[ 
https://issues.apache.org/jira/browse/LUCENE-8638?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17374089#comment-17374089
 ] 

Michael Sokolov commented on LUCENE-8638:
-----------------------------------------

re: {{FixBrokenOffsetsFilter}} I don't think we have a replacement? I guess the 
deprecation comment ("fix the broken filters") implies that none of the filters 
we ship ever produce broken offsets, and if a user makes one,  it is their 
responsibility to "fix broken offsets" within their filter.  Yet our API 
doesn't make it possible for filter writers to produce offsets that are both 
precise and correct in the face of character level changes, applied by other 
filters, that can change token lengths. And because IndexWriter enforces 
correct offsets by throwing exceptions, it can lead to catastrophic failure if 
you don't have something like this in place to catch and fix up errors in 
offsets. As a practical matter I think users will still need such a filter. If 
we remove this, I guess they can create their own, but what's the cost of 
providing it?

Also, I don't think {{WordDelimiterGraphFilter}} is a full replacement for 
{{WordDelimiterFilter}} since it can't be used in conjunction with other 
filters that consume or produce graphs, like {{SynonymGraphFilter}}. The same 
comment applies to {{SynonymFilter}}. If we remove these, I think we need to 
fix all the {{-Graph}} filters so they can consume filter graphs?

 

> Remove deprecated code in master
> --------------------------------
>
>                 Key: LUCENE-8638
>                 URL: https://issues.apache.org/jira/browse/LUCENE-8638
>             Project: Lucene - Core
>          Issue Type: Improvement
>            Reporter: Alan Woodward
>            Assignee: Alan Woodward
>            Priority: Blocker
>             Fix For: main (9.0)
>
>          Time Spent: 1h 20m
>  Remaining Estimate: 0h
>
> There are a number of deprecations in master that should be removed. This 
> issue is to keep track of deprecations as a whole, some individual 
> deprecations may require their own issues.
>  
> Work on this issue should be pushed to the `master-deprecations` branch on 
> gitbox



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org

Reply via email to