comp.lang.c++.moderated Google Group

Re: help on round robin tournament

2009-11-09T02:28:14Z

Il Sun, 08 Nov 2009 16:29:38 -0600, red floyd ha scritto:
Well, I know no comp.algorithms newsgroup...

Re: #include file optimizer

2009-11-09T02:27:12Z

In my personal opinion this is rather unsatisfactory.
With self-contained headers (headers that include other headers as
needed) a new header dependency only triggers client recompilation, but
no editing of clients required (provided interfaces don't change).
A good practice they use in boost to ensure that headers are

Re: help on round robin tournament

2009-11-08T22:29:38Z

Sorry. Then you're better off asking in comp.algorithms. There's
nothing C++ specific in your question, other than your use of a deque.

Re: operator+() code for objects

2009-11-08T22:29:18Z

Huh?
Get a good book. Many things changed.
Several things here:
1. Make one-argument constructor 'explicit' to avoid accidental
construction.
2. Do not create a virtual destructor for objects you don't plan to use
polymorphically.
3. You #include <iostream> but use printf() for which you didn't include any

Re: operator+() code for objects

2009-11-08T20:22:16Z

In article
<bea3b28d-235b-4b0d-8421-9fef3 ecb8...@a31g2000yqn.googlegrou ps.com>,
...
And here is your problem. Your copy constructor should take its
argument by constant reference:
vec(const vec &rv) : mx(rv.mx) { printf("in copy constr\n"); };
Currently you cannot copy const objects, and so the compiler complains

Run-time overhead of text-based storage formats for numerical data

2009-11-08T20:22:11Z

Hi all.
A couple of weeks ago I posted a question on comp.lang.c++ about some
technicality
about binary file IO. Over the course of the discussion, I discovered
to my
amazement - and, quite frankly, horror - that there seems to be a
school of
thought that text-based storage formats are universally preferable to

Re: Check .lib is complete

2009-11-08T20:22:09Z

On Nov 2, 11:53 am, Louis Lavery <Lo...@LaverREMOVE.demon.co.uk >
wrote:
google dependency walker. It's a free download. That will go a long
way toward telling you. But even that will sometimes give an external
dependency that is not needed.
HTH

Re: operator+() code for objects

2009-11-08T20:16:45Z

My comments are inline with te code:
Your use of printf() (why not use iostreams?) eans that you should
include the header it s declared in:
This is your main problem. A copy constructor should take a const
reference as it's parameter - you want:
vec( const vec &rv) : mx(rv.mx) { printf("in copy constr\n"); };

Re: static_cast(signed type);

2009-11-08T20:18:54Z

That's what it was. Now it's a hexadecimal constant prefix.

Re: operator+() code for objects

2009-11-08T20:17:27Z

silentway schrieb:
<snip>
I hope so. The problem is you copy contructor taking a refernce to not
const vec. The term A + B creates a temporary and a temporary can't be
bound to a reference to not const vec.
vec(const vec &rv) : mx(rv.mx) { printf("in copy constr\n"); };
^^^^^
will solve the problem.

Re: #include file optimizer

2009-11-08T20:18:53Z

In the headers-don't-include-other-he aders style, the "prerequisites"
comment will now read
//
// prerequisites:
// <string>
// <list>
//
and all clients are required to update their source code accordingly.
Hmm. I agree that it's not an ideal situation, but it seems to me
that it's basically the same as that for any other change in class x

Re: operator+() code for objects

2009-11-08T20:17:28Z

The problem here is not actually that you cannot return a local variable, but
rather that you do not have a copy constructor that can perform the required
copy initialization of F from the value returned by your overload operator+()
(see below).
I was not seriously coding in C++ 10 years ago and have never used the Zortech

Re: operator+() code for objects

2009-11-08T20:16:40Z

Copy constructor should have parameter passed by const reference
like this:
vec(const vec& rv) : mx(rv.mx) { /* wathever */ }
There is no need (I can think of) for returning const value and yes,
this method should rather be const (you do not whant to change value
in this vector while calculating its sum with A)

Re: operator+() code for objects

2009-11-08T20:16:40Z

[snip]
Your copy constructor takes a reference-to-NON-CONST. This probably
explains the errors you're getting.
You should get rid of your copy ctor, assignment operator *and*
destructor (unless you want it to be virtual for some reason). The
compiler-generated ones are correct and should be preferred.

Re: operator+() code for objects

2009-11-08T20:17:33Z

silentway schrieb:
Let's see...
Your copy constructor takes a non-const reference.
Your operator+ returns a temporary. If you do
vec v = a + b;
the temporary returned by op+ cannot be used to copy-construct v,
because temporaries can only be bound to const references.
So change your copy ctor to
vec(const vec &rv)