[
https://issues.apache.org/jira/browse/MAPREDUCE-1927?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Greg Roelofs updated MAPREDUCE-1927:
------------------------------------
Attachment: MR-1927.v12.trunk-hadoop-mapreduce.patch
Updated (final?) patch. This corrects the internals of the gzip extra field in
one of the test files (not relevant to the test, but matter of
spec-compliance), changes a native-libs-only assertion to a warning to work
around Apache Hudson's lack of support, and removes all but three commented-out
code blocks.
The three remaining blocks are all a result of other bugs: the aforementioned
Hudson native-library one (lack of coverage of a key infrastructure component
is a problem, and it's getting worse with every new native codec); HADOOP-6799
(broken reset() on gzip compression); and HADOOP-6852 (broken bzip2
decompression, apparently). I realize commented-out code is frowned upon as a
style issue, but I would rather not remove useful tests that are currently
blocked by things outside my (immediate) control. I'll defer to committer's
opinion, however.
This passed "ant test-patch" on a 10-node cluster, so I'm going to "submit
patch" to see if it gets past Apache's Hudson, too.
> unit test for HADOOP-6835 (concatenated gzip support)
> -----------------------------------------------------
>
> Key: MAPREDUCE-1927
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1927
> Project: Hadoop Map/Reduce
> Issue Type: Test
> Components: test
> Affects Versions: 0.22.0
> Reporter: Greg Roelofs
> Assignee: Greg Roelofs
> Priority: Minor
> Fix For: 0.22.0
>
> Attachments: construct-all-gz-header-types-20100812.zip,
> MR-1927.v12.trunk-hadoop-mapreduce.patch,
> MR-1927.v9.trunk-hadoop-mapreduce.patch
>
>
> More extensive test of concatenated gzip (and bzip2) decoding support for
> HADOOP-6835 (and HADOOP-4012 and HADOOP-6852).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.