Re: Please wait... indefinitely

2009/9/22 Francois Daoust <fd@w3.org>:
> Hi Danny,
>
> Danny Ayers wrote:
>>
>> First I must say: great to see the service! Could do with a bit more promo
>> though, I only stumbled on it, when I mentioned it on Twitter quite a few
>> people I'd have expected to have known about it were (pleasantly) surprised.
>
> Yes, we love secrets... No, I'm just kidding ;).

Oh yes, W3C speciality :)

> Getting the word out is not easy. For instance, today, on the W3C twitter
> account @w3c, there was a link to a blog post entitled "Let's Make Every Day
> One Web Day!". The post encouraged people to use the mobileOK Checker among
> other tools:
>  http://www.w3.org/QA/2009/09/lets_make_every_day_one_web_da.html

Think that's where I heard.

> Is it enough? No. We try to mention such tools as often as we can. You can
> find an incomplete (sigh) list of tools developed at W3C by searching for
> the "Open Source Software" link on the W3C home page. That should lead you
> to:
>  http://www.w3.org/Status
>
> Anyway, glad to hear that you like the tool!
>
>
>
>> Unfortunately when I try on the URI:
>>
>> http://dannyayers.com
>>
>> it stays on "The address is being checked. Please wait..." (about half an
>> hour this time).
>
> Hmmm. I think I know what happens. That does involve a bug in the mobileOK
> Checker as well as a weird thing happening on your side though.
>
> Accessing your Web site with a regular Web browser seems to start an
> apparent endless loop. When I try to access the Web page on my local
> machine, I eventually manage to receive an HTTP status code 502 after a few
> minutes, probably returned by some intermediary proxy that gets tired of
> waiting for a response.

Right - it's fallen over (thanks for the bug report!). I'm running
apache2 with a pointer to some Scala stuff on port 8080 (which in
effect is Java, major memory hog - if I get the chance today will set
it to write blog page as a file for Apache to serve). The 502 would
have come from a proxy, my error-handling is minimal right now, not
even doing 404s yet...

> The mobileOK Checker has the same problem. It tries to retrieve the page,
> but this takes a few minutes. There is an internal mechanism that makes the
> mobileOK Checker forget about tasks that have run for too long, to protect
> the server from endless loops.
>
> This mechanism has the exact opposite effect in practice (ooops). It drops
> the task as expected, but the task is automatically re-created by the client
> who started the check (and who polls the server every few seconds for a
> status update).

heh, oops :)

> In short, if you leave your browser open, the mobileOK Checker will just run
> endlessly trying to test your page, as you experienced. I created a bug to
> track this down and will fix this as soon as possible.

Cool.

> [ See bug 7702:
> http://www.w3.org/Bugs/Public/show_bug.cgi?id=7702 ]
>
> That said, why does your page fail to return a response in a timely fashion?
> (if you do not experience that issue, could you make sure that  the access
> to your Web site is not restricted to a given IP address or something
> similar?)

I *thought* it was ok when I ran into the mobileOK issue (I had
checked just previously, maybe unlucky coincidence it fell over when I
tried to validate), but will try and sort & retry my end.

Many thanks,
Danny.



-- 
http://danny.ayers.name

Received on Wednesday, 23 September 2009 08:13:47 UTC