Re: C++ templates and overloading

For the record, (it has only been implied in the text of interaction
here so far) the result type does not participate in C++ name mangling
or function reference resolution; atleast in products that I have
used. (Would love to hear otherwise in products or evolving
standards).


Also note that posting
From: Robert A Duff bobduff@world.std.com
Date: 4 May 2000 17:17:06 -0400
<<
C++ allows various implicit type conversions, which complicates the
type-matching.


>>
Is a wonderful practice of understatement. Getting the compiler to
match and gen is one trick: getting your tool user to see how that one
takes priority is another trick.


  The original poster is interested in resources that talk about the
issues of overloading (and polymorphism). So the C/C++ standard itself
may be a good source.


In the original post
From: Ugo dallago@sci.uniud.it
Date: 1 May 2000 13:20:26 -0400


<< I'm looking for some resources which talks about this problem, i.e.
polymorphism/overloading in IMPERATIVE languages. >>


I wonder if the languages discussed, C++ and Ada, ARE
'imperative'. The original poster is rather emphatic. Have we missed
something?


In a bygone era 'functional' languages like C were not considered
'imperative'. I am not mincing words here, I hope. What is the
meaning the original poster attaches to the word?


Ada is not my thing; but it is neither 'imperative', nor, as I
understand it, 'functional'.


Such distinctions were all the rage once upon a time. Object
orientation is definitely being jammed into all of the old traditional
languages that were once rigorously distinguished as 'imperative'. So
maybe there is something special in the original questions emphasis.


By cc I invite the original poster to send email if there is a need
for assistance getting to the C standard as a study. (Smalltalk and
Java have some of these aspects, too: so standards can be a guide to
atleast the arcane vocabulary and itemization of possible
'complications'.)


Best Wishes,
Bob Rayhawk
RKRayhawk@aol.com
[The type matching rules for C++ aren't likely to change, the standards
effort is just standarizing the current practice, in which the result
type isn't part of the name but the argument types are. -John]

Return to the comp.compilers page.
Search the comp.compilers archives again.