lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ishan Chattopadhyaya (JIRA)" <>
Subject [jira] [Commented] (SOLR-10011) Refactor PointField & TrieField to share common code
Date Sat, 21 Jan 2017 22:21:26 GMT


Ishan Chattopadhyaya commented on SOLR-10011:

bq. Kind of... Anyone using TrieFields.TrieTypes, TrieTypes.getType() or TrieTypes.type will
have their code failing. So I don't know if we consider those kind of internal, or if we are
OK with the change anyway since it's easy to fix.

AFAIR, we've cared about such internal changes only on a best effort basis. Since it should
be fairly apparent to a developer as to why their plugin fails, I suggest we go ahead with
the refactoring even for 6x. But I shall defer to your judgement. [~hossman], do you have
any opinion?

> Refactor PointField & TrieField to share common code
> ----------------------------------------------------
>                 Key: SOLR-10011
>                 URL:
>             Project: Solr
>          Issue Type: Improvement
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Ishan Chattopadhyaya
>         Attachments: SOLR-10011.patch, SOLR-10011.patch, SOLR-10011.patch, SOLR-10011.patch
> We should eliminate PointTypes and TrieTypes enum to have a common enum for both. That
would enable us to share a lot of code between the two field types.
> In the process, fix the bug:
> PointFields with indexed=false, docValues=true seem to be using (Int|Double|Float|Long)Point.newRangeQuery()
for performing exact matches and range queries. However, they should instead be using DocValues
based range query.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message