[
https://issues.apache.org/jira/browse/SOLR-8578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15111742#comment-15111742
]
Mark Miller commented on SOLR-8578:
-----------------------------------
Seems only Binary will actually return a content length header rather than
using chunked encoding.
This seems to be because of some Writer nonsense. I *think* that we put a
Writer around the outputstream. We can't and don't want to close it, that will
close the outputstream. We can't do nothing or everything stays in the buffer.
So we flush. I think that causes Jetty to use chunked encoding. I think what we
would like to do is Writer#flushBuffer, but it's package protected. Binary just
doesn't have this Writer issue.
> ConcurrentUpdateSolrServer with BinaryResponseParser leaves 40-42 bytes
> unconsumed from the request response even without errors.
> ---------------------------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-8578
> URL: https://issues.apache.org/jira/browse/SOLR-8578
> Project: Solr
> Issue Type: Improvement
> Reporter: Mark Miller
> Priority: Minor
>
> Does not seem to happen with XML response parser.
> Not the largest deal because HttpClient appears to consume unread bytes from
> the stream for us, but something seems off.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
