[jira] [Commented] (HIVE-15205) Create ReplDumpTask/ReplDumpWork for dumping out metadata

2016-11-15 Thread Sergey Shelukhin (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-15205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15668294#comment-15668294
 ] 

Sergey Shelukhin commented on HIVE-15205:
-

Theres' no patch so I cannot tell :)

> Create ReplDumpTask/ReplDumpWork for dumping out metadata
> -
>
> Key: HIVE-15205
> URL: https://issues.apache.org/jira/browse/HIVE-15205
> Project: Hive
>  Issue Type: Sub-task
>  Components: repl
>Reporter: Vaibhav Gumashta
>Assignee: Vaibhav Gumashta
>
> The current bootstrap code generates dump metadata during semantic analysis 
> which breaks security and task/work abstraction. It also uses existing 
> classes (from Export/Import world) for code reuse purpose, but as a result 
> ends up dealing with a lot if-then-elses. It makes sense to have a cleaner 
> abstraction which uses ReplDumpTask and ReplDumpWork (to configure the Task). 
> Also perhaps worth evaluating ReplLoadTask/ReplLoadWork for load side.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-15205) Create ReplDumpTask/ReplDumpWork for dumping out metadata

2016-11-15 Thread Vaibhav Gumashta (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-15205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15668286#comment-15668286
 ] 

Vaibhav Gumashta commented on HIVE-15205:
-

[~sershe] This jira doesn't touch ImportSemanticAnalyzer etc. It is resolving 
some of the problems you are pointing to in original jira. All the classes here 
are new (unless you have created ones with similar names). 

> Create ReplDumpTask/ReplDumpWork for dumping out metadata
> -
>
> Key: HIVE-15205
> URL: https://issues.apache.org/jira/browse/HIVE-15205
> Project: Hive
>  Issue Type: Sub-task
>  Components: repl
>Reporter: Vaibhav Gumashta
>Assignee: Vaibhav Gumashta
>
> The current bootstrap code generates dump metadata during semantic analysis 
> which breaks security and task/work abstraction. It also uses existing 
> classes (from Export/Import world) for code reuse purpose, but as a result 
> ends up dealing with a lot if-then-elses. It makes sense to have a cleaner 
> abstraction which uses ReplDumpTask and ReplDumpWork (to configure the Task). 
> Also perhaps worth evaluating ReplLoadTask/ReplLoadWork for load side.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (HIVE-15205) Create ReplDumpTask/ReplDumpWork for dumping out metadata

2016-11-15 Thread Sergey Shelukhin (JIRA)

[ 
https://issues.apache.org/jira/browse/HIVE-15205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15668018#comment-15668018
 ] 

Sergey Shelukhin commented on HIVE-15205:
-

Temporary -1
See the comment in parent JIRA "Is it possible to do work in the branch? This 
causes immense conflicts with hive-14535 branch, and I see tons of comments 
that purport with FIXMEs and stuff to move code around and refactor this and 
that.
I think this should be done on the branch and merged once "

> Create ReplDumpTask/ReplDumpWork for dumping out metadata
> -
>
> Key: HIVE-15205
> URL: https://issues.apache.org/jira/browse/HIVE-15205
> Project: Hive
>  Issue Type: Sub-task
>  Components: repl
>Reporter: Vaibhav Gumashta
>Assignee: Vaibhav Gumashta
>
> The current bootstrap code generates dump metadata during semantic analysis 
> which breaks security and task/work abstraction. It also uses existing 
> classes (from Export/Import world) for code reuse purpose, but as a result 
> ends up dealing with a lot if-then-elses. It makes sense to have a cleaner 
> abstraction which uses ReplDumpTask and ReplDumpWork (to configure the Task). 
> Also perhaps worth evaluating ReplLoadTask/ReplLoadWork for load side.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)