You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Sergey Kalashnikov (JIRA)" <ji...@apache.org> on 2017/06/08 18:54:18 UTC
[jira] [Comment Edited] (IGNITE-5204) The Unicode character in the
value of a field which are included in an un-unique index will cause "stack
overhead" exception
[ https://issues.apache.org/jira/browse/IGNITE-5204?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16043237#comment-16043237 ]
Sergey Kalashnikov edited comment on IGNITE-5204 at 6/8/17 6:54 PM:
--------------------------------------------------------------------
Reproduced with attached java test.
Just added 'ccfg.setSqlIndexMaxInlineSize(4)' line to my initial java test.
was (Author: skalashnikov):
Reproduced
> The Unicode character in the value of a field which are included in an un-unique index will cause "stack overhead" exception
> ----------------------------------------------------------------------------------------------------------------------------
>
> Key: IGNITE-5204
> URL: https://issues.apache.org/jira/browse/IGNITE-5204
> Project: Ignite
> Issue Type: Bug
> Components: cache, sql
> Affects Versions: 2.0
> Environment: windows server 2012, JDK 1.8, X64
> Reporter: Chris Wang
> Assignee: Sergey Kalashnikov
> Priority: Critical
> Fix For: 2.1
>
> Attachments: IgniteSqlIssue5204Test.java
>
>
> When put "草DX009090" as the value of BillId, which is a field of entity Bill. If I define a index includes the BillId, and execute the query like "select * from Bill where BillId=’草DX009090‘ in the H2 debug console, there throws an exception by the H2 with a code 5000.
> another scenario is, I have two entities, "Bill" and "Detail", both have field "BillId". If either of them have value like "草DX009090" and execute the query like "select bill.* from bill left join detail on bill.billid=detail.billid", the whole ignite cache node will halt ( suppose there should be an stack overhead exception, dead loop).
> ======================
> I think the issue should relate to hash computing on the unicode character.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)