[
https://issues.apache.org/jira/browse/DERBY-4260?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13022154#comment-13022154
]
Bryan Pendleton commented on DERBY-4260:
----------------------------------------
Typically, the Derby community prefers to work on multiple smaller patches,
rather than a single larger patch.
Each patch should address a single problem. We agree with you that this makes
the SVN history easier
to comprehend.
So, in the cases that you identify, we'd prefer to address the test case
refactoring separately
from the non-English locale issues, as two patches rather than one.
Similarly, we try to avoid doing things like removing unneeded "import"
statements or changing whitespace
or formatting/indentation of the code as part of a patch. The Derby code
unfortunately has a significant
number of whitespace issues, as it is a old and mature codebase that has been
worked on by many
programmers over the years, so we just accept those problems and try to avoid
those changes in our patches.
Meanwhile, Tiago, if you get a chance, can you confirm that this test now works
properly for you in
a non-English locale? Then I think you could close this job?
> Make derbynet/NetworkServerControlClientCommandTest run regardless of the
> locale
> --------------------------------------------------------------------------------
>
> Key: DERBY-4260
> URL: https://issues.apache.org/jira/browse/DERBY-4260
> Project: Derby
> Issue Type: Improvement
> Components: Test
> Affects Versions: 10.6.1.0
> Reporter: Tiago R. Espinha
> Assignee: Houx Zhang
> Priority: Minor
> Labels: gsoc, gsoc2011
> Fix For: 10.9.0.0
>
> Attachments: 4260-1.patch
>
>
> The NetworkServerControlClientCommandTest has in it a check that will only
> run the test if the locale is set to 'en'.
> It is open to suggestion the how to achieve this and whether it should be
> done. Since the test is a pretty generic one, I think it should be ran
> regardless of the locale.
> Knut suggested we'd force the locale to 'en' on the setUp() and back to its
> previous value on the tearDown(). On a first analysis this seems like the
> ideal method to go about it. Just removing the check for the locale
> altogether might not be wise because of the calls to assertFailedPing() that
> involve some hardcoded strings.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
