Memory corruption in mysql_real_query

[Vercantez]

What happens?

When scanning my very small MySQL server I consistently get a SIGABRT in mysql_real_query. It looks like it's due to mysql_real_query not being protected by a lock as it should be according to the libmysql docs

In particular, you must ensure that between calls to mysql_real_query() or mysql_query() and mysql_store_result() in one thread, no other thread uses the same connection.

macOS crash report attached. Two threads are clearly simultaneously in mysql_real_query(). Python-2024-11-12-142406.ips.zip

To Reproduce

Different on macOS vs Amazon linux.

        db = duckdb.connect(":memory:")
        db.execute("INSTALL MYSQL;")
        db.execute(f"ATTACH 'database={database_connection_info['database']} user={database_connection_info['user']} password={database_connection_info['password']} host={database_connection_info['host']} port={database_connection_info['port']}' AS mysqldb (TYPE MYSQL, READ_ONLY);")
        db.execute(f"USE mysqldb.{database_connection_info['database']};")
        db.execute("SET enable_external_access = false;")
        db.execute("SHOW;")

OS:

macOS 15.1 and Amazon Linux 2023

MySQL Version:

8.0.35

DuckDB Version:

v1.1.2 f680b7d08f

DuckDB Client:

Python

Full Name:

Miguel Angel Salinas

Affiliation:

camelAI

Have you tried this on the latest main branch?

• [X] I agree

Have you tried the steps to reproduce? Do they include all relevant data and configuration? Does the issue you report still appear there?

• [X] I agree