Path: utzoo!attcan!uunet!wuarchive!udel!rochester!newkodak!kodak!ektools!randolph
From: randolph@ektools.UUCP (Gary L. Randolph)
Newsgroups: comp.lang.c++
Subject: Re: implicit conv to int in subscript
Message-ID: <2670@ektools.UUCP>
Date: 30 May 90 14:22:36 GMT
References: <684@dyndata.UUCP>
Sender: randolph@ektools (Gary L. Randolph)
Reply-To: randolph@ektools.UUCP (Gary L. Randolph)
Organization: Eastman Kodak, Dept. 47, Rochester NY
Lines: 57

In article <684@dyndata.UUCP> dan@dyndata.UUCP (Dan Everhart) writes:
#
#Consider the following program:
#
#class Index
#   {
#public:
#   Index (int i) : v (i) {};
#   operator int ()   { return v; };
#private:
#   int   v;
#   };
#
#Index x (0);
#char  ac [10];
#
#f ()
#   {
#   ac [x] = 'a';                // this line fails
                                    ^^^^^^^^^^^^^^^It shouldn't
#   
#   ac [(int) x] = 'a';
#   ac [int (x)] = 'a';
#   }
#
#
#On the marked line, I expected the compiler to make an implicit
#conversion of x to int, since the Index class has an operator int ()
#function defined.  Zortech 2.06 gives the error message:

Your expectation is valid!

#"opint.cpp", line 17 Syntax error: illegal pointer arithmetic
#Had: <near *><char>
#and: <struct Index>

#The subsequent two lines, with their explicit conversions, compile
#fine. 

You should not have to cast!

#Although I could not find any example in The C++ Primer of exactly
#this nature, there is a close one on p. 283 where Lippman defines a
#SmallInt class, with similar conversion operators and says, "A
#SmallInt class object can now be used *anywhere* in int object can be
#used."  (Emphasis mine.)  

Lippman is correct.


#So, should this work?

Yes.
I have tested your code using Sun C++ 2.0 and all is as you had
hoped!  (minor point: Obviously f() should be defined as a void function)

Gary