[SOLR-13255] LanguageIdentifierUpdateProcessor broken for documents sent with SolrJ/javabin

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 7.7
• Fix Version/s: 7.7.1, 8.0
• Component/s: contrib - LangId
• Assignee: Noble Paul
• Reporter: Andreas Hubold
• Created: 15/Feb/19 09:01
• Updated: 08/Jun/19 14:57
• Resolved: 20/Feb/19 11:46

Description

7.7 changed the object type of string field values that are passed to UpdateRequestProcessor implementations from java.lang.String to ByteArrayUtf8CharSequence. SOLR-12992 was mentioned on solr-user as cause.

The LangDetectLanguageIdentifierUpdateProcessor still expects String values, does not work for CharSequences, and logs warnings instead. For example:

2019-02-14 13:14:47.537 WARN  (qtp802600647-19) [   x:studio] o.a.s.u.p.LangDetectLanguageIdentifierUpdateProcessor Field name_tokenized not a String value, not including in detection

I'm not sure, but there could be further places where the changed type for string values needs to be handled. (Our custom UpdateRequestProcessor are broken as well since 7.7 and it would be great to have a proper upgrade note as part of the release notes)

Issue Links

• causes: SOLR-13331 Atomic Update Multivalue remove does not work
• Dependent: SOLR-12983 JavabinLoader should avoid creating String Objects and create UTF8CharSequence fields from byte[]
• is related to: SOLR-13249 ByteArrayUtf8CharSequence.getStringOrNull returns null
• relates to: SOLR-13249 ByteArrayUtf8CharSequence.getStringOrNull returns null