See also: IRC log
<trackbot> Date: 08 April 2015
Steven: I'm a bit cross that they didn't talk to us first... but I guess they are expecting us to work from out of the community group
\... until they recharter us next month.
scribe: my guess anyway, because
as I say, this was a bit out of the blue.
... I'll be contacting them, and keep you in contact
[No link yet]
[No link yet]
Steven: Anything to say yet?
Erik: I haven't looked at his stuff yet, but I am studying the issue at the moment. We haven't implemented it yet; that is in progress, to give us some experience so that we can give him some better feedback.
fails
https://lists.w3.org/Archives/Public/public-forms/2015Apr/0008
Erik: Alain points out some
places where it can go wrong.
... we could have some rules on recovering
... defaulting to string and so on
... if we do actually want error handling, that stops
serialization.
... Four ways we could do it
... 1. XPath dynamic error
... (in the spec now)
... similar to an exception
... 2. Return an empty sequence
... 3. return an error document
... 4. Dispatch an event
... similar to dynamic error
... We would like to specify the granularity of error
handling
... we don't have the mechanisms really
... next best is to follow XPath
... that will interrupt the action,
... which would generate an error event
... None of this is ideal.
Steven: What do you do Alain?
Alain: It is a lax function, but
I don't think that is the right approach
... it is an author problem.
... not a runtime problem.
Steven: I imagined that we would
just do a best-possible serialization.
... if you hacve <a type="number">abc</a> then just
serialise it as a: abc
Erik: Or ignore noes in different namespaces, let the author manipulate extra nodes if that is useful.
Steven: I agree with that.
... The spec says "this is how you convert JSON to XML, and
serialization is the reverse", we could be more specific
Erik: At least mention that extra nodes are ignored.
Steven: Would we be happy with a best-try version, which does what it can without complaining?
Erik: That would be OK
Alain: Yes, I'm fine, just makes debugging harder, that's all.
<scribe> ACTION: Steven to make serialization of JSON more explicit. [recorded in http://www.w3.org/2015/04/08-forms-minutes.html#action01]
<trackbot> Error finding 'Steven'. You can review and register nicknames at <https://www.w3.org/2005/06/tracker/xforms/users>.
Steven: Damn. That means the issue tracker is messed up...
Erik: "If serialization fails, a dynamic error is raised" in the spec.
Steven: I don't think that will happen.
http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#Date_and_Time_Functions
Steven: All date examples use dates in 2006. Leave as is, or make more modern?
Erik: I think 2015 or 16 would be good.
Alain: Agree
Steven: Section 4.5.3 http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#The_id.28.29_Function
"Note:
When an ID is assigned to a attribute by asigning any type derived from xsd:ID by an XML schema, or the type model item property that inform the ID type information is provided to the XPath data model. And consequently the is-id accessor will return true."
Steven: I don't understand
this.
... What is the is-id accessor?
<ebruchez> http://www.w3.org/TR/xpath-datamodel/#dm-is-id
Steven: Then we need to link to
that.
... So this is saying that a thing is an id thanks to its type
and not any other reason.
s/So/
Steven: 4.5.5 http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#The_bind.28.29_Function
" It is an error to specify a non existing bind ID."
Steven: What does that mean?
Erik: Same arguments as we just
had for serialize
... I think it would be reasonable to return an empty
sequence
... The bind attribute says that the binding becomes empty, and
the bind-error event is dispatched.
Steven: So we will return the empty sequence
Erik: Yes.
Steven: 4.6.3
http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#The_parse.28.29_Function
... Doesn't return an error, only the empty sequence. OK?
Erik: I was looking at the DOM's
parse function. You would expect an exception, but it doesn't
do that.
... they return a funny XML document.
... There are several possible approaches, this is one. At
least you know there is an error.
Steven: Leave as is?
Erik: Yes.
Alain: OK
Steven: 4.8.1 http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#The_location-uri.28.29_Function
"NOTE: If an implementation is not on the client, the fragment might not be returned."
Steven: What is being suggested here?
Erik: That is related to how we
implement things on the server.
... some implementations may not know what was after the #
Steven: 4.11
http://www.w3.org/MarkUp/Forms/wiki/XPath_Expressions_Module#Error_Indications
... It refers to exceptions, but that text should go
Steven: Anything?
[none]
[ADJOURN]
This is scribe.perl Revision: 1.140 of Date: 2014-11-06 18:16:30 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/3/3./ Succeeded: s/sould/would/ FAILED: s/So// Succeeded: s/SO/So/ No ScribeNick specified. Guessing ScribeNick: Steven Inferring Scribes: Steven Default Present: Steven_, [IPcaller], ebruchez Present: Steven_ [IPcaller] ebruchez Alain Regrets: Nick Philip Agenda: https://lists.w3.org/Archives/Public/public-forms/2015Apr/0010 Found Date: 08 Apr 2015 Guessing minutes URL: http://www.w3.org/2015/04/08-forms-minutes.html People with action items: steven[End of scribe.perl diagnostic output]