Hello,

 In Stroustrup's C++ 3rd Ed., he mentions that the standards
committee has deprecated the use of keyword 'static' to limit a
declaration/definition to file (translation unit) scope.  Instead, he
advises the use of anonymous namespaces.  Earlier in the text, he mentions
that namespaces are open, in that declarations/definitions may be added at
any point in time by "reopening" the namespace with "namespace <name>".
These factors combined have lead me to design the following style of code:

---some_file.cc---

namespace {
   void some_function();
}

int main( int, char*[] ) {

 // ...
 some_function();
 // ...
}

namespace {
   void some_function() { /* ... */ }
}

---end file---

However, my compiler will not link this.  I was then told that each
anonymous namespace was unique, which implies that anonymous namespaces are
closed.  Is this true?  If so, doesn't that suck?  This forces us to
forward-define all functions if we wish to avoid using 'static'; like this:

---some_file.cc---

namespace {
   void some_function() { /* ... */ }
}

int main( int, char*[] ) {
 // ...
 ...

---end file---

This would seem to detract from style of coding much more than the use of
'static'!  Can anyone definitively advise me on what is the correct
behavior of anonymous namespaces?  I am hoping that they are open for the
scope of the entire translation unit.

thanks,
-bp
--
B. James Phillippe <bryan@terran.org>
Linux Software Engineer, WGT Inc.
http://earth.terran.org/~bryan



[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://reality.sgi.com/austern_mti/std-c++/faq.html              ]

In article 100000@earth.terran.org, "B. James Phillippe" <bryan@terran.org> writes:
>
> In Stroustrup's C++ 3rd Ed., he mentions that the standards
>committee has deprecated the use of keyword 'static' to limit a
>declaration/definition to file (translation unit) scope.  Instead, he
>advises the use of anonymous namespaces.  Earlier in the text, he mentions
>that namespaces are open, in that declarations/definitions may be added at
>any point in time by "reopening" the namespace with "namespace <name>".
>These factors combined have lead me to design the following style of code:
>
>---some_file.cc---
>
>namespace {
>   void some_function();
>}
>
>int main( int, char*[] ) {
>
> // ...
> some_function();
> // ...
>}
>
>namespace {
>   void some_function() { /* ... */ }
>}
>
>---end file---
>
>However, my compiler will not link this.  I was then told that each
>anonymous namespace was unique, which implies that anonymous namespaces are
>closed.  Is this true?

No, your compiler is wrong. The unnamed namespace acts like it has
a unique name in each different compilation unit, but the same
unique name within one compilation unit. Reference: draft standard
section 7.3.1.1, "Unnamed namespaces".

Your code is OK as you show it.

---
Steve Clamage, stephen.clamage@sun.com



[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://reality.sgi.com/austern_mti/std-c++/faq.html              ]