MirrorHub/construct
0
0
Fork 0
mirror of https://github.com/matrix-construct/construct synced 2024-12-27 07:54:05 +01:00
Code Issues Releases Wiki Activity

ircd::utf8: Add multiple lane overloads for length(); internal template; minor comments.

Browse source
This commit is contained in:
Jason Volk 2020-07-02 15:27:43 -07:00
parent 35bee76625
commit c9c61124e7
2 changed files with 38 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
include/ircd/utf.h
View file

@ -20,7 +20,11 @@ namespace ircd::utf
/// Unicode Transformation Format (8-bit)
namespace ircd::utf8
{
// Get the utf8-encoded length from decoded codepoints.
u32x16 length(const u32x16 codepoints) noexcept;
u32x8 length(const u32x8 codepoints) noexcept;
u32x4 length(const u32x4 codepoints) noexcept;
u32x16 encode(const u32x16 codepoints) noexcept;
u32x16 decode(const u8x16 string) noexcept;
}
@ -38,7 +42,6 @@ namespace ircd::utf16
u8x16 find_surrogate(const u8x16 input) noexcept;
// scan for utf-16 surrogates including incomplete sequences truncated
// by the end of the input; also matches a single trailing slash.
u8x16 find_surrogate_partial(const u8x16 input) noexcept;
}

39
ircd/utf.cc
View file

@ -29,6 +29,8 @@ ircd::utf16::truncation_table
~shl<0x08>(~full_mask),
};
/// scan for utf-16 surrogates including incomplete sequences truncated
/// by the end of the input; also matches a single trailing slash.
ircd::u8x16
ircd::utf16::find_surrogate_partial(const u8x16 input)
noexcept
@ -264,21 +266,48 @@ noexcept
;
}
/// Determine the utf-8 encoding length of multiple codepoints in parallel.
/// The input vector char32_t codepoints and the output yields an integer
/// of 0-4 for each lane.
namespace ircd::utf8
{
template<class u32xN> static u32xN _length(const u32xN codepoint) noexcept;
}
ircd::u32x4
ircd::utf8::length(const u32x4 codepoint)
noexcept
{
return _length(codepoint);
}
ircd::u32x8
ircd::utf8::length(const u32x8 codepoint)
noexcept
{
return _length(codepoint);
}
ircd::u32x16
ircd::utf8::length(const u32x16 codepoint)
noexcept
{
const u32x16
return _length(codepoint);
}
/// Determine the utf-8 encoding length of multiple codepoints in parallel.
/// The input vector char32_t codepoints and the output yields an integer
/// of 0-4 for each lane.
template<class u32xN>
u32xN
ircd::utf8::_length(const u32xN codepoint)
noexcept
{
const u32xN
length_1 { codepoint <= 0x7f },
length_2 { codepoint <= 0x7ff && codepoint > 0x7f },
length_3_lo { codepoint <= 0xd7ff && codepoint > 0x7ff },
length_3_hi { codepoint <= 0xffff && codepoint > 0xdfff },
length_4 { codepoint <= 0x10ffff && codepoint > 0xffff };
[[gnu::unused]] const u32x16 // Preserved here for future reference
[[gnu::unused]] const u32xN // Preserved here for future reference
length_3_err { codepoint <= 0xdfff && codepoint > 0xd7ff },
length_err { (codepoint > 0x10ffff) | length_3_err };