Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!mcvax!ukc!pyrltd!slxsys!jclark!jjc
From: jjc@jclark.UUCP (James Clark)
Newsgroups: comp.lang.c++
Subject: overloaded functions
Summary: rules could be smarter?
Message-ID: <JJC.89May30110723@jclark.UUCP>
Date: 30 May 89 10:07:23 GMT
Sender: jjc@jclark.UUCP
Distribution: comp
Organization: None
Lines: 31

I was wondering whether the C++ rules for overloaded function names
could not be a little smarter. For example, the following won't
compile (with cfront 1.2 or g++):

struct SubString {
};

struct String {
  String(SubString &);
  operator char *();
};

String operator+(String &, String &);
String operator+(String &, const char *);

String foo(SubString &s, String &t)
{
  return s + t;
}
  
The choice here is between:

- converting s to a String and doing nothing to t; and

- converting s to a String and converting t to a char *.

It would be nice if C++ was able to do the right thing, and choose the
first. Is there some good reason why it doesn't?

James Clark
jjc@jclark.uucp