[libc++ patch] (and/or standard bug?): Let std::stof() call strtod / wcstod instead of strtof / wcstof

[Marshall Clow]

On Jun 4, 2014, at 5:49 PM, Nico Weber <thakis at chromium.org> wrote:

> On Thu, Jun 5, 2014 at 2:21 AM, Richard Smith <richard at metafoo.co.uk> wrote:
> On Wed, Jun 4, 2014 at 3:27 PM, Joerg Sonnenberger <joerg at britannica.bec.de> wrote:
> On Wed, Jun 04, 2014 at 02:51:49AM +0200, Nico Weber wrote:
> > 21.5p4 and 21.5p11 say that std::stof() and std::stod() are both supposed
> > to call strtod() (for char*) or wcstod() (for wchar_t*). libc++ currently
> > calls strtof() / wcstof() from std::stof(), so the attached patch fixes
> > this.
> 
> This looks like a bug in the standard.
> 
> It is:
> 
> http://lwg.github.io/issues/lwg-active.html#2009
> 
> Cool, thanks. I hope that proposed resolution gets a slightly more detailed text. Similar wording in the C standard apparently implies something along the lines of "`1.e60` is a valid IEEE 754 spelling of INFINITY, and INFINITY is a representable value" ( http://www.open-std.org/jtc1/sc22/wg14/www/docs/dr_025.html ) (which as far as I understand is the interpretation that e.g. musl is using, so their strtof doesn't set ERANGE on this input.)

Nico —

Since this is an active LWG bug, I’d like to hold off on your patch until after the next committee meeting (which is the week after next).

So it doesn’t get lost, can you open a bug, and attach the patch there.
(You can assign the bug to me)

Thanks!

— Marshall


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20140605/d2047ffb/attachment.html>