Mail Archives: djgpp/2000/05/17/16:38:17
Eli Zaretskii wrote:
>
> On Wed, 17 May 2000, Alexei A. Frounze wrote:
>
> > > > IMHO there must no be any C++ library, just a C++ compiler itself, if people
> > > > don't have standard for C++ library.
> > >
> > > I don't think this is possible. All the C++ classes are implemented
> > > in a library (libstcxx.a, in the DJGPP case). You *must* have that
> > > library, and you *must* include the appropriate headers, if you want
> > > to use data types other than the basic ones (int, char, etc.).
> >
> > I can define my own classes instead.
>
> Sure. And you can define your own complex type as well, assuming that
> you write all the code that implements it.
>
> > So why should I use C++ library, if it's not standartized?
>
> Because it *is* standardized.
Is standard or is <being> standartized?
> The recently-adopted ANSI/ISO C++ Standard
> includes the description of a Standard C++ Class Library.
Okay, what does it state about the complex class (what is the date of the
standard release)?
> > Btw, what so I need size_t for, if both size_t and
> > int equal the same machine word?
>
> You need it for portability.
>
> size_t is not equal to int. Its precise definition depends on the
> implementation. For example, a 64-bit machine could use unsigned long
> (64-bit) for size_t. There are library functions that accept or return
> size_t, and if you use int instead, you will get either warnings or bugs.
I said they equals machine word. I.e. 32-bit on i386+. For sure developer
can setup them differently... But this way seems to be common.
bye.
Alexei A. Frounze
-----------------------------------------
Homepage: http://alexfru.chat.ru
Mirror: http://members.xoom.com/alexfru
- Raw text -