[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [wsrp-interfaces] Re: [wsrp] Re: [wsrp-interfaces] Re: [wsrp]Enable caching of resource content on the client
See my reply to Stefan's e-mail. I agree with you its something that
needs to be looked at beyond 2.0 but don't think it should be addressed
in 2.0 as the portlet has the option of using http proxying. -Mike- Richard Jacob wrote: ok, if this is the case, we really need to look into the caching issue. For efficient client side rendering using gR it seems to be crucial for me that the responses are cacheable by the infrastructure. Mit freundlichen Gruessen / best regards, Richard Jacob ______________________________________________________ IBM Lab Boeblingen, Germany Dept. 2289, WebSphere Portal Server Development 1 WSRP Team Lead WSRP Architecture & Standardization Phone: ++49 7031 16-3469 - Fax: ++49 7031 16-4888 Email: mailto:richard.jacob@de.ibm.com Michael Freedman <michael.freedman @oracle.com> To wsrp-interfaces@lists.oasis-open.or 02/15/07 08:25 PM g, WSRP TC <wsrp@lists.oasis-open.org> cc Subject [wsrp] Re: [wsrp-interfaces] Re: [wsrp] Enable caching of resource content on the client I am opposed to dropping support for using portletURLs in a resource response as I expect our XPR/full-protocol Ajax support to be too cumbersome for many Ajax use cases. I.e. most portlet use cases solved with WSRP 1.0. -Mike- Richard Jacob wrote:I think option 1 excludes the most common use case where e.g. JS libs, CSS classes, etc. are loaded. We should still be optimized for a caching infrastructure and offload a) the producer and b) the consumer itself. I dont see why we should weaken that and provide a solution that from a plain infrastructure and performance viewpoint is worse than 1.0. I think the caching use cases are important enough to cleanly supportthem.Alternative 2 sounds feasible, in that case we would need to modify the EBNF of the resource URL. Another third solution I see is to drop the support for having portletURLsin resources' markups. We talked today about the symmetrical behavior of gR and HTTP resource serving via the Consumer and the importance of having the same functionality for both. So enabling them in the one case while dissallowing them in the otherseemsinconsistent to me. What was the ability added for again? Afaik we tried to solve some Ajaxusecases with it, right? On the other and we seem to say that a solution for Ajax won't be part of the 2.0 spec and we need to understand it more. So here again it would seem consistent to me that we dropped that capability and continue the Ajax discussion and add the necessary items post-2.0. Mit freundlichen Gruessen / best regards, Richard Jacob ______________________________________________________ IBM Lab Boeblingen, Germany Dept. 2289, WebSphere Portal Server Development 1 WSRP Team Lead WSRP Architecture & Standardization Phone: ++49 7031 16-3469 - Fax: ++49 7031 16-4888 Email: mailto:richard.jacob@de.ibm.comStefan Hepper<sthepper@hursley.ibm.com> ToWSRP TC <wsrp@lists.oasis-open.org>02/15/07 04:30 PM ccSubject[wsrp] Enable caching of resourcecontent on the clientAs mentioned in one of the WSRP calls currently caching of content served through resource URLs on the client is basically not possible for resource that are served through gR. Reason for this is that the resource URL that the consumer produces needs to contain all nav state of all portlets on the page in order to allow the gR call triggered with the URL to produce markup that contains WSRP URLs again. This means that the URL will likely change for each request and there will always be a cache miss on the client. So basically there are two different use cases: 1. render a resource that contains no or only resource WSRP tokens 2. render a resource that may contain any WSRP URL token Solutions I see: 1. say that the requiresRewrite is used to distinguish between these two different use cases drawbacks: excludes use cases that only use namespacing or resouce URLs and would normally be cachable 2. introduce a new attribute wsrp-markupContainsPortletURLs to distinguish between these two use cases drawback: needs a spec change If we decide to open the spec for fixing gR I would opt for 2. if not stick with a clarification to allow 1. Comments? Stefan |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]