doc/html/proto/appendices.html

   1 <html>
   2 <head>
   3 <meta http-equiv="Content-Type" content="text/html; charset=US-ASCII">
   4 <title>Appendices</title>
   5 <link rel="stylesheet" href="../../../doc/src/boostbook.css" type="text/css">
   6 <meta name="generator" content="DocBook XSL Stylesheets V1.78.1">
   7 <link rel="home" href="../index.html" title="The Boost C++ Libraries BoostBook Documentation Subset">
   8 <link rel="up" href="../proto.html" title="Chapter&#160;23.&#160;Boost.Proto">
   9 <link rel="prev" href="../Transform.html" title="Concept Transform">
  10 <link rel="next" href="../boost_random.html" title="Chapter&#160;24.&#160;Boost.Random">
  11 </head>
  12 <body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
  13 <table cellpadding="2" width="100%"><tr>
  14 <td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../boost.png"></td>
  15 <td align="center"><a href="../../../index.html">Home</a></td>
  16 <td align="center"><a href="../../../libs/libraries.htm">Libraries</a></td>
  17 <td align="center"><a href="http://www.boost.org/users/people.html">People</a></td>
  18 <td align="center"><a href="http://www.boost.org/users/faq.html">FAQ</a></td>
  19 <td align="center"><a href="../../../more/index.htm">More</a></td>
  20 </tr></table>
  21 <hr>
  22 <div class="spirit-nav">
  23 <a accesskey="p" href="../Transform.html"><img src="../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../proto.html"><img src="../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="../boost_random.html"><img src="../../../doc/src/images/next.png" alt="Next"></a>
  24 </div>
  25 <div class="section">
  26 <div class="titlepage"><div><div><h2 class="title" style="clear: both">
  27 <a name="proto.appendices"></a><a class="link" href="appendices.html" title="Appendices">Appendices</a>
  28 </h2></div></div></div>
  29 <div class="toc"><dl class="toc">
  30 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.release_notes">Appendix A: Release
  31       Notes</a></span></dt>
  32 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.history">Appendix B: History</a></span></dt>
  33 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.rationale">Appendix C: Rationale</a></span></dt>
  34 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.implementation">Appendix D: Implementation
  35       Notes</a></span></dt>
  36 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.acknowledgements">Appendix E:
  37       Acknowledgements</a></span></dt>
  38 </dl></div>
  39 <div class="section">
  40 <div class="titlepage"><div><div><h3 class="title">
  41 <a name="boost_proto.appendices.release_notes"></a><a class="link" href="appendices.html#boost_proto.appendices.release_notes" title="Appendix A: Release Notes">Appendix A: Release
  42       Notes</a>
  43 </h3></div></div></div>
  44 <h5>
  45 <a name="boost_proto.appendices.release_notes.h0"></a>
  46         <span class="phrase"><a name="boost_proto.appendices.release_notes.boost_1_51"></a></span><a class="link" href="appendices.html#boost_proto.appendices.release_notes.boost_1_51">Boost
  47         1.51</a>
  48       </h5>
  49 <p>
  50         <span class="bold"><strong>Unpacking Expressions</strong></span>
  51       </p>
  52 <p>
  53         In Boost 1.51, Proto got simple unpacking patterns. When working with Proto
  54         transforms, unpacking expressions are useful for unpacking the children of
  55         an expression into a function call or an object constructor, while optionally
  56         applying some transformations to each child in turn.
  57       </p>
  58 <p>
  59         See the <a class="link" href="users_guide.html#boost_proto.users_guide.back_end.expression_transformation.unpacking_expressions" title="Unpacking Expressions">Unpacking
  60         Expressions</a> section for more information.
  61       </p>
  62 <h5>
  63 <a name="boost_proto.appendices.release_notes.h1"></a>
  64         <span class="phrase"><a name="boost_proto.appendices.release_notes.boost_1_44"></a></span><a class="link" href="appendices.html#boost_proto.appendices.release_notes.boost_1_44">Boost
  65         1.44</a>
  66       </h5>
  67 <p>
  68         <span class="bold"><strong>Behavior Change: proto::and_&lt;&gt;</strong></span>
  69       </p>
  70 <p>
  71         In Boost 1.44, the behavior of <code class="computeroutput"><a class="link" href="../boost/proto/and_.html" title="Struct template and_">proto::and_&lt;&gt;</a></code>
  72         as a transform changed. Previously, it only applied the transform associated
  73         with the last grammar in the set. Now, it applies all the transforms but
  74         only returns the result of the last. That makes it behave like C++'s comma
  75         operator. For example, a grammar such as:
  76       </p>
  77 <pre class="programlisting"><span class="identifier">proto</span><span class="special">::</span><span class="identifier">and_</span><span class="special">&lt;</span> <span class="identifier">G0</span><span class="special">,</span> <span class="identifier">G1</span><span class="special">,</span> <span class="identifier">G2</span> <span class="special">&gt;</span>
  78 </pre>
  79 <p>
  80         when evaluated with an expression <code class="computeroutput"><span class="identifier">e</span></code>
  81         now behaves like this:
  82       </p>
  83 <pre class="programlisting"><span class="special">((</span><span class="keyword">void</span><span class="special">)</span><span class="identifier">G0</span><span class="special">()(</span><span class="identifier">e</span><span class="special">),</span> <span class="special">(</span><span class="keyword">void</span><span class="special">)</span><span class="identifier">G1</span><span class="special">()(</span><span class="identifier">e</span><span class="special">),</span> <span class="identifier">G2</span><span class="special">()(</span><span class="identifier">e</span><span class="special">))</span>
  84 </pre>
  85 <div class="note"><table border="0" summary="Note">
  86 <tr>
  87 <td rowspan="2" align="center" valign="top" width="25"><img alt="[Note]" src="../../../doc/src/images/note.png"></td>
  88 <th align="left">Note</th>
  89 </tr>
  90 <tr><td align="left" valign="top"><p>
  91           Why the void casts? It's to avoid argument-dependent lookup, which might
  92           find an overloaded comma operator.
  93         </p></td></tr>
  94 </table></div>
  95 <p>
  96         <span class="bold"><strong>Behavior Change: proto::as_expr() and proto::as_child()</strong></span>
  97       </p>
  98 <p>
  99         The functions <code class="computeroutput"><a class="link" href="../boost/proto/as_expr_idp227626032.html" title="Function as_expr">proto::as_expr()</a></code> and <code class="computeroutput"><a class="link" href="../boost/proto/as_child_idp227648000.html" title="Function as_child">proto::as_child()</a></code>
 100         are used to guarantee that an object is a Proto expression by turning it
 101         into one if it is not already, using an optionally specified domain. In previous
 102         releases, when these functions were passed a Proto expression in a domain
 103         different to the one specified, they would apply the specified domain's generator,
 104         resulting in a twice-wrapped expression. This behavior was surprising to
 105         some users.
 106       </p>
 107 <p>
 108         The new behavior of these two functions is to always leave Proto expressions
 109         alone, regardless of the expressions' domains.
 110       </p>
 111 <p>
 112         <span class="bold"><strong>Behavior Change: proto::(pod_)generator&lt;&gt; and
 113         proto::basic_expr&lt;&gt;</strong></span>
 114       </p>
 115 <p>
 116         Users familiar with Proto's extension mechanism have probably used either
 117         <code class="computeroutput"><a class="link" href="../boost/proto/generator.html" title="Struct template generator">proto::generator&lt;&gt;</a></code> or <code class="computeroutput"><a class="link" href="../boost/proto/pod_generator.html" title="Struct template pod_generator">proto::pod_generator&lt;&gt;</a></code>
 118         with a wrapper template when defining their domain. In the past, Proto would
 119         instantiate your wrapper template with instances of <code class="computeroutput"><a class="link" href="../boost/proto/expr.html" title="Struct template expr">proto::expr&lt;&gt;</a></code>.
 120         In Boost 1.44, Proto now instantiates your wrapper template with instances
 121         of a new type: <code class="computeroutput"><a class="link" href="../boost/proto/basic_expr.html" title="Struct template basic_expr">proto::basic_expr&lt;&gt;</a></code>.
 122       </p>
 123 <p>
 124         For instance:
 125       </p>
 126 <pre class="programlisting"><span class="comment">// An expression wrapper</span>
 127 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">Expr</span><span class="special">&gt;</span>
 128 <span class="keyword">struct</span> <span class="identifier">my_expr_wrapper</span><span class="special">;</span>
 129
 130 <span class="comment">// A domain</span>
 131 <span class="keyword">struct</span> <span class="identifier">my_domain</span>
 132   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">&lt;</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">generator</span><span class="special">&lt;</span> <span class="identifier">my_expr_wrapper</span> <span class="special">&gt;</span> <span class="special">&gt;</span>
 133 <span class="special">{};</span>
 134
 135 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">Expr</span><span class="special">&gt;</span>
 136 <span class="keyword">struct</span> <span class="identifier">my_expr_wrapper</span>
 137   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr_wrapper</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">&gt;,</span> <span class="identifier">my_domain</span><span class="special">&gt;</span>
 138 <span class="special">{</span>
 139   <span class="comment">// Before 1.44, Expr was an instance of proto::expr&lt;&gt;</span>
 140   <span class="comment">// In 1.44, Expr is an instance of proto::basic_expr&lt;&gt;</span>
 141 <span class="special">};</span>
 142 </pre>
 143 <p>
 144         The motivation for this change was to improve compile times. <code class="computeroutput"><a class="link" href="../boost/proto/expr.html" title="Struct template expr">proto::expr&lt;&gt;</a></code>
 145         is an expensive type to instantiate because it defines a host of member functions.
 146         When defining your own expression wrapper, the instance of <code class="computeroutput"><a class="link" href="../boost/proto/expr.html" title="Struct template expr">proto::expr&lt;&gt;</a></code>
 147         sits as a hidden data member function in your wrapper and the members of
 148         <code class="computeroutput"><a class="link" href="../boost/proto/expr.html" title="Struct template expr">proto::expr&lt;&gt;</a></code> go unused. Therefore,
 149         the cost of those member functions is wasted. In contrast, <code class="computeroutput"><a class="link" href="../boost/proto/basic_expr.html" title="Struct template basic_expr">proto::basic_expr&lt;&gt;</a></code>
 150         is a very lightweight type with no member functions at all.
 151       </p>
 152 <p>
 153         The vast majority of programs should recompile without any source changes.
 154         However, if somewhere you are assuming that you will be given instances specifically
 155         of <code class="computeroutput"><a class="link" href="../boost/proto/expr.html" title="Struct template expr">proto::expr&lt;&gt;</a></code>, your code will break.
 156       </p>
 157 <p>
 158         <span class="bold"><strong>New Feature: Sub-domains</strong></span>
 159       </p>
 160 <p>
 161         In Boost 1.44, Proto introduces an important new feature called "sub-domains".
 162         This gives you a way to spcify that one domain is compatible with another
 163         such that expressions in one domain can be freely mixed with expressions
 164         in another. You can define one domain to be the sub-domain of another by
 165         using the third template parameter of <code class="computeroutput"><a class="link" href="../boost/proto/domain.html" title="Struct template domain">proto::domain&lt;&gt;</a></code>.
 166       </p>
 167 <p>
 168         For instance:
 169       </p>
 170 <pre class="programlisting"><span class="comment">// Not shown: define some expression</span>
 171 <span class="comment">// generators genA and genB</span>
 172
 173 <span class="keyword">struct</span> <span class="identifier">A</span>
 174   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">&lt;</span> <span class="identifier">genA</span><span class="special">,</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">_</span> <span class="special">&gt;</span>
 175 <span class="special">{};</span>
 176
 177 <span class="comment">// Define a domain B that is the sub-domain</span>
 178 <span class="comment">// of domain A.</span>
 179 <span class="keyword">struct</span> <span class="identifier">B</span>
 180   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">&lt;</span> <span class="identifier">genB</span><span class="special">,</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">_</span><span class="special">,</span> <span class="identifier">A</span> <span class="special">&gt;</span>
 181 <span class="special">{};</span>
 182 </pre>
 183 <p>
 184         Expressions in domains <code class="computeroutput"><span class="identifier">A</span></code>
 185         and <code class="computeroutput"><span class="identifier">B</span></code> can have different
 186         wrappers (hence, different interfaces), but they can be combined into larger
 187         expressions. Without a sub-domain relationship, this would have been an error.
 188         The domain of the resulting expression in this case would be <code class="computeroutput"><span class="identifier">A</span></code>.
 189       </p>
 190 <p>
 191         The complete description of sub-domains can be found in the reference sections
 192         for <code class="computeroutput"><a class="link" href="../boost/proto/domain.html" title="Struct template domain">proto::domain&lt;&gt;</a></code> and <code class="computeroutput"><a class="link" href="../boost/proto/deduce_domain.html" title="Struct deduce_domain">proto::deduce_domain</a></code>.
 193       </p>
 194 <p>
 195         <span class="bold"><strong>New Feature: Domain-specific as_expr() and as_child()</strong></span>
 196       </p>
 197 <p>
 198         Proto has always allowed users to customize expressions post-hoc by specifying
 199         a Generator when defining their domain. But it has never allowed users to
 200         control how Proto assembles sub-expressions in the first place. As of Boost
 201         1.44, users now have this power.
 202       </p>
 203 <p>
 204         Users defining their own domain can now specify how <code class="computeroutput"><a class="link" href="../boost/proto/as_expr_idp227626032.html" title="Function as_expr">proto::as_expr()</a></code>
 205         and <code class="computeroutput"><a class="link" href="../boost/proto/as_child_idp227648000.html" title="Function as_child">proto::as_child()</a></code> work in their domain. They
 206         can do this easily by defining nested class templates named <code class="computeroutput"><span class="identifier">as_expr</span></code> and/or <code class="computeroutput"><span class="identifier">as_child</span></code>
 207         within their domain class.
 208       </p>
 209 <p>
 210         For example:
 211       </p>
 212 <pre class="programlisting"><span class="keyword">struct</span> <span class="identifier">my_domain</span>
 213   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">&lt;</span> <span class="identifier">my_generator</span> <span class="special">&gt;</span>
 214 <span class="special">{</span>
 215   <span class="keyword">typedef</span>
 216       <span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">&lt;</span> <span class="identifier">my_generator</span> <span class="special">&gt;</span>
 217   <span class="identifier">base_domain</span><span class="special">;</span>
 218
 219   <span class="comment">// For my_domain, as_child does the same as</span>
 220   <span class="comment">// what as_expr does by default.</span>
 221   <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">T</span><span class="special">&gt;</span>
 222   <span class="keyword">struct</span> <span class="identifier">as_child</span>
 223     <span class="special">:</span> <span class="identifier">base_domain</span><span class="special">::</span><span class="identifier">as_expr</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">&gt;</span>
 224   <span class="special">{};</span>
 225 <span class="special">};</span>
 226 </pre>
 227 <p>
 228         In the above example, <code class="computeroutput"><span class="identifier">my_domain</span><span class="special">::</span><span class="identifier">as_child</span><span class="special">&lt;&gt;</span></code> simply defers to <code class="computeroutput"><span class="identifier">proto</span><span class="special">::</span><span class="identifier">domain</span><span class="special">::</span><span class="identifier">as_expr</span><span class="special">&lt;&gt;</span></code>. This has the nice effect of causing
 229         all terminals to be captured by value instead of by reference, and to likewise
 230         store child expressions by value. The result is that expressions in <code class="computeroutput"><span class="identifier">my_domain</span></code> are safe to store in <code class="computeroutput"><span class="keyword">auto</span></code> variables because they will not have
 231         dangling references to intermediate temporary expressions. (Naturally, it
 232         also means that expression construction has extra runtime overhead of copying
 233         that the compiler may or may not be able to optimize away.)
 234       </p>
 235 <h5>
 236 <a name="boost_proto.appendices.release_notes.h2"></a>
 237         <span class="phrase"><a name="boost_proto.appendices.release_notes.boost_1_43"></a></span><a class="link" href="appendices.html#boost_proto.appendices.release_notes.boost_1_43">Boost
 238         1.43</a>
 239       </h5>
 240 <p>
 241         In Boost 1.43, the recommended usage of <code class="computeroutput"><a class="link" href="../boost/proto/extends.html" title="Struct template extends">proto::extends&lt;&gt;</a></code>
 242         changed slightly. The new usage looks like this:
 243       </p>
 244 <pre class="programlisting"><span class="comment">// my_expr is an expression extension of the Expr parameter</span>
 245 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Expr</span><span class="special">&gt;</span>
 246 <span class="keyword">struct</span> <span class="identifier">my_expr</span>
 247   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">&gt;,</span> <span class="identifier">my_domain</span><span class="special">&gt;</span>
 248 <span class="special">{</span>
 249     <span class="identifier">my_expr</span><span class="special">(</span><span class="identifier">Expr</span> <span class="keyword">const</span> <span class="special">&amp;</span><span class="identifier">expr</span> <span class="special">=</span> <span class="identifier">Expr</span><span class="special">())</span>
 250       <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr</span><span class="special">,</span> <span class="identifier">my_domain</span><span class="special">&gt;(</span><span class="identifier">expr</span><span class="special">)</span>
 251     <span class="special">{}</span>
 252
 253     <span class="comment">// NEW: use the following macro to bring</span>
 254     <span class="comment">// proto::extends::operator= into scope.</span>
 255     <span class="identifier">BOOST_PROTO_EXTENDS_USING_ASSIGN</span><span class="special">(</span><span class="identifier">my_expr</span><span class="special">)</span>
 256 <span class="special">};</span>
 257 </pre>
 258 <p>
 259         The new thing is the use of the <code class="literal"><code class="computeroutput"><a class="link" href="../BOOST_PROTO_EXTENDS_USING_ASSIGN.html" title="Macro BOOST_PROTO_EXTENDS_USING_ASSIGN">BOOST_PROTO_EXTENDS_USING_ASSIGN</a></code>()</code>
 260         macro. To allow assignment operators to build expression trees, <code class="computeroutput"><a class="link" href="../boost/proto/extends.html" title="Struct template extends">proto::extends&lt;&gt;</a></code> overloads the assignment
 261         operator. However, for the <code class="computeroutput"><span class="identifier">my_expr</span></code>
 262         template, the compiler generates a default copy assignment operator that
 263         hides the ones in <code class="computeroutput"><a class="link" href="../boost/proto/extends.html" title="Struct template extends">proto::extends&lt;&gt;</a></code>. This is often not desired
 264         (although it depends on the syntax you want to allow).
 265       </p>
 266 <p>
 267         Previously, the recommended usage was to do this:
 268       </p>
 269 <pre class="programlisting"><span class="comment">// my_expr is an expression extension of the Expr parameter</span>
 270 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Expr</span><span class="special">&gt;</span>
 271 <span class="keyword">struct</span> <span class="identifier">my_expr</span>
 272   <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">&gt;,</span> <span class="identifier">my_domain</span><span class="special">&gt;</span>
 273 <span class="special">{</span>
 274     <span class="identifier">my_expr</span><span class="special">(</span><span class="identifier">Expr</span> <span class="keyword">const</span> <span class="special">&amp;</span><span class="identifier">expr</span> <span class="special">=</span> <span class="identifier">Expr</span><span class="special">())</span>
 275       <span class="special">:</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr</span><span class="special">,</span> <span class="identifier">my_domain</span><span class="special">&gt;(</span><span class="identifier">expr</span><span class="special">)</span>
 276     <span class="special">{}</span>
 277
 278     <span class="comment">// OLD: don't do it like this anymore.</span>
 279     <span class="keyword">using</span> <span class="identifier">proto</span><span class="special">::</span><span class="identifier">extends</span><span class="special">&lt;</span><span class="identifier">Expr</span><span class="special">,</span> <span class="identifier">my_expr</span><span class="special">,</span> <span class="identifier">my_domain</span><span class="special">&gt;::</span><span class="keyword">operator</span><span class="special">=;</span>
 280 <span class="special">};</span>
 281 </pre>
 282 <p>
 283         While this works in the majority of cases, it still doesn't suppress the
 284         implicit generation of the default assignment operator. As a result, expressions
 285         of the form <code class="computeroutput"><span class="identifier">a</span> <span class="special">=</span>
 286         <span class="identifier">b</span></code> could either build an expression
 287         template or do a copy assignment depending on whether the types of <code class="computeroutput"><span class="identifier">a</span></code> and <code class="computeroutput"><span class="identifier">b</span></code>
 288         happen to be the same. That can lead to subtle bugs, so the behavior was
 289         changed.
 290       </p>
 291 <p>
 292         The <code class="literal"><code class="computeroutput"><a class="link" href="../BOOST_PROTO_EXTENDS_USING_ASSIGN.html" title="Macro BOOST_PROTO_EXTENDS_USING_ASSIGN">BOOST_PROTO_EXTENDS_USING_ASSIGN</a></code>()</code>
 293         brings into scope the assignment operators defined in <code class="computeroutput"><a class="link" href="../boost/proto/extends.html" title="Struct template extends">proto::extends&lt;&gt;</a></code>
 294         as well as suppresses the generation of the copy assignment operator.
 295       </p>
 296 <p>
 297         Also note that the <code class="computeroutput"><a class="link" href="../boost/proto/literal.html" title="Struct template literal">proto::literal&lt;&gt;</a></code> class template, which
 298         uses <code class="computeroutput"><a class="link" href="../boost/proto/extends.html" title="Struct template extends">proto::extends&lt;&gt;</a></code>, has been chaged to use
 299         <code class="literal"><code class="computeroutput"><a class="link" href="../BOOST_PROTO_EXTENDS_USING_ASSIGN.html" title="Macro BOOST_PROTO_EXTENDS_USING_ASSIGN">BOOST_PROTO_EXTENDS_USING_ASSIGN</a></code>()</code>.
 300         The implications are highlighted in the sample code below:
 301       </p>
 302 <pre class="programlisting"><span class="identifier">proto</span><span class="special">::</span><span class="identifier">literal</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span> <span class="identifier">a</span><span class="special">(</span><span class="number">1</span><span class="special">),</span> <span class="identifier">b</span><span class="special">(</span><span class="number">2</span><span class="special">);</span> <span class="comment">// two non-const proto literals</span>
 303 <span class="identifier">proto</span><span class="special">::</span><span class="identifier">literal</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span> <span class="keyword">const</span> <span class="identifier">c</span><span class="special">(</span><span class="number">3</span><span class="special">);</span> <span class="comment">// a const proto literal</span>
 304
 305 <span class="identifier">a</span> <span class="special">=</span> <span class="identifier">b</span><span class="special">;</span> <span class="comment">// No-op. Builds an expression tree and discards it.</span>
 306        <span class="comment">// Same behavior in 1.42 and 1.43.</span>
 307
 308 <span class="identifier">a</span> <span class="special">=</span> <span class="identifier">c</span><span class="special">;</span> <span class="comment">// CHANGE! In 1.42, this performed copy assignment, causing</span>
 309        <span class="comment">// a's value to change to 3. In 1.43, the behavior is now</span>
 310        <span class="comment">// the same as above: build and discard an expression tree.</span>
 311 </pre>
 312 </div>
 313 <div class="section">
 314 <div class="titlepage"><div><div><h3 class="title">
 315 <a name="boost_proto.appendices.history"></a><a class="link" href="appendices.html#boost_proto.appendices.history" title="Appendix B: History">Appendix B: History</a>
 316 </h3></div></div></div>
 317 <div class="variablelist">
 318 <p class="title"><b></b></p>
 319 <dl class="variablelist">
 320 <dt><span class="term">August 13, 2010</span></dt>
 321 <dd><p>
 322               Boost 1.44: Proto gets sub-domains and per-domain control of <code class="computeroutput"><a class="link" href="../boost/proto/as_expr_idp227626032.html" title="Function as_expr">proto::as_expr()</a></code> and <code class="computeroutput"><a class="link" href="../boost/proto/as_child_idp227648000.html" title="Function as_child">proto::as_child()</a></code> to meet the needs
 323               of Phoenix3.
 324             </p></dd>
 325 <dt><span class="term">August 11, 2008</span></dt>
 326 <dd><p>
 327               Proto v4 is merged to Boost trunk with more powerful transform protocol.
 328             </p></dd>
 329 <dt><span class="term">April 7, 2008</span></dt>
 330 <dd><p>
 331               Proto is accepted into Boost.
 332             </p></dd>
 333 <dt><span class="term">March 1, 2008</span></dt>
 334 <dd><p>
 335               Proto's Boost review begins.
 336             </p></dd>
 337 <dt><span class="term">January 11, 2008</span></dt>
 338 <dd><p>
 339               Boost.Proto v3 brings separation of grammars and transforms and a "round"
 340               lambda syntax for defining transforms in-place.
 341             </p></dd>
 342 <dt><span class="term">April 15, 2007</span></dt>
 343 <dd><p>
 344               Boost.Xpressive is ported from Proto compilers to Proto transforms.
 345               Support for old Proto compilers is dropped.
 346             </p></dd>
 347 <dt><span class="term">April 4, 2007</span></dt>
 348 <dd><p>
 349               Preliminary submission of Proto to Boost.
 350             </p></dd>
 351 <dt><span class="term">December 11, 2006</span></dt>
 352 <dd><p>
 353               The idea for transforms that decorate grammar rules is born in a private
 354               email discussion with Joel de Guzman and Hartmut Kaiser. The first
 355               transforms are committed to CVS 5 days later on December 16.
 356             </p></dd>
 357 <dt><span class="term">November 1, 2006</span></dt>
 358 <dd><p>
 359               The idea for <code class="computeroutput"><span class="identifier">proto</span><span class="special">::</span><span class="identifier">matches</span><span class="special">&lt;&gt;</span></code> and the whole grammar facility
 360               is hatched during a discussion with Hartmut Kaiser on the spirit-devel
 361               list. The first version of <code class="computeroutput"><span class="identifier">proto</span><span class="special">::</span><span class="identifier">matches</span><span class="special">&lt;&gt;</span></code> is checked into CVS 3 days
 362               later. Message is <a href="http://osdir.com/ml/parsers.spirit.devel/2006-11/msg00003.html" target="_top">here</a>.
 363             </p></dd>
 364 <dt><span class="term">October 28, 2006</span></dt>
 365 <dd><p>
 366               Proto is reborn, this time with a uniform expression types that are
 367               POD. Announcement is <a href="http://lists.boost.org/Archives/boost/2006/10/112453.php" target="_top">here</a>.
 368             </p></dd>
 369 <dt><span class="term">April 20, 2005</span></dt>
 370 <dd><p>
 371               Proto is born as a major refactorization of Boost.Xpressive's meta-programming.
 372               Proto offers expression types, operator overloads and "compilers",
 373               an early formulation of what later became transforms. Announcement
 374               is <a href="http://lists.boost.org/Archives/boost/2005/04/85256.php" target="_top">here</a>.
 375             </p></dd>
 376 </dl>
 377 </div>
 378 </div>
 379 <div class="section">
 380 <div class="titlepage"><div><div><h3 class="title">
 381 <a name="boost_proto.appendices.rationale"></a><a class="link" href="appendices.html#boost_proto.appendices.rationale" title="Appendix C: Rationale">Appendix C: Rationale</a>
 382 </h3></div></div></div>
 383 <div class="toc"><dl class="toc">
 384 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.rationale.static_initialization">Static
 385         Initialization</a></span></dt>
 386 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.rationale.preprocessor">Why
 387         Not Reuse MPL, Fusion, et cetera?</a></span></dt>
 388 </dl></div>
 389 <div class="section">
 390 <div class="titlepage"><div><div><h4 class="title">
 391 <a name="boost_proto.appendices.rationale.static_initialization"></a><a class="link" href="appendices.html#boost_proto.appendices.rationale.static_initialization" title="Static Initialization">Static
 392         Initialization</a>
 393 </h4></div></div></div>
 394 <p>
 395           Proto expression types are PODs (Plain Old Data), and do not have constructors.
 396           They are brace-initialized, as follows:
 397         </p>
 398 <pre class="programlisting"><span class="identifier">terminal</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;::</span><span class="identifier">type</span> <span class="keyword">const</span> <span class="identifier">_i</span> <span class="special">=</span> <span class="special">{</span><span class="number">1</span><span class="special">};</span>
 399 </pre>
 400 <p>
 401           The reason is so that expression objects like <code class="computeroutput"><span class="identifier">_i</span></code>
 402           above can be <span class="emphasis"><em>statically initialized</em></span>. Why is static
 403           initialization important? The terminals of many embedded domain-specific
 404           languages are likely to be global const objects, like <code class="computeroutput"><span class="identifier">_1</span></code>
 405           and <code class="computeroutput"><span class="identifier">_2</span></code> from the Boost Lambda
 406           Library. Were these object to require run-time initialization, it might
 407           be possible to use these objects before they are initialized. That would
 408           be bad. Statically initialized objects cannot be misused that way.
 409         </p>
 410 </div>
 411 <div class="section">
 412 <div class="titlepage"><div><div><h4 class="title">
 413 <a name="boost_proto.appendices.rationale.preprocessor"></a><a class="link" href="appendices.html#boost_proto.appendices.rationale.preprocessor" title="Why Not Reuse MPL, Fusion, et cetera?">Why
 414         Not Reuse MPL, Fusion, et cetera?</a>
 415 </h4></div></div></div>
 416 <p>
 417           Anyone who has peeked at Proto's source code has probably wondered, "Why
 418           all the dirty preprocessor gunk? Couldn't this have been all implemented
 419           cleanly on top of libraries like MPL and Fusion?" The answer is that
 420           Proto could have been implemented this way, and in fact was at one point.
 421           The problem is that template metaprogramming (TMP) makes for longer compile
 422           times. As a foundation upon which other TMP-heavy libraries will be built,
 423           Proto itself should be as lightweight as possible. That is achieved by
 424           prefering preprocessor metaprogramming to template metaprogramming. Expanding
 425           a macro is far more efficient than instantiating a template. In some cases,
 426           the "clean" version takes 10x longer to compile than the "dirty"
 427           version.
 428         </p>
 429 <p>
 430           The "clean and slow" version of Proto can still be found at http://svn.boost.org/svn/boost/branches/proto/v3.
 431           Anyone who is interested can download it and verify that it is, in fact,
 432           unusably slow to compile. Note that this branch's development was abandoned,
 433           and it does not conform exactly with Proto's current interface.
 434         </p>
 435 </div>
 436 </div>
 437 <div class="section">
 438 <div class="titlepage"><div><div><h3 class="title">
 439 <a name="boost_proto.appendices.implementation"></a><a class="link" href="appendices.html#boost_proto.appendices.implementation" title="Appendix D: Implementation Notes">Appendix D: Implementation
 440       Notes</a>
 441 </h3></div></div></div>
 442 <div class="toc"><dl class="toc">
 443 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.implementation.sfinae">Quick-n-Dirty
 444         Type Categorization</a></span></dt>
 445 <dt><span class="section"><a href="appendices.html#boost_proto.appendices.implementation.function_arity">Detecting
 446         the Arity of Function Objects</a></span></dt>
 447 </dl></div>
 448 <div class="section">
 449 <div class="titlepage"><div><div><h4 class="title">
 450 <a name="boost_proto.appendices.implementation.sfinae"></a><a class="link" href="appendices.html#boost_proto.appendices.implementation.sfinae" title="Quick-n-Dirty Type Categorization">Quick-n-Dirty
 451         Type Categorization</a>
 452 </h4></div></div></div>
 453 <p>
 454           Much has already been written about dispatching on type traits using SFINAE
 455           (Substitution Failure Is Not An Error) techniques in C++. There is a Boost
 456           library, Boost.Enable_if, to make the technique idiomatic. Proto dispatches
 457           on type traits extensively, but it doesn't use <code class="computeroutput"><span class="identifier">enable_if</span><span class="special">&lt;&gt;</span></code> very often. Rather, it dispatches
 458           based on the presence or absence of nested types, often typedefs for void.
 459         </p>
 460 <p>
 461           Consider the implementation of <code class="computeroutput"><span class="identifier">is_expr</span><span class="special">&lt;&gt;</span></code>. It could have been written as
 462           something like this:
 463         </p>
 464 <pre class="programlisting"><span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T</span><span class="special">&gt;</span>
 465 <span class="keyword">struct</span> <span class="identifier">is_expr</span>
 466   <span class="special">:</span> <span class="identifier">is_base_and_derived</span><span class="special">&lt;</span><span class="identifier">proto</span><span class="special">::</span><span class="identifier">some_expr_base</span><span class="special">,</span> <span class="identifier">T</span><span class="special">&gt;</span>
 467 <span class="special">{};</span>
 468 </pre>
 469 <p>
 470           Rather, it is implemented as this:
 471         </p>
 472 <pre class="programlisting"><span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">Void</span> <span class="special">=</span> <span class="keyword">void</span><span class="special">&gt;</span>
 473 <span class="keyword">struct</span> <span class="identifier">is_expr</span>
 474   <span class="special">:</span> <span class="identifier">mpl</span><span class="special">::</span><span class="identifier">false_</span>
 475 <span class="special">{};</span>
 476
 477 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T</span><span class="special">&gt;</span>
 478 <span class="keyword">struct</span> <span class="identifier">is_expr</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">T</span><span class="special">::</span><span class="identifier">proto_is_expr_</span><span class="special">&gt;</span>
 479   <span class="special">:</span> <span class="identifier">mpl</span><span class="special">::</span><span class="identifier">true_</span>
 480 <span class="special">{};</span>
 481 </pre>
 482 <p>
 483           This relies on the fact that the specialization will be preferred if <code class="computeroutput"><span class="identifier">T</span></code> has a nested <code class="computeroutput"><span class="identifier">proto_is_expr_</span></code>
 484           that is a typedef for <code class="computeroutput"><span class="keyword">void</span></code>.
 485           All Proto expression types have such a nested typedef.
 486         </p>
 487 <p>
 488           Why does Proto do it this way? The reason is because, after running extensive
 489           benchmarks while trying to improve compile times, I have found that this
 490           approach compiles faster. It requires exactly one template instantiation.
 491           The other approach requires at least 2: <code class="computeroutput"><span class="identifier">is_expr</span><span class="special">&lt;&gt;</span></code> and <code class="computeroutput"><span class="identifier">is_base_and_derived</span><span class="special">&lt;&gt;</span></code>, plus whatever templates <code class="computeroutput"><span class="identifier">is_base_and_derived</span><span class="special">&lt;&gt;</span></code>
 492           may instantiate.
 493         </p>
 494 </div>
 495 <div class="section">
 496 <div class="titlepage"><div><div><h4 class="title">
 497 <a name="boost_proto.appendices.implementation.function_arity"></a><a class="link" href="appendices.html#boost_proto.appendices.implementation.function_arity" title="Detecting the Arity of Function Objects">Detecting
 498         the Arity of Function Objects</a>
 499 </h4></div></div></div>
 500 <p>
 501           In several places, Proto needs to know whether or not a function object
 502           <code class="computeroutput"><span class="identifier">Fun</span></code> can be called with
 503           certain parameters and take a fallback action if not. This happens in
 504           <code class="computeroutput"><a class="link" href="../boost/proto/context/callable_context.html" title="Struct template callable_context">proto::callable_context&lt;&gt;</a></code>
 505           and in the <code class="computeroutput"><a class="link" href="../boost/proto/call.html" title="Struct template call">proto::call&lt;&gt;</a></code> transform. How does
 506           Proto know? It involves some tricky metaprogramming. Here's how.
 507         </p>
 508 <p>
 509           Another way of framing the question is by trying to implement the following
 510           <code class="computeroutput"><span class="identifier">can_be_called</span><span class="special">&lt;&gt;</span></code>
 511           Boolean metafunction, which checks to see if a function object <code class="computeroutput"><span class="identifier">Fun</span></code> can be called with parameters of
 512           type <code class="computeroutput"><span class="identifier">A</span></code> and <code class="computeroutput"><span class="identifier">B</span></code>:
 513         </p>
 514 <pre class="programlisting"><span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Fun</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">A</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">B</span><span class="special">&gt;</span>
 515 <span class="keyword">struct</span> <span class="identifier">can_be_called</span><span class="special">;</span>
 516 </pre>
 517 <p>
 518           First, we define the following <code class="computeroutput"><span class="identifier">dont_care</span></code>
 519           struct, which has an implicit conversion from anything. And not just any
 520           implicit conversion; it has a ellipsis conversion, which is the worst possible
 521           conversion for the purposes of overload resolution:
 522         </p>
 523 <pre class="programlisting"><span class="keyword">struct</span> <span class="identifier">dont_care</span>
 524 <span class="special">{</span>
 525     <span class="identifier">dont_care</span><span class="special">(...);</span>
 526 <span class="special">};</span>
 527 </pre>
 528 <p>
 529           We also need some private type known only to us with an overloaded comma
 530           operator (!), and some functions that detect the presence of this type
 531           and return types with different sizes, as follows:
 532         </p>
 533 <pre class="programlisting"><span class="keyword">struct</span> <span class="identifier">private_type</span>
 534 <span class="special">{</span>
 535     <span class="identifier">private_type</span> <span class="keyword">const</span> <span class="special">&amp;</span><span class="keyword">operator</span><span class="special">,(</span><span class="keyword">int</span><span class="special">)</span> <span class="keyword">const</span><span class="special">;</span>
 536 <span class="special">};</span>
 537
 538 <span class="keyword">typedef</span> <span class="keyword">char</span> <span class="identifier">yes_type</span><span class="special">;</span>      <span class="comment">// sizeof(yes_type) == 1</span>
 539 <span class="keyword">typedef</span> <span class="keyword">char</span> <span class="special">(&amp;</span><span class="identifier">no_type</span><span class="special">)[</span><span class="number">2</span><span class="special">];</span> <span class="comment">// sizeof(no_type)  == 2</span>
 540
 541 <span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T</span><span class="special">&gt;</span>
 542 <span class="identifier">no_type</span> <span class="identifier">is_private_type</span><span class="special">(</span><span class="identifier">T</span> <span class="keyword">const</span> <span class="special">&amp;);</span>
 543
 544 <span class="identifier">yes_type</span> <span class="identifier">is_private_type</span><span class="special">(</span><span class="identifier">private_type</span> <span class="keyword">const</span> <span class="special">&amp;);</span>
 545 </pre>
 546 <p>
 547           Next, we implement a binary function object wrapper with a very strange
 548           conversion operator, whose meaning will become clear later.
 549         </p>
 550 <pre class="programlisting"><span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Fun</span><span class="special">&gt;</span>
 551 <span class="keyword">struct</span> <span class="identifier">funwrap2</span> <span class="special">:</span> <span class="identifier">Fun</span>
 552 <span class="special">{</span>
 553     <span class="identifier">funwrap2</span><span class="special">();</span>
 554     <span class="keyword">typedef</span> <span class="identifier">private_type</span> <span class="keyword">const</span> <span class="special">&amp;(*</span><span class="identifier">pointer_to_function</span><span class="special">)(</span><span class="identifier">dont_care</span><span class="special">,</span> <span class="identifier">dont_care</span><span class="special">);</span>
 555     <span class="keyword">operator</span> <span class="identifier">pointer_to_function</span><span class="special">()</span> <span class="keyword">const</span><span class="special">;</span>
 556 <span class="special">};</span>
 557 </pre>
 558 <p>
 559           With all of these bits and pieces, we can implement <code class="computeroutput"><span class="identifier">can_be_called</span><span class="special">&lt;&gt;</span></code> as follows:
 560         </p>
 561 <pre class="programlisting"><span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Fun</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">A</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">B</span><span class="special">&gt;</span>
 562 <span class="keyword">struct</span> <span class="identifier">can_be_called</span>
 563 <span class="special">{</span>
 564     <span class="keyword">static</span> <span class="identifier">funwrap2</span><span class="special">&lt;</span><span class="identifier">Fun</span><span class="special">&gt;</span> <span class="special">&amp;</span><span class="identifier">fun</span><span class="special">;</span>
 565     <span class="keyword">static</span> <span class="identifier">A</span> <span class="special">&amp;</span><span class="identifier">a</span><span class="special">;</span>
 566     <span class="keyword">static</span> <span class="identifier">B</span> <span class="special">&amp;</span><span class="identifier">b</span><span class="special">;</span>
 567
 568     <span class="keyword">static</span> <span class="keyword">bool</span> <span class="keyword">const</span> <span class="identifier">value</span> <span class="special">=</span> <span class="special">(</span>
 569         <span class="keyword">sizeof</span><span class="special">(</span><span class="identifier">no_type</span><span class="special">)</span> <span class="special">==</span> <span class="keyword">sizeof</span><span class="special">(</span><span class="identifier">is_private_type</span><span class="special">(</span> <span class="special">(</span><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span><span class="identifier">b</span><span class="special">),</span> <span class="number">0</span><span class="special">)</span> <span class="special">))</span>
 570     <span class="special">);</span>
 571
 572     <span class="keyword">typedef</span> <span class="identifier">mpl</span><span class="special">::</span><span class="identifier">bool_</span><span class="special">&lt;</span><span class="identifier">value</span><span class="special">&gt;</span> <span class="identifier">type</span><span class="special">;</span>
 573 <span class="special">};</span>
 574 </pre>
 575 <p>
 576           The idea is to make it so that <code class="computeroutput"><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span><span class="identifier">b</span><span class="special">)</span></code> will
 577           always compile by adding our own binary function overload, but doing it
 578           in such a way that we can detect whether our overload was selected or not.
 579           And we rig it so that our overload is selected if there is really no better
 580           option. What follows is a description of how <code class="computeroutput"><span class="identifier">can_be_called</span><span class="special">&lt;&gt;</span></code> works.
 581         </p>
 582 <p>
 583           We wrap <code class="computeroutput"><span class="identifier">Fun</span></code> in a type that
 584           has an implicit conversion to a pointer to a binary function. An object
 585           <code class="computeroutput"><span class="identifier">fun</span></code> of class type can be
 586           invoked as <code class="computeroutput"><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">)</span></code> if it has such a conversion operator,
 587           but since it involves a user-defined conversion operator, it is less preferred
 588           than an overloaded <code class="computeroutput"><span class="keyword">operator</span><span class="special">()</span></code>, which requires no such conversion.
 589         </p>
 590 <p>
 591           The function pointer can accept any two arguments by virtue of the <code class="computeroutput"><span class="identifier">dont_care</span></code> type. The conversion sequence
 592           for each argument is guaranteed to be the worst possible conversion sequence:
 593           an implicit conversion through an ellipsis, and a user-defined conversion
 594           to <code class="computeroutput"><span class="identifier">dont_care</span></code>. In total,
 595           it means that <code class="computeroutput"><span class="identifier">funwrap2</span><span class="special">&lt;</span><span class="identifier">Fun</span><span class="special">&gt;()(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">)</span></code>
 596           will always compile, but it will select our overload only if there really
 597           is no better option.
 598         </p>
 599 <p>
 600           If there is a better option --- for example if <code class="computeroutput"><span class="identifier">Fun</span></code>
 601           has an overloaded function call operator such as <code class="computeroutput"><span class="keyword">void</span>
 602           <span class="keyword">operator</span><span class="special">()(</span><span class="identifier">A</span> <span class="identifier">a</span><span class="special">,</span> <span class="identifier">B</span> <span class="identifier">b</span><span class="special">)</span></code> ---
 603           then <code class="computeroutput"><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">)</span></code> will resolve to that one instead. The
 604           question now is how to detect which function got picked by overload resolution.
 605         </p>
 606 <p>
 607           Notice how <code class="computeroutput"><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">)</span></code> appears in <code class="computeroutput"><span class="identifier">can_be_called</span><span class="special">&lt;&gt;</span></code>: <code class="computeroutput"><span class="special">(</span><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">),</span> <span class="number">0</span><span class="special">)</span></code>.
 608           Why do we use the comma operator there? The reason is because we are using
 609           this expression as the argument to a function. If the return type of <code class="computeroutput"><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span> <span class="identifier">b</span><span class="special">)</span></code> is <code class="computeroutput"><span class="keyword">void</span></code>,
 610           it cannot legally be used as an argument to a function. The comma operator
 611           sidesteps the issue.
 612         </p>
 613 <p>
 614           This should also make plain the purpose of the overloaded comma operator
 615           in <code class="computeroutput"><span class="identifier">private_type</span></code>. The return
 616           type of the pointer to function is <code class="computeroutput"><span class="identifier">private_type</span></code>.
 617           If overload resolution selects our overload, then the type of <code class="computeroutput"><span class="special">(</span><span class="identifier">fun</span><span class="special">(</span><span class="identifier">a</span><span class="special">,</span>
 618           <span class="identifier">b</span><span class="special">),</span>
 619           <span class="number">0</span><span class="special">)</span></code>
 620           is <code class="computeroutput"><span class="identifier">private_type</span></code>. Otherwise,
 621           it is <code class="computeroutput"><span class="keyword">int</span></code>. That fact is used
 622           to dispatch to either overload of <code class="computeroutput"><span class="identifier">is_private_type</span><span class="special">()</span></code>, which encodes its answer in the size
 623           of its return type.
 624         </p>
 625 <p>
 626           That's how it works with binary functions. Now repeat the above process
 627           for functions up to some predefined function arity, and you're done.
 628         </p>
 629 </div>
 630 </div>
 631 <div class="section">
 632 <div class="titlepage"><div><div><h3 class="title">
 633 <a name="boost_proto.appendices.acknowledgements"></a><a class="link" href="appendices.html#boost_proto.appendices.acknowledgements" title="Appendix E: Acknowledgements">Appendix E:
 634       Acknowledgements</a>
 635 </h3></div></div></div>
 636 <p>
 637         I'd like to thank Joel de Guzman and Hartmut Kaiser for being willing to
 638         take a chance on using Proto for their work on Spirit-2 and Karma when Proto
 639         was little more than a vision. Their requirements and feedback have been
 640         indespensable.
 641       </p>
 642 <p>
 643         Thanks also to Thomas Heller and again to Hartmut for their feedback and
 644         suggestions during the redesign of Phoenix. That effort yielded several valuable
 645         advanced features such as sub-domains, external transforms, and per-domain
 646         <code class="computeroutput"><span class="identifier">as_child</span></code> customization.
 647       </p>
 648 <p>
 649         Thanks to Daniel James for providing a patch to remove the dependence on
 650         deprecated configuration macros for C++0x features.
 651       </p>
 652 <p>
 653         Thanks to Joel Falcou and Christophe Henry for their enthusiasm, support,
 654         feedback, and humor; and for volunteering to be Proto's co-maintainers.
 655       </p>
 656 <p>
 657         Thanks to Dave Abrahams for an especially detailed review, and for making
 658         a VM with msvc-7.1 available so I could track down portability issues on
 659         that compiler.
 660       </p>
 661 <p>
 662         Many thanks to Daniel Wallin who first implemented the code used to find
 663         the common domain among a set, accounting for super- and sub-domains. Thanks
 664         also to Jeremiah Willcock, John Bytheway and Krishna Achuthan who offered
 665         alternate solutions to this tricky programming problem.
 666       </p>
 667 <p>
 668         Thanks also to the developers of <a href="http://acts.nersc.gov/formertools/pete/index.html" target="_top">PETE</a>.
 669         I found many good ideas there.
 670       </p>
 671 </div>
 672 </div>
 673 <table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
 674 <td align="left"></td>
 675 <td align="right"><div class="copyright-footer">Copyright &#169; 2008 Eric Niebler<p>
 676         Distributed under the Boost Software License, Version 1.0. (See accompanying
 677         file LICENSE_1_0.txt or copy at <a href="http://www.boost.org/LICENSE_1_0.txt" target="_top">http://www.boost.org/LICENSE_1_0.txt</a>)
 678       </p>
 679 </div></td>
 680 </tr></table>
 681 <hr>
 682 <div class="spirit-nav">
 683 <a accesskey="p" href="../Transform.html"><img src="../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../proto.html"><img src="../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="../boost_random.html"><img src="../../../doc/src/images/next.png" alt="Next"></a>
 684 </div>
 685 </body>
 686 </html>