I know this sounds stupid, but look at this simple example (working dir should have more than one item):
#define BOOST_FILESYSTEM_VERSION 3
#include <boost/filesystem.hpp>
#include <cassert>
int main()
{
using namespace boost::filesystem;
directory_iterator it("./");
directory_iterator it_copy = it;
++it;
assert(it_copy != it);
return 0;
}
it_copy
is modified together with it
! (boost 1.45) What considerations could lead to such design (directory_iterator
is something like smart ptr)?
I just need to save a copy of directory_iterator
to use it later.
If you take a look at the reference you will notice that it is advertised to be a boost::single_pass_traversal_tag
.
This is the equivalent (in boost terminology) of the Input Iterator in the STL (think of it as an iterator delivering packets from a network connection, you cannot rewind).
Also note (from this same page):
i == j
does not imply that++i == ++j
.
At this point, one may wonder why it can be copied. The reason is that STL algorithms have set the norm taking their arguments by copy. Therefore it would not be usable with STL algorithms if it could not be copied.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With