Re: template argument required

[Larry I Smith]

Patrick Rammelt wrote:
> Larry I Smith wrote:
>
> > 'class' should not be required.  In fact, it may
> > confuse the issue (on some compilers).  Try:
> >
> >   A<T> & foo (X p) {
> >     ...
> >   }
>
> Yes, that works too. Thanks again. Now I have three working and one 
> non-working version:
>
> Working:
> class A<T>& foo (X p) { ... } // 1)
> A<T>& foo (X p) {... }        // 2)
> A& foo (X p) { ... }          // 3)
>
> Not Working:
> class A& foo (X p) { ... }    // 4)
>
> So is it a gcc-bug? In my opinion 3) and 4) should not behave 
> differently (I do not really understand why "class" confuses my new 
> little gcc).
>
> Ciao,
> Patrick

I'm going to go out on a limb here; I'm sure that smarter
folks will correct me if I'm wrong...

1) 'class' is redundant and un-necesary
2) declares that foo() returns a ref to an 'A<T>'
3) declares that foo() returns a ref to an 'A' whose 'T' is not
   specified (similar to a base class??).  not very useful?

You might get better (correct??) answers in the newsgroup:

 comp.lang.c++

Some smart folks hang out there.

Regards,
Larry

--
Anti-spam address, change each 'X' to '.' to reply directly.