W3C home > Mailing lists > Public > public-qa-dev@w3.org > October 2004

Re: [meeting] Notes and log 1004-10-12

From: Ville Skyttä <ville.skytta@iki.fi>
Date: Fri, 15 Oct 2004 12:31:26 +0300
To: QA Dev <public-qa-dev@w3.org>
Message-Id: <1097832686.26996.176.camel@bobcat.mine.nu>

On Fri, 2004-10-15 at 07:13, Bjoern Hoehrmann wrote:

> >We were planning to add a link to 
> >bugzilla from this error message, but thought of something a bit more 
> >sophisticated, such as a cgi recording relevant data about the browser, 
> >referer, query etc. The catch is that in order to do so, we will have 
> >to switch off CGI::Carp (perl cgi error handler).
> 
> Only if we want this for both Apache-level and CGI::Carp level fatal
> errors and a simple way to share code between these two.

IMHO, having two "levels" of fatal errors is a bug.  Anyway, it might be
possible to keep CGI::Carp, the need for nuking it was my hunch which I
couldn't (and still cannot) remember the details for, will post more
info to list later once I get around to look into it.

>  How common
> are Apache-level fatal errors on validator.w3.org?

Dunno.  I peeked at the logs (which I found only available from
yesterday onwards), and did not find a single "500" there.

The problem is of course that some errors, like the malformed POST one,
cannot be debugged further with the information we currently get in
error_log.

People seem to be somewhat attached to CGI::Carp, is it due to the
"fatalsToBrowser" functionality, or the error logging part?
Received on Friday, 15 October 2004 09:31:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 19 August 2010 18:12:45 GMT