[
https://issues.apache.org/jira/browse/LUCENE-6156?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14263465#comment-14263465
]
Michael McCandless commented on LUCENE-6156:
--------------------------------------------
Hmm, that stack trace is from Lucene 4.9.x not 4.10.x. Maybe you saw this with
Elasticsearch 1.3.x?
Lucene's regexp parsing/building is recursive, meaning it consumes one java
stack frame per regexp operation (|, &, etc.). This hasn't changed recently,
e.g. it was the same way before LUCENE-5752, so I'm not sure why you saw it
working with previous Lucene version (btw, 10.6 is not a valid Lucene
version... can you re-check which Elasticsearch/Lucene version you saw this
working on?).
Making this code non-recursive is likely not a great option ... it makes the
code more complex. For example, isFinite can also hit StackOverflowError, but
we abandoned making it non-recursive in LUCENE-5659 because it increased code
complexity.
Maybe we should add a "max regexp parts" limit so you get a more sane exception
for too-large regexps? Similar to what we did for determinize in LUCENE-6046
...
> StackOverFlow error during parsing of a request
> -----------------------------------------------
>
> Key: LUCENE-6156
> URL: https://issues.apache.org/jira/browse/LUCENE-6156
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/queryparser
> Affects Versions: 4.10.2
> Environment: windows 2008, osx yosemite with java 1.7.0_60
> Reporter: Aurelien PISU
> Priority: Critical
>
> during parsing of a query send to lucene via elasticSearch 1.4.2, i encounter
> that stackOverFlow exception on RegExp.
> here the stack trace
> Caused by: java.lang.StackOverflowError
> at java.util.BitSet.(BitSet.java:154)
> at org.apache.lucene.util.automaton.Automaton.(Automaton.java:75)
> at org.apache.lucene.util.automaton.Automata.makeString(Automata.java:273)
> at org.apache.lucene.util.automaton.RegExp.toAutomaton(RegExp.java:518)
> at org.apache.lucene.util.automaton.RegExp.findLeaves(RegExp.java:553)
> at org.apache.lucene.util.automaton.RegExp.findLeaves(RegExp.java:550)
> at org.apache.lucene.util.automaton.RegExp.findLeaves(RegExp.java:551)
> at org.apache.lucene.util.automaton.RegExp.findLeaves(RegExp.java:551)
> Note : the regular expression is quite large and contains only ascii
> character and '|' character. let me know, If you need the value of the
> regular expression. This issue was firstly raise to elastic search component
> on github that use the 4.10.2 version of lucene and identify as a lucene
> issue.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
