[huacnlee]

Description

A malformed hunk source line beginning with a multibyte UTF-8 character panics before the parser can return DiffParseError. parse_hunk sliced &content[1..] before checking whether the first byte was a valid hunk prefix.

Validate the ASCII prefix first, then remove it only for space, + and -. Invalid prefixes retain the existing error message and one-based patch line. Empty context lines stripped by editors/mail transports remain supported.

Minimal input: --- a/a\n+++ b/a\n@@ -1 +1 @@\néinvalid\n.

Regression coverage

• Invalid 2-, 3- and 4-byte Unicode prefixes (é, 中, 😀) and an invalid ASCII prefix return the line-4 parse error.
• Valid context, deleted and added lines preserve their Unicode source.
• Existing empty-context and malformed-hunk tests remain unchanged.

Public API

None. No Base-layer, dependency or UI changes.

Validation

• Independent source review completed.
• Patch application and whitespace checks passed against main 369f7a8f5a6602da980f008cf726467588ad517f.
• Local Rust tests, rustfmt and clippy were not run: the saved coding environment failed to start, and the available workspace has no Rust toolchain.
• CI run 38328586529 passed all 15 jobs for head 2846b8d612f96ebd5dda65c1348bf104200d5149, including standard macOS/Linux/Windows tests, GPUI Fast, lint and Shell checks.
• The Linux test log records both new Unicode regression tests passing; the Component unit suite passed 708 tests.
• The PR remains a draft for maintainer review. Focused rerun: cargo test -p gpui-component --lib diff::parser_tests.

AI-assisted: prepared and independently source-reviewed with OpenAI Codex.