Meeting minutes
The Control Element
Steven: What should happen if the bindings mismatch?
Erik: xforms-binding-error
Steven: Should they both get it?
Alain: Just the parent
Erik: Well, some processing has happened in the child at that point, so I'm not sure...
Steven: You will have initialised the instances in the child already
Alain: The error should be in the parent, and the child just dies
Erik: The default of an error is halting, with recovery if you cancel the default.
Steven: I think recovery is something you might want to do.
Erik: In all cases where processing halts, there is an error event first.
Steven: So both parent and child should get an event
Alain: There might be a case where the child offers to share, but the parent doesn't supply anything.
Steven: I'm not convinced
Steven: Moving to the <shared/> element
… do we agree that the @initial is not an AVT?
Erik: It is part of initialising
Alain: I don't see a use case...
Steven: Do we include attributes in the sharing?
Alain: Yes
Erik: Yes
Alain: Namespaces?
Erik: We have specified how we handle namespaces. There are a couple of ways to do it. You could import all inscope namespaces, or only the ones used.
Alain: In the embedded form, the namespaces should be declared
Erik: The document could be entirely new...
… with thousands of new namespaces
Alain: There could be conflicts
Erik: It should be fine
Steven: Do we think this needs more thought?
Erik: I need to see an example that might be a problem
Steven: Why is this different from submission replace= for a sub part of an instance?
Erik: Seems similar
Alain: The author has to watch out for namespace issues
Steven: On to <signal/>
… should we make it exactly like <dispatch/>
Alain: Make it exactly like dispatch.
Erik: Consistency is good
Steven: When signalling the parent, is it required that there be a <shared/> element?
… and if so what should the error be?
Erik: Can we embed a form without a <shared/> element?
Steven: Yes.
Erik: I think we should require a shared element.
Steven: On the other hand, if it's embedded, even without a shared element, why not?
Erik: Not a really important issue.
Steven: I suggest we allow it.
Alain: OK
Steven: I'll produce a new version
XPath 4.0 functions for node construction
https://
Steven: Anything we need to do?
Erik: Depends on whether the issue is still open.
Alain: We could use the same names...
Erik: Having functions to do this is good. I like it
… but we have them anyway
… element() and attribute()
… we could look to see if we need more
Alain: We should avoid name conflicts
Erik: I see some new comments on the issue
… consistency is not quite there.
… We can chime in with opinions
Steven: No actions then, just keep our eyes open
ACTION: Steven produce new <control/> spec
<trackbot> Created ACTION-2343 - Produce new <control/> spec [on Steven Pemberton - due 2023-09-15].
Close action-2330
<trackbot> Closed action-2330.
close action-2314
<trackbot> Closed action-2314.
AOB
Steven: I'm teaching XForms at the XML Summer School next week.
Steven: No call for two weeks.
[ADJOURN]