Hello, I'm looking for some insight for some odd behavior with some unit tests, which I think stems from some misunderstanding of how these classes actually work.
1. For TestCase, the Django docs say that it does matter what order in which the tests are run. But elsewhere in the same doc, it says that TestCase methods are individually wrapped in a single transaction block that is rolled back after the test completes. So... does each test method start 'fresh' or not? I currently have a test method that is supposed to query two records, but so far it finds either 4 or 6 with no apparent pattern; so I don't know if it's my code that is causing problems or the order in which my tests are running. 2. The strangest behavior comes from a test method that passes by itself, fails when run with all the others, but the code under test actually does seem to work. As one might expect, the section under test takes some records created especially for that purpose and modifies a few fields; the test calls the function, queries the test database, and assert-tests the results. The test fails, saying that only some of the fields were successfully changed. I thought that I perhaps forgot to call save() in one of my conditional branches, but I looked and found that save() was in indeed where it belonged. I stepped through the code and everything appeared to work as expected. The values looked correct. Finally, I stuck a query into the code under test itself immediately before the return statement, and printed the results; and the query proves that the database field values themselves did indeed change correctly. But after the return statement (when it goes back to the test method itself), an identical query gets different results. So I'm rather puzzled. I do use select_for_update in my section under test, which I know sometimes causes problems with TestCase, but all I'm doing is testing for correct results. 3. I thought subclassing TransactionTestCase instead might solve both problems, but set-up data that works flawlessly with TestCase breaks every single test with TransactionTestCase--get() complains about non-existent records, the db complains about broken constraints, etc. I use both create() and bulk_create() in setting up my data, but neither appears to work at all. I tried using them in setUpTestData(), setUpClass(), and setUp(), and not a single test would run with any of them. The docs say that the two classes 'are identical except for the manner in which the database is reset to a known state', so I'm not sure where I went wrong. Using Django 2.0. And unfortunately we use MS SQL Server, which probably isn't helping. Thank you! Heather -- You received this message because you are subscribed to the Google Groups "Django users" group. To unsubscribe from this group and stop receiving emails from it, send an email to django-users+unsubscr...@googlegroups.com. To post to this group, send email to django-users@googlegroups.com. Visit this group at https://groups.google.com/group/django-users. To view this discussion on the web visit https://groups.google.com/d/msgid/django-users/671d7dc0-ce5a-4e90-9a44-70357ecd8f6b%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
