Re: [wsrp] Navigational parameter restrictions

[Michael Freedman <michael.freedman@oracle.com>]

Well, you have me smiling somewhat. I raised, we discussed, and we 
decided to take no action on helping producers to manage request scoped 
data within a wsrp (action) lifecycle. We decided that the producer can 
handle this without further support from us.  I believe this to be true 
though note that improvements could be made by formalizing the 
multi-window on a portlet instance case, and potentially an outOfScope 
event case.  Anyway,  NavigationalParameters are not (and should not be) 
designed to address this problem.  NavigationalParameters are public 
navigational state.  Navigational state is managed at a different 
lifecycle then a wsrp (action) lifecycle.
    -Mike-

Subbu Allamaraju wrote:

> I agree with the theme, but a problem in JSR286 is that portlets can't 
> pass on objects from action to render.
>
> One example is the spring integration issues you posted to the JSR286 
> EG (pardon me folks on this list). Similar needs come up with most 
> bridge portlets. While these issues are specific to JSR286, the issues 
> may be relevant to general purpose producers as well.
>
> I'm not favoring one way or the other, but there are certainly some 
> real-world use cases to consider.
>
> Subbu
>
> Stefan Hepper wrote:
>
>> I would be also in favor of this approach. It would make the 
>> alignment with JSR 286 simpler as in the JSR 286 we only support 
>> Strings as navigational parameters.
>>
>> Stefan
>>
>> Michael Freedman wrote:
>>
>>> Should we restrict the returned navgiationalParameters in 
>>> UpdateResponse to only be values representing xml scalar types [int, 
>>> string, etc.]?  The issue here is similar to the one that caused us 
>>> to restrict what the producer can encode in the URLs it writes into 
>>> markup, namely many consumers will prefer to manage these values in 
>>> the same way they manage navigational state.  As this is commonly 
>>> handled by writing the state into every URL in the page, shouldn't 
>>> we make it easy for a consumer to do so?  If so, should we back 
>>> track all together and only support scalar types?  Or do we preserve 
>>> the full open model for those properties that aren't set by 
>>> producers and might presumably either be computed by the consumer 
>>> vs. being encoded?
>>>    -Mike-
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe from this mail list, you must leave the OASIS TC that
>>> generates this mail.  You may a link to this group and all your TCs 
>>> in OASIS
>>> at:
>>> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>>>
>>>
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe from this mail list, you must leave the OASIS TC that
>> generates this mail.  You may a link to this group and all your TCs 
>> in OASIS
>> at:
>> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  You may a link to this group and all your TCs in 
> OASIS
> at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php