ump45nose opened a new pull request, #11638: URL: https://github.com/apache/arrow-rs/pull/11638
# Which issue does this PR close? - Closes #11576. # Rationale for this change `UnionFields` derived `Deserialize`, so deserializing could build a value with duplicate or negative type ids, skipping the checks that #11533 added to the other constructors. # What changes are included in this PR? `Serialize` is still derived (transparent), but `Deserialize` is now implemented by hand: it reads the same sequence of `(type_id, field)` pairs and builds the value with `UnionFields::try_new`, returning a serde error if validation fails. The wire format doesn't change, so anything serialized before still deserializes as long as it was valid. I kept serde rather than removing it, since `DataType` (and therefore `Field` and `Schema`) derive serde and need `UnionFields` to support it. # Are these changes tested? Yes, two new tests in `fields.rs` (behind the `serde` feature, using `postcard` like the existing `Field` serde tests): a round trip of a valid `UnionFields`, and deserializing input with a duplicate type id and with a negative type id, both of which now fail. The second test fails without this change. # Are there any significant user-facing changes? Deserializing an invalid `UnionFields` (or a `DataType::Union` containing one) now returns an error instead of succeeding. AI disclosure: I used an AI coding assistant to write the change, the tests and this description, and I reviewed all of it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
