[
https://issues.apache.org/jira/browse/SOLR-578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12598179#action_12598179
]
Noble Paul commented on SOLR-578:
---------------------------------
It is by design that SOLR keep the requesthandlers and response writers loosely
coupled. The responsibilities are clearly defined
* RequestHandler : Takes in the request parameters , do the necessary
operations and prepare a response to be written back to the client
* ResponseWriter : Take in the response *object* and write it down in a format
of client's choice
If your Handler can write out data only using a specific writer, you have the
flexibility of overriding the 'wt' in the handler. Register your own writer in
solrconfig.xml.
We are implementing SOLR-561 using the same strategy .
* We have a handler which reads an index file (or a part of it) and puts the
stream object into the SolrQueryResponse rsp.add("stream", inputstreamobject)
* It sets a custom writer . (solrparams.Params.set("wt", "filestream")).
* The custom binary writer then take the inputstream object (rsp.get("stream"))
and pipe the data to the outputstream.
> Binary stream response for request
> ----------------------------------
>
> Key: SOLR-578
> URL: https://issues.apache.org/jira/browse/SOLR-578
> Project: Solr
> Issue Type: Improvement
> Affects Versions: 1.3
> Reporter: Jason Rutherglen
>
> Allow sending binary response back from request. This is not the same as
> encoding in binary such as BinaryQueryResponseWriter. Simply need access to
> servlet response stream for sending something like a Lucene segment.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
