Why can't some sort of architectural "checksum" be calculated on
the nodes of the template object graph as the template is first
built in the appendToResponse, stored in the WOContext for that
page and compared against any subsequent rebuilding of the
template, throwing an exception with a clear message when the two
don't compare? The answer to this question is left as an exercise
for the student, I'm up to my ears right now preparing for a roll-out.
While not fully addressing this issue, Wonder actually does have a
cool feature where it computes hash codes for items in a WORepetition
and modifies the element id to include this hash, which allows
structural (i.e. positional) changes to the elements of the
repetition while still allowing actions and bindings to direct to the
correct item. For instance, if you are editing the first item in a
repetition, and before you submit back a new "first item" gets
inserted into the array, when you submit your form, it will properly
update the correct EO rather than getting confused and trying to
update (symbolically) "the first item" (which has now changed). It
does seem like you could somehow come up with some equivalent
trickery in computing a WOConditional's contribution to the element
id ... There are probably some oddities in binding management to
contend with, but I'm sure it could be done.
ms
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com
This email sent to [EMAIL PROTECTED]
