This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c++/32240] wstring::length() does not work properly with --short-wchar
- From: "pcarlini at suse dot de" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 7 Jun 2007 08:51:09 -0000
- Subject: [Bug c++/32240] wstring::length() does not work properly with --short-wchar
- References: <bug-32240-14667@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #1 from pcarlini at suse dot de 2007-06-07 08:51 -------
(In reply to comment #0)
> std::wstring wstr = L"Hello, World!";
A couple of clarifications. First, your way of using --short-wchar cannot work
in principle, because std::string functions are exported from the installed
.so, which is built *without* --short-wchar. At minimum, for consistency, the
library itself should be rebuilt at installation time with that switch. In the
specific case above, where you are using the constructor taking a "C" string,
it forwards to the underlying glibc wcslen in order to compute the length of
the initializer "C" string, which is computing the length incorrectly: we
alredy knew that, glibc doesn't work completely fine together with
--short-wchar: the docs are pretty clear that on GNU systems a wchar_t is
always assumed to be 32 bits. I don't think there is a way to deal portably
with such issues within C++03: the new C++0x will include char16_t and char23_t
types and everything will be straightforward. In the meanwhile, as far as I
know, most of the library has chances to work fine if you provide replacement
std::char_traits not forwarding to underlying C functions.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32240