The XFER user tests are fairly exhaustive, because that was my focus. They are all :min_version_3_2 because an unrelated behaviour changed in 3.2 (implied RETURN (SPECIAL-USE) in LIST), and so as written the tests don't pass on 3.0. They can be tinkered with to prove the fixes work for 3.0 as well, but I didn't feel like duplicating them just for that.
The XFER mailbox test is singular for now, just to make sure nothing unexpected broke. It is :max_version_3_4 because it turns out 3.5 can create an intermediate on this code path, which is bad! That'll need fixing separately, so it'll get a test case separately too.
Also adds a convenience detail to IMAPMessageStore (if we know the real uid and uidvalidity when appending a message, update the local copy to match), and fixes a test in Pop3 that had a bogus assumption that the improved IMAPMessageStore exposed.
This does not yet add any tests for XFER where the name is a mailbox pattern, nor a partition name. I don't expect those cases to be affected by the bugs I've been addressing so far, so I'll write those tests later, when I'm working on those things.
This adds a bunch of tests for XFER of a user, and one test for XFER of a single mailbox. These tests are for the fixes in https://github.com/cyrusimap/cyrus-imapd/pull/3693
The XFER user tests are fairly exhaustive, because that was my focus. They are all
:min_version_3_2
because an unrelated behaviour changed in 3.2 (impliedRETURN (SPECIAL-USE)
in LIST), and so as written the tests don't pass on 3.0. They can be tinkered with to prove the fixes work for 3.0 as well, but I didn't feel like duplicating them just for that.The XFER mailbox test is singular for now, just to make sure nothing unexpected broke. It is
:max_version_3_4
because it turns out 3.5 can create an intermediate on this code path, which is bad! That'll need fixing separately, so it'll get a test case separately too.Also adds a convenience detail to IMAPMessageStore (if we know the real uid and uidvalidity when appending a message, update the local copy to match), and fixes a test in Pop3 that had a bogus assumption that the improved IMAPMessageStore exposed.
This does not yet add any tests for XFER where the name is a mailbox pattern, nor a partition name. I don't expect those cases to be affected by the bugs I've been addressing so far, so I'll write those tests later, when I'm working on those things.