Ishan Chattopadhyaya created SOLR-17839:
-------------------------------------------
Summary: PreAnalyzedField functionality removal in Solr 10
Key: SOLR-17839
URL: https://issues.apache.org/jira/browse/SOLR-17839
Project: Solr
Issue Type: Improvement
Reporter: Ishan Chattopadhyaya
PreAnalyzedUpdateProcessorTest and PreAnalyzedFieldTest are failing after
upgrading to Lucene 10. [~gerlowskija] described the situation the best:
[https://github.com/apache/solr/pull/3053#issuecomment-2820948089:|https://github.com/apache/solr/pull/3053#issuecomment-2820948089]
{quote}PreAnalyzedUpdateProcessorTest and PreAnalyzedFieldTest are failing
because they both rely on some quirky behavior in our PreAnalyzedField class
that is now disabled by Lucene's Field API starting in 10.0. Specifically:
having a custom value and a TokenStream on the same field. This was explicitly
prevented by apache/lucene#12053 (see the comment here), and results in
exceptions of the form:
java.lang.IllegalArgumentException: cannot change value type from String to
TokenStream
at org.apache.lucene.document.Field.setTokenStream(Field.java:406)
~[lucene-core-10.1.0.jar:10.1.0 884954006de769dc43b811267230d625886e6515 -
2024-12-17 16:15:44]
at
org.apache.solr.schema.PreAnalyzedField.fromString(PreAnalyzedField.java:260)
~[main/:?]
at
org.apache.solr.schema.PreAnalyzedField.createField(PreAnalyzedField.java:128)
~[main/:?]
{quote}
{quote}Maybe someone can find a way around this change in Lucene, but given the
obscurity of the URP in question, I wonder whether this might make it a
reasonable candidate for removal. Curious what everyone thinks...
{quote}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]