[CAUSEWAY-414] When a stack overflow kills thread, then HSQLDB in-memory session locks up on subsequent request.

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: viewer-wicket-1.2.0, objectstore-jdo-1.1.0
• Component/s: Persistence JDO, Viewer Wicket
• Assignee: Daniel Keir Haywood
• Reporter: Daniel Keir Haywood
• Created: 28/May/13 14:39
• Updated: 31/May/13 13:10
• Resolved: 28/May/13 15:10

Description

The stack overflowing on a previous request means that the session is not closed properly. This missing housekeeping breaks any subsequent request that uses the same connection.

Not sure if this is a problem with HSQLDB (2.2.9), or whether Isis should deal with the issue. As things stand, a localized problem (stack overflow) becomes a global issue - need to recycle web server.