On 18/02/15 19:07 +, Jonathan Wakely wrote:
While working on PR64797 I discovered that the codecvt
specialization was, erm, completely broken when creating UTF-16
surrogate pairs.
This fixes it and adds a test, based on the char32_t one I added to
the testsuite yesterday. Tested x86_64-linux
On Wed, Feb 18, 2015 at 07:07:05PM +, Jonathan Wakely wrote:
> While working on PR64797 I discovered that the codecvt
> specialization was, erm, completely broken when creating UTF-16
> surrogate pairs.
>
> This fixes it and adds a test, based on the char32_t one I added to
> the testsuite yes
While working on PR64797 I discovered that the codecvt
specialization was, erm, completely broken when creating UTF-16
surrogate pairs.
This fixes it and adds a test, based on the char32_t one I added to
the testsuite yesterday. Tested x86_64-linux (little-endian) and
powerpc64-linux (big-endian)