[PATCH 3/2] t5309: mark delta-cycle failover tests as passing

t5309 checks that we detect and fail when there is a cycle of deltas (i.e., A is a delta on B, which is a delta on A). It also checks the cases where we have such a cycle, but we have an extra copy of the object either in the same pack or in another one. These cases are recoverable, because we can use the alternate copy to reconstruct the delta object.

However, it did not work in practice (and the tests were marked as expect_failure) because it led to us trying to resolve some deltas multiple times, something we assert() should not happen. E.g., we'd see the full base A, use it to resolve delta B on top of A, then use that to resolve the delta A on top of B, and then finally try resolving B on top of A again.

Since the previous commit, we handle duplicate bases more gracefully, and the tests pass. While we're here, let's also make the tests a little more robust:

  1. For the case of finding the extra base in the same
     pack, we do not need to use --fix-thin; it should (and
     does) work without it.

  2. Confirm that the resulting pack has a duplicate object.
     In the case of using --fix-thin to pull the object from
     another pack, the duplicate is added to the pack by us
     (this is a good thing, as otherwise future readers
     would encounter the cycle again).

Signed-off-by: Jeff King <peff@peff.net>
---
An obvious follow-on to the other two patches.

The implications of this make me slightly nervous, though. In the --fix-thin case, the resulting pack will have 3 objects:

  - A as a delta on B
  - B as a delta on A
  - a full copy of either A (or B) provided by --fix-thin

We create a .idx that has duplicate entries for A. If a reader is trying to reconstruct B and they find the full copy of A, they're fine. If they find the delta copy, what happens?

Ideally the reader would say "hey, I can't reconstruct A here, let me try to find another copy". But I am not sure if that happens, or if we are even capable of finding another copy of A (certainly we can find one in another pack, but I do not think we are smart enough to find a duplicate in the same pack).

By definition, we have a copy in another pack here (that is where the --fix-thin base came from), but there is nothing to guarantee that the other pack lives on.

 t/t5309-pack-delta-cycles.sh | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)