Frank Maas wrote:
> You conclude somehwere that IE is not honouring the Expires header,
> but that need not be completely true. The only thing that header is
> telling IE is that it may not _unconditionally_ reuse the same data
> the next time it is needed.  And most probably IE complies to it.
> Since the Expires header tells IE that  the data is "expired" it
> will request the image again. But, different to wget, it will provide
> Apache with the Etag and timestamp of the image it has downloaded the
> last time. Apache then will reply a 304 meaning 'the data you have is
> the same as I have'. This response is based on ETag, URI and timestamp
> of the downloaded image. As you can see in the result of both requests
> (see your data) that information is the same both times.

Sounds plausible.  (If it becomes necessary, perhaps I can snoop the LAN with
Ethereal, etc., to see the differences between wget and IE.)


> What you can try is to update the Last-Modified header on each
> request. Bear in mind that this won't work for requests that are made
> within the same second. To probably state the obvious: it is necessary
> that you set the header to the timestamp of the request, not to the
> timestamp of the underlying image!

I'd prefer to get no_cache(1) working, but will keep the various date-forcing
ideas as backups.


Thanks!

David

Reply via email to