[
https://issues.apache.org/jira/browse/SOLR-12775?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16693283#comment-16693283
]
Uwe Schindler commented on SOLR-12775:
--------------------------------------
bq. TokenizerFactory.create() returns a Tokenizer, so we can't wrap there
either unfortunately.
You are right, I remember. There is still a way without code duplication. Just
define the Tokenizer as a new class without any code, except:
In the constructor create the "real tokenizer" and wrapping tokenfiter chain
using the given attributesource/attributefactory and store in final field. Then
also override incrementToken(), end() and just delegate to this factory. I did
this for a customer a while back, should work.
Nevertheless, I am fine with the current patch. We should just figure out why
there is no deprecation warning printed automatically. In the past this always
worked!
> Add a deprecated implementation of LowerCaseTokenizer
> -----------------------------------------------------
>
> Key: SOLR-12775
> URL: https://issues.apache.org/jira/browse/SOLR-12775
> Project: Solr
> Issue Type: Task
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Alan Woodward
> Priority: Blocker
> Fix For: master (8.0)
>
> Attachments: SOLR-12775.patch, SOLR-12775.patch, SOLR-12775.patch,
> SOLR-12775.patch
>
>
> LUCENE-8498 will remove LowerCaseTokenizer and LowerCaseTokenizerFactory from
> lucene 8. To make upgrading from Solr 7.x to Solr 8 easier for users who
> have schemas that use LowerCaseTokenizerFactory, we should add a deprecated
> copy of the code to the {{org.apache.solr.analysis}} package.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
