Comparing the behaviour of 1.2 and 1.3rc1 I made a very interesting discovery.
- When I execute test with 1.3rc1 after setting up the tables it fails.
- When I execute test with 1.2 after setting up the tables it shows no failures.
- When I execute test with 1.3rc1 after a successful run with 1.2 it also does not show failures with 1.3rc1.
This leaded me to take a look at the tables and compare the records after a test execution with 1.2 with the setup script. Unfortunately it apears that test at 1.2 modified the record with ID=1 of table test1355_golf_course where it sets CITY_ID to NULL. Therefore the test at 1.2 has never executed successful but instead modified records which it should not have done. In contrast behaviour of 1.3rc1 does not modify records but also does not show the expected behaviour.
According to the stacktrace it seams that the problem at hand is related to lazy loading of city entities.
Comparing the behaviour of 1.2 and 1.3rc1 I made a very interesting discovery.
This leaded me to take a look at the tables and compare the records after a test execution with 1.2 with the setup script. Unfortunately it apears that test at 1.2 modified the record with ID=1 of table test1355_golf_course where it sets CITY_ID to NULL. Therefore the test at 1.2 has never executed successful but instead modified records which it should not have done. In contrast behaviour of 1.3rc1 does not modify records but also does not show the expected behaviour.
According to the stacktrace it seams that the problem at hand is related to lazy loading of city entities.
- When I execute test with 1.3rc1 after setting up the tables it fails.
- When I execute test with 1.2 after setting up the tables it shows no failures.
- When I execute test with 1.3rc1 after a successful run with 1.2 it also does not show failures with 1.3rc1.
This leaded me to take a look at the tables and compare the records after a test execution with 1.2 with the setup script. Unfortunately it apears that test at 1.2 modified the record with ID=1 of table test1355_golf_course where it sets CITY_ID to NULL. Therefore the test at 1.2 has never executed successful but instead modified records which it should not have done. In contrast behaviour of 1.3rc1 does not modify records but also does not show the expected behaviour. According to the stacktrace it seams that the problem at hand is related to lazy loading of city entities.