[
https://issues.apache.org/jira/browse/HADOOP-15546?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16550079#comment-16550079
]
Steve Loughran commented on HADOOP-15546:
-----------------------------------------
HADOOP-15546 patch 008
* review ITestAzureBlobFileSystemListStatus
* move unicode test in ITestAzureBlobFileSystemRename into parameterized test
suite ITestAzureBlobFileSystemRenameUnicode; isolates failures better. (left
the old one there for now)
* ITestReadWriteAndSeek -> ITestAbfsReadWriteAndSeek
* helper methods touch(path), write(path, data) in AbstractAbfsIntegrationTest;
use where appropriate
* all tests of > a megabyte or two tagged as scale, so splitting up "basic fast
set" from "the full diligence set". People must still due that scale test
before a final commit, but it means that basic dev cycles are shorter
Ignoring the parallel test issue (which I am), I'm pretty much done here. If
things are failing, I don't think it's because this patch breaks stuff, it just
surfaces some problems.
Test results
* ITestAzureBlobFileSystemRenameUnicode fails
* All the wasb tests are failing with an NPE inside wasb using the login
details which I am using to run the abfs tests. (as shown above)
* new test which compare LocatedFileStatus to VersionedFileStatus fail if it
VersionedFileStatus is on the left of the .equals(), as it requires a matching
type and version. To date, all FileStatus subclasses just compare on path. *I
don't know if this is going to be a problem*.
I think this is ready to go in. Yes, stuff fails, but this gets the tests
runnable by others, so we can make progress on the branch. Reviews please!
> ABFS: tune imports & javadocs; stabilise tests
> ----------------------------------------------
>
> Key: HADOOP-15546
> URL: https://issues.apache.org/jira/browse/HADOOP-15546
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Affects Versions: HADOOP-15407
> Reporter: Steve Loughran
> Assignee: Steve Loughran
> Priority: Major
> Attachments: HADOOP-15546-001.patch,
> HADOOP-15546-HADOOP-15407-001.patch, HADOOP-15546-HADOOP-15407-002.patch,
> HADOOP-15546-HADOOP-15407-003.patch, HADOOP-15546-HADOOP-15407-004.patch,
> HADOOP-15546-HADOOP-15407-005.patch, HADOOP-15546-HADOOP-15407-006.patch,
> HADOOP-15546-HADOOP-15407-006.patch, HADOOP-15546-HADOOP-15407-007.patch,
> HADOOP-15546-HADOOP-15407-008.patch
>
>
> Followup on HADOOP-15540 with some initial review tuning
> h2. Tuning
> * ordering of imports
> * rely on azure-auth-keys.xml to store credentials (change imports,
> docs,.gitignore)
> * log4j -> info
> * add a "." to the first sentence of all the javadocs I noticed.
> * remove @Public annotations except for some constants (which includes some
> commitment to maintain them).
> * move the AbstractFS declarations out of the src/test/resources XML file
> into core-default.xml for all to use
> * other IDE-suggested tweaks
> h2. Testing
> Review the tests, move to ContractTestUtil assertions, make more consistent
> to contract test setup, and general work to make the tests work well over
> slower links, document, etc.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
