator-displayname</code></dt>
<dd>
This is a (dead) property that is generated from Subversion's
concept of the "user" who made a particular change.
<p></p>
</dd>
<dt><code>DAV:creationdate</code></dt>
<dd>
This is a read-only live property created by the server at
commit time.
</dd>
</dl>
<p>
The history for a specified file will be generated using the
<code>REPORT</code> method and a
<code>DAV:property-report</code> report. A typical history will
fetch the three properties mentioned above for each version of
the file/directory.
</p>
<p>
Based on the client design, it may be important to specify other
read-only live properties for information about versions. For
examply, how many lines were added/removed in a particular
checkin for a file? Creating these live properties will be quite
straight-forward, and driven by the client design over time.
</p>
<blockquote>
Note: if we do this, however, then we'd end up tying the client
to the server. Of course, if the client were run against another
DeltaV server which didn't report these properties, then we'd
simply not display them in the UI. (e.g. graceful degradation of
functionality)
</blockquote>
<hr width="80&#37;">
<h2>Fetching Status and Updates</h2>
<p>
After the initial checkout, the client can request a status
report (what has been changed on the client, pending a commit;
what has been changed on the server, pending an update). The
update process is similar, except that we also fetch the changes
from the server.
</p>
<p>
The local changes can be handled entirely on the client
side. The Working Copy library can easily handle the detection
and reporting of these changes. We're concerned with efficiently
detecting what has changed on the server.
</p>
<p>
While it would be possible to traverse the repository, fetching
the current state, and comparing that to the client state, it
would not be efficient. The Subversion design enables the server
to easily compute what has changed (relative to the client), if
it is given a description of the client state.
</p>
<p>
The core of the <em>status</em> and <em>update</em> commands is
based on a custom Subversion-specific WebDAV report. This custom
report will transmit the state of the working copy to the
server, and the server response will specify which resources
will need to be updated (fetched).
</p>
<p>
The request is a standard <code>REPORT</code> request, with a
custom XML body. The body will use the standard Subversion
technique of reporting a top-level revision number, and then
only reporting children that have different revisions. The
result of the report will use the same technique of reporting
only the resources where a change is found. If a change is
found, the server will provide a URL to the version resource to
fetch for the changed resource. The server will also report the
current revision number.
</p>
<blockquote class="comment">
The XML DTDs for the request and response are TBD.
</blockquote>
<blockquote>
<small>
The custom report will tie the client to only those servers
which support the report, but a future version of the software
will contain a fallback codepath, a graceful degradation, to
support other DeltaV servers.
</small>
</blockquote>
<p>
When an updated is performed, the client will fetch each of the
URLs (using <code>GET</code> requests) provided in the server
response.
</p>
<p>
<code>GET</code> (and <code>PUT</code>) operations will transfer
content in a "diff" format when possible. The mechanics of this
will follow the Internet Draft, titled
<a href="http://www.ietf.org/internet-drafts/draft-mogul-http-delta-10.txt">Delta Encoding in HTTP</a>.
</p>
<h3>Entity Tags (etags)</h3>
<p>
Etags are required to be unique across all versions of a
resource. Luckily, this
is very easy for a version control system. Each etag will be
simply be the repository's <em>node-id</em> for the resource.
</p>
<p>
Etags are used to generate diffs, following the guidelines in
the aforementioned draft:
<a href="http://www.ietf.org/internet-drafts/draft-mogul-http-delta-10.txt">Delta Encoding in HTTP</a>.
The problem then becomes how to get the etag for each file
stored on the client (we don't need etags for directories since
we never fetch them). During a <em>checkout</em> or
<em>update</em> process, this is easy: the etag is provided in
the HTTP response headers for each file retrieved.
</p>
<p>
The other part of the problem is getting the etag after a
<em>commit</em> has occurred. The <code>MERGE</code> response
provides a way to request properties from the version resources
which are created as part of the checkin of the activity. The
etag (and other properties) can be fetched using that mechanism.
</p>
<hr width="80&#37;">
<h2>Tags and Branches</h2>
<p>
Tags and branches within Subversion are performed by copying
from one area to another. For example:
</p>
<blockquote>
<pre>
[.../src/my-project]$ svn cp trunk tags/1.0.3-rc4
[.../src/my-project]$ svn commit
</pre>
</blockquote>
<p>
In the above example, <code>tags/1.0.3-rc4</code> should now be
considered readonly and will always reflect the status of
<code>trunk</code>.
</p>
<p>
These copies are handled just like a regular commit. An activity
is created with <span class="method">MKACTIVITY</span>, a
working resource is created via <span
class="method">CHECKOUT</span> (for the target directory;
<code>tags/</code> in our example above), and then a <span
class="method">COPY</span> is performed. The activity is then
merged back into the repository with a <span
class="method">MERGE</span> request.
</p>
<hr width="80&#37;">
<h2>Server Requirements</h2>
<p class="comment">
<strong>Warning:</strong> this section is out of date. The
DeltaV draft has gone through a number of revisions, and our use
of DeltaV has changed some.
</p>
<h3>DAV Methods</h3>
<p>
The server will need to implement the following WebDAV methods
for proper operation:
</p>
<ul>
<li><span class="method">OPTIONS</span></li>
<li><span class="method">GET</span></li>
<li><span class="method">DELETE</span></li>
<li><span class="method">COPY</span></li>
<li><span class="method">PROPPATCH</span></li>
<li><span class="method">PROPFIND</span></li>
<li><span class="method">MKACTIVITY</span></li>
<li><span class="method">CHECKOUT</span></li>
<li><span class="method">MERGE</span></li>
<li><span class="method">REPORT</span></li>
</ul>
<p>
The following methods are not required by Subversion at this
time:
</p>
<ul>
<li><span class="method">CHECKIN</span></li>
<li><span class="method">UNCHECKOUT</span></li>
<li><span class="method">UPDATE</span></li>
<li><span class="method">LABEL</span></li>
<li><span class="method">VERSION-CONTROL</span></li>
<li><span class="method">BASELINE-CONTROL</span></li>
<li><span class="method">MKWORKSPACE</span></li>
</ul>
<h3>DAV Properties</h3>
<p>
The following DeltaV properties will be implemented:
</p>
<ul>
<!-- resource properties -->
<li><span class="prop">DAV:comment</span></li>
<li><span class="prop">DAV:creator-displayname</span></li>
<li><span class="prop">DAV:supported-method-set</span></li>
<li><span class="prop">DAV:supported-live-property-set</span></li>
<li><span class="prop">DAV:supported-report-set</span></li>
<li><span class="prop">DAV:version-controlled-configuration</span></li>
<!-- version-controlled resource properties -->
<li><span class="prop">DAV:checked-in</span></li>
<li>
<span class="prop">DAV:auto-version</span> is a readonly,
empty element (auto versioning not supported).
</li>
<!-- checked-out resource properties -->
<li><span class="prop">DAV:checked-out</span></li>
<li>
<span class="prop">DAV:predecessor-set</span>
<br>
<i>
Note: the Subversion design document is not clear on the
mechanics of how multiple predecessors are merged to create
a single, new revision. When this clarifies, then
<code>DAV:predecessor-set</code> may end up containing more
than zero or one predecessor URLs
</i>
</li>
<!-- version resource properties -->
<!-- predecessor-set -->
<li>
<span class="prop">DAV:version-name</span> is simply the
(global) revision number.
</li>
<li><span class="prop">DAV:checkout-fork</span></li>
<li><span class="prop">DAV:checkin-fork</span></li>
<!-- checked-out (working resource) properties -->
<!-- checkout-fork, checkin-fork -->
<li><span class="prop">DAV:auto-update</span></li>
<li>
<span class="prop">DAV:subbaseline-set</span> is a readonly,
empty property (sub-baselines not supported).
</li>
<li>
<span class="prop">DAV:unreserved</span> is set to
<code>F</code>.
</li>
<!-- vcc properties -->
<li><span class="prop">DAV:baseline-controlled-collection</span></li>
<!-- baseline properties -->
<!-- subbaseline-set -->
<li><span class="prop">DAV:baseline-collection</span></li>
<!-- activity properties -->
<li>
<span class="prop">DAV:subactivity-set</span> is a readonly,
empty property (sub-activities not supported).
</li>
<!-- version-controlled collection properties -->
<li>
<span class="prop">DAV:eclipsed-set</span> is always empty
(internal members can never be eclipsed).
</li>
</ul>
<p>
Contrary to the DeltaV specification, the following required
properties will not be implemented:
</p>
<ul>
<li>
<span class="prop">DAV:successor-set</span> - this may be an
expensive operation to synthesize this value.
</li>
<li>
<span class="prop">DAV:checkout-set</span> - we do not record
what has actually been checked out, but use the working
resource URL to provide the necessary information; thus we
have no record of the data to populate this property.
</li>
<li>
<span class="prop">DAV:merge-set</span> - our <span
class="method">MERGE</span> method is solely to support a
commit. It doesn't really support the arbitrary merging
defined within the spec.
</li>
<li>
<span class="prop">DAV:auto-merge-set</span> - same as for
<span class="prop">DAV:merge-set</span>.
</li>
<li>
<span class="prop">DAV:activity-version-set</span> -
activities are only used for working resources, so versions
cannot be part of an activity.
<br>
<i>maybe this should be defined as the empty set?</i>
</li>
<li>
<span class="prop">DAV:activity-checkout-set</span>
activities are only used for working resources, and we do not
record what working resources "exist".
</li>
<li>
<span class="prop">DAV:activity-set</span> -
activities are only used for working resources, so versions
cannot be part of an activity.
<br>
<i>maybe this should be defined as the empty set?</i>
</li>
<li>
<span class="prop">DAV:version-controlled-binding-set</span> -
we do not have version history resources to include in the
property.
</li>
</ul>
<h3>OPTIONS</h3>
<p>
The <span class="method">OPTIONS</span> request will signal that
it supports the following DAV features:
</p>
<ul>
<li><code>1</code></li>
<li><code>2</code></li>
<li><code>version-control</code></li>
<li><code>checkout</code></li>
<li><code>working-resource</code></li>
<li><code>merge</code></li>
<li><code>baseline</code></li>
<li><code>activity</code></li>
<li><code>version-controlled-collection</code></li>
</ul>
<h3>Reports</h3>
<p>
The <span class="prop">DAV:supported-report-set</code> property
will signal support for the following reports:
</p>
<ul>
<li>svn:update-report</li>
<li>svn:log-report</li>
</ul>
<p>
These reports are available only on the "public" resources (the
VCRs). They are not available on the resources within the
<span class="url">$svn/</span> area.
</p>
<h3>Notes, reminders</h3>
<p>
Discuss timeouts and auto-purge of activities (and the related
working resources).
<br>
Discuss the activity database maintained by mod_dav_svn.
<br>
Discuss other implementation details of ra_dav and mod_dav_svn.
</p>
<h2><a name="rationale">Appendix A: Rationale</a></h2>
<p>
Several times, people have asked, "Why choose
HTTP/WebDAV/DeltaV? That seems awfully bloated and
ill-suited. Why didn't you design a custom, well-tuned protocol?
Or maybe use the CVS protocol?" Listed below are a number of
reasons for our choice of WebDAV as our network protocol.
</p>
<blockquote>
<small>
While this list could certainly be expanded with more reasons
(and to be fair, with a list of reasons why WebDAV was a poor
choice), it certainly demonstrates the basic reasons for our
choice.
</small>
</blockquote>
<blockquote>
<small>
Note: this list came from an email note, so the tone and point
of view might be a bit off. Further word-smithing is
welcome...
</small>
</blockquote>
<h3>Builtin web browsing of the repository</h3>
<blockquote>
For example, take a
look at:
<a href="http://svn.collab.net/repos/svn/trunk/HACKING">http://svn.collab.net/repos/svn/trunk/HACKING</a>
(that's the HEAD right there; we also have URLs for every
previous revision of every file)
</blockquote>
<h3>DAV-based browsing</h3>
<blockquote>
Use Web Folders or WebDrive or somesuch on
your Windows box (or Windows XP's native DAV mounts) to browse
the SVN repository with Windows Explorer. Mac OS X has builtin
DAV server mounting. Nautilus has DAV capabilities. Then you
have your Open Source tools such as cadaver, Goliath, etc.
</blockquote>
<h3>People can use existing libraries</h3>
<blockquote>
I couldn't even begin to count the number of HTTP tools and
libraries available. If we had designed our own protocol, then
we would have /none/ of those benefits. Heck, two HTTP library
implementors (Joe Orton of Neon, and Daniel Stenberg of CURL)
are regulars here. we wouldn't get that benefit. I've used
Python's httplib (and a davlib of my own) to do a lot of testing
of our server. No need to go and roll new protocol libraries.
</blockquote>
<h3>Existing tools</h3>
<blockquote>
One word: Ethereal :-) When we capture network traces, Ethereal
already knows about HTTP. It's quite nice, but I know there are
even better ones out there. But we also have other tools like
squid and other (caching) proxies (see the next item).
</blockquote>
<h3>Caching proxies</h3>
<blockquote>
Subversion will work great with caching proxies. There is no
longer a need for specialized tools like "cvsup". Just drop in a
caching proxy, and you've already got your distributed read-only
repository. That european dev team can just drop in the cache
between them and the US server and their checkouts/updates will
get cached for the benefit of the other team members. Commits
will flow through, back to the US-based server.
</blockquote>
<h3>Sophisticated and broad-choice authentication</h3>
<blockquote>
We don't have to reimplement an authentication scheme for a new
protocol. We can use all of the various schemes that have been
defined for HTTP. Ever look at the CVS protocol? Ever see the "I
Love You" or "I Hate You" lines? :-) That is all part of
creating a new authentication scheme. But we get to use SSL and
certificate-based auth if we want. Kerberos. NTLM. or even just
simple Basic or Digest. And our users can come from text files,
database, LDAP, or PAM. We don't have to reinvent the wheel cuz
it is all available for Apache already.
</blockquote>
<h3>Awesome network server</h3>
<blockquote>
We don't have to worry about how to portably set TCP_CORK for
optimal network packets. We don't have to worry about when
sendfile() makes sense, or if it is available. We don't have to
worry about dropped client connections, how to best use threads
and processes to scale, request management, monitoring, logging,
etc. Apache gives us all of that and a ton more. I *really*
would not want to do that through xinetd. I mean... setting
TCP_CORK on stdout? freaky :-)
</blockquote>
<h3>Well-defined on-wire compression</h3>
<blockquote>
We already have on-wire compression, similar to CVS's "-z#"
switch. And we didn't do anything. The client library and server
that we use just support it automatically for us, according to
RFC 2616.
</blockquote>
<h3>Future interoperability</h3>
<blockquote>
In the future, we'll be able to interoperate with a multitude of
IDEs and other WebDAV/DeltaV clients. As DeltaV becomes more
prevalent, IDEs could very well use it for source code
management, and we'll be right there without needing to write
some MS/SCC library to interface to the tool.
</blockquote>
<hr>
<address><a href="mailto:gstein@lyra.org">Greg Stein</a></address>
<!-- Created: Thu Aug 10 19:14:20 PDT 2000 -->
<!-- hhmts start -->
Last modified: Fri Jan 25 12:54:20 PST 2002
<!-- hhmts end -->
</body>
</html>