[
https://issues.apache.org/jira/browse/SOLR-1824?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12847166#action_12847166
]
Hoss Man commented on SOLR-1824:
--------------------------------
Scratch that -- i get it now:
* IndexSchem uses anonymous subclasses of AbstractPluginLoader to instantiate a
variety of differnet things
* AbstractPluginLoader processes things in a loop, recording errors in
SolrConfig.severeErrors when a particular instance can't be inited, but
creating the rest of the objects just fine.
* when abortOnConfigurationError=false this results in solr using a schema with
missing filters (or missing fields, etc...) .. the only thing that protects
people when abortOnConfigurationError=true is that SolrDispatchFilter pays
attention to both abortOnConfigurationError and SolrConfig.severeErrors
(someone using embedded Solr might never notice the error at all, even if the
config did say abortOnConfigurationError=true)
> partial field types created on error
> ------------------------------------
>
> Key: SOLR-1824
> URL: https://issues.apache.org/jira/browse/SOLR-1824
> Project: Solr
> Issue Type: Bug
> Affects Versions: 1.1.0
> Reporter: Yonik Seeley
> Priority: Minor
>
> When abortOnConfigurationError=false, and there is a typo in one of the
> filters in a chain, the field type is still created by omitting that
> particular filter. This is particularly dangerous since it will result in
> incorrect indexing.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
