blob: f01e2d491b163280dfea7771e2510a1bcca40226 [file] [log] [blame]
<section xmlns="http://docbook.org/ns/docbook" version="5.0"
xml:id="std.util.memory.auto_ptr" xreflabel="auto_ptr">
<?dbhtml filename="auto_ptr.html"?>
<info><title>auto_ptr</title>
<keywordset>
<keyword>ISO C++</keyword>
<keyword>auto_ptr</keyword>
</keywordset>
</info>
<section xml:id="auto_ptr.limitations"><info><title>Limitations</title></info>
<para>Explaining all of the fun and delicious things that can
happen with misuse of the <classname>auto_ptr</classname> class
template (called <acronym>AP</acronym> here) would take some
time. Suffice it to say that the use of <acronym>AP</acronym>
safely in the presence of copying has some subtleties.
</para>
<para>
The AP class is a really
nifty idea for a smart pointer, but it is one of the dumbest of
all the smart pointers -- and that's fine.
</para>
<para>
AP is not meant to be a supersmart solution to all resource
leaks everywhere. Neither is it meant to be an effective form
of garbage collection (although it can help, a little bit).
And it can <emphasis>not</emphasis>be used for arrays!
</para>
<para>
<acronym>AP</acronym> is meant to prevent nasty leaks in the
presence of exceptions. That's <emphasis>all</emphasis>. This
code is AP-friendly:
</para>
<programlisting>
// Not a recommend naming scheme, but good for web-based FAQs.
typedef std::auto_ptr&lt;MyClass&gt; APMC;
extern function_taking_MyClass_pointer (MyClass*);
extern some_throwable_function ();
void func (int data)
{
APMC ap (new MyClass(data));
some_throwable_function(); // this will throw an exception
function_taking_MyClass_pointer (ap.get());
}
</programlisting>
<para>When an exception gets thrown, the instance of MyClass that's
been created on the heap will be <function>delete</function>'d as the stack is
unwound past <function>func()</function>.
</para>
<para>Changing that code as follows is not <acronym>AP</acronym>-friendly:
</para>
<programlisting>
APMC ap (new MyClass[22]);
</programlisting>
<para>You will get the same problems as you would without the use
of <acronym>AP</acronym>:
</para>
<programlisting>
char* array = new char[10]; // array new...
...
delete array; // ...but single-object delete
</programlisting>
<para>
AP cannot tell whether the pointer you've passed at creation points
to one or many things. If it points to many things, you are about
to die. AP is trivial to write, however, so you could write your
own <code>auto_array_ptr</code> for that situation (in fact, this has
been done many times; check the mailing lists, Usenet, Boost, etc).
</para>
</section>
<section xml:id="auto_ptr.using"><info><title>Use in Containers</title></info>
<para>
</para>
<para>All of the <link linkend="std.containers">containers</link>
described in the standard library require their contained types
to have, among other things, a copy constructor like this:
</para>
<programlisting>
struct My_Type
{
My_Type (My_Type const&amp;);
};
</programlisting>
<para>
Note the const keyword; the object being copied shouldn't change.
The template class <code>auto_ptr</code> (called AP here) does not
meet this requirement. Creating a new AP by copying an existing
one transfers ownership of the pointed-to object, which means that
the AP being copied must change, which in turn means that the
copy ctors of AP do not take const objects.
</para>
<para>
The resulting rule is simple: <emphasis>Never ever use a
container of auto_ptr objects</emphasis>. The standard says that
<quote>undefined</quote> behavior is the result, but it is
guaranteed to be messy.
</para>
<para>
To prevent you from doing this to yourself, the
<link linkend="manual.ext.compile_checks">concept checks</link> built
in to this implementation will issue an error if you try to
compile code like this:
</para>
<programlisting>
#include &lt;vector&gt;
#include &lt;memory&gt;
void f()
{
std::vector&lt; std::auto_ptr&lt;int&gt; &gt; vec_ap_int;
}
</programlisting>
<para>
Should you try this with the checks enabled, you will see an error.
</para>
</section>
</section>