Meeting minutes
ACTION-2309: Research xpath3 function definitions (Erik)
https://
Erik: Could we use XQuery function definition? (XPath doesn't have a syntact for it)
… In XForms we can declare a function with a signature using XPath/XQuery syntax
… and the body inside the element
… with an extension because XPath2 doesn't have variables
… XPath 3 has <let>
… so we define variables local to the function.
… This compatible with XPath 1 and 2. Unneeded for XPath 3
… because XPath 3 supports it natively
… If we use the XQuery syntac, we would still have to have an enclosing element
… So I think it might not give us much
… but since we want to support other expression languages, I don't think it gives us much.
… We would lose the ability to support functions properly in earlier versions of XPath
Alain: I don't have a strong opinion either.
… I'd like more time to think about it.
Steven: We'll leave it at that now, and come back in the near future.
ACTION-2312: Devise text for submission using @value (Steven)
https://
Steven: No progress this week.
Detecting/Ignoring unsaved data
https://
Alain: I was wondering how to do this.
… maybe if the data has been submitted, maybe it's OK.
Alain: We have "Model-destruct" but we can't cancel it.
Erik: We do this at a higher level that we have
… we track the data status. Use events on the main instance.
… we also have auto-save
… browsers have b/f caches
… you register a handler when the data is dirty
Steven: We either need something better than model-destruct, or redefine model-destruct
Erik: We get one xforms-ready per model
… in JS that is before-unload
… which is not that useful.
Steven: If we defined xforms-done, could we do anything?
Alain: Existing webpages just cancel the exit action
Erik: We do this, I need to look it up
Steven: If the processor was listening for the exit event, and then dispatches an XForms-done event' if that is cancelled then the processor doesn't do the exit. Wouldn;t that be enough?
… The XForms can deal with messages.
[Erik looks at his code]
Erik: The code is simple.
Erik: No prevent default; you register and deregister the handler.
… I'm not convinced using XForms events would work well
Steven: According to the spec it looks like we can use AVTs in event attributes.
… so that would work.
Erik: I'm not sure if it would be possible to support though.
Steven: In brief -- needs more thought.
Erik: We could define actions that reports to the processor that it shouldn't automatically exit
Alain: We don't have an isdirty function.
Steven: It's quite simple to do yourself. I have a tutorial somewhere
Alain: It's less declarative
Erik: You could add a function that tells you whether a ref is dirty or not.
Steven: You would also need a function to clear the dirty bit.
ACTION: Steven research the alternatives for 'dirty' data, and produce code in all of them for comparison
<trackbot> Created ACTION-2313 - Research the alternatives for 'dirty' data, and produce code in all of them for comparison [on Steven Pemberton - due 2022-01-28].
Timescale Plans
https://
Steven: Anything you would particularly like to comment on?
Erik: Need more time.
Steven: Next week on the agenda specifically <control/> and mirroring.
AOB
[None]
[ADJOURN]