perf(parse): cut allocations and validate UTF-8 in the scan

Per-statement work no longer allocates where it does not need to: parseKeyPath returns the first key component directly, absolute paths and key buffers sit on the parser, the definition maps are created on first write, arrays presize, and a string without escapes is copied out in one piece instead of built byte by byte. Repeated keys share one string across array-of-tables elements through a parser-local intern table whose lookup works on the bytes, so a repeated key costs no allocation.

UTF-8 validity is no longer a whole-input pass before the parser runs: the scan validates where it meets a multi-byte sequence, and a SyntaxError now names the line the invalid byte sits on. TestParseRejectsInvalidUTF8 covers nine positions; docs/API.md describes the new reporting.

Long document: 67 664 to 31 765 allocations per parse, about 51 to about 106 MB/s; representative document: 157 to 104 allocations.

docs/API.md

@@ -19,7 +19,8 @@ Decodes a TOML document into a [Document](#documents): the values, the order the
 keys were written in, whether a table was written inline, and the comments.
 The values follow the mapping in the [Decoding](#decoding) section below.
 Returns `*SyntaxError` on a malformed document. Input that is not valid UTF-8
-is rejected before the parser runs. Equivalent to
+is rejected with a `SyntaxError` naming the line where the invalid byte
+appears, because validity is checked during the scan. Equivalent to
 `ParseContext(context.Background(), data)`.
 
 ```go