Timeline for Non-transactional Database and Integration Testing
Current License: CC BY-SA 3.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 16, 2011 at 19:16 | comment | added | user3792 | @ryanzec - I think I might see what you mean. If application A requires 10 records in the table to pass, but app B inserted a few extra at the same time, so A's test fails. In this case, you will need to write your tests differently. I would provide some sort of transaction number, so that I am not looking for any 10 records in a table for a test, but 10 records with an associated transaction number. | |
| Aug 16, 2011 at 19:13 | comment | added | user3792 | @ryanzec - I do not quite follow. If you have a test to make sure that 10 records are in the table and you get a result of 10, then it seems the test will pass. | |
| Aug 16, 2011 at 19:01 | comment | added | ryanzec | That happens already without transaction (at the end, the tables are just emptied since I always start my tests with empty tables) but it doesn't solve the issue of multiple tests running at the same time. If I am running integration tests for application A and application B at the same and they both insert 10 records in to database.table_a, if I have a test to make sure that 10 records are in the table but I get a result of 10, the test would fail even though that test is working. This is the case I am trying to avoid. | |
| Aug 16, 2011 at 18:51 | history | answered | user3792 | CC BY-SA 3.0 |