Commit ab6065407d74fc8d63d398c60f9fe653374d9d6d | zig

author	daurnimator <quae@daurnimator.com> 2019-12-28 14:35:53+11:00
committer	daurnimator <quae@daurnimator.com> 2019-12-28 16:49:59+11:00
log	ab6065407d74fc8d63d398c60f9fe653374d9d6d
tree	f8d53192f40c96e6738d83264ebc683aee91ce87
parent	5843a6e3bc1a6353f76ebca57d6099337b90139a
signature	Commit is signed but in an unrecognized format.

std: simplify utf8ToUtf16Le

Also faster, on my machine unicode/throughput_test.zig now gives e.g.
> original utf8ToUtf16Le: elapsed: 1048 ns (0 ms)
> new utf8ToUtf16Le: elapsed: 971 ns (0 ms)

1 files changed, 14 insertions(+), 26 deletions(-)

lib
- std
  - unicode.zig

lib/std/unicode.zig+14-26

 @@ -576,33 +576,21 @@ pub fn utf8ToUtf16Le(utf16le: []u16, utf8: []const u8) !usize {
     var dest_i: usize = 0;
     var src_i: usize = 0;
     while (src_i < utf8.len) {
-        const byte = utf8[src_i];
+        const n = utf8ByteSequenceLength(utf8[src_i]) catch return error.InvalidUtf8;
-        const n = @clz(u8, ~byte);
+        const next_src_i = src_i + n;
-        switch (n) {
+        const codepoint = utf8Decode(utf8[src_i..next_src_i]) catch return error.InvalidUtf8;
-            0 => {
+        if (codepoint < 0x10000) {
-                utf16le[dest_i] = byte;
+            const short = @intCast(u16, codepoint);
-                dest_i += 1;
+            utf16le[dest_i] = mem.nativeToLittle(u16, short);
-                src_i += 1;
+            dest_i += 1;
-                continue;
+        } else {
-            },
+            const high = @intCast(u16, (codepoint - 0x10000) >> 10) + 0xD800;
-            2, 3, 4 => {
+            const low = @intCast(u16, codepoint & 0x3FF) + 0xDC00;
-                const next_src_i = src_i + n;
+            utf16le[dest_i] = mem.nativeToLittle(u16, high);
-                const codepoint = utf8Decode(utf8[src_i..next_src_i]) catch return error.InvalidUtf8;
+            utf16le[dest_i + 1] = mem.nativeToLittle(u16, low);
-                if (codepoint < 0x10000) {
+            dest_i += 2;
-                    const short = @intCast(u16, codepoint);
-                    utf16le[dest_i] = mem.nativeToLittle(u16, short);
-                    dest_i += 1;
-                } else {
-                    const high = @intCast(u16, (codepoint - 0x10000) >> 10) + 0xD800;
-                    const low = @intCast(u16, codepoint & 0x3FF) + 0xDC00;
-                    utf16le[dest_i] = mem.nativeToLittle(u16, high);
-                    utf16le[dest_i + 1] = mem.nativeToLittle(u16, low);
-                    dest_i += 2;
-                }
-                src_i = next_src_i;
-            },
-            else => return error.InvalidUtf8,
         }
+        src_i = next_src_i;
     }
     return dest_i;
 }