[
https://issues.apache.org/jira/browse/PHOENIX-4671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16413401#comment-16413401
]
Lars Hofhansl commented on PHOENIX-4671:
----------------------------------------
In this case, yeah, it adds runtime and does not guard against future problems
like this.
This was obviously wrong and the change obviously fixes it.
I don't think we need a test for every single class that ever has a clear()
method and then verify that all relevant members are reset. A unit test here
would just check that the value is reset to 0.
Perhaps you mean a bigger _integration_ test that does the scenario I ran into,
not sure when I'll be able to get to that.
> Fix minor size accounting bug for MutationSize
> ----------------------------------------------
>
> Key: PHOENIX-4671
> URL: https://issues.apache.org/jira/browse/PHOENIX-4671
> Project: Phoenix
> Issue Type: Bug
> Reporter: Lars Hofhansl
> Priority: Minor
> Attachments: 4671.txt
>
>
> Just ran into a bug where UPSERT INTO table ... SELECT ... FROM table would
> fail due to "Error: ERROR 730 (LIM02): MutationState size is bigger than
> maximum allowed number of bytes (state=LIM02,code=730)" even with auto commit
> on.
> Ran it through a debugger, just a simple accounting bug.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
