Bryan Ischo wrote:
Yes, but since the class which contains the constructor which has the
argument which has the default is:
basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> >
Then shouldn't npos be identified as:
std::basic_string<wchar_t, td::char_traits<wchar_t>,
std::allocator<wchar_t>
::npos
instead of :
std::basic_string<_CharT, _Traits, _Alloc>::npos
???
In other words, shouldn't the type of npos be identified by the actual
template parameter values used in declaring the template instance in
which npos is being referenced, instead of by the variable names used in
the template definition itself?
I think what happens is that GCC's expression-to-string conversion is
designed to be used during error message construction. GCC error
messages look like this:
error: initializing argument 1 of ‘std::basic_string<_CharT, _Traits,
_Alloc>::basic_string(const _CharT*, const _Alloc&) [with _CharT = char,
_Traits = std::char_traits<char>, _Alloc = std::allocator<char>]’
(the code tried to construct std::string(1) which is an error). Note
that the prototype is given with the template formal parameters instead
of the actual arguments. Then the rest of the message maps the
parameters to their values. GCC-XML is using the same
expression-to-string functionality from GCC. This is not the only
limitation of the expression-to-string conversion that people have
encountered. In general I think the "default" attribute is useful
mostly for human reference.
-Brad
_______________________________________________
gccxml mailing list
[email protected]
http://www.gccxml.org/mailman/listinfo/gccxml
