b91926@fsgi01.fnal.gov (David Sachs) writes:

>This reminds me of another problem with Template functions. Suppose
>You have a section of code like this:
>
>...
>template <class T> inline T max(T a, T b)
>{
>  return a>b?a:b;
>}
>
>long max(long, long);
>....
>
>
>What does the last declaration declare?
>
>A. An external function, long max(long,long)
>
>-- or ---
>
>B. Instantiate the template function max<long> and allow
>conversion of its arguments.

ARM 14.4: "Declaring a functino with the same name as a functino
template with a matching type constitutes a declaration of
a specific template function".   What does this mean?  Well,
it means that the compiler will decide at link time which
of (A) and (B) is correct.  If a definition of an external
function `long max(long,long)' is found at link-time, then
that function is considered a template specialization, otherwise
code will be generated using the template.

--
Fergus Henderson - fjh@munta.cs.mu.oz.au

Article 7478 of 7482, Sat 13:48.
Subject: Re: template implications for deprecated float->int conversion
maxtal@physics.su.OZ.AU (John Max Skaller @ School of Physics, University of Sydney, Australia) writes

 > >> In article <CMI5r7.26q@cbnewse.cb.att.com> grumpy@cbnewse.cb.att.com (Paul J Lucas) writes:
 > >>> I just finished reading Bjarne's paper, "New Casts Revisited";
 > >>> specifically, Chapter 20, "Deprecated Features."  There is says
 > >>> that conversions from floating-point to integer are deprecated.
 >
 >      ...
 >
 >  Well, Bjarnes book is wrong. Presumably because the text
 > was written in expectation of a vote to deprecate such conversions.

Please note that Paul was referring to a PAPER I wrote for the committee.
It contained a proposal and only part of that proposal was accepted.
He did not refer to any of my BOOKs.

I try very hard to be accurate in my writings and to the best of my
knowledge I have never written anything depending on the expected
outcome of a vote (maybe something on the form ``the following will
be voted on at such and such meeting ...'' which does not mislead).

Also, none of my books have a chapter 20.

Statements of the form ``X's book is wrong'' cast seriously doubt on
``X's book'' as a source of information so please be very careful with
such statements.

 - Bjarne

In article <CMw68G.73r@cbnewse.cb.att.com> grumpy@cbnewse.cb.att.com (Paul J Lucas) writes:
>From article <CMqqLB.E5v@ucc.su.OZ.AU>, by maxtal@physics.su.OZ.AU (John Max Skaller):
>> In article <CMI5r7.26q@cbnewse.cb.att.com> grumpy@cbnewse.cb.att.com (Paul J Lucas) writes:
>>> I just finished reading Bjarne's paper, "New Casts Revisited";
>>> specifically, Chapter 20, "Deprecated Features."  There is says
>>> that conversions from floating-point to integer are deprecated.
>>> Does this mean that:
>>>
>>>  template<class T> T Max( T, T );
>>>
>>>  int i = 1;
>>>  double d = 2.0;
>>>
>>>  double m = Max( i, d );
>>>
>>> will now work because the only sensible thing to do is to
>>> promote 'i' to double?
>>
>>  Doesnt matter. Deprecation of old style casts was rejected.
>
> And what's that got to do with anything?

 Well, Bjarnes book is wrong. Presumably because the text
was written in expectation of a vote to deprecate such conversions.
But the resolution on new style casts explicitly doesnt
deprecate any old style casts at all. That may change before
the Standard is published.

--
        JOHN (MAX) SKALLER,         INTERNET:maxtal@suphys.physics.su.oz.au
 Maxtal Pty Ltd,      CSERVE:10236.1703
        6 MacKay St ASHFIELD,     Mem: SA IT/9/22,SC22/WG21
        NSW 2131, AUSTRALIA

I have a stupid question regarding array of classes.
There is an array of student class.
class comp{
private:
 element student[20];
 ....
};
There is an another class which holds a student's information such as
its name.
class element{
private:
 char *name;
 ....
};

Say if I have to assign Joe's information (stored in the element class) to the
student[2], in what way I can do this.
- A C++ learner.

This reminds me of another problem with Template functions. Suppose
You have a section of code like this:


...
template <class T> inline T max(T a, T b)
{
  return a>b?a:b;
}

long max(long, long);
....


What does the last declaration declare?

A. An external function, long max(long,long)

-- or ---

B. Instantiate the template function max<long> and allow
conversion of its arguments.

 I just finished reading Bjarne's paper, "New Casts Revisited";
 specifically, Chapter 20, "Deprecated Features."  There is says
 that conversions from floating-point to integer are deprecated.
 Does this mean that:

  template<class T> T Max( T, T );

  int i = 1;
  double d = 2.0;

  double m = Max( i, d );

 will now work because the only sensible thing to do is to
 promote 'i' to double?
--
 - Paul J. Lucas
   AT&T Bell Laboratories
   Naperville, IL

In article <CMI5r7.26q@cbnewse.cb.att.com> grumpy@cbnewse.cb.att.com (Paul J Lucas) writes:
> I just finished reading Bjarne's paper, "New Casts Revisited";
> specifically, Chapter 20, "Deprecated Features."  There is says
> that conversions from floating-point to integer are deprecated.
> Does this mean that:
>
>  template<class T> T Max( T, T );
>
>  int i = 1;
>  double d = 2.0;
>
>  double m = Max( i, d );
>
> will now work because the only sensible thing to do is to
> promote 'i' to double?

 Doesnt matter. Deprecation of old style casts was rejected.

--
        JOHN (MAX) SKALLER,         INTERNET:maxtal@suphys.physics.su.oz.au
 Maxtal Pty Ltd,      CSERVE:10236.1703
        6 MacKay St ASHFIELD,     Mem: SA IT/9/22,SC22/WG21
        NSW 2131, AUSTRALIA