On Tue, Jun 25, 2013 at 7:15 PM, Branko Čibej <brane_at_wandisco.com> wrote:
> On 25.06.2013 15:37, Ivan Zhakov wrote:
>> On Tue, Jun 25, 2013 at 4:15 PM, Ivan Zhakov <ivan_at_visualsvn.com> wrote:
>>> On Sun, Jun 23, 2013 at 3:43 AM, Greg Stein <gstein_at_gmail.com> wrote:
>>>> On Fri, Jun 21, 2013 at 11:36 AM, Ivan Zhakov <ivan_at_visualsvn.com> wrote:
>>>>> The fix seems to be pretty simple and could be safely backported (see
>>>>> patch). The only problem that probably we cannot backport addition of
>>>>> new configuration option due our backward compatibility policy. But
>>>>> may be I'm wrong.
>>>> The SVN_CONFIG_OPTION_HTTP_FORCE_HTTP10 symbol cannot be added.
>>>> Anything using and compiled against that could not be compiled against
>>>> But: we don't have to add the public symbol. We can certainly read/use
>>>> the config option in a 1.8.1 release.
>>>> But as I said else-thread, I think we should only disable chunked
>>>> requests. Not force HTTP/1.0 completely. And call it busted-proxy :-)
>>> I agree that force-http10 is not good name and semantic. Actually
>>> these proxies is not busted: it's allowed to HTTP/1.1 proxies to
>>> require content-length if they want. And strictly speaking proxies may
>>> have different behavior for different requests.
>>> So I'm going to commit my patch with 'http-disable-chunked-requests'
>>> and nominate for backport. This should provide workaround for current
>>> users. And then continue discussion about full/automatic solution.
>> Added 'htttp-chunked-requests' configuration in r1496470.
> I'm really not a fan of this config knob. Anyone who carries their
> laptop around will effectively have to set this as the default, because
> you never know when the next weird proxy will pop up in front of your
> server. And disabling chunked requests by default is a lot worse than
> the extra non-pipelined request for broken proxies, IMO.
These proxies actually are not broken. Such behavior allowed by RFC.
Btw we already have knob to disabling compression or control maximum
number of concurrent connections to server.
Anyway I consider this configuration option as workaround, because any
full solution other than disable chunked requests by default will
require significant effort and time.
CTO | VisualSVN | http://www.visualsvn.com
Received on 2013-06-25 17:42:43 CEST