[
https://issues.apache.org/jira/browse/CASSANDRA-17964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17679638#comment-17679638
]
Ekaterina Dimitrova edited comment on CASSANDRA-17964 at 1/22/23 8:32 PM:
--------------------------------------------------------------------------
{quote}I believe that nobody is looking at this as it is a regression. It would
be a regression if a respective test was passing and it started to fail. But
these tests are technically failing already, we are just making it visible.
{quote}
That's what I think too.
{quote}[~e.dimitrova] We may indeed not ignore them if that is what you prefer.
Nothing against that. The obvious consequence of doing so is that they would
start to be visible as failures in the test results pretty much constantly
until they are fixed.
{quote}
I will leave the final decisions to you and the reviewers. It is indeed a
matter of visibility. What I was wondering is why adding a precedent, to
silence failing tests? What would be the rationality? This reminds me of the
case when [~dcapwell] found a long-term issue with DTests framework which led
to more failing tests after the bug was fixed. Some of those failing tests are
still out there with associated tickets for pre-4.0, for example? I believe
there was something around a repair bug being exposed for example.
was (Author: e.dimitrova):
{quote}I believe that nobody is looking at this as it is a regression. It would
be a regression if a respective test was passing and it started to fail. But
these tests are technically failing already, we are just making it visible.
{quote}
That's what I think too.
{quote}[~e.dimitrova] We may indeed not ignore them if that is what you prefer.
Nothing against that. The obvious consequence of doing so is that they would
start to be visible as failures in the test results pretty much constantly
until they are fixed.
{quote}
I will leave the final decisions to you and the reviewers. It is indeed a
matter of visibility. What I was wondering is why adding a precedent, to
silence failing tests? What would be the rationality? This reminds me of the
case when [~dcapwell] found a long-term issue with DTests framework which led
to more failing tests. Some of those failing tests are still out there with
associated tickets for pre-4.0, for example? I believe there was something
around a repair bug being exposed for example.
> Some tests are never executed due to naming violation - fix it and add
> checkstyle where applicable
> --------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-17964
> URL: https://issues.apache.org/jira/browse/CASSANDRA-17964
> Project: Cassandra
> Issue Type: Task
> Components: Test/unit
> Reporter: Ruslan Fomkin
> Assignee: Stefan Miklosovic
> Priority: Normal
> Fix For: 3.0.x, 3.11.x, 4.0.x, 4.1.x, 4.x
>
> Time Spent: 2h 50m
> Remaining Estimate: 0h
>
> [BatchTests|https://github.com/apache/cassandra/blob/trunk/test/unit/org/apache/cassandra/cql3/BatchTests.java]
> doesn't follow naming convention to be run as unit tests and, thus, is never
> run.
> The rule in build expects names as `*Test`.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
