Add quick_erase to the unordered containers. Refs #3966.

[SVN r60754]
This commit is contained in:
Daniel James
2010-03-22 00:42:07 +00:00
parent 1d02663275
commit ee034e23bb
5 changed files with 155 additions and 28 deletions

View File

@ -117,4 +117,15 @@ First official release.
Add missing `std` qualifier to `ptrdiff_t`.
* Some code formatting changes to fit almost all lines into 80 characters.
[h2 Boost 1.43.0]
* [@http://svn.boost.org/trac/boost/ticket/3966 Ticket 3966]:
`erase_return_void` is now `quick_erase`, which is the
[@http://home.roadrunner.com/~hinnant/issue_review/lwg-active.html#579
current forerunner for resolving the slow erase by iterator], although
there's a strong possibility that this may change in the future. The old
method name remains for backwards compatibility but is considered deprecated
and will be removed in a future release.
[endsect]

View File

@ -464,8 +464,8 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
When the number of elements is a lot smaller than the number of buckets
this function can be very inefficient as it has to search through empty
buckets for the next element, in order to return the iterator.
As a temporary workaround, the container has the method
<methodname>erase_return_void</methodname> which will be faster.
The method <methodname>quick_erase</methodname> is faster, but has yet
to be standardized.
</para>
</notes>
</method>
@ -503,6 +503,30 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
</method>
<method name="quick_erase">
<parameter name="position">
<paramtype>const_iterator</paramtype>
</parameter>
<type>void</type>
<description>
<para>Erase the element pointed to by <code>position</code>.</para>
</description>
<throws>
<para>Only throws an exception if it is thrown by <code>hasher</code> or <code>key_equal</code>.</para>
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
<notes>
<para>
This method is faster than <methodname>erase</methodname> as
it doesn't have to find the next element in the container -
a potentially costly operation.
</para>
<para>
As it hasn't been standardized, it's likely that this may
change in the future.
</para>
</notes>
</method>
<method name="erase_return_void">
<parameter name="position">
<paramtype>const_iterator</paramtype>
@ -517,10 +541,10 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
</throws>
<notes>
<para>
This is a temporary workaround for the inefficient
<methodname>erase</methodname> method. Hopefully, in a future
version the signature of <methodname>erase</methodname> will
be changed and this will be deprecated.
This method is now deprecated, use
<methodname>quick_return</methodname> instead. Although be
warned that as that isn't standardized yet, it could also
change.
</para>
</notes>
</method>
@ -1327,8 +1351,8 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
When the number of elements is a lot smaller than the number of buckets
this function can be very inefficient as it has to search through empty
buckets for the next element, in order to return the iterator.
As a temporary workaround, the container has the method
<methodname>erase_return_void</methodname> which will be faster.
The method <methodname>quick_erase</methodname> is faster, but has yet
to be standardized.
</para>
</notes>
</method>
@ -1366,6 +1390,30 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
</method>
<method name="quick_erase">
<parameter name="position">
<paramtype>const_iterator</paramtype>
</parameter>
<type>void</type>
<description>
<para>Erase the element pointed to by <code>position</code>.</para>
</description>
<throws>
<para>Only throws an exception if it is thrown by <code>hasher</code> or <code>key_equal</code>.</para>
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
<notes>
<para>
This method is faster than <methodname>erase</methodname> as
it doesn't have to find the next element in the container -
a potentially costly operation.
</para>
<para>
As it hasn't been standardized, it's likely that this may
change in the future.
</para>
</notes>
</method>
<method name="erase_return_void">
<parameter name="position">
<paramtype>const_iterator</paramtype>
@ -1380,10 +1428,10 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
</throws>
<notes>
<para>
This is a temporary workaround for the inefficient
<methodname>erase</methodname> method. Hopefully, in a future
version the signature of <methodname>erase</methodname> will
be changed and this will be deprecated.
This method is now deprecated, use
<methodname>quick_return</methodname> instead. Although be
warned that as that isn't standardized yet, it could also
change.
</para>
</notes>
</method>
@ -2204,8 +2252,8 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
When the number of elements is a lot smaller than the number of buckets
this function can be very inefficient as it has to search through empty
buckets for the next element, in order to return the iterator.
As a temporary workaround, the container has the method
<methodname>erase_return_void</methodname> which will be faster.
The method <methodname>quick_erase</methodname> is faster, but has yet
to be standardized.
</para>
</notes>
</method>
@ -2243,6 +2291,30 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
</method>
<method name="quick_erase">
<parameter name="position">
<paramtype>const_iterator</paramtype>
</parameter>
<type>void</type>
<description>
<para>Erase the element pointed to by <code>position</code>.</para>
</description>
<throws>
<para>Only throws an exception if it is thrown by <code>hasher</code> or <code>key_equal</code>.</para>
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
<notes>
<para>
This method is faster than <methodname>erase</methodname> as
it doesn't have to find the next element in the container -
a potentially costly operation.
</para>
<para>
As it hasn't been standardized, it's likely that this may
change in the future.
</para>
</notes>
</method>
<method name="erase_return_void">
<parameter name="position">
<paramtype>const_iterator</paramtype>
@ -2257,10 +2329,10 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
</throws>
<notes>
<para>
This is a temporary workaround for the inefficient
<methodname>erase</methodname> method. Hopefully, in a future
version the signature of <methodname>erase</methodname> will
be changed and this will be deprecated.
This method is now deprecated, use
<methodname>quick_return</methodname> instead. Although be
warned that as that isn't standardized yet, it could also
change.
</para>
</notes>
</method>
@ -3116,8 +3188,8 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
When the number of elements is a lot smaller than the number of buckets
this function can be very inefficient as it has to search through empty
buckets for the next element, in order to return the iterator.
As a temporary workaround, the container has the method
<methodname>erase_return_void</methodname> which will be faster.
The method <methodname>quick_erase</methodname> is faster, but has yet
to be standardized.
</para>
</notes>
</method>
@ -3155,6 +3227,30 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
</method>
<method name="quick_erase">
<parameter name="position">
<paramtype>const_iterator</paramtype>
</parameter>
<type>void</type>
<description>
<para>Erase the element pointed to by <code>position</code>.</para>
</description>
<throws>
<para>Only throws an exception if it is thrown by <code>hasher</code> or <code>key_equal</code>.</para>
<para>In this implementation, this overload doesn't call either function object's methods so it is no throw, but this might not be true in other implementations.</para>
</throws>
<notes>
<para>
This method is faster than <methodname>erase</methodname> as
it doesn't have to find the next element in the container -
a potentially costly operation.
</para>
<para>
As it hasn't been standardized, it's likely that this may
change in the future.
</para>
</notes>
</method>
<method name="erase_return_void">
<parameter name="position">
<paramtype>const_iterator</paramtype>
@ -3169,10 +3265,10 @@ file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
</throws>
<notes>
<para>
This is a temporary workaround for the inefficient
<methodname>erase</methodname> method. Hopefully, in a future
version the signature of <methodname>erase</methodname> will
be changed and this will be deprecated.
This method is now deprecated, use
<methodname>quick_return</methodname> instead. Although be
warned that as that isn't standardized yet, it could also
change.
</para>
</notes>
</method>

View File

@ -369,6 +369,11 @@ namespace boost
return iterator(table_.erase_range(get(first), get(last)));
}
void quick_erase(const_iterator position)
{
table_.erase(get(position));
}
void erase_return_void(const_iterator position)
{
table_.erase(get(position));
@ -907,6 +912,11 @@ namespace boost
return iterator(table_.erase_range(get(first), get(last)));
}
void quick_erase(const_iterator position)
{
table_.erase(get(position));
}
void erase_return_void(const_iterator position)
{
table_.erase(get(position));

View File

@ -361,6 +361,11 @@ namespace boost
return iterator(table_.erase_range(get(first), get(last)));
}
void quick_erase(const_iterator position)
{
table_.erase(get(position));
}
void erase_return_void(const_iterator position)
{
table_.erase(get(position));
@ -849,6 +854,11 @@ namespace boost
return iterator(table_.erase_range(get(first), get(last)));
}
void quick_erase(const_iterator position)
{
table_.erase(get(position));
}
void erase_return_void(const_iterator position)
{
table_.erase(get(position));

View File

@ -116,7 +116,7 @@ void erase_tests1(Container*,
BOOST_TEST(x.erase(x.begin(), x.end()) == x.begin());
}
std::cerr<<"erase_return_void(begin()).\n";
std::cerr<<"quick_erase(begin()).\n";
{
test::random_values<Container> v(1000, generator);
Container x(v.begin(), v.end());
@ -126,7 +126,7 @@ void erase_tests1(Container*,
BOOST_DEDUCED_TYPENAME Container::key_type
key = test::get_key<Container>(*x.begin());
std::size_t count = x.count(key);
x.erase_return_void(x.begin());
x.quick_erase(x.begin());
--size;
BOOST_TEST(x.count(key) == count - 1);
BOOST_TEST(x.size() == size);
@ -134,7 +134,7 @@ void erase_tests1(Container*,
BOOST_TEST(x.empty());
}
std::cerr<<"erase_return_void(random position).\n";
std::cerr<<"quick_erase(random position).\n";
{
test::random_values<Container> v(1000, generator);
Container x(v.begin(), v.end());
@ -155,7 +155,7 @@ void erase_tests1(Container*,
BOOST_DEDUCED_TYPENAME Container::key_type
key = test::get_key<Container>(*pos);
std::size_t count = x.count(key);
x.erase_return_void(pos);
x.quick_erase(pos);
--size;
if(size > 0)
BOOST_TEST(index == 0 ? next == x.begin() :