Thanks everyone for their comments on these issues. Thanks too to Deepa
for posting these issues from 10.1 client/ 10.2 server testing.
After a rough triage I think the following:
DERBY-1448 - Not a bug. Improved error message in sysinfo. Bryan
added release note to DERBY-668 to clarify.
DERBY-1449 - The checkDataSource test is trying to read a system
property "framework". It is not clear why this is, It may be related
xaHelper, but there is no smoking gun of a product issue here, so I will
not change the priority, probably could be changed to Minor.
DERBY-1450 - Not a bug. Expected behavior change due to DERBY-822 but I
think we need a release note for DERBY-822. I added a comment to that
issue.
DERBY-1452 - Not a bug. Same issue as 1450, plus there is some output
differences for Decimal output that looks ok.
DERBY-1453 - This one I raised to critical for further evaluation. It
is a lock time-out where there was none before. It may be a DERBY-822
side effect but I can't really tell. This also occurs just running the
10.1 tests with 10.2 client and 10.2 server, so is a general
compatibility issue, not limited to mixed version client and server
environments.
So short story, just one new critical issue to be evaluated in this set.
This testing I think is been very useful in identifying potential
compatibility issues, but I suppose these tests will be put on the
don't run list now that they fail. For 10.3 it would be great to think
about sustainable compatibility testing. For 10.2 this is what we
have. It is very useful to see how the 10.1 tests (simulating an
existing application) react when faced with 10.2. To further understand
compat issues with 10.2 I did a derbyall run of 10.1 tests against
10.2 ( server and client).
http://people.apache.org/~kmarsden/10_1_tests_10_2_derby/derbyall.sum
Kathey
