William Burns <wburns> updated the status of jira ISPN-3518 to Coding In Progress
Radim Vansa <rvansa> made a comment on jira ISPN-3518 I know it's already fixed, but is it really a bug? The described scenario does not break the serializability property of the transactions - the outcome is the same as if tx2 was executed before tx1. Could you describe the scenario where it does break some expected property?