comp.lang.c++.moderated Google Group

Re: C++ FAQ

lr...@superlink.net (LR) — Sun, 07 July 2009 09:01:02 UT

[snippage]
int e = -r / 2 ;
if ( e >= 0 ) e -= --x ; // I think required for r == 1
//> while (true);
while(true) {
This time it's for me to note that there is some similarity in our code.
I think yours was more readable. I took the liberty of making a few
minor corrections.

lying typedef's

g3r...@gmail.com (something_fit) — Sun, 07 July 2009 09:03:35 UT

i was writing some code when vc++08 gave me a strange error once i
made a struct inside a namespace of the name of a namespace
have i found a bug in the compiler or is this undefined behavior?
because g++ works as expected
struct dummy{};
namespace people{
struct base{
virtual ~base(){}
void say(){

Re: Enumerations with scope

gmrehb...@gmail.com (yeroen) — Sun, 07 July 2009 09:02:04 UT

On Jul 3, 12:18 pm, Francis Glassborow
If your current compiler supports it, why not just use C++0x strongly
typed enums? The strongly typed enum construct (declared as enum class
{...}) provides just what you're looking for. I know on *nix, strongly
typed enums are supported as of GCC 4.4.

C++ typed interface to hundreds of C functions

graham.k...@gmail.com (PGK) — Sun, 07 July 2009 08:53:28 UT

Hi there,
I'm handling pointers to "objects" described by C-style
struct (e.g. Base), and I manipulate them through a low
level C API. Sometimes a third-party binary function is
applied to two of these, giving a pointer to a third.
extern "C" Base *foo1_A1_B1_C1(Base *, Base *);
extern "C" Base *foo1_A1_B1_C2(Base *, Base *);

Re: Is this safe?

joshuamaur...@gmail.com (Joshua Maurice) — Sun, 07 July 2009 08:48:15 UT

First thing to note: in all libraries, aka all code which isn't "main
()", if you use a namespace member or static class member to implement
exposed functionality, a namespace member of static class member
during its construction in initialization of statics may try to access
your interface, and thus access your namespace member or static class

Re: std::fstream and buffer sizes

b...@gmb.dk (Bo Persson) — Sat, 07 July 2009 21:06:20 UT

It doesn't say that you have to to anything, so just supply a new
buffer. You will have to keep the buffer alive, of course, as long as
it is used.
Bo Persson

Re: Const return type with non-class type

b...@ingen.ddns.info (Bart van Ingen Schenau) — Sat, 07 July 2009 21:06:37 UT

litb wrote:
As far as I can see, it is neither.
Based on the rules for writing standardese, the statement in question
("non-class rvalues always have cv-unqualified types.") is not a
requirement on the implementation or on C++ programs, but it is a
statement of fact. It is irrefutably true (unless there is a defect in

Re: C++ FAQ

lr...@superlink.net (LR) — Sat, 07 July 2009 21:01:14 UT

In which case, we might mention that speed is the thing that we're
optimizing for. Optimizations for space of the executable code and
speed may conflict. Optimizing for readability will likely conflict
with optimizations for space and speed.
I agree. But would you agree that readability and space of the

Re: Mis-feature?

and...@tomazos.com (Andrew Tomazos) — Sat, 07 July 2009 20:56:30 UT

You can't appeal to "bad design" whenever C++ doesn't have a feature
to support something people want to do. I have already provided a
good example, and the design is fine. Many times you have an
interface that is only designed for some subset of classes, and not
just everyone.
Your suggested workaround using inner classes and friend is not

Re: C++ FAQ

r...@zedat.fu-berlin.de (Stefan Ram) — Sat, 07 July 2009 20:54:44 UT

In a language without exceptions:
main_loop( r, s, f )
{ while( ... )
{ if( some error )
/* just return and cleanup will happen */ }}
main_open( r, s )
{ f = open( ... );
if( f )
{ main_loop( r, s, f );
close( f ); }}
main_connect( s )
{ r = connect( ... );
if( r )
{ main_open( r, s );

Re: C++ FAQ

francis.glassbo...@btinternet.com (Francis Glassborow) — Sat, 07 July 2009 20:55:58 UT

Yes and we are in the subjective area. What appeals to one person has
the reverse effect on another
s = socket(...);
if (s >= 0){
r = connect(...);
if (r >= 0){
f = open(...);
if (f >= 0){
while (...) {
if (some error) {
// process error;

Re: Buffer overflow c/c++

tact...@gmail.com (Tactoth) — Sat, 07 July 2009 20:57:02 UT

You may use boundarycheck, it's the most popular one.
-----------
[link]

Re: Buffer overflow c/c++

loufo...@gmail.com (Mathias Gaunard) — Sat, 07 July 2009 20:53:59 UT

There is a great tool to find memory errors, valgrind.

Re: C++ FAQ

francis.glassbo...@btinternet.com (Francis Glassborow) — Sat, 07 July 2009 20:55:30 UT

Here it is
template < class F > inline
void circle_octant ( int r, F f ){
int x = r ; // note that r needs to be larger than 0
int y = 0 ; // I am not sure whether both x and y
// should be unsigned
if(r & 1){
--e;
f(x,y);
e+= ++y;
}
while (true);

Re: how save binary file into variable ??

sjdonnell...@gmail.com (Steve Donnelly) — Sat, 07 July 2009 12:55:30 UT

If you don't know the size of the file beforehand, you can determine
the length at runtime and allocate an array that fits the file
exactly.
The following example was taken from:
[link]
// load a file into memory
using namespace std;
int main () {
int length;