On Sat, Oct 11, 2025 at 1:21 AM Jonathan Wakely <[email protected]> wrote:
> This should check for c <= 0x7f not x < 0x7f, because 0x7f is an ASCII
> character (DEL).
>
> libstdc++-v3/ChangeLog:
>
> * include/bits/unicode.h (__is_single_code_unit): Fix check for
> 7-bit ASCII characters.
> ---
>
> Tested x86_64-linux.
>
LGTM.
>
> There's no new test for this as I couldn't get std::format to produce
> wrong results with e.g. DEL as a fill character. It gives the right
> result, just using a slower path that converts the fill character to
> UTF-32 and then constructs a basic_string from _Utf32_view. With the
> fix, we avoid that slow path.
>
> libstdc++-v3/include/bits/unicode.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libstdc++-v3/include/bits/unicode.h
> b/libstdc++-v3/include/bits/unicode.h
> index 88e97d41a9eb..00efbe89ca8e 100644
> --- a/libstdc++-v3/include/bits/unicode.h
> +++ b/libstdc++-v3/include/bits/unicode.h
> @@ -61,7 +61,7 @@ namespace __unicode
> __is_single_code_unit(char32_t __c)
> {
> if constexpr (__gnu_cxx::__int_traits<_CharT>::__max <= 0xFF)
> - return __c < 0x7F; // ASCII character
> + return __c <= 0x7F; // ASCII character
> else
> return __c < __gnu_cxx::__int_traits<_CharT>::__max
> && __is_scalar_value(__c);
> --
> 2.51.0
>
>
