Re: [Boost-users] Newbie -- Regex, Unicode, XCode and MacOSX

From: John Maddock (john_at_[hidden])
Date: 2004-07-23 05:18:30

> > Are these symbols correctly defined?
> I'm not exactly sure which symbols you're referring to, but...  The
> BOOST_NO_CWCHAR is defined subsequent to a comment in macos.hpp which
> states: "// Using the Mac OS X system BSD-style C library"
> Thus, I'm
> not sure if BOOST_NO_CWCHAR is defined "correctly" since I'm not sure
> what the justification is based on the #define's preceding comment.

I don't have access to that platform, so I just go what people tell me,
BOOST_NO_CWCHAR should be defined only if the platform is missing a
conforming version of the <cwchar> header.

> As
> for the boost::wregex and boost::wcmatch symbols, if BOOST_NO_CWCHAR is
> defined, then BOOST_NO_WREGEX is subsequently defined, which results in
> the typedefs for wcregex and wcmatch in <boost/regex/v4/regex.hpp>
> being skipped, so these symbols are not defined (again, I'm not sure if
> this is "correct" behavior since I"m not sure of the reasoning behind
> the BOOST_NO_CWCHAR in the first place).

That's correct, if there is no <cwchar> header then regex can't provide wide
character support.

> > Can you comment them out and still get
> > a clean compile?
> If I comment out the #define BOOST_NO_CWCHAR line regex compiles with
> no errors using the gcc.mak file in <libs/regex/build> directory.

OK, so can you use wregex etc now?

If not check to see if BOOST_NO_CWCTYPE or BOOST_NO_STD_WSTRING are defined,
and if so comment out these defines as well...

John.