See also: IRC log
Norm proposes to add C14N to the agenda
Accepted as amended.
Norm: Thank you, Henry, for
... And scribing!
None regrets heard.
Norm: Anyone looked at it?
Henry: Not I, sorry.
Norm: I put this on because there
was a thread, but I think it all resolved itself.
... The p:filter step needs to throw an error if you get non-nodes.
Vojtech: Yes, and I wanted to reuse an XD error but we should use a component error.
Vojtech: Can the p:error step throw err:XD* errors?
Norm: I think it should be allowed to.
Vojtech: I think it makes it possible for a step to re-throw errors, which could be usefl.
Norm: Anyone opposed?
Proposal: there are no constraints on what errors p:error can throw.
Rui: What about static errors?
Norm: The fact that you'd have to throw an error if someone attempted to throw an error that they're not allowed to throw is sufficiently bizarre that I think we shouldn't bother.
Norm attempt to summarize the minutes
Norm: For V1.0, let's leave it up to implementors to use an implementation-defined method.
Alex: But if someone wanted to do it, they'd need to use extension attributes
Some discussion about whether or not extension attributes are allowed on p:serialization. They are.
Alex: I'm ok with this story.
Proposal: Do nothing for this in V1.0
Norm observes that Mohamed isn't here, but hopefull that'll be ok.
Alex: There were some questions about http-request.
Vojtech: If @detailed is false, is the c:body element generated or not?
Alex: If it's not specified, you get the raw content for XML media types, otherwise you get a c:body with data inside.
Vojtech: It seems clear to me now.
Alex: You can put multipart
inside multipart, but I don't think we should try to deal with
... I think we should only do the first level.
Vojtech: But there's no way you can dive deeper into the multipart-mixed
Alex: There's a hole there, but I don't think I want to get http-request to dig deeper
Norm: I'm comfortable with the answer.
Vojtech: I was just wondering
Proposal: We're only going to support one-level of de-multiparting.
<scribe> ACTION: Alex to add a note to the spec to indicate that we'll only do one level. [recorded in http://www.w3.org/2008/06/19-xproc-minutes.html#action01]
Some discussion of content-type vs. encoding.
Vojtech: Then maybe what's missing is some discussion of what to do if the encoding is base64
Norm: Maybe an example that uses base64 encoding would help.
Vojtech: The test suite could cover this.
<scribe> ACTION: Norm+Alex to construct examples for the test suite [recorded in http://www.w3.org/2008/06/19-xproc-minutes.html#action02]
Some discussion of the note at the end of the "Converting Response Entity Bodies" section (7.1.9/10.4)
Alex: If you send the data with a non-Unicode character encoding, that's not our problem.
Proposal: Strike the note.