[
https://issues.apache.org/jira/browse/HTTPCLIENT-2362?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17927296#comment-17927296
]
Matthew E. Dawson edited comment on HTTPCLIENT-2362 at 2/14/25 10:33 PM:
-------------------------------------------------------------------------
Thanks for the prompt reply and solution, [~olegk]. Is this behavior formally
documented anywhere, and if so, can you point me to it? I found it *incredibly*
confusing that a method that I'd previously used to force 1.0 semantics
silently changed on me, and I didn't see anything obvious in the release notes
or API documentation. (I also can't remember breakfast, so it's possible that I
skipped right over it and forgot it.)
was (Author: pavlov112):
Thanks for the prompt reply and solution, [~olegk]. Is this formally behavior
documented anywhere, and if so, can you point me to it? I found it *incredibly*
confusing that a method that I'd previously used to force 1.0 semantics
silently changed on me, and I didn't see anything obvious in the release notes
or API documentation. (I also can't remember breakfast, so it's possible that I
skipped right over it and forgot it.)
> Regression: Versions after 5.3.1 silently ignore attempts to perform HTTP/1.0
> requests
> --------------------------------------------------------------------------------------
>
> Key: HTTPCLIENT-2362
> URL: https://issues.apache.org/jira/browse/HTTPCLIENT-2362
> Project: HttpComponents HttpClient
> Issue Type: Bug
> Affects Versions: 5.4-alpha1, 5.4-alpha2, 5.4, 5.4.1, 5.4.2
> Reporter: Matthew E. Dawson
> Priority: Major
>
> Every attempt I've made to use the classic API to invoke an HTTP/1.0 request
> is silently converted into HTTP/1.1:
> {code:java}
> final var request = new HttpGet(uri);
> request.setVersion(HttpVersion.HTTP_1_0);
> client.execute(host, request, responseHandler);
> {code}
> There's no error on the client side, but everything that I see on the server
> side shows HTTP/1.1, whether I use the classic or async interfaces. I've
> drilled down into the source and I think that it's because
> DefaultHttpRequestWriter entirely ignores the request's version in favor of
> that returned by its internal config object (which always points back at
> HttpVersion.HTTP_1_1), which would more properly make this a core defect
> rather than client, but I have no idea if the problem is confined to the core
> library.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@hc.apache.org
For additional commands, e-mail: dev-h...@hc.apache.org
