[
https://issues.apache.org/jira/browse/LUCENE-9212?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17044009#comment-17044009
]
David Smiley commented on LUCENE-9212:
--------------------------------------
> Automatons can be defined in both binary and unicode space, and there's no
>way of telling which it is when it comes to compiling them
Isn't that a problem with our API -- more of a root cause? I've been bitten by
the un-typed nature of byte vs char automatons.
> Intervals.multiterm() should take a CompiledAutomaton
> -----------------------------------------------------
>
> Key: LUCENE-9212
> URL: https://issues.apache.org/jira/browse/LUCENE-9212
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Alan Woodward
> Assignee: Alan Woodward
> Priority: Major
> Fix For: 8.5
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> LUCENE-9028 added a `multiterm` factory method for intervals that accepts an
> arbitrary Automaton, and converts it internally into a CompiledAutomaton.
> This isn't necessarily correct behaviour, however, because Automatons can be
> defined in both binary and unicode space, and there's no way of telling which
> it is when it comes to compiling them. In particular, for automatons
> produced by FuzzyTermsEnum, we need to convert them to unicode before
> compilation.
> The `multiterm` factory should just take `CompiledAutomaton` directly, and we
> should deprecate the methods that take `Automaton` and remove in master.
--
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
