W3C home > Mailing lists > Public > public-browser-tools-testing@w3.org > July to September 2014

Re: Status indicating intermediary server problems

From: Simon Stewart <simon.m.stewart@gmail.com>
Date: Fri, 26 Sep 2014 16:16:11 +0100
Message-ID: <CAOrAhYFYa2o_=1sguujLUyFnfxD8fSH+EmvUWFVQSpPgrT7dcA@mail.gmail.com>
To: David Burns <dburns@mozilla.com>
Cc: Andreas Tolfsen <ato@mozilla.com>, public-browser-tools-testing <public-browser-tools-testing@w3.org>
>From the end user's point of view, what's the difference? And how is this
different from a normal HTTP request where a proxy chokes? As far as each
step on the path from local to remote end goes, the "next hop" _is_ the
remote end, so it'd be hard for an intermediary to know whether or not the
problem was caused by the remote end failing to respond or another node?

Would a 502 or 504 HTTP response be appropriate?

Simon

On Fri, Sep 26, 2014 at 3:18 PM, David Burns <dburns@mozilla.com> wrote:

> Just for clarification, this error would not come from the UA part of the
> Remote End but from the HTTP Endpoint if something goes wrong.
>
> David
>
>
> On 24/09/2014 18:39, Andreas Tolfsen wrote:
>
>> There is currently no way to determine where a command fails if you
>> have one or more intermediary servers sitting in between the local and
>> remote ends.
>>
>> If for example you use a proxy like a Selenium remote server between
>> your client and your driver, there's no way to find out whether an
>> unsuccessful response/error was caused by the driver or the proxy.
>>
>> I'm currently in a situation where I have multiple proxies sitting
>> between the local and final remote end where if one of the proxies
>> have an internal problem or the next remote fails to reply, it would
>> be useful for it to return a response indicating the problem is caused
>> by the intermediary rather than the driver.
>>
>> For this reason I'm proposing to introduce a new status called “proxy
>> error” or “intermediate error”.
>>
>> This error would not be possible to use for the driver
>> implementations, but could be used by intermediary remotes, such as
>> proxies, to say that an error occurred in their domain, and that the
>> command never made it all the way to the driver.
>>
>>
>
>
Received on Friday, 26 September 2014 15:16:38 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 12 February 2015 15:39:03 UTC