<!doctype HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!--
(C) Copyright 2002-4 Robert Ramey - http://www.rrsd.com . 
Use, modification and distribution is subject to the Boost Software
License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at
http://www.boost.org/LICENSE_1_0.txt)
-->
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<link rel="stylesheet" type="text/css" href="../../../boost.css">
<link rel="stylesheet" type="text/css" href="style.css">
<title>Serialization - Dataflow Iterators</title>
</head>
<body link="#0000ff" vlink="#800080">
<table border="0" cellpadding="7" cellspacing="0" width="100%" summary="header">
  <tr> 
    <td valign="top" width="300"> 
      <h3><a href="../../../index.htm"><img height="86" width="277" alt="C++ Boost" src="../../../boost.png" border="0"></a></h3>
    </td>
    <td valign="top"> 
      <h1 align="center">Serialization</h1>
      <h2 align="center">Dataflow Iterators</h2>
    </td>
  </tr>
</table>
<hr>
<h3>Motivation</h3>
Consider the problem of translating an arbitrary length sequence of 8 bit bytes 
to base64 text. Such a process can be summarized as:
<p>
source =&gt; 8 bit bytes =&gt; 6 bit integers =&gt; encode to base64 characters =&gt; insert line breaks =&gt; destination
<p>
We would prefer the solution that is:
<ul>
  <li>Decomposable. so we can code, test, verify and use each (simple) stage of the conversion 
  independently.
  <li>Composable. so we can used this composite as a new component somewhere else.
  <li>Efficient, so we're not required to re-implement it again.
  <li>Scalable, so that it works well for short and arbitrarily long sequences.
</ul>
The approach that comes closest to meeting these requirements is that described
and implemented with <a href="../../iterator/doc/index.html">Iterator Adaptors</a>.
The fundamental feature of an Iterator Adaptor template that makes in interesting to
us is that it takes as a parameter a base iterator from which it derives its
input. This suggests that something like the following might be possible.
<pre><code>
typedef 
    insert_linebreaks&lt;         // insert line breaks every 72 characters
        base64_from_binary&lt;    // convert binary values ot base64 characters
            transform_width&lt;   // retrieve 6 bit integers from a sequence of 8 bit bytes
                const char *,
                6,
                8
            &gt;
        &gt; 
        ,72
    &gt; 
    base64_text; // compose all the above operations in to a new iterator

std::copy(
    base64_text(address),
    base64_text(address + count),
    ostream_iterator&lt;CharType&gt;(os)
);
</code></pre>
Indeed, this seems to be exactly the kind of problem that iterator adaptors are 
intended to address.  The Iterator Adaptor library already includes
modules which can be configured to implement some of the operations above.  For example,
included is <a target="transform_iterator" href="../../iterator/doc/transform_iterator.html">
transform_iterator</a>, which can be used to implement 6 bit integer =&gt; base64 code.

<h3>Dataflow Iterators</h3>
Unfortunately, not all iterators which inherit from Iterator Adaptors are guarenteed
to meet the composability goals stated above.  To accomplish this purpose, they have
to be written with some additional considerations in mind.

We define a Dataflow Iterator as an class inherited from <code style="white-space: normal">iterator_adaptor</code> which
fulfills as a small set of additional requirements.

<h4>Templated Constructors</h4>
<p>
Templated constructor have the form:
<pre><code>
template&lt;class T&gt;
dataflow_iterator(T start) :
    iterator_adaptor(Base(start))
{}
</code></pre>
When these constructors are applied to our example of above, the following code is generated:
<pre><code>
std::copy(
    insert_linebreaks(
        base64_from_binary(
            transform_width(
                address
            ),
        )
    ),
    insert_linebreaks(
        base64_from_binary(
            transform_width(
                address + count
            )
        )
    )
    ostream_iterator&lt;char&gt;(os)
);
</code></pre>
The recursive application of this template is what automatically generates the
constructor <code style="white-space: normal">base64_text(const char *)</code>  in our example above.  The original
Iterator Adaptors include a <code style="white-space: normal">make_xxx_iterator</code> to fulfill this function.
However, I believe these are unwieldy to use compared to the above solution usiing
Templated constructors.
<p>
Unfortunately, some systems which fail to properly support partial function template
ordering cannot support the concept of a templated constructor as implemented above.  
A special"wrapper"  macro has been created to work around this problem. With this "wrapper" 
the above example is modified to:
<pre><code>
std::copy(
    base64_text(BOOST_MAKE_PFTO_WRAPPER(address)),
    base64_text(BOOST_MAKE_PFTO_WRAPPER(address + count)),
    ostream_iterator&lt;char&gt;(os)
);
</code></pre>
This macro is defined in <a target="pfto" href="../../../boost/serialization/pfto.hpp">&lt;boost/serialization/pfto.hpp&gt;</a>.
For more information about this topic, check the source.

<h4>Dereferencing</h4>
Dereferencing some iterators can cause problems.  For example, a natural
way to write a <code style="white-space: normal">remove_whitespace</code> iterator is to increment past the initial 
whitespaces when the iterator is constructed.  This will fail if the iterator passed to the
constructor "points" to  the end of a string.  The 
<a target="filter_iterator" href="../../iterator/doc/filter_iterator.html">
<code style="white-space: normal">filter_iterator</code></a> is implemented
in this way so it can't be used in our context. So, for implementation of this iterator, 
space removal is deferred until the iterator actually is dereferenced.

<h4>Comparison</h4>
The default implementation of iterator equality of <code style="white-space: normal">iterator_adaptor</code> just
invokes the equality operator on the base iterators.  Generally this is satisfactory.
However, this implies that other operations (E. G. dereference) do not prematurely 
increment the base iterator.  Avoiding this can be surprisingly tricky in some cases.
(E.G. transform_width)

<p>
Iterators which fulfill the above requirements should be composable and the above sample
code should implement our binary to base64 conversion.

<h3>Iterators Included in the Library</h3>
Dataflow iterators for the serialization library are all defined in the hamespace
<code style="white-space: normal">boost::archive::iterators</code> included here are:
<dl class="index">
  <dt><a target="base64_from_binary" href="../../../boost/archive/iterators/base64_from_binary.hpp">
  base64_from_binary</a></dt>
  <dd>transforms a sequence of integers to base64 text</dd>

  <dt><a target="base64_from_binary" href="../../../boost/archive/iterators/binary_from_base64.hpp">
  binary_from_base64</a></dt>
  <dd>transforms a sequence of base64 characters to a sequence of integers</dd>

  <dt><a target="insert_linebreaks" href="../../../boost/archive/iterators/insert_linebreaks.hpp">
  insert_linebreaks</a></dt>
  <dd>given a sequence, creates a sequence with newline characters inserted</dd>

  <dt><a target="mb_from_wchar" href="../../../boost/archive/iterators/mb_from_wchar.hpp">
  mb_from_wchar</a></dt>
  <dd>transforms a sequence of wide characters to a sequence of multi-byte characters</dd>

  <dt><a target="remove_whitespace" href="../../../boost/archive/iterators/remove_whitespace.hpp">
  remove_whitespace</a></dt>
  <dd>given a sequence of characters, returns  a sequence with the white characters
  removed.  This is  a derivation from the <code style="white-space: normal">boost::filter_iterator</code></dd>

  <dt><a target="transform_width" href="../../../boost/archive/iterators/transform_width.hpp">
  transform_width</a></dt>
  <dd>transforms a sequence of x bit elements into a sequence of y bit elements.  This
  is a key component in iterators which translate to and from base64 text.</dd>

  <dt><a target="wchar_from_mb" href="../../../boost/archive/iterators/wchar_from_mb.hpp">
  wchar_from_mb</a></dt>
  <dd>transform a sequence of multi-byte characters in the current locale to wide characters.</dd>

  <dt><a target="xml_escape" href="../../../boost/archive/iterators/xml_escape.hpp">
  xml_escape</a></dt>
  <dd>escapes xml meta-characters from xml text</dd>

  <dt><a target="xml_unescape" href="../../../boost/archive/iterators/xml_unescape.hpp">
  xml_unescape</a></dt>
  <dd>unescapes xml escape sequences to create a sequence of normal text<dd>
</dl>
<p>
The standard stream iterators don't quite work for us.  On systems which implement <code style="white-space: normal">wchar_t</code>
as unsigned short integers (E.G. VC 6) they didn't function as I expected. I also made some
adjustments to be consistent with our concept of Dataflow Iterators.  Like the rest of our
iterators, they are found in the namespace <code style="white-space: normal">boost::archive::interators</code> to avoid
conflict the standard library version.
<dl class = "index">
  <dt><a target="istream_iterator" href="../../../boost/archive/iterators/istream_iterator.hpp">
  istream_iterator</a></dt>
  <dt><a target="ostream_iterator" href="../../../boost/archive/iterators/ostream_iterator.hpp">
  ostream_iterator</a></dt>
</dl>

<hr>
<p><i>&copy; Copyright <a href="http://www.rrsd.com">Robert Ramey</a> 2002-2004. 
Distributed under the Boost Software License, Version 1.0. (See
accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
</i></p>
</body>
</html>