Release 13.0 Improve handling of invalid UTF-8 byte sequences

[m3hm3t]

SQLancer generates invalid UTF-8 byte sequences, resulting in errors such as:

ERROR: invalid byte sequence for encoding "UTF8": 0xee 0x22 0x20

Steps to Reproduce:

Insert a string with an invalid UTF-8 sequence:

INSERT INTO test_table (name) VALUES (E'\xee\x22\x20');

Expected Behavior:

Queries may validate input data and reject invalid byte sequences gracefully.

Suggested Solution:
Add input validation logic to reject malformed UTF-8 sequences.

Severity:
Not a blocker: This edge case affects robustness but doesn't disrupt overall functionality. While improving data validation or rejecting malformed data would enhance usability, it isn’t critical for the release.

[m3hm3t]

This problem also occurs in the release-12.1 tests.

[m3hm3t]

release-13 test:
Update sqlancer-test-pipelines_2.yml · m3hm3t/citus-13.0@7d7bef7

to compare release-12.1 test:
Update sqlancer-test-pipelines_2.yml · m3hm3t/citus-13.0@7d7bef7

[onurctirtir]

Do we get a different error when the same insert is run on a Postgres table?

[m3hm3t]

Do we get a different error when the same insert is run on a Postgres table?

No, we do not get a different error. This error originates from PostgreSQL's handling of invalid UTF-8 sequences. When the same INSERT query with an invalid UTF-8 sequence is run on a plain PostgreSQL table, PostgreSQL raises the exact same error.

[onurctirtir]

I think we can close the issue then.