You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hc.apache.org by "Michael Osipov (Jira)" <ji...@apache.org> on 2020/07/29 09:01:00 UTC
[jira] [Commented] (HTTPCORE-641) Consider segmenting large writes
for the out of order response check
[ https://issues.apache.org/jira/browse/HTTPCORE-641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17167070#comment-17167070 ]
Michael Osipov commented on HTTPCORE-641:
-----------------------------------------
I do not think that this is a wise idea. This could significantly hurt performance. If someone selects a large buffer, we should respect that. Please see WAGON-537 for such a case.
> Consider segmenting large writes for the out of order response check
> --------------------------------------------------------------------
>
> Key: HTTPCORE-641
> URL: https://issues.apache.org/jira/browse/HTTPCORE-641
> Project: HttpComponents HttpCore
> Issue Type: New Feature
> Components: HttpCore
> Affects Versions: 5.0.1
> Reporter: Carter Kozak
> Priority: Major
>
> The blocking client checks for an out-of-order response every 8kb, however content-length requests may write large buffers that aren't chunked before being written to the socket, allowing them to bypass this protection.
> Should the OutputStream which checks for early responses also split large writes in order to check more frequently?
> Note that I don't think chunked encoding is impacted due the the (default) 2 KiB chunk size.
> Fixing this may prevent socket exceptions, but could come at a slight performance cost due to more frequent 1ms reads.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org