[ 
https://issues.apache.org/jira/browse/HADOOP-4044?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12637974#action_12637974
 ] 

Doug Cutting commented on HADOOP-4044:
--------------------------------------

Stepping back a bit, I don't believe this dispute is fundamentally about 
programming style.  That, like Owen's arguments about RPCs, are supporting 
arguments.  The primary issue is that we're changing the FileSystem API in a 
fundamental way.  Instead of direct functions that map parameters to values, we 
wish to implement conditional functions, that do one of two things, and 
indicate in their response which path was taken.  This is a very different 
style, one that is forced on us for performance reasons.  The alternative is to 
double or more the number of RPCs per file operation, which is unacceptable.

If we're using a different style, then it should look different.  It should not 
look like the direct mapping of the existing API.  We should not attempt to 
hide this, as that will lead to confusion.  We should make it apparent to 
readers of the code that these methods return not a single type of value, but 
one of two kinds of values.  This is not a matter of exception style, it's a 
matter of expressiveness and readability.  Code should make clear what it does. 
 Style guidelines derive from this.  The reason exceptions are discouraged for 
normal control flow is that they hide control flow.  We are factoring our 
FileSystem API in a different way, and that factoring should be apparent to all 
as an obvious, natural attribute of the API.



> 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.

Reply via email to