Removed dead file.

[SVN r43086]
This commit is contained in:
John Maddock
2008-02-04 09:18:30 +00:00
parent 6ba37d939d
commit 11e3e918ac

View File

@ -1,699 +0,0 @@
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
<title>Background and Tutorial</title>
<link rel="stylesheet" href="../../../../../../doc/html/boostbook.css" type="text/css">
<meta name="generator" content="DocBook XSL Stylesheets Vsnapshot_2006-12-17_0120">
<link rel="start" href="../../index.html" title="Chapter<65>1.<2E>Boost.TypeTraits">
<link rel="up" href="../category.html" title="Type Traits by Category">
<link rel="prev" href="../category.html" title="Type Traits by Category">
<link rel="next" href="value_traits.html" title="Type Traits that Describe the Properties of a Type">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table cellpadding="2" width="100%"><tr>
<td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../../../../boost.png"></td>
<td align="center"><a href="../../../../../../index.htm">Home</a></td>
<td align="center"><a href="../../../../../libraries.htm">Libraries</a></td>
<td align="center"><a href="../../../../../../people/people.htm">People</a></td>
<td align="center"><a href="../../../../../../more/faq.htm">FAQ</a></td>
<td align="center"><a href="../../../../../../more/index.htm">More</a></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="../category.html"><img src="../../../../../../doc/html/images/prev.png" alt="Prev"></a><a accesskey="u" href="../category.html"><img src="../../../../../../doc/html/images/up.png" alt="Up"></a><a accesskey="h" href="../../index.html"><img src="../../../../../../doc/html/images/home.png" alt="Home"></a><a accesskey="n" href="value_traits.html"><img src="../../../../../../doc/html/images/next.png" alt="Next"></a>
</div>
<div class="section" lang="en">
<div class="titlepage"><div><div><h3 class="title">
<a name="boost_typetraits.category.background"></a><a href="background.html" title="Background and Tutorial"> Background and
Tutorial</a>
</h3></div></div></div>
<p>
The following is an updated version of the article "C++ Type traits"
by John Maddock and Steve Cleary that appeared in the October 2000 issue
of <a href="http://www.ddj.com" target="_top">Dr Dobb's Journal</a>.
</p>
<p>
Generic programming (writing code which works with any data type meeting
a set of requirements) has become the method of choice for providing reusable
code. However, there are times in generic programming when "generic"
just isn't good enough - sometimes the differences between types are too
large for an efficient generic implementation. This is when the traits technique
becomes important - by encapsulating those properties that need to be considered
on a type by type basis inside a traits class, we can minimize the amount
of code that has to differ from one type to another, and maximize the amount
of generic code.
</p>
<p>
Consider an example: when working with character strings, one common operation
is to determine the length of a null terminated string. Clearly it's possible
to write generic code that can do this, but it turns out that there are much
more efficient methods available: for example, the C library functions <code class="computeroutput"><span class="identifier">strlen</span></code> and <code class="computeroutput"><span class="identifier">wcslen</span></code>
are usually written in assembler, and with suitable hardware support can
be considerably faster than a generic version written in C++. The authors
of the C++ standard library realized this, and abstracted the properties
of <code class="computeroutput"><span class="keyword">char</span></code> and <code class="computeroutput"><span class="keyword">wchar_t</span></code>
into the class <code class="computeroutput"><span class="identifier">char_traits</span></code>.
Generic code that works with character strings can simply use <code class="computeroutput"><span class="identifier">char_traits</span><span class="special">&lt;&gt;::</span><span class="identifier">length</span></code> to determine the length of a null
terminated string, safe in the knowledge that specializations of <code class="computeroutput"><span class="identifier">char_traits</span></code> will use the most appropriate
method available to them.
</p>
<a name="boost_typetraits.category.background.type_traits"></a><h5>
<a name="id437674"></a>
<a href="background.html#boost_typetraits.category.background.type_traits">Type Traits</a>
</h5>
<p>
Class <code class="computeroutput"><span class="identifier">char_traits</span></code> is a classic
example of a collection of type specific properties wrapped up in a single
class - what Nathan Myers termed a <span class="emphasis"><em>baggage class</em></span><a href="../../">[1]</a>. In the Boost type-traits library,
we<a href="../../">[2]</a> have written a set of
very specific traits classes, each of which encapsulate a single trait from
the C++ type system; for example, is a type a pointer or a reference type?
Or does a type have a trivial constructor, or a const-qualifier? The type-traits
classes share a unified design: each class inherits from a the type <a href="../reference/integral_constant.html" title="integral_constant">true_type</a> if
the type has the specified property and inherits from <a href="../reference/integral_constant.html" title="integral_constant">false_type</a>
otherwise. As we will show, these classes can be used in generic programming
to determine the properties of a given type and introduce optimizations that
are appropriate for that case.
</p>
<p>
The type-traits library also contains a set of classes that perform a specific
transformation on a type; for example, they can remove a top-level const
or volatile qualifier from a type. Each class that performs a transformation
defines a single typedef-member <code class="computeroutput"><span class="identifier">type</span></code>
that is the result of the transformation. All of the type-traits classes
are defined inside namespace <code class="computeroutput"><span class="identifier">boost</span></code>;
for brevity, namespace-qualification is omitted in most of the code samples
given.
</p>
<a name="boost_typetraits.category.background.implementation"></a><h5>
<a name="id437797"></a>
<a href="background.html#boost_typetraits.category.background.implementation">Implementation</a>
</h5>
<p>
There are far too many separate classes contained in the type-traits library
to give a full implementation here - see the source code in the Boost library
for the full details - however, most of the implementation is fairly repetitive
anyway, so here we will just give you a flavor for how some of the classes
are implemented. Beginning with possibly the simplest class in the library,
<code class="computeroutput"><span class="identifier">is_void</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">&gt;</span></code> inherits
from <code class="computeroutput"><a href="../reference/integral_constant.html" title="integral_constant">true_type</a></code>
only if <code class="computeroutput"><span class="identifier">T</span></code> is <code class="computeroutput"><span class="keyword">void</span></code>.
</p>
<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>
<span class="keyword">struct</span> <a href="../reference/is_void.html" title="is_void">is_void</a> <span class="special">:</span> <span class="keyword">public</span> <a href="../reference/integral_constant.html" title="integral_constant">false_type</a><span class="special">{};</span>
<span class="keyword">template</span> <span class="special">&lt;&gt;</span>
<span class="keyword">struct</span> <a href="../reference/is_void.html" title="is_void">is_void</a><span class="special">&lt;</span><span class="keyword">void</span><span class="special">&gt;</span> <span class="special">:</span> <span class="keyword">public</span> <a href="../reference/integral_constant.html" title="integral_constant">true_type</a><span class="special">{};</span>
</pre>
<p>
Here we define a primary version of the template class <code class="computeroutput"><a href="../reference/is_void.html" title="is_void">is_void</a></code>,
and provide a full-specialization when <code class="computeroutput"><span class="identifier">T</span></code>
is <code class="computeroutput"><span class="keyword">void</span></code>. While full specialization
of a template class is an important technique, sometimes we need a solution
that is halfway between a fully generic solution, and a full specialization.
This is exactly the situation for which the standards committee defined partial
template-class specialization. As an example, consider the class <code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">is_pointer</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">&gt;</span></code>:
here we needed a primary version that handles all the cases where T is not
a pointer, and a partial specialization to handle all the cases where T is
a pointer:
</p>
<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>
<span class="keyword">struct</span> <a href="../reference/is_pointer.html" title="is_pointer">is_pointer</a> <span class="special">:</span> <span class="keyword">public</span> <a href="../reference/integral_constant.html" title="integral_constant">false_type</a><span class="special">{};</span>
<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>
<span class="keyword">struct</span> <a href="../reference/is_pointer.html" title="is_pointer">is_pointer</a><span class="special">&lt;</span><span class="identifier">T</span><span class="special">*&gt;</span> <span class="special">:</span> <span class="keyword">public</span> <a href="../reference/integral_constant.html" title="integral_constant">true_type</a><span class="special">{};</span>
</pre>
<p>
The syntax for partial specialization is somewhat arcane and could easily
occupy an article in its own right; like full specialization, in order to
write a partial specialization for a class, you must first declare the primary
template. The partial specialization contains an extra &lt;...&gt; after
the class name that contains the partial specialization parameters; these
define the types that will bind to that partial specialization rather than
the default template. The rules for what can appear in a partial specialization
are somewhat convoluted, but as a rule of thumb if you can legally write
two function overloads of the form:
</p>
<pre class="programlisting">
<span class="keyword">void</span> <span class="identifier">foo</span><span class="special">(</span><span class="identifier">T</span><span class="special">);</span>
<span class="keyword">void</span> <span class="identifier">foo</span><span class="special">(</span><span class="identifier">U</span><span class="special">);</span>
</pre>
<p>
Then you can also write a partial specialization of the form:
</p>
<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>
<span class="keyword">class</span> <span class="identifier">c</span><span class="special">{</span> <span class="comment">/*details*/</span> <span class="special">};</span>
<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>
<span class="keyword">class</span> <span class="identifier">c</span><span class="special">&lt;</span><span class="identifier">U</span><span class="special">&gt;{</span> <span class="comment">/*details*/</span> <span class="special">};</span>
</pre>
<p>
This rule is by no means foolproof, but it is reasonably simple to remember
and close enough to the actual rule to be useful for everyday use.
</p>
<p>
As a more complex example of partial specialization consider the class <code class="computeroutput"><span class="identifier">remove_extent</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">&gt;</span></code>.
This class defines a single typedef-member <code class="computeroutput"><span class="identifier">type</span></code>
that is the same type as T but with any top-level array bounds removed; this
is an example of a traits class that performs a transformation on a type:
</p>
<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>
<span class="keyword">struct</span> <a href="../reference/remove_extent.html" title="remove_extent">remove_extent</a>
<span class="special">{</span> <span class="keyword">typedef</span> <span class="identifier">T</span> <span class="identifier">type</span><span class="special">;</span> <span class="special">};</span>
<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="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">N</span><span class="special">&gt;</span>
<span class="keyword">struct</span> <a href="../reference/remove_extent.html" title="remove_extent">remove_extent</a><span class="special">&lt;</span><span class="identifier">T</span><span class="special">[</span><span class="identifier">N</span><span class="special">]&gt;</span>
<span class="special">{</span> <span class="keyword">typedef</span> <span class="identifier">T</span> <span class="identifier">type</span><span class="special">;</span> <span class="special">};</span>
</pre>
<p>
The aim of <code class="computeroutput"><a href="../reference/remove_extent.html" title="remove_extent">remove_extent</a></code>
is this: imagine a generic algorithm that is passed an array type as a template
parameter, <code class="computeroutput"><a href="../reference/remove_extent.html" title="remove_extent">remove_extent</a></code>
provides a means of determining the underlying type of the array. For example
<code class="computeroutput"><span class="identifier">remove_extent</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">[</span><span class="number">4</span><span class="special">][</span><span class="number">5</span><span class="special">]&gt;::</span><span class="identifier">type</span></code> would evaluate to the type <code class="computeroutput"><span class="keyword">int</span><span class="special">[</span><span class="number">5</span><span class="special">]</span></code>. This example also shows that the number
of template parameters in a partial specialization does not have to match
the number in the default template. However, the number of parameters that
appear after the class name do have to match the number and type of the parameters
in the default template.
</p>
<a name="boost_typetraits.category.background.optimized_copy"></a><h5>
<a name="id490244"></a>
<a href="background.html#boost_typetraits.category.background.optimized_copy">Optimized
copy</a>
</h5>
<p>
As an example of how the type traits classes can be used, consider the standard
library algorithm copy:
</p>
<pre class="programlisting">
<span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">Iter1</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">Iter2</span><span class="special">&gt;</span>
<span class="identifier">Iter2</span> <span class="identifier">copy</span><span class="special">(</span><span class="identifier">Iter1</span> <span class="identifier">first</span><span class="special">,</span> <span class="identifier">Iter1</span> <span class="identifier">last</span><span class="special">,</span> <span class="identifier">Iter2</span> <span class="identifier">out</span><span class="special">);</span>
</pre>
<p>
Obviously, there's no problem writing a generic version of copy that works
for all iterator types <code class="computeroutput"><span class="identifier">Iter1</span></code>
and <code class="computeroutput"><span class="identifier">Iter2</span></code>; however, there
are some circumstances when the copy operation can best be performed by a
call to <code class="computeroutput"><span class="identifier">memcpy</span></code>. In order
to implement copy in terms of <code class="computeroutput"><span class="identifier">memcpy</span></code>
all of the following conditions need to be met:
</p>
<div class="itemizedlist"><ul type="disc">
<li>
Both of the iterator types <code class="computeroutput"><span class="identifier">Iter1</span></code>
and <code class="computeroutput"><span class="identifier">Iter2</span></code> must be pointers.
</li>
<li>
Both <code class="computeroutput"><span class="identifier">Iter1</span></code> and <code class="computeroutput"><span class="identifier">Iter2</span></code> must point to the same type - excluding
const and volatile-qualifiers.
</li>
<li>
The type pointed to by <code class="computeroutput"><span class="identifier">Iter1</span></code>
must have a trivial assignment operator.
</li>
</ul></div>
<p>
By trivial assignment operator we mean that the type is either a scalar type<a href="../../">[3]</a> or:
</p>
<div class="itemizedlist"><ul type="disc">
<li>
The type has no user defined assignment operator.
</li>
<li>
The type does not have any data members that are references.
</li>
<li>
All base classes, and all data member objects must have trivial assignment
operators.
</li>
</ul></div>
<p>
If all these conditions are met then a type can be copied using <code class="computeroutput"><span class="identifier">memcpy</span></code> rather than using a compiler generated
assignment operator. The type-traits library provides a class <code class="computeroutput"><a href="../reference/has_trivial_assign.html" title="has_trivial_assign">has_trivial_assign</a></code>,
such that <code class="computeroutput"><span class="identifier">has_trivial_assign</span><span class="special">&lt;</span><span class="identifier">T</span><span class="special">&gt;::</span><span class="identifier">value</span></code> is true only if T has a trivial assignment
operator. This class "just works" for scalar types, but has to
be explicitly specialised for class/struct types that also happen to have
a trivial assignment operator. In other words if <a href="../reference/has_trivial_assign.html" title="has_trivial_assign">has_trivial_assign</a>
gives the wrong answer, it will give the "safe" wrong answer -
that trivial assignment is not allowable.
</p>
<p>
The code for an optimized version of copy that uses <code class="computeroutput"><span class="identifier">memcpy</span></code>
where appropriate is given in <a href="../../">the examples</a>.
The code begins by defining a template function <code class="computeroutput"><span class="identifier">do_copy</span></code>
that performs a "slow but safe" copy. The last parameter passed
to this function may be either a <code class="computeroutput"><a href="../reference/integral_constant.html" title="integral_constant">true_type</a></code>
or a <code class="computeroutput"><a href="../reference/integral_constant.html" title="integral_constant">false_type</a></code>.
Following that there is an overload of do<span class="underline">copy
that uses `memcpy`: this time the iterators are required to actually be pointers
to the same type, and the final parameter must be a `</span>_true_type<code class="computeroutput"><span class="special">.</span> <span class="identifier">Finally</span><span class="special">,</span> <span class="identifier">the</span> <span class="identifier">version</span> <span class="identifier">of</span>
</code>copy<code class="computeroutput"> <span class="identifier">calls</span> </code>do<span class="underline">copy`, passing `</span>_has_trivial_assign&lt;value_type&gt;()`
as the final parameter: this will dispatch to the optimized version where
appropriate, otherwise it will call the "slow but safe version".
</p>
<a name="boost_typetraits.category.background.was_it_worth_it_"></a><h5>
<a name="id490786"></a>
<a href="background.html#boost_typetraits.category.background.was_it_worth_it_">Was
it worth it?</a>
</h5>
<p>
It has often been repeated in these columns that "premature optimization
is the root of all evil" <a href="../../">[4]</a>.
So the question must be asked: was our optimization premature? To put this
in perspective the timings for our version of copy compared a conventional
generic copy<a href="../../">[5]</a> are shown in
table 1.
</p>
<p>
Clearly the optimization makes a difference in this case; but, to be fair,
the timings are loaded to exclude cache miss effects - without this accurate
comparison between algorithms becomes difficult. However, perhaps we can
add a couple of caveats to the premature optimization rule:
</p>
<div class="itemizedlist"><ul type="disc">
<li>
If you use the right algorithm for the job in the first place then optimization
will not be required; in some cases, memcpy is the right algorithm.
</li>
<li>
If a component is going to be reused in many places by many people then
optimizations may well be worthwhile where they would not be so for a single
case - in other words, the likelihood that the optimization will be absolutely
necessary somewhere, sometime is that much higher. Just as importantly
the perceived value of the stock implementation will be higher: there is
no point standardizing an algorithm if users reject it on the grounds that
there are better, more heavily optimized versions available.
</li>
</ul></div>
<div class="table">
<a name="id490853"></a><p class="title"><b>Table<EFBFBD>1.1.<2E>Time taken to copy 1000 elements using `copy&lt;const
T*, T*&gt;` (times in micro-seconds)</b></p>
<div class="table-contents"><table class="table" summary="Time taken to copy 1000 elements using `copy&lt;const
T*, T*&gt;` (times in micro-seconds)">
<colgroup>
<col>
<col>
<col>
</colgroup>
<thead><tr>
<th>
<p>
Version
</p>
</th>
<th>
<p>
T
</p>
</th>
<th>
<p>
Time
</p>
</th>
</tr></thead>
<tbody>
<tr>
<td>
<p>
"Optimized" copy
</p>
</td>
<td>
<p>
char
</p>
</td>
<td>
<p>
0.99
</p>
</td>
</tr>
<tr>
<td>
<p>
Conventional copy
</p>
</td>
<td>
<p>
char
</p>
</td>
<td>
<p>
8.07
</p>
</td>
</tr>
<tr>
<td>
<p>
"Optimized" copy
</p>
</td>
<td>
<p>
int
</p>
</td>
<td>
<p>
2.52
</p>
</td>
</tr>
<tr>
<td>
<p>
Conventional copy
</p>
</td>
<td>
<p>
int
</p>
</td>
<td>
<p>
8.02
</p>
</td>
</tr>
</tbody>
</table></div>
</div>
<br class="table-break"><a name="boost_typetraits.category.background.pair_of_references"></a><h5>
<a name="id491010"></a>
<a href="background.html#boost_typetraits.category.background.pair_of_references">Pair
of References</a>
</h5>
<p>
The optimized copy example shows how type traits may be used to perform optimization
decisions at compile-time. Another important usage of type traits is to allow
code to compile that otherwise would not do so unless excessive partial specialization
is used. This is possible by delegating partial specialization to the type
traits classes. Our example for this form of usage is a pair that can hold
references <a href="../../">[6]</a>.
</p>
<p>
First, let us examine the definition of <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span></code>,
omitting the comparison operators, default constructor, and template copy
constructor for simplicity:
</p>
<pre class="programlisting">
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T1</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">T2</span><span class="special">&gt;</span>
<span class="keyword">struct</span> <span class="identifier">pair</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="identifier">T1</span> <span class="identifier">first_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">T2</span> <span class="identifier">second_type</span><span class="special">;</span>
<span class="identifier">T1</span> <span class="identifier">first</span><span class="special">;</span>
<span class="identifier">T2</span> <span class="identifier">second</span><span class="special">;</span>
<span class="identifier">pair</span><span class="special">(</span><span class="keyword">const</span> <span class="identifier">T1</span> <span class="special">&amp;</span> <span class="identifier">nfirst</span><span class="special">,</span> <span class="keyword">const</span> <span class="identifier">T2</span> <span class="special">&amp;</span> <span class="identifier">nsecond</span><span class="special">)</span>
<span class="special">:</span><span class="identifier">first</span><span class="special">(</span><span class="identifier">nfirst</span><span class="special">),</span> <span class="identifier">second</span><span class="special">(</span><span class="identifier">nsecond</span><span class="special">)</span> <span class="special">{</span> <span class="special">}</span>
<span class="special">};</span>
</pre>
<p>
Now, this "pair" cannot hold references as it currently stands,
because the constructor would require taking a reference to a reference,
which is currently illegal <a href="../../">[7]</a>.
Let us consider what the constructor's parameters would have to be in order
to allow "pair" to hold non-reference types, references, and constant
references:
</p>
<div class="table">
<a name="id491374"></a><p class="title"><b>Table<EFBFBD>1.2.<2E>Required Constructor Argument Types</b></p>
<div class="table-contents"><table class="table" summary="Required Constructor Argument Types">
<colgroup>
<col>
<col>
</colgroup>
<thead><tr>
<th>
<p>
Type of <code class="computeroutput"><span class="identifier">T1</span></code>
</p>
</th>
<th>
<p>
Type of parameter to initializing constructor
</p>
</th>
</tr></thead>
<tbody>
<tr>
<td>
<p>
T
</p>
</td>
<td>
<p>
const T &amp;
</p>
</td>
</tr>
<tr>
<td>
<p>
T &amp;
</p>
</td>
<td>
<p>
T &amp;
</p>
</td>
</tr>
<tr>
<td>
<p>
const T &amp;
</p>
</td>
<td>
<p>
const T &amp;
</p>
</td>
</tr>
</tbody>
</table></div>
</div>
<br class="table-break"><p>
A little familiarity with the type traits classes allows us to construct
a single mapping that allows us to determine the type of parameter from the
type of the contained class. The type traits classes provide a transformation
<a href="../reference/add_reference.html" title="add_reference">add_reference</a>,
which adds a reference to its type, unless it is already a reference.
</p>
<div class="table">
<a name="id491508"></a><p class="title"><b>Table<EFBFBD>1.3.<2E>Using add_reference to synthesize the correct constructor
type</b></p>
<div class="table-contents"><table class="table" summary="Using add_reference to synthesize the correct constructor
type">
<colgroup>
<col>
<col>
<col>
</colgroup>
<thead><tr>
<th>
<p>
Type of <code class="computeroutput"><span class="identifier">T1</span></code>
</p>
</th>
<th>
<p>
Type of <code class="computeroutput"><span class="keyword">const</span> <span class="identifier">T1</span></code>
</p>
</th>
<th>
<p>
Type of <code class="computeroutput"><span class="identifier">add_reference</span><span class="special">&lt;</span><span class="keyword">const</span>
<span class="identifier">T1</span><span class="special">&gt;::</span><span class="identifier">type</span></code>
</p>
</th>
</tr></thead>
<tbody>
<tr>
<td>
<p>
T
</p>
</td>
<td>
<p>
const T
</p>
</td>
<td>
<p>
const T &amp;
</p>
</td>
</tr>
<tr>
<td>
<p>
T &amp;
</p>
</td>
<td>
<p>
T &amp; [8]
</p>
</td>
<td>
<p>
T &amp;
</p>
</td>
</tr>
<tr>
<td>
<p>
const T &amp;
</p>
</td>
<td>
<p>
const T &amp;
</p>
</td>
<td>
<p>
const T &amp;
</p>
</td>
</tr>
</tbody>
</table></div>
</div>
<br class="table-break"><p>
This allows us to build a primary template definition for <code class="computeroutput"><span class="identifier">pair</span></code>
that can contain non-reference types, reference types, and constant reference
types:
</p>
<pre class="programlisting">
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">typename</span> <span class="identifier">T1</span><span class="special">,</span> <span class="keyword">typename</span> <span class="identifier">T2</span><span class="special">&gt;</span>
<span class="keyword">struct</span> <span class="identifier">pair</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="identifier">T1</span> <span class="identifier">first_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">T2</span> <span class="identifier">second_type</span><span class="special">;</span>
<span class="identifier">T1</span> <span class="identifier">first</span><span class="special">;</span>
<span class="identifier">T2</span> <span class="identifier">second</span><span class="special">;</span>
<span class="identifier">pair</span><span class="special">(</span><span class="identifier">boost</span><span class="special">::</span><a href="../reference/add_reference.html" title="add_reference">add_reference</a><span class="special">&lt;</span><span class="keyword">const</span> <span class="identifier">T1</span><span class="special">&gt;::</span><span class="identifier">type</span> <span class="identifier">nfirst</span><span class="special">,</span>
<span class="identifier">boost</span><span class="special">::</span><a href="../reference/add_reference.html" title="add_reference">add_reference</a><span class="special">&lt;</span><span class="keyword">const</span> <span class="identifier">T2</span><span class="special">&gt;::</span><span class="identifier">type</span> <span class="identifier">nsecond</span><span class="special">)</span>
<span class="special">:</span><span class="identifier">first</span><span class="special">(</span><span class="identifier">nfirst</span><span class="special">),</span> <span class="identifier">second</span><span class="special">(</span><span class="identifier">nsecond</span><span class="special">)</span> <span class="special">{</span> <span class="special">}</span>
<span class="special">};</span>
</pre>
<p>
Add back in the standard comparison operators, default constructor, and template
copy constructor (which are all the same), and you have a <code class="computeroutput"><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span></code>
that can hold reference types!
</p>
<p>
This same extension could have been done using partial template specialization
of <code class="computeroutput"><span class="identifier">pair</span></code>, but to specialize
<code class="computeroutput"><span class="identifier">pair</span></code> in this way would require
three partial specializations, plus the primary template. Type traits allows
us to define a single primary template that adjusts itself auto-magically
to any of these partial specializations, instead of a brute-force partial
specialization approach. Using type traits in this fashion allows programmers
to delegate partial specialization to the type traits classes, resulting
in code that is easier to maintain and easier to understand.
</p>
<a name="boost_typetraits.category.background.conclusion"></a><h5>
<a name="id492124"></a>
<a href="background.html#boost_typetraits.category.background.conclusion">Conclusion</a>
</h5>
<p>
We hope that in this article we have been able to give you some idea of what
type-traits are all about. A more complete listing of the available classes
are in the boost documentation, along with further examples using type traits.
Templates have enabled C++ uses to take the advantage of the code reuse that
generic programming brings; hopefully this article has shown that generic
programming does not have to sink to the lowest common denominator, and that
templates can be optimal as well as generic.
</p>
<a name="boost_typetraits.category.background.acknowledgements"></a><h5>
<a name="id492161"></a>
<a href="background.html#boost_typetraits.category.background.acknowledgements">Acknowledgements</a>
</h5>
<p>
The authors would like to thank Beman Dawes and Howard Hinnant for their
helpful comments when preparing this article.
</p>
<a name="boost_typetraits.category.background.references"></a><h5>
<a name="id492191"></a>
<a href="background.html#boost_typetraits.category.background.references">References</a>
</h5>
<div class="orderedlist"><ol type="1">
<li>
Nathan C. Myers, C++ Report, June 1995.
</li>
<li>
The type traits library is based upon contributions by Steve Cleary, Beman
Dawes, Howard Hinnant and John Maddock: it can be found at www.boost.org.
</li>
<li>
A scalar type is an arithmetic type (i.e. a built-in integer or floating
point type), an enumeration type, a pointer, a pointer to member, or a
const- or volatile-qualified version of one of these types.
</li>
<li>
This quote is from Donald Knuth, ACM Computing Surveys, December 1974,
pg 268.
</li>
<li>
The test code is available as part of the boost utility library (see algo_opt_examples.cpp),
the code was compiled with gcc 2.95 with all optimisations turned on, tests
were conducted on a 400MHz Pentium II machine running Microsoft Windows
98.
</li>
<li>
John Maddock and Howard Hinnant have submitted a "compressed_pair"
library to Boost, which uses a technique similar to the one described here
to hold references. Their pair also uses type traits to determine if any
of the types are empty, and will derive instead of contain to conserve
space -- hence the name "compressed".
</li>
<li>
This is actually an issue with the C++ Core Language Working Group (issue
#106), submitted by Bjarne Stroustrup. The tentative resolution is to allow
a "reference to a reference to T" to mean the same thing as a
"reference to T", but only in template instantiation, in a method
similar to multiple cv-qualifiers.
</li>
<li>
For those of you who are wondering why this shouldn't be const-qualified,
remember that references are always implicitly constant (for example, you
can't re-assign a reference). Remember also that "const T &amp;"
is something completely different. For this reason, cv-qualifiers on template
type arguments that are references are ignored.
</li>
</ol></div>
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
<td align="left"></td>
<td align="right"><small>Copyright <20> 2000, 2006 Adobe Systems Inc, David Abrahams,
Steve Cleary, Beman Dawes, Aleksey Gurtovoy, Howard Hinnant, Jesse Jones, Mat
Marcus, Itay Maman, John Maddock, Alexander Nasonov, Thorsten Ottosen, Robert
Ramey and Jeremy Siek</small></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="../category.html"><img src="../../../../../../doc/html/images/prev.png" alt="Prev"></a><a accesskey="u" href="../category.html"><img src="../../../../../../doc/html/images/up.png" alt="Up"></a><a accesskey="h" href="../../index.html"><img src="../../../../../../doc/html/images/home.png" alt="Home"></a><a accesskey="n" href="value_traits.html"><img src="../../../../../../doc/html/images/next.png" alt="Next"></a>
</div>
</body>
</html>