const inconsistency in g++ 2.95.2 sparc-sun-solaris2.7

[Alexandre Oliva]

On Apr  6, 2000, Tom Ace <crux@QNET.COM> wrote:

> I've noted what seems to be an inconsistency in gcc.

Nope.  If it's an inconsistency, it's in the C++ Standard.  But I
don't think it is.

> test.cc:4: assignment to `char *' from `const char *' discards qualifiers

> char  *c;
[snip]
> void t3(int x) { c = x ? "t" : (char *) 0; }     

The type of `"t"' is `const char [2]', so, after array-to-pointer
decaying, it would be `const char*'.  However, for backward
compatibility, the C++ Standard allows a string literal to decay to
non-const `char*', but this conversion is deprecated.  That's why `c =
"t"' works.  But when the string literal appears in a ?: expression,
there's no reason why it should decay to `char*' instead of `const
char*'.  Furthermore `(char*)0' can be implicitly const-qualified, so
the result type of the ?: expression is `const char*'.  The
work-around is to explicitly cast the string literal to `char*'.  But
The Right Thing (TM) is to const-qualify `c'.

-- 
Alexandre Oliva    Enjoy Guaraná, see http://www.ic.unicamp.br/~oliva/
Cygnus Solutions, a Red Hat company        aoliva@{redhat, cygnus}.com
Free Software Developer and Evangelist    CS PhD student at IC-Unicamp
oliva@{lsd.ic.unicamp.br, gnu.org}   Write to mailing lists, not to me