W3C home > Mailing lists > Public > www-forms@w3.org > January 2002

Re: Order of xforms:revalidate and xforms:recalculate

From: Alistair Coles <anc@hplb.hpl.hp.com>
Date: Wed, 02 Jan 2002 17:25:51 +0000
Message-ID: <3C33429F.3FC12BAF@hplb.hpl.hp.com>
To: www-forms@w3.org

> From: Jérôme Nègr <jerome.negre@e-xmlmedia.fr>
> To: <tvraman@almaden.ibm.com>
> Cc: <www-forms@w3.org>
> Date: Thu, 20 Dec 2001 09:41:29 +0100
> Subject: Re: Order of xforms:revalidate and xforms:recalculate
> To sum up, the default processing for chapter 11.3.6 should be :
> 1. If the value from the form control meets all validity constraints:
>     a- It is reflected in the instance data.
>     b- Event recalculate has been dispatched to element model.
>     c- Event revalidate has been dispatched to element model.
> If not,
>     a- The instance data remains as it was before processing this event.
>     b- The form control is marked as invalid.
> 2. Event refresh has been dispatched to element model.
> Is this correct ?

May I add a follow-up question: how would this processing (if
confirmed as correct) differ from that for

Perhaps there is no difference intended, but if that is the case I am
wondering how I could listen for changes to an instance data element
(as opposed to the form control) and distinguish between instance data
changes that were the result of a valueChanged vs. a valueChanging


> ----- Original Message -----
> From: "T. V. Raman" <tvraman@almaden.ibm.com>
> To: "Jérôme Nègre" <jerome.negre@e-xmlmedia.fr>
> Cc: <www-forms@w3.org>
> Sent: Wednesday, December 19, 2001 11:35 PM
> Subject: Order of xforms:revalidate and xforms:recalculate
> The thought underlying the sequence is:
> You type a value into a field --
> first you need to check if that is valid.
> Once you've checked it is valid, you run the recalculate to
> update fields that may depend on the one you filled in.
> That in turn may change fields --hence the revalidate after
> calculate.
> Thus, depending on how the constraints are authored, it
> might be possible to type in a value that is "locally valid"
> but turns out to be invalid with respect to the form.
>   -- 
Alistair Coles                          anc@hplb.hpl.hp.com
Hewlett-Packard Laboratories, 
Filton Road, Bristol, BS34 8QZ, UK
Tel:+44 117 312 8750  Telnet:312 8750  Fax:+44 117 312 9286
Received on Wednesday, 2 January 2002 12:17:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:50 GMT