[
https://issues.apache.org/jira/browse/SOLR-18116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18057686#comment-18057686
]
David Smiley commented on SOLR-18116:
-------------------------------------
It's the old Pull vs Push API choice in computer science.
ContentStream is a "pull" API, it gives you something to consume – an
InputStream. On the SolrJ (client) side this created an extra step to turn
around and write (push) that / transfer to the outgoing destination (to a Solr
server). Pull conversion to Push; ugh. SolrJ side, it's better to simply have
a Push API.
But server side, it's the inverse. The server needs to pull the data from the
client. See {{org.apache.solr.request.SolrQueryRequest#getContentStreams}}
It'd be very awkward if it received a ContentWriter (Push) ... it'd need to be
inverted, maybe with another thread. No way. All servers have a pull model.
> Remove ContentStream from SolrJ (move to server)
> ------------------------------------------------
>
> Key: SOLR-18116
> URL: https://issues.apache.org/jira/browse/SOLR-18116
> Project: Solr
> Issue Type: Task
> Components: SolrJ
> Reporter: David Smiley
> Priority: Major
> Labels: newdev
>
> ContentStream related methods in SolrJ have been deprecated for multiple
> major releases (mostly if not every case). It's time to move ContentStream
> out of SolrJ into solr-core, and then un-deprecate.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
