users@jersey.java.net

Re: [Jersey] Is setMaxAge needed at notModified()?

From: Paul Sandoz <Paul.Sandoz_at_Sun.COM>
Date: Mon, 28 Jul 2008 12:10:11 +0200

Hi Markus,

If you are consistently sending the same headers to different browsers
than i doubt it is your fault.

I really do not know much about how browsers conform. Perhaps you can
try searching for such conformance issues. Or send email to the Yahoo
group rest-discuss, where i think you may get a more informed answer
than at least i can provide you.

Paul.

Markus KARG wrote:
> Hi Jersey Community,
>
> I am doing some experiments with Jersey-0.8 and HTTP Caching. Explicitly
> I am checking how to efficiently implement support for conditional
> requests in a JSR-311 resource to reduce the actual workload in
> subsequent processes like JPA / JDBC etc.
>
> In some test line I noticed that when using "return
> Response.notModified().build();" then the clients behave differently:
> Some browsers' chaches interpret it in a way that the cached entity is
> "fresh" again for the max-age defined in the ORIGINAL response, while
> other browsers' caches interpret it in a way that the cached entity is
> revalidated but keeps "stale" status. As a result some browsers do not
> validate once more for a another max-age period, while others still
> revalidate each single subsequent request.
>
> I know that it is not 100% on-topic in this forum, but I am new to the
> HTTP/1.1 protocol on this technical level and like to know whether I
> have to send "return
> Response.notModified().cacheControl(myMaxAgeCacheControl).build();"
> instead of "return Response.notModified().build();" to get the same
> result in all browsers, or whether I am doing something wrong. Certainly
> I want that all browsers behave the same. Is it my fault that they
> behave differently, or is it a bug in one of the browsers (or in Jersey)?
>
> Thanks
> Markus
>

-- 
| ? + ? = To question
----------------\
    Paul Sandoz
         x38109
+33-4-76188109