Quoting Mike Orr :
> I looked through the HTTP RFC and http_response.py, and it looks like
> it doesn't cache by default but somehow it is.  self.cache is 0, which
> seems to translate to "Expires: -1", but my browser (Firefox) says
> "Expires: not specified".  I'd like to put "Expires: 0" but it looks
> like I'd have to hack the Publisher to do so.  I figured I'd ask here
> first.

Well, not to tell you how to arrange your app :), but I think rearranging your
URL space (as David suggested, or some similar approach) may be preferable to
modifying the Qx Expires header behavior for a couple of reasons:

  1.  Your URLs are not currently 'shareable', or strictly speaking, URIs.  I
can't copy a link and email to someone, and be sure they'll see the same thing
I saw when I copied the link.  In fact, when I just tried to use the link
http://www.incidentnews.gov/1/entries in a browser that hadn't already
established a session, it caused an error.

  2.  If you do something that forces your whole site to always instantly expire
(cache-wise) your visitor's browsers will always have to request fresh copies
of everything.  Unless you 'hack the hack,' that will include all images,
including any 'static' images (header, footer, etc) that are served by Quixote.
 Also any external stylesheets, js files, etc.

J