RE: xforms:choose within xforms:repeat

Hi Eric,

We've actually given this a lot of thought in the group, 
and when you get right down to it, it is an error of oversight
that switch in repeat should not be defined because of some 
id problem.

The question arises, for example, how should setfocus behave
when the identified control is in a repeat.  Of course it
should set the focus to the identified item *in the currently
indexed row* of the repeat.

Switch should behave the same way, and we're working on defining
that for 1.1.  Moreover, the language about expecting implementer
feedback means that you can make your 1.0 implementation do this,
but just don't expect everyone to catch up on an aggressive
schedule (our implementation works, for example).

So, to be clearer, when you use *any* construct of XForms that
is based on id, it is best to obtain the identified element,
then drill up through its ancestral chain to find all repeats
in which it may be nested.

Then, for each ancestor repeat starting with the outermost one,
use the current index of that repeat to select a row.  Then,
if another repeat was in the ancestral chain, choose the one
in that row, use its index and keep going until you've 
resolved the row in the repeat containing the identified element.

Note that this should have an affect not just for toggling a 
case of a repeat, but setting or getting the index of an inner 
repeat, setting focus to a repeated control, etc.

Best regards,
John Boyer


-----Original Message-----
From: www-forms-request@w3.org [mailto:www-forms-request@w3.org]On
Behalf Of Erik Bruchez
Sent: Tuesday, August 16, 2005 3:36 PM
To: www-forms@w3.org
Subject: xforms:choose within xforms:repeat



All,

 From the 1.0 spec, section 9.3.10:

"A necessary consequence of this is that XForms 1.0 does not specify the 
behavior of construct switch within element repeat. Future versions of 
XForms may specify the behavior of switch inside repeat based on 
implementation experience and user feedback."

Any thoughts about what possible behaviors could be? I see two:

1. All the repeated switches show the same case element and switch at 
the same time.

2. Somehow, a particular switch only switches, maybe based on the source 
of the initial event causing the toggle action.

I don't think #1 is very satisfying, and I am not too sure about the 
consequences of #2. Any other thoughts on the subject? Should the toggle 
action take one or more repeat ids / indexes as parameters?

Actual users are likely to hit this snag. See for example this initial 
post on ops-users:

   http://mail-archive.objectweb.org/ops-users/2005-08/msg00069.html

-Erik

Received on Tuesday, 16 August 2005 23:59:34 UTC