[
https://issues.apache.org/jira/browse/SOLR-18620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SOLR-18620:
----------------------------------
Labels: pull-request-available (was: )
> ValueSourceAugmenter: invalid preFetchDocs returns 500; stale prefetch cache
> when a context has no DocList
> ----------------------------------------------------------------------------------------------------------
>
> Key: SOLR-18620
> URL: https://issues.apache.org/jira/browse/SOLR-18620
> Project: Solr
> Issue Type: Bug
> Components: search
> Affects Versions: 9.9, 10.0
> Reporter: Ian Springer
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> SOLR-17775 added prefetching of function values to ValueSourceAugmenter. That
> change has three problems:
> 1. preFetchDocs is read with Integer.parseInt and is never validated.
> - preFetchDocs=abc throws a NumberFormatException. SolrReturnFields only
> catches SyntaxError, so the request fails with a 500.
> - preFetchDocs=-1 causes new int[-1] in setContext, which throws a
> NegativeArraySizeException, also a 500.
> Both should return 400 BAD_REQUEST.
> 2. The public constructor now calls qparser.getParam(...). Code that passes a
> null QParser used to work, because the parser was only stored, and now gets
> an NPE.
> 3. When setContext gets a context with no DocList (for example from RTG), it
> returns before replacing cachedValuesById. If the same transformer instance
> was used before, values cached against the earlier searcher can be returned
> for docids on the new one.
> The existing TestValueSourceAugmenterPrefetchLimit only checks numFound and
> score. It never checks the function value itself, so none of these problems
> are caught.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
