问题
I've been up and down stackoverflow and even the very, very nice Dr. Dobbs article but I can't find a definitive answer to the question.
A section of the answer to the question What are the shortcomings of std::reverse_iterator? says that it might not be possible at all.
std::list::reverse_iterator it = list.rbegin();
while( it != list.rend() )
{
int value=*it;
if( some_cond_met_on(value) )
{
++it;
list.erase( it.base() );
}
else
{
++it;
}
}
PS: I do know there are other alternatives, such as erase_if(), but I'm looking for an answer to this specific question.
回答1:
It should just be
std::list<int>::reverse_iterator it = list.rbegin();
while( it != list.rend() )
{
int value=*it;
if( some_cond_met_on(value) )
{
++it;
it= reverse_iterator(list.erase(it.base()); // change to this!
}
else
{
++it;
}
}
回答2:
Most erase()
implementations I have seen return the next iterator in the sequence for exactly this kind of situation, eg:
std::list<int>::reverse_iterator it = list.rbegin();
while( it != list.rend() )
{
int value = *it;
if( some_cond_met_on(value) )
{
it = list.erase( it );
}
else
{
++it;
}
}
来源:https://stackoverflow.com/questions/8621426/how-do-you-erase-and-continue-using-a-stdreverse-iterator