Fix rune iteration bugs

[jakebailey]

I wrote a lint rule to complain on all "ranges over string", then looked at each case. These three were bugs.

[Copilot]

Pull request overview

This PR fixes a few incorrect string-iteration behaviors in the Go port that stem from accidentally iterating by runes (UTF-8 code points) when the code expects byte indexing, and updates the affected fourslash baseline output accordingly.

Changes:

• Fix an off-by-one error in TruncateByRunes so truncation occurs after maxLength runes, not before.
• Make stack-trace whitespace scanning use byte indexing rather than rune iteration.
• Fix glob * backtracking to iterate byte positions (not rune start offsets), aligning with the matcher’s byte-oriented indexing.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
testdata/baselines/reference/fourslash/documentSymbols/navigationBarItemsMultilineStringIdentifiers3.baseline	Updates expected output to reflect corrected truncation behavior.
internal/stringutil/util.go	Corrects TruncateByRunes truncation boundary (>= → >).
internal/lsp/stack_sanitizer.go	Uses byte-index loop when skipping leading whitespace on stack trace lines.
internal/glob/glob.go	Changes * backtracking loop to iterate by byte offset rather than rune boundaries.