Hello, I think it is a problem of two-phase name lookup which this link should help explain : http://gcc.gnu.org/ml/gcc-help/2004-04/msg00382.html Regards, PH --- Leon Pollak <leon.pollak@xxxxxxxxx> a écrit : > I should agree with you in the case when I should > want to refer > ANOTHER i (from another template parameter value). > But in this case I > have based class B on THE SAME value of template > parameter. I was sure > that this indicates very clear that I want the same > 'i' which resides > in A<T> instantiation...:-( > > On 5/9/06, Blake Huff <stangmechanic@xxxxxxxxx> > wrote: > > I'm pretty bad with this, but here's my guess: > What if B had > > multiple templates, i.e., A, C, D, and E, which > each had their own > > i? Compiler needs to know where to go. Someone > else might have a > > better technical explanation (or maybe I'm wrong > altogether!). > > > > Blake > > > > > > On May 9, 2006, at 6:27 AM, Leon Pollak wrote: > > > > > Hello, all. > > > I shall be very thankful for the help with the > following compilation > > > problem. The code example is: > > > > > > template <typename T> > > > class A { > > > protected: int i; > > > }; > > > > > > template <typename T> > > > class B : public A<T> { > > > public: B() {i = 0;} > > > }; > > > > > > My gcc version 4.0.2 20051125 (Red Hat 4.0.2-8) > produces: > > > error: 'i' was not declared in this scope. > Writing: > > > public: B() {A<T>::i = 0;} > > > solves the problem, but why do I need to specify > this? Isn't it > > > obvious? > > > > > > Many thanks ahead. > > > > > > Leon. > > > > Blake Huff > > stangmechanic@xxxxxxxxx > > > > > > > > > > > > > > > > > > > ___________________________________________________________________________ Yahoo! Mail réinvente le mail ! Découvrez le nouveau Yahoo! Mail et son interface révolutionnaire. http://fr.mail.yahoo.com
