This is the mail archive of the
mailing list for the Cygwin project.
Re: strtold() availability
- From: VÃclav Zeman <vhaisman at gmail dot com>
- To: cygwin at cygwin dot com
- Date: Mon, 10 Mar 2014 10:11:05 +0100
- Subject: Re: strtold() availability
- Authentication-results: sourceware.org; auth=none
- References: <CAKw7uVhC5X10rCZW7HY3LuAE0VUP9Ds5EWHpN9Qs08jkC7ocXw at mail dot gmail dot com> <20140307200832 dot GC2744 at calimero dot vinschen dot de>
On 7 March 2014 21:08, Corinna Vinschen wrote:
> On Mar 7 17:04, VÃclav Zeman wrote:
>> This bit from /usr/include/stdlib.h hides the `strtold()` function
>> even though the `strtold` symbol appears to be exported from
>> /* On platforms where long double equals double. */
>> #ifdef _LDBL_EQ_DBL
>> #if !defined(__STRICT_ANSI__) || (__STDC_VERSION__ >= 199901L) ||
>> (__cplusplus >= 201103L)
>> extern long double strtold (const char *__restrict, char **__restrict);
>> #endif /* _LDBL_EQ_DBL */
>> This in turn breaks my source because I cannot use `_GLIBCXX_USE_C99`
>> to expose some C++11 features like `std::vsnprintf()`.
>> Shouldn't the `_LDBL_EQ_DBL` guard be removed and the function exposed
>> regardless of `double` and `long double` sizes?
> I fixed that in newlib after a discussion with my co-maintainer.
> strtold now gets defined if _HAVE_LONG_DOUBLE is defined. This is
> the case for Cygwin, for instance.
I think that similar fix needs to be applied to `wchar.h` and `wcstold()`.
Problem reports: http://cygwin.com/problems.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple