jsr344-experts@javaserverfaces-spec-public.java.net

[jsr344-experts] Re: [jsr344-experts mirror] Re: PPS Discussion

From: Andy Schwartz <andy.schwartz_at_oracle.com>
Date: Thu, 31 May 2012 09:34:16 -0400

On 5/31/12 8:35 AM, Kito Mann wrote:
> +1 on this. It was actually a client requirement that drove its
> implementation in PrimeFaces. The issue is currently targeted for 2.2
> (http://java.net/jira/browse/JAVASERVERFACES-1841),

The corresponding spec issue is here:

http://java.net/jira/browse/JAVASERVERFACES_SPEC_PUBLIC-723

Note that Ted raises an additional requirement:

> tedgoddard added a comment - 21/Apr/11 10:12 AM
> This is a very good feature provided that there is also a "whitelist"
> of additional parameters that must be submitted with every Ajax post.
> Alternatively, a JavaScript callback could be registered to allow
> frameworks to add required parameters to the POST before submission.

Do either MyFaces or PrimeFaces address this?

Minor API comment: I would prefer that we go with an enum-based value
rather than a boolean - ie. I would prefer something like:

- submit = partial/full

Over:

- pps = true/false

Andy

> so we just need to make sure it's completed for 2.2
>
> On Thu, May 31, 2012 at 5:48 AM, Çağatay Çivici
> <cagatay.civici_at_gmail.com <mailto:cagatay.civici_at_gmail.com>> wrote:
>
> Hi,
>
> The general feedback I received from community after adding
> partialsubmit to PrimeFaces is that "This should be a core JSF
> feature" so I totally agree with Werner.
>
> Here is a working example of what Werner described;
>
> http://www.primefaces.org/showcase-labs/ui/partialSubmit.jsf
>
> Regards,
>
> On 31.May.2012, at 10:05, Werner Punz wrote:
>
>> Hello Everyone
>>
>> Just a minor question because Primefaces added that feature and
>> MyFaces had it as a config option for ages.
>> But I personally think this is a feature which should be part of
>> the spec not part of implementation features or component
>> set features.
>>
>> PPS Ajax submit
>>
>> As for now we encode the entire form for the execute parts and
>> send it down.
>> Wouldn´t just from a minor performance and traffic perspective it
>> being better just to send the components
>> which actually are executed.
>>
>> aka
>>
>> <h:form>
>> <h:inputText id="component1" .../>
>> <h:inputText id="component2" .../>
>> <h:inputText id="component3" .../>
>>
>> ....
>> <h:commandButton ...>
>> <f:ajax execute="component1" render="..."
>> </h:commandButton>
>>
>> this currently encodes component1, component2 and component 3 +
>> the viewstate and whatever is in the form
>>
>> While in the pps case we only would encode component1 and the
>> viewstate
>>
>> Now this should not be the default behavior because it would
>> probably break the backwards compatibility in some corner cases
>> but how about <f:ajax ... pps="true" /> which enables this.
>> I think if you have a bigger number of input components then this
>> could give some performance improvement.
>> The good side also is that most of the heavy lifting would be on
>> the Javascript side of things.
>>
>>
>>
>> Werner
>>
>>
>
> *Çağatay Çivici*
> Principal Consultant
> PrimeFaces Lead | JSF EG Member
>
> Prime Teknoloji
> Bilkent Cyberpark, A-303d
> 06800 Ankara/Turkey
> Tel: +90 312 265 05 07 <tel:%2B90%20312%20265%2005%2007>
> http://www.prime.com.tr
>
>