[reporter="jvs", created="Tue, 5 Sep 2006 13:08:01 -0500 (GMT-05:00)"]
In eigenchange 7557, I set codeCacheMaxBytes to min for this test. Without that, it was tripping over OutOfMemory errors when run with my refVerifyConstraints changes. I do not think those changes were the cause of the problem; I think they were just a straw that broke the camel's back. FarragoTrace.log shows steady increase in heap size for successive tests, leading up to about 50M before semijoin.sql starts. semijoin.sql itself runs a lot of statements, tying up a lot of memory in the code cache.
[reporter="jvs", created="Tue, 5 Sep 2006 13:08:01 -0500 (GMT-05:00)"] In eigenchange 7557, I set codeCacheMaxBytes to min for this test. Without that, it was tripping over OutOfMemory errors when run with my refVerifyConstraints changes. I do not think those changes were the cause of the problem; I think they were just a straw that broke the camel's back. FarragoTrace.log shows steady increase in heap size for successive tests, leading up to about 50M before semijoin.sql starts. semijoin.sql itself runs a lot of statements, tying up a lot of memory in the code cache.