[jira] [Commented] (PHOENIX-1763) Support building with HBase-1.1.0

["Jonathan Lawlor (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PHOENIX-1763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14486063#comment-14486063 ] 

Jonathan Lawlor commented on PHOENIX-1763:
------------------------------------------

[~enis] Just an FYI, HBASE-13421 just went in on branch-1+ of HBASE to address the increased object creations that resulted from NextState introduction in HBASE-11544. The fix from HBASE-13421 introduces ScannerContext into the RegionScanner/InternalScanner interfaces to encapsulate the scanner limits and progress towards those limits. With the introduction of ScannerContext, it made the most sense to move NextState into the ScannerContext and revert the return type of RegionScanner/InternalScanner to a boolean that indicates whether or not more rows exist (I apologize for this as you have already done work to incorporate the NextState...). I made sure to make ScannerContext match the visibility of RegionScanner and I would be happy to provide any help in regards to any issues you run into when incorporating this change.

Again, I apologize for yet another change in the RegionScanner interface! However, this latest change from HBASE-13421 will ensure that all limit logic is nicely encapsulated inside ScannerContext and will also ensure that we do not have a bunch of additional object creations inside the hot code paths. Please let me know if I can help in any way or provide some additional context behind the changes.

> Support building with HBase-1.1.0 
> ----------------------------------
>
>                 Key: PHOENIX-1763
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-1763
>             Project: Phoenix
>          Issue Type: Improvement
>            Reporter: Enis Soztutar
>             Fix For: 5.0.0
>
>         Attachments: phoenix-1763_v1.patch, phoenix-1763_v2.patch
>
>
> HBase-1.1 is in the works. However, due to HBASE-11544 and possibly HBASE-12972 and more, we need some changes for supporting HBase-1.1 even after PHOENIX-1642. 
> We can decide on a plan to support (or not) HBase-1.1 on which branches by the time it comes out. Let's use subtasks to keep progress for build support for 1.1.0-SNAPSHOT. 



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