[
https://issues.apache.org/jira/browse/HADOOP-4044?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12637335#action_12637335
]
Sanjay Radia commented on HADOOP-4044:
--------------------------------------
Owen says:
>My preference is toward not using exceptions for nominal cases, and I agree
>with Doug that links seem like nominal cases. Clearly, however, it isn't a
>slam dunk.
>At the fear of crossing streams, I think that the classes like FSLinkBoolean
>are problematic. I think we are better off creating a type for each method's
>return type. I think that a lot of these return types should implement a
>class/interface (Owen ducks) that provides the common interface for
>dereferencing the links.
What you loose with owen's class/interface suggestion is that the signature
not longer indicates
that the method returns a boolean or an array of locatedBlocks or an file
descriptor etc.
I think Dhruba was trying to preserve that aspect of the original interfaces
where by reading the name of the method and the names and types of the
input and output parameter, once can easily guess the semantics of the method.
(the eraser I threw at Owen bounced off the screen).
But Owen is right in that his approach is a little more future proof (and
exceptions are more future proof :-)
Owen what do you propose for the internal interfaces in the name node? -- do
you suggest that we remove the symlink-related-exception and use your
structure every where in the FS and HDFS code?
If not, why the inconsistency?
> Create symbolic links in HDFS
> -----------------------------
>
> Key: HADOOP-4044
> URL: https://issues.apache.org/jira/browse/HADOOP-4044
> Project: Hadoop Core
> Issue Type: New Feature
> Components: dfs
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
> Attachments: symLink1.patch, symLink1.patch, symLink4.patch,
> symLink5.patch, symLink6.patch, symLink8.patch, symLink9.patch
>
>
> HDFS should support symbolic links. A symbolic link is a special type of file
> that contains a reference to another file or directory in the form of an
> absolute or relative path and that affects pathname resolution. Programs
> which read or write to files named by a symbolic link will behave as if
> operating directly on the target file. However, archiving utilities can
> handle symbolic links specially and manipulate them directly.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
