Mailing-List: contact dev-help@hc.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "HttpComponents Project" <dev@hc.apache.org>
Message-ID: <1044327948.1257971079705.JavaMail.jira@brutus>
Date: Wed, 11 Nov 2009 20:24:39 +0000 (UTC)
From: =?utf-8?Q?Ortwin_Gl=C3=BCck_=28JIRA=29?= <jira@apache.org>
To: dev@hc.apache.org
Subject: [jira] Commented: (HTTPCLIENT-889) Should USE_EXPECT_CONTINUE be
 false by default?
In-Reply-To: <1694514010.1257959379577.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/HTTPCLIENT-889?page=3Dcom.atlas=
sian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D=
12776636#action_12776636 ]=20

Ortwin Gl=C3=BCck commented on HTTPCLIENT-889:
-----------------------------------------

Yeah, for "small" request entities (such as form data) it doesn't make much=
 of a difference. It's really only useful for
a) large request entities, such as image uploads over a low-bandwidth conne=
ction
b) request entities that are not cached/repeatable, i.e. only producible on=
ce (such as streamed data)

Judge for yourself if those are "the common" use case for HttpClient.

Mind you, reasons why a server would reject a request whose body it hasn't =
even seen are mainly:
- not authenticated
- not authorized
- content too large
- missing content-length information (chunked TE not supported/allowed)
- content type not expected/supported

> Should USE_EXPECT_CONTINUE be false by default?
> -----------------------------------------------
>
>                 Key: HTTPCLIENT-889
>                 URL: https://issues.apache.org/jira/browse/HTTPCLIENT-889
>             Project: HttpComponents HttpClient
>          Issue Type: Wish
>          Components: HttpClient
>    Affects Versions: 4.0 Final
>            Reporter: Ren Waldura
>            Priority: Minor
>             Fix For: 4.1 Alpha1
>
>
> It seems the point of USE_EXPECT_CONTINUE is to improve performance when =
posting large data.=20
> http://www.w3.org/Protocols/rfc2616/rfc2616-sec8.html says:
> << The purpose of the 100 (Continue) status (see section 10.1.1) is to al=
low a client that is sending a request message with a request body to deter=
mine if the origin server is willing to accept the request (based on the re=
quest headers) before the client sends the request body. In some cases, it =
might either be inappropriate or highly inefficient for the client to send =
the body if the server will reject the message without looking at the body.=
 >>
> There's nothing wrong with HttpClient performing well by default, however=
, every other HTTP client library I've used does not behave like this (PHP =
curl, Perl LWP). The default is always to do one request, including the bod=
y. Maybe dumb, but simple.
> It seems to me HttpClient's default behavior should the simplest, most co=
mpatible with all HTTP-speaking services out there. "100 Continue" is somew=
hat advanced, and may not be correctly implemented by all services. (That's=
 of course how I found out about it -- my server doesn't implement it.)
> If USE_EXPECT_CONTINUE is used only for performance reasons, it seems lik=
e it would be simpler (and therefore maybe more "correct") to have it "off"=
 by default. And only enable it when needed, when there is a good reason to=
.
> Just my thoughts. And a wish. Thanks!=20

--=20
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org