I was able to reproduce this issue in IDLE using the vanilla 0.3.2 connector (I have a subclass/wrapper with custom behavior so I wanted to test without it).
The fact that the exact text affects it is starting to make me thing that it might be caching-related. I added SQL_NO_CACHE to my selects to see if that fixes it. If this is the case, it must be a bug with that version of mysql and failing to properly deal with large, improperly cached result sets. The PROCESSLIST on the server will show the query as complete, and the connection as sleeping while the client is hung. I'm thinking that maybe the mysql thread for the client has some kind of fart related to the cache and fails to send the goodbye handshake? I"m not sure why it would happen in the middle of result retrieval (adding traces show that some packets are received during fetchall).
I was able to reproduce this issue in IDLE using the vanilla 0.3.2 connector (I have a subclass/wrapper with custom behavior so I wanted to test without it).
The fact that the exact text affects it is starting to make me thing that it might be caching-related. I added SQL_NO_CACHE to my selects to see if that fixes it. If this is the case, it must be a bug with that version of mysql and failing to properly deal with large, improperly cached result sets. The PROCESSLIST on the server will show the query as complete, and the connection as sleeping while the client is hung. I'm thinking that maybe the mysql thread for the client has some kind of fart related to the cache and fails to send the goodbye handshake? I"m not sure why it would happen in the middle of result retrieval (adding traces show that some packets are received during fetchall).