[prog] School problem, C++ and cout.precision(2);

Robert J. Hansen cortana at earthlink.net
Sat Nov 23 20:21:29 EST 2002


> > sometimes ignored by compiler authors.  For instance, Sun's compiler was
> > for years infamously bad with templates; the most trivial use of
> > templates could bloat code size by over a meg.
> 
> Hmm...How do Microsoft's compiler fare ? Do they add their /custom/
> extensions ?

Visual C++ 7 is a first-rate compiler, probably a little bit better than
GCC 3.2.  Visual C++ 6 and previous weren't C++ compilers.  No, I'm not
kidding; the language Visual C++ 6 supported was almost identical to
C++, except that there were lots of places where Microsoft would let you
do things that were absolutely forbidden in the C++ specification.

VC6 also had the worst STL implementation I've ever seen.  It was just
tragically, comically bad.  You could write code that was one hundred
percent perfect and get over 4,000 erroneous warnings--I know: it
happened to me.

> So, is GLIBC the library that implements C functions like *printf() ? I

Essentially.

> find KDE to be real fast under FreeBSD. Is it because FreeBSD does not
> use GLIBC ?

I don't know; my FreeBSD experience is minimal.  Installed it a couple
of times, keep on coming back to Linux.

> Do you still deal with complex pointer arithematic ? I dont mean simple

Hardly ever.  That's one of the big wins of the STL; you very rarely
have to do pointer arithmetic more complex than increment and decrement.

> This is greek to me.

:)  It's Greek to a lot of C++ programmers, too.  A lot of corporate C++
programmers are C programmers who took a weeklong course in C++ and
became "the office C++ guy".  Really learning C++ thoroughly takes
several years.  I'm still learning new stuff today after thirteen years.

> So which GUI framework are you using ?

Gtkmm.  I prefer Qt, API-wise -- Gtkmm is almost a caricature of How To
Make A Bad C++ Library By Insisting On Using Every Last Feature Of C++
-- but my desktop is GNOME, so...





More information about the Programming mailing list