[
https://issues.apache.org/jira/browse/MINDEXER-108?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16265041#comment-16265041
]
Andreas Sewe commented on MINDEXER-108:
---------------------------------------
> What versions of lucene accepts * as prefix? Current master uses Lucene 5.x
> and it does not accept it...
I am using Lucene 5.4.4 (haven’t changed that from the current master). After
uncommenting the above {{ParseException}}-throwing in {{indexer-core}} code, it
happily runs queries like {{(#g:*.maven)}}, which I build like this
{noformat}
queryCreator.constructQuery(MAVEN.GROUP_ID, new
SearchTypedStringSearchExpression("*.maven", SearchType.EXACT))
{noformat}
and run using {{indexer.searchIterator(...)}}.
That being said, the Maven Indexer still logs the following:
{noformat}
[DEBUG] Executing unbounded search, and fitting topHitCounts to 12537, an OOMEx
might follow. To avoid OOM use narrower queries or limit your expectancy with
request.setCount() method where appropriate. See MINDEXER-14 for details.
{noformat}
AFAICT, the {{IteratorSearchResponse}} contains all results, so that warning
seems inappropriate; MINDEXER-14 has been marked as fixed, after all.)
> Lift restriction on leading wildcard queries
> --------------------------------------------
>
> Key: MINDEXER-108
> URL: https://issues.apache.org/jira/browse/MINDEXER-108
> Project: Maven Indexer
> Issue Type: Improvement
> Affects Versions: 6.0
> Reporter: Andreas Sewe
> Fix For: 6.0
>
>
> The {{DefaultQueryCreator}} prevents queries with leading wildcard:
> {noformat}
> if ( query.startsWith( "*" ) || query.startsWith( "?" ) )
> {
> throw new ParseException( "Query cannot start with '*' or '?'!" );
> }
> {noformat}
> While this was necessary in older versions of Lucene, the version used now
> happily executes such queries. The restriction can thus be lifted.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
