On Mon, Jun 23, 2003 at 03:24:46PM -0500, Jason Sibre wrote:
> Was there something in particular about the approach I took in addressing
> this that you aren't comfortable with?  Existing code, perhaps, or some
> other short-sighted glitch on my end?  If so, let me know, and I'll revisit
> my modification.

Please review the attached patch.  It is basically your patch with some
minor changes.  I've already changed http_response.py to handle 'cache'
a bit differently.  If it is set to None then the Expires header will
not be included in the response.  Sometimes it doesn't make sense to set
the Expires header.

For static files, I've changed the IMS code to only perform exact
matches (i.e. the browser needs to preserve the value send in the
Last-Modified header).  That makes things simpler and more robust (your
code would die if the client sent a badly formated IMS header).  AFAIK,
IE 6 and Mozilla both get this right.

I've made an incompatible change to StaticFile in that the Expires
header will no longer be set to -1 by default.  That seems like a silly
thing to do for static files.  Now the default is to not set an Expires
header but always set a Last-Modified header.  I that matches more
closely what Apache does and will allow browser and proxies to cache
static files at their discretion.

  Neil