httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@covalent.net>
Subject Re: cvs commit: httpd-2.0 STATUS
Date Wed, 10 Oct 2001 21:06:56 GMT
From: "Joshua Slive" <joshua@slive.ca>
Sent: Wednesday, October 10, 2001 3:10 PM


> > From: William A. Rowe, Jr. [mailto:wrowe@covalent.net]
> > Let me make this _perfectly_clear_, the only reason .asis worked
> > before is that
> > _every_ filename extension worked before.  .bak, .old, didn't
> > matter.  Everything
> > was served, and many users (>100) complained that there was some
> > 'bug' in their
> > apache installation.  You may call that a misconfiguration, but
> > our users disagree.
> 
> Out of curiosity mostly, can you tell me where that user data comes from?
> If you are talking about the problem with the "it worked" page in recent
> versions,  I don't see that as directly related to the problem we are
> discussing.

It is entirely related, please review the code until you grok it.
The messages were initiated in several bug reports, and most of the
conversations were initiated on security@ because _THAT_ is the kind
of bug that users believed they were confronting.

Sometimes such long debugging sessions get mis-cc'ed over the course of
time; Bill Stoddard was able to retrieve sufficient postings 
to overload his brain^[ddi
to convince him the entire patchset was required to resolve all of the
aspects of these tangled problems.

> Selecting a "non-negotiable" variant when "negotiable" variants
> are available was obviously wrong.  But selecting a "non-negotiable" variant
> when nothing else is available seems sensible.

What's so wrong with that, if the rest of your theory holds up?

That is more complex behavior.  So you suggest that requesting
index.html should serve up index.html.bak?  I don't think so.
 
> I think it is great that you have done all this work to clean up the
> negotiation issues.  Things seem to be in MUCH better shape now.  What is
> left is a relatively minor issue about a feature of multiviews that,
> admittedly, has nothing to do with negotiation.  If you want to add a
> directive to turn on or off this part of multiviews, that is fine with me.
> But I see no good reason to get rid of this feature entirely.

And I didn't insist on doing so.  As you say, it has nothing to do with
negotiation and doesn't belong there.  For the sake of argument, we will
leave this in there.

I will also offer this variant;

MultiviewsMatch Any

to behave as pre-1.3.20.  If the administrator turns that on, woe be to
him if he creates index.html.bak :)

Bill



Mime
View raw message