[
https://issues.apache.org/jira/browse/HADOOP-10047?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13809495#comment-13809495
]
Gopal V commented on HADOOP-10047:
----------------------------------
I will redo the original WIP patch's test-case for this interface - it had a
decompressor loop.
As for the method visibility, If only inflateDirect() had to be exposed, I
would have gone down that path.
I read the code again just now and the shim class is rather odd and kludgey.
For the sake of readability & clarity, I think I should move the entire method
from the shim into a package private method in ZlibDecompressor and get rid of
the extra class overhead.
I will put this into PA tomorrow once the tests are done and they pass.
> Add a directbuffer Decompressor API to hadoop
> ---------------------------------------------
>
> Key: HADOOP-10047
> URL: https://issues.apache.org/jira/browse/HADOOP-10047
> Project: Hadoop Common
> Issue Type: New Feature
> Components: io
> Reporter: Gopal V
> Assignee: Gopal V
> Labels: compression
> Attachments: DirectCompressor.html, DirectDecompressor.html,
> HADOOP-10047-redo-WIP.patch, HADOOP-10047-WIP.patch,
> HADOOP-10047-with-tests.patch
>
>
> With the Zero-Copy reads in HDFS (HDFS-5260), it becomes important to perform
> all I/O operations without copying data into byte[] buffers or other buffers
> which wrap over them.
> This is a proposal for adding a DirectDecompressor interface to the
> io.compress, to indicate codecs which want to surface the direct buffer layer
> upwards.
> The implementation should work with direct heap/mmap buffers and cannot
> assume .array() availability.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
