[
https://issues.apache.org/jira/browse/DERBY-2551?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12496081
]
Knut Anders Hatlen commented on DERBY-2551:
-------------------------------------------
> Using the ClientXid is ok since it only stores the transaction identification
> and does not contain any client driver related code.
Sounds OK. Perhaps you could add a comment about that in the test.
> Shouldn't we write the tests more sloppy that also a garbage collection would
> be tested?
Definitely not! :) We have already tried that, and it only made the JUnit tests
fail intermittently and randomly. The most common problem is that tearDown() is
not allowed to drop a table because some open object depends on the table and
hasn't been garbage collected yet. I think there are some tests for memory
leaks which implicitly test that garbage collection work.
> Global Xid value garbled in syscs_diag.transaction_table.
> ---------------------------------------------------------
>
> Key: DERBY-2551
> URL: https://issues.apache.org/jira/browse/DERBY-2551
> Project: Derby
> Issue Type: Bug
> Components: Services
> Affects Versions: 10.3.0.0
> Reporter: Julius Stroffek
> Assigned To: Julius Stroffek
> Fix For: 10.2.2.1, 10.3.0.0
>
> Attachments: d2551-round2.diff, d2551-round2.stat, d2551.diff
>
>
> The value of global xid is dumped without leading zeros on
> GlobalTransactionId and BranchQualifier byte arrays. Thus, it is impossible
> to reconstruct the xid value from the transaction_table.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
