supabase / etl / 28147442615

75%

Build Type

push

github

Committed by web-flow

Commit Message

fix(core): avoid bytea hex UTF-8 boundary panics (#849)

- Update crates/etl/src/conversions/hex.rs to avoid str slicing on
potentially non-ASCII input in parse_bytea_hex_string.

- Switch parsing to byte-based prefix stripping and chunk validation
so malformed UTF-8 boundary input cannot panic.

- Add regression tests for boundary cases: aé, \\é, \\x🤔🤔.

- Add a fuzzing harness under fuzz/ with a dedicated target for
parse_bytea_hex_string and seed corpus entries for the above cases.

- Add a fuzzing crate feature and a #[doc(hidden)] fuzz-only shim to
expose the parser for fuzz input without changing public crate API
semantics.

Coverage Stats

7 of 10 new or added lines in 2 files covered. (70.0%)

16 existing lines in 2 files now uncovered.

39243 of 52478 relevant lines covered (74.78%)

35510.82 hits per line