Content-Location non-bug/bug dilemma

[Dawit Alemayehu]

On Friday 23 December 2005 08:58, Thiago Macieira wrote:
> Dawit Alemayehu wrote:
> >Hello,
> >
> >The following patch is meant to address BR# 82747:
> >http://bugs.kde.org/show_bug.cgi?id=82747
> >
> >It reverts a patch that fixed BR# 51185:
> >http://bugs.kde.org/show_bug.cgi?id=51185
> >
> >Following a standard is worthless if we are the only ones that do it.
> > Every other major browser seems to break this. At least Mozilla/Firefox
> > does. Anyways, I want to bring this up for discussion here because this
> > breaks starndard compliant behavior.
>
> According to the Mozilla bug report on the issue, they decided to back out
> their support given the high number of IIS servers being stupid. The rate
> was as high as 9.5% in Jan 1st, 2004, but it has dropped to 0.25% the
> last month (compare
> http://www.securityspace.com/s_survey/data/man.200511/firewalled_cloc.html
> to
> http://www.securityspace.com/s_survey/data/man.200312/firewalled_cloc.html)
>
> This means those servers are slowly being fixed. Which is why I think this
> feature should stay.

Well it really does not matter if it breaks 0.25% or 9.5% websites. All the 
websites in our bug report still do not work. And I do not know how many 
countless ones are out there. Also this is not only an issue of buggy IIS 
servers. It can result from buggy implementation as well. The code is 
currently commented out since none of the major browsers support it. If they 
do not support it, no one can and will use that feature! Anyways, the code 
can be uncommented when all the clients and servers properly deal with this 
issue.

> If at all, maybe the Opera solution would suit better
> (https://bugzilla.mozilla.org/show_bug.cgi?id=109553#c14): use
> Content-Location if the protocol, host and port are the same.

Doing this IMHO is complemently useless since it also breaks the spec. They 
are doing what is not specified in the spec. We should either implement or 
not just like the many other parts of HTTP/1.1 we do not yet support.

-- 
Regards,
Dawit A.
"Practice what you preach, preach what you practice"