See also: IRC log
<trackbot> Date: 25 September 2013
<nvdbleek> I'm calling in a minute
<unl> zakim doesn't like me...
http://www.balisage.net/Proceedings/vol10/html/Pemberton01/BalisageVol10-Pemberton01.html
sniffing
http://lists.w3.org/Archives/Public/public-xformsusers/2013Sep/0000.html
sniffing
http://lists.w3.org/Archives/Public/public-xformsusers/2013Sep/0000.html
Steven: Maybe replace the exception with an error?
Erik: I had an action item to
report on our error handling
... for instance an xpath expression error just halts the
processor
... we don't dispatch binding-exception nor
compute-exception
<ebruchez> http://www.w3.org/2012/02/15-forms-minutes.html#action02
<ebruchez> <trackbot> Created ACTION-1868 - Bruchez to summarize problems with error handling and three options for variable type handling.
action-1868?
<trackbot> action-1868 -- Erik Bruchez to Bruchez to summarize problems with error handling and three options for variable type handling. -- due 2012-02-22 -- OPEN
<trackbot> https://www.w3.org/2005/06/tracker/xforms/actions/1868
Erik: There are two areas where
you might want to catch errors: xpath exprs and actions
... having an answer to that might help this problem
... we might also have a way to test for a function
Steven: If you had a way of testing, how might you alter the form's behaviour? What's the use case?
<nvdbleek> http://www.w3.org/TR/xslt-30/#func-function-available
Erik: His use case is for testing implementations
<nvdbleek> This is the XSLT 3.0 function to test if a function is available
Erik: We could just import that function.
<ebruchez> http://www.w3.org/TR/xpath-functions-30/#func-function-lookup
Erik: That example uses the fact that XPath 3 has functions as returnable values
<nvdbleek> erik: This is not possible in XPath 2.0
Erik: function-available is in XSLT, not XPath
<nvdbleek> http://www.w3.org/TR/xslt20/#function-function-available
Erik: so it is in XSLT 2
... So we could import the xslt 2 function with the same
signature and behavior
Steven: Do we think it's worth doing?
Erik: It is valuable.
... but maybe we should check with Eric. It looks like it could
help.
Steven: Or make the exception an error.
<nvdbleek> http://www.w3.org/TR/xslt-30/#element-try
Erik: That would be rather heavy
to use.
... In our implementation we change the error events a bit
comapred with the spec.
... I would be happy to send a suggestion for changes.
Steven: If we made bad decisions in the past with error events, we should fix them.
Erik: Eric agreed that we should try to fix the problem areas in a chat I had with him
Steven: What are we going to do?
Erik: 1) reply to him and mention the choices 2) make errors recoverable 3) import the XSLT function
Nick: the error handling is going
to be a challenge, and a lot of work
... for actions it is easy, but not for bindings.
Steven: it doesn't sound like we can fix error handling before last call.
Erik: I can send a message on how we do it, and what impact it would have
Steven: Please do, that is action-1868
Erik: I can answer Eric's message
<scribe> ACTION: Erik to answer Eric vd Vlist's message on identifying available functions [recorded in http://www.w3.org/2013/09/25-forms-minutes.html#action01]
<trackbot> Created ACTION-1956 - Answer eric vd vlist's message on identifying available functions [on Erik Bruchez - due 2013-10-02].
code markup
http://lists.w3.org/Archives/Public/public-forms/2013Sep/0003.html
Steven: Looks OK.
http://lists.w3.org/Archives/Public/public-forms/2013Sep/0004.html
<ebruchez> http://www.w3.org/MarkUp/Forms/wiki/XForms_2.0#Typed_Values
Erik: Although the expression
language is a separate module, there is a core xpath element in
the language
... it leaves space for other expression languages, but it is
hard to know what another language might require, so hard to
say anything concrete about it.
... Expressions needs access to types.
... We have three ways to add type information (bind, xsi: and
schema)
... but some off-the-shelf schema validators don't annotate
values in some cases
... This means that we need text that says "may" for those
cases.
... so it is unclear if we *can* mandate the types being
available in all cases
... but this text covers built-in types that are always
available.
... I listed all types that *must* be provided
... an XPath processor doesn't know about the xforms variants
of the xsd basic types
... so the data model must provide the xsd type to the xpath
processor
... And the last part, which is what happens if the expression
doesn't match the type, an error case
... a run-time error
... that can be recovered from
... so this defines what happens there
... the result is an empty string
... and then we consider that constraints fail.
... or it is as if the expression hasn't run.
Steven: Does an event get proagated?
Erik: There is no event for that, the only one that we have stops processing.
Steven: So we do need one, so
that you can catch the error
... So how to take this further?
Erik: I already have an action item for the error handling.
Steven: Can I ask you Nick to summarise for next week where we are in respect of getting to LC?
Nick: OK.
s;s/Erik:/Erik:/G;;
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/bind-/compute-/ Succeeded: s/vir/cior/ Succeeded: s/cior/vior/ Succeeded: s/Erik/Eric/ Succeeded: s/Erik:/Eric:/G Succeeded: s/ do it / fix error handling / Succeeded: s/Eric:/Eric:/G Succeeded: s/ric:/rik:/G No ScribeNick specified. Guessing ScribeNick: Steven Inferring Scribes: Steven Default Present: ebruchez, Steven, [IPcaller], nvdbleek, unl Present: ebruchez Steven [IPcaller] nvdbleek unl Agenda: http://lists.w3.org/Archives/Public/public-forms/2013Sep/0005 Found Date: 25 Sep 2013 Guessing minutes URL: http://www.w3.org/2013/09/25-forms-minutes.html People with action items: erik[End of scribe.perl diagnostic output]