W3C home > Mailing lists > Public > www-validator@w3.org > February 2005

Re: checklink: timeout bug?

From: Ville Skyttä <ville.skytta@iki.fi>
Date: Sat, 05 Feb 2005 11:06:48 +0200
To: www-validator@w3.org
Cc: Shadi Abou-Zahra <shadi@w3.org>
Message-Id: <1107594408.30769.70.camel@bobcat.mine.nu>

On Sat, 2005-02-05 at 11:34 +0900, Olivier Thereaux wrote:
> On Fri, Feb 04, 2005, Shadi Abou-Zahra wrote:
> > ref: <http://www.w3.org/wai/policy,checklink>
> > fyi, times out after a while and does not work.
> 
> Strictly speaking, there is no bug... For pages with a lot of links to
> check, he link checker takes a long time, and if used in summary mode
> where no permanent output is produced, some browsers tend to think the
> service has crashed and decide to time out.

Yep, but there seems to an additional catch: among the links in that
page, http://accessability.noie.gov.au/ times out currently.  The
default timeout of the link checker is one minute, but the above failure
causes a 2 minute delay before the checking proceeds and new output gets
sent to the browser, also in non-summary-only mode.  I'm guessing this
is due to trying to fetch /robots.txt, then the actual document, both
timing out after a minute.

I'm not sure what to do about this.  The default timeout of one minute
sounds somewhat generous, maybe drop that to 30 seconds?  Also, the
emit-whitespace hack we apply to summary only mode could in theory be
applied to the normal output mode too.  But that'd require some changes
to the way progress is currently reported; emitting whitespace in a
<pre> block will have its effects on what's shown browsers (not that I
would be really fond of the <pre> anyway).

Thoughts, suggestions welcome.
Received on Saturday, 5 February 2005 09:06:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:18 GMT