Fix UTF-8 boundary panics in two locations not covered by PR #2154: 1. telegram.rs:3112 - OTP/approval message preview truncation - Changed from byte-based `&raw_args[..260]` to char-based truncation - Uses existing `truncate_with_ellipsis` utility 2. detection.rs:222 - Tool output hash prefix - Changed from `&output[..4096]` to UTF-8-safe boundary - Uses existing `floor_utf8_char_boundary` utility Added test: hash_output_utf8_boundary_safe() verifies no panic on CJK text. Fixes #2276 |
||
|---|---|---|
| .. | ||
| context.rs | ||
| detection.rs | ||
| execution.rs | ||
| history.rs | ||
| parsing.rs | ||