Re: erase overload for reverse_iterators
In article <1192629682.968273.97840@i13g2000prf.googlegroups.com>,
Abhishek Padmanabh <abhishek.padmanabh@gmail.com> wrote:
For the STL containers that support reverse iterators, why aren't
there overloads for erase() [or probably other functions that take a
(forward) iterator] that take them as an argument? Atleast for
consistency in the interface?
First a reverse iterator requires the 'regular' iterator to be
bidirectional not forward. Most stl containers have at least
bidirectional iterators.
When designing such classes is it considered that these overloads
would make the class interface bulkier and probably redundant (going
by the fact that generic algorithms can be written as non-member, non-
friends)? But then, what is the purpose of reverse iterators if
functionalities like that are not supported on them? Reverse iteration
can surely be implemented in terms of forward iterators. Isn't it?
The missing overloads might be a historical oversight, but since
reverse iterators are convertable [via base() member] to a forward
moving iterator there is no real loss.
The purpose of reverse iterators is to allow forward moving
algorithms [such as find, etc] to work backward, nothing else.
the typedef's in the stl containers are to simplify some template
programming.
--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]