| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Comments:
<0> :P
<1> I mean I have this: template< typename A, typename B > struct R{}; template< typename A, typename B > struct S; and I want to specialize S where B is any R
<1> it might not be possible
<0> hey!
<0> that's exactly what i need!!!
<1> I doubt it
<0> dude.. tell me if u get the answer
<0> i'm trying to build a Gram matrix
<0> of all permutated combinations of a finite set of objects
<0> permuted
<0> and i want the user to p*** in a templated Functor
<0> for each term in my Gram matrix..
<1> functors shouldn't be a problem
<0> however, the type for such a functor really "fixes" the cl*** definition
<1> they're just a type that you can use operator() on
<0> but the functor itself.. is a template..
<1> shouldn't matter
<0> template<cl*** T1, cl*** T2> cl*** TBinOp
<0> {public: virtual T2 operator()(T1 &a, T1 &b) = 0;};
<0> that's my functor for Binary Operations
<1> template <typename T> call_functor( T functor ){ functor(); }
<1> throw a void in there for return type
<1> just look at how the standard library does it
<1> check out for_each, and all of those
<0> yeah, but notice that u used T already for the functor parameters? that affects the Gram matrix builder that accepts the Functor
<1> not really
<1> I only templated a function
<1> which is the same sorta thing we just tried to show you how to do
<0> Tensor<T> EigenSolver<T>::buildAffMatrix(T2 *v, int n, TBinOp<T2, T2> &binOp)
<1> struct A{ template <typename T> void call_functor( T functor ){ functor(); } };
<1> there
<1> the functors you use don't affect A at all
<1> yeah?
<0> but look here:
<1> template< typename T > struct A{ template <typename FunctorType> void call_functor( FunctorType functor ){ functor(); } };
<0> void call_functor( T functor ) << u already specified that this member function only accepts T type functors.. but i'd like that T to be some variant.. like. decided at runtime
<1> well there's your problem
<1> templates aren't a runtime thing
<1> runtime polymorphism usually == virtual functions
<1> templates are a generic programming tool
<0> OK.. i see what you're doing now
<0> so you're saying.. u can localize the template specialization by declaring the template args at the member function level.. this way it won't affect the carrying cl***
<1> no, it's not a specialization
<1> it's simply a template member function
<0> wait.. is that even possible?
<1> template member functions? yes, I just showed you one
<0> i can declare the cl*** without <T1, T2>
<0> and use a member function with <T1 T2>
<0> ?
<1> that's what we've been showing you how to do for awhile
<0> hmm
<1> template< typename T > struct A{ template< typename T2 > void foo(){} }; int main(){ A<int> a; a.foo<float>(); a.foo<char>(); }
<2> hi all
<2> should strcat work also in C?
<1> also?
<1> strcat originated with C
<2> so it's ok..
<2> ?
<1> this is #C++, by the way
<1> C is in #C
<2> what's their key today?
<2> i dont know
<1> should be main
<1> /join #C main
<1> you're welcome
<1> oh, figured out my nonsense anyway
<0> thanks~~ rdragon you're the best
<0> rdragon
<0> template< typename T > struct A{ template< typename T2 > void foo(){} }; int main(){ A<int> a; a.foo<float>(); a.foo<char>(); }
<1> yes?
<1> i think you need to hit the books
<1> like AC++, or meyers
<3> What's AC++?
<3> hello
<1> http://www.rudbek.com/books.html
<1> AC++ = accelerated C++
<1> I'm off to bed
<3> ok thx :)
<4> hi
<0> is template nesting even legal?
<0> damnit vc6 sucx
<5> its that time of year for nesting ok
<5> in the northern hemisphere
<0> Is it possible to specialise a member function of a cl*** template without specialising the whole template?
<5> yes
<0> k
<5> not with VC6
<0> yup. i just discovered
<5> which is why we said dump vc6
<5> its just not worth the pain
<6> In case we didn't mention it, yet: dump vc6
<0> haha
<0> okok already
<6> skip over vc 7.0
<5> vc8 is free
<6> Use 7.1 or 8.0
<0> is there some service pack for vc6 that will enable template member specialization
<5> nuke your p0rn collection and free up the space
<0> or some cheat code
<6> dextre1: no
<0> hey. leave my porn collection out of this
<5> not a chance
<4> why get a service pack for vc6, if vc8 is free?
<4> and i deleted half my porn collection lately, i feel guilty :/
<5> if you want to write something approaching c++ dump vc6 :) just in case we forgot to mention it
<0> u'll be missing it
<6> Although there were several sevice packs, none of them fixed the most important problems of VC6.
<0> why is vc a half baked piece of crap
<0> vc6
<5> its very old
<5> nearly 10 years
<4> the better question is why do people still use it?
<5> an eternity in computing
<0> is there a way to tell if a cl***'s method is being used as an LHS or RHS in code?
<0> is it possible to p*** a functor that carries more information than specified in the operator() arguments?
<0> like.. in the constructor? i find it hard to tell apart the syntax for "operator()" and an ordinary constructor.
<0> for instance.. if i want to build a functor that calculates scaled euclidean distance.. in addition to the 2 arguments that are p***ed as arguments into the functor operator(), there needs an extra standard deviation factor to scale the difference.. is there a way to "preset" the functor to know my stdDev in advance so i don't have to p*** it in the operator() arguments?
<6> Yes, that is called binding.
<6> You should check out Boost.Bind at http://boost.org/
<0> wait. i asked a stupid question
<0> got it thanks
<0> i thought i was calling the () when i p***ed the functor.. but in fact i was using the default constructor
<5> having a brain fart moment:
<5> struct base {//stuff}; struct other : public base {//more stuff};
<5> dosome(base& thing)
<5> can I p*** a other into this ?
<5> w/o slicing
<0> what is slicing.. i see that a lot
<5> bits of the cl*** getting chopped off
<5> sort of
<0> that sounds horrible
<0> why does it happen
<6> bealtine: a reference doesn't cause slicing. The slicing will happen if you copy.
<5> copy as in ?
<6> dextre1: it happens when you do something like this: derived d; base b = d; // b will not be a complete copy of d
<5> i need to get an "other" to p*** and sometimes a base
<6> References don't create new object, so p***ing by reference doesn't cause any slicing.
<0> i see!
<5> however when I p*** the "other" it seems to slice
<6> However, if you copy later then using a reference won't help. E.g. vector<base> v; v.push_back(d); // will cause slicing although push_back takes a reference.
<0> roll your own operator=
<6> dextre1: no, that won't work well.
<5> I dont think that will work
<5> i'm just trying to make this silly C style protocol work "nicely"
<7> Who Can Help Me?!
<5> did you ask a question?
<7> I Want To Learn The Fonctionnement Of Graphe In langage C
<5> joic #c
<7> can You Give me Some Pages Web Traiting It??
<5> ./join #c main
Return to
#c++
or
Go to some related
logs:
#london
install urpmi +ubuntu
kndels
#MissKitten
tequila Cure For Shyness
Mrxvt screenrc
what does works toilet cleaner contain
#AllNiteCafe
#c
dan_feld
|
|