[
https://issues.apache.org/jira/browse/SOLR-14437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17097137#comment-17097137
]
David Smiley commented on SOLR-14437:
-------------------------------------
bq. can you explain a bit more on the getApis()
Put simply, if a request handler is startup=lazy, then Solr doesn't know at
what V2 Apis (paths) to register it. In V1 it's trivially the name of the
request handler.
> Remove/refactor "ApiSupport" interface? (for V2 API)
> ----------------------------------------------------
>
> Key: SOLR-14437
> URL: https://issues.apache.org/jira/browse/SOLR-14437
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Components: v2 API
> Affects Versions: master (9.0)
> Reporter: David Smiley
> Priority: Major
>
> ApiSupport.java is an interface relating to the V2 API that is implemented by
> all request handlers, both those at a core level and others. It's
> essentially this: (comments removed)
> {code:java}
> public interface ApiSupport {
> Collection<Api> getApis();
> default Boolean registerV1() { return Boolean.TRUE; }
> default Boolean registerV2() { return Boolean.FALSE; }
> }
> {code}
> Firstly, let's always assume that the handler will always be registered in
> V2. All implementations I've seen explicitly return true here; maybe I'm
> missing something though.
> Secondly, getApis() seems problematic for the ability to lazily load request
> handlers. Can we assume, at least for core level request handlers, that
> there is exactly one API and where necessary rely on the "spec" JSON
> definition -- see org.apache.solr.api.ApiBag#registerLazy ?
--
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
