How can make the following code inside of for_each more readable? void f(char*); typedef std::map<int,char> TMap; TMap a; std::for_each(a.begin(),a.end(),boost::bind(f,boost::bind(&TMap::value_type::second,_1))); Is it possible to convert boost::bind(&TMap::value_type::second,_1) to a function object or a typedef name? -- Regards, Alexander. http://sjcomp.com
On 9/15/06, Alexander Shyrokov <shirokov@unh.edu> wrote:
Is it possible to convert boost::bind(&TMap::value_type::second,_1) to a function object or a typedef name?
Though it's not exactly what you're asking for, one nice way would be to use an iterator adapter that returns base_itr->second when dereferenced. I hear that it's simple to make one with Boost.Iterator. You'd then just do std::for_each( seconder(a.begin()), seconder(a.end()), boost::bind(f,_1) );
Alexander Shyrokov wrote:
How can make the following code inside of for_each more readable?
void f(char*); typedef std::map<int,char> TMap; TMap a;
std::for_each(a.begin(),a.end(),boost::bind(f,boost::bind(&TMap::value_type::second,_1)));
I use indentation: std::for_each( a.begin(), a.end(), boost::bind( f, boost::bind( &TMap::value_type::second, _1 ) ) ); or a custom for_each for maps: template<class It, class F> void for_each_pair( It first, It last, F f ) { for( ; first != last; ++first ) { f( first->first, first->second ); } } for_each_pair( a.begin(), a.end(), boost::bind( f, _2 ) );
template<class It, class F> void for_each_pair( It first, It last, F f ) { for( ; first != last; ++first ) { f( first->first, first->second ); } }
for_each_pair( a.begin(), a.end(), boost::bind( f, _2 ) ); Could you explain why _2 is used? It looks like f parameter should have two parameters and we are using bind to use only 2nd parameter and ignore the first one? Is it correct?
The second question: why last is not a constant and maybe taken by reference? void for_each_pair( It first, const It& last, F f ) Thanks. -- Regards, Alexander. http://sjcomp.com
On 9/15/06, Alexander Shyrokov <shirokov@unh.edu> wrote:
Could you explain why _2 is used? It looks like f parameter should have two parameters and we are using bind to use only 2nd parameter and ignore the first one? Is it correct?
boost::bind( f, _2 ) is generating a binary functor that, when called, calls the unary function f and passes it the second argument to the functor.
The second question: why last is not a constant and maybe taken by reference? void for_each_pair( It first, const It& last, F f )
Iterators are supposed to be cheap to copy and as far as I can tell it's just the usual way to pass the end iterator. There might also be some weird issues with argument deduction in some compilers. ( I seem to recall something about that... ) HTH, Scott
Alexander Shyrokov <shirokov@unh.edu> writes:
Is it possible to convert boost::bind(&TMap::value_type::second,_1) to a function object or a typedef name?
Yes: boost::function< void (TMap::value_type) > proc = boost::bind(f, boost::bind(&TMap::value_type::second, _1)); would do it if the signatures did match (f is taking char*, not char). Also, some boost:: q are redundant (ADL): boost::function< void (TMap::value_type) > proc = bind(f, bind(&TMap::value_type::second, _1)); Regards, Jens
participants (4)
-
Alexander Shyrokov -
Jens Theisen -
me22 -
Peter Dimov