<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE manualpage SYSTEM "../style/manualpage.dtd">
<?xml-stylesheet type="text/xsl" href="../style/manual.en.xsl"?>
<!-- $LastChangedRevision$ -->

<!--
 Licensed to the Apache Software Foundation (ASF) under one or more
 contributor license agreements.  See the NOTICE file distributed with
 this work for additional information regarding copyright ownership.
 The ASF licenses this file to You under the Apache License, Version 2.0
 (the "License"); you may not use this file except in compliance with
 the License.  You may obtain a copy of the License at

     http://www.apache.org/licenses/LICENSE-2.0

 Unless required by applicable law or agreed to in writing, software
 distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
-->

<manualpage metafile="filters.xml.meta">
<parentdocument href="./">Developer Documentation</parentdocument>

<title>How filters work in Apache 2.0</title>

<summary>
    <note type="warning"><title>Warning</title>
      <p>This is a cut 'n paste job from an email
      (&lt;022501c1c529$f63a9550$7f00000a@KOJ&gt;) and only reformatted for
      better readability. It's not up to date but may be a good start for
      further research.</p>
    </note>
</summary>

<section id="types"><title>Filter Types</title>
    <p>There are three basic filter types (each of these is actually broken
    down into two categories, but that comes later).</p>

    <dl>
    <dt><code>CONNECTION</code></dt>
    <dd>Filters of this type are valid for the lifetime of this connection.
    (<code>AP_FTYPE_CONNECTION</code>, <code>AP_FTYPE_NETWORK</code>)</dd>

    <dt><code>PROTOCOL</code></dt>
    <dd>Filters of this type are valid for the lifetime of this request from
    the point of view of the client, this means that the request is valid
    from the time that the request is sent until the time that the response
    is received. (<code>AP_FTYPE_PROTOCOL</code>,
    <code>AP_FTYPE_TRANSCODE</code>)</dd>

    <dt><code>RESOURCE</code></dt>
    <dd>Filters of this type are valid for the time that this content is used
    to satisfy a request.  For simple requests, this is identical to
    <code>PROTOCOL</code>, but internal redirects and sub-requests can change
    the content without ending the request. (<code>AP_FTYPE_RESOURCE</code>,
    <code>AP_FTYPE_CONTENT_SET</code>)</dd>
    </dl>

    <p>It is important to make the distinction between a protocol and a
    resource filter.  A resource filter is tied to a specific resource, it
    may also be tied to header information, but the main binding is to a
    resource.  If you are writing a filter and you want to know if it is
    resource or protocol, the correct question to ask is:  "Can this filter
    be removed if the request is redirected to a different resource?"  If
    the answer is yes, then it is a resource filter.  If it is no, then it
    is most likely a protocol or connection filter.  I won't go into
    connection filters, because they seem to be well understood. With this
    definition, a few examples might help:</p>

    <dl>
    <dt>Byterange</dt>
    <dd>We have coded it to be inserted for all requests, and it is removed
    if not used.  Because this filter is active at the beginning of all
    requests, it can not be removed if it is redirected, so this is a
    protocol filter.</dd>

    <dt>http_header</dt>
    <dd>This filter actually writes the headers to the network.  This is
    obviously a required filter (except in the asis case which is special
    and will be dealt with below) and so it is a protocol filter.</dd>

    <dt>Deflate</dt>
    <dd>The administrator configures this filter based on which file has been
    requested.  If we do an internal redirect from an autoindex page to an
    index.html page, the deflate filter may be added or removed based on
    config, so this is a resource filter.</dd>
    </dl>

    <p>The further breakdown of each category into two more filter types is
    strictly for ordering.  We could remove it, and only allow for one
    filter type, but the order would tend to be wrong, and we would need to
    hack things to make it work.  Currently, the <code>RESOURCE</code> filters
    only have one filter type, but that should change.</p>
</section>

<section id="howinserted"><title>How are filters inserted?</title>
    <p>This is actually rather simple in theory, but the code is
    complex.  First of all, it is important that everybody realize that
    there are three filter lists for each request, but they are all
    concatenated together:</p>
    <ul>
        <li><code>r->output_filters</code> (corresponds to RESOURCE)</li>
        <li><code>r->proto_output_filters</code> (corresponds to PROTOCOL)</li>
        <li><code>r->connection->output_filters</code> (corresponds to CONNECTION)</li>
    </ul> 
    
    <p>The problem previously, was that we used a singly linked list to create the filter stack, and we
    started from the "correct" location.  This means that if I had a
    <code>RESOURCE</code> filter on the stack, and I added a
    <code>CONNECTION</code> filter, the <code>CONNECTION</code> filter would
    be ignored. This should make sense, because we would insert the connection
    filter at the top of the <code>c->output_filters</code> list, but the end
    of <code>r->output_filters</code> pointed to the filter that used to be
    at the front of <code>c->output_filters</code>. This is obviously wrong.
    The new insertion code uses a doubly linked list. This has the advantage
    that we never lose a filter that has been inserted. Unfortunately, it comes
    with a separate set of headaches.</p>

    <p>The problem is that we have two different cases were we use subrequests.
    The first is to insert more data into a response. The second is to
    replace the existing response with an internal redirect. These are two
    different cases and need to be treated as such.</p>

    <p>In the first case, we are creating the subrequest from within a handler
    or filter.  This means that the next filter should be passed to
    <code>make_sub_request</code> function, and the last resource filter in the
    sub-request will point to the next filter in the main request.  This
    makes sense, because the sub-request's data needs to flow through the
    same set of filters as the main request.  A graphical representation
    might help:</p>

<example>
<pre>
Default_handler --> includes_filter --> byterange --> ...
</pre>
</example>

    <p>If the includes filter creates a sub request, then we don't want the
    data from that sub-request to go through the includes filter, because it
    might not be SSI data.  So, the subrequest adds the following:</p>

<example>
<pre>
Default_handler --> includes_filter -/-> byterange --> ...
                                    /
Default_handler --> sub_request_core
</pre>
</example>

    <p>What happens if the subrequest is SSI data?  Well, that's easy, the
    <code>includes_filter</code> is a resource filter, so it will be added to
    the sub request in between the <code>Default_handler</code> and the
    <code>sub_request_core</code> filter.</p>

    <p>The second case for sub-requests is when one sub-request is going to
    become the real request.  This happens whenever a sub-request is created
    outside of a handler or filter, and NULL is passed as the next filter to
    the <code>make_sub_request</code> function.</p>

    <p>In this case, the resource filters no longer make sense for the new
    request, because the resource has changed.  So, instead of starting from
    scratch, we simply point the front of the resource filters for the
    sub-request to the front of the protocol filters for the old request.
    This means that we won't lose any of the protocol filters, neither will
    we try to send this data through a filter that shouldn't see it.</p>

    <p>The problem is that we are using a doubly-linked list for our filter
    stacks now. But, you should notice that it is possible for two lists to
    intersect in this model.  So, you do you handle the previous pointer?
    This is a very difficult question to answer, because there is no "right"
    answer, either method is equally valid.  I looked at why we use the
    previous pointer.  The only reason for it is to allow for easier
    addition of new servers.  With that being said, the solution I chose was
    to make the previous pointer always stay on the original request.</p>

    <p>This causes some more complex logic, but it works for all cases.  My
    concern in having it move to the sub-request, is that for the more
    common case (where a sub-request is used to add data to a response), the
    main filter chain would be wrong.  That didn't seem like a good idea to
    me.</p>
</section>

<section id="asis"><title>Asis</title>
    <p>The final topic.  :-)  Mod_Asis is a bit of a hack, but the
    handler needs to remove all filters except for connection filters, and
    send the data.  If you are using <module>mod_asis</module>, all other
    bets are off.</p>
</section>

<section id="conclusion"><title>Explanations</title>
    <p>The absolutely last point is that the reason this code was so hard to
    get right, was because we had hacked so much to force it to work.  I
    wrote most of the hacks originally, so I am very much to blame.
    However, now that the code is right, I have started to remove some
    hacks.  Most people should have seen that the <code>reset_filters</code>
    and <code>add_required_filters</code> functions are gone.  Those inserted
    protocol level filters for error conditions, in fact, both functions did
    the same thing, one after the other, it was really strange. Because we
    don't lose protocol filters for error cases any more, those hacks went away.
    The <code>HTTP_HEADER</code>, <code>Content-length</code>, and
    <code>Byterange</code> filters are all added in the
    <code>insert_filters</code> phase, because if they were added earlier, we
    had some interesting interactions.  Now, those could all be moved to be
    inserted with the <code>HTTP_IN</code>, <code>CORE</code>, and
    <code>CORE_IN</code> filters.  That would make the code easier to
    follow.</p>
</section>
</manualpage>