QT4 CG Meeting 149 Minutes 2026-01-20

Regrets: JWL, JK.

• [X] David J Birnbaum (DB)
• [X] Reece Dunn (RD)
• [X] Christian Grün (CG)
• [ ] Joel Kalvesmaki (JK)
• [X] Michael Kay (MK)
• [X] Juri Leino (JLO)
• [ ] John Lumley (JWL)
• [X] Alan Painter (AP
• [X] Wendell Piez (WP)
• [X] Norm Tovey-Walsh (NW) Scribe. Chair.

Proposal: Accept the agenda.

Accepted.

Proposal: Accept the minutes of the previous meeting.

Accepted.

The next meeting is planned for 27 January 2026.

JWL gives regrets for 27 January.

• [ ] QT4CG-143-02: MK to try to recover the ability to extract formal equivalences into tests
• [ ] QT4CG-143-03: JK to look for C14N test suites.
• [ ] QT4CG-144-01: MK to consider if any now lost value comparisons should be added as examples.
• [X] QT4CG-148-01: NW to draft a PR that marks #2351 items as “at risk” with appropriate definitions at the top of the document
• [ ] QT4CG-148-02: NW to publish a dated draft after QT4CG-148-01 is complete.
• [X] QT4CG-148-03: NW to revise his attempt at #2313 by using the text “node” and a link to the appropriate text

This section summarizes all of the issues and pull requests that need to be resolved before we can finish. See Technical Agenda below for the focus of this meeting.

See #2313.

(To briefly discuss CG’s comment.)

CG introduces his comment.

• CG: In my original PR, I tried to unify all the function calls, and on the other hand I fixed a few small bugs. I’d like to get rid of all the special rules. I’d like to have one general rule.
• CG: A general rule that says whenever the input is empty, the default is applied. Then we could avoid having to write all the detailed cases.
• CG: With this rule, we could possibly generate test cases for absent vs default.
• MK: For the text example where we say that the value is 0, empty, or absent. In the XSLT spec, we talk about the value after applying defaults.
  • … We could also have markup for noting that omitting the parameter is the same as setting it to empty. But I couldn’t find a good notation for it.
• CG: But if we have one specific rule, that might be enough.
• JLO: Is this about optional parameters?
• CG: It’s about the cases when an empty sequence is supplied but it should be replaced by the default value.
  • … The original purpose of the PR was to distinguish between an absent argument and an argument that’s present but is the empty sequence.
• JLO: That sounds great, but sometimes we want to say that you want to use the empty sequence and not the default value.
• RD: If I understand correctly, you’ve changed the type signature of the defaults so that the “?” is removed from the type signature and replaced it with a default.
  • … I’m kind of hesitant with that because the empty sequence means “zero”.
  • … Even if we do have this rule, specifying a zero-item object can be passed to the function.
• CG: That’s what NW’s (: Note :) markup is for.
• RD: But in the function signatures, the types have changed.
  • … I’m not happy with start as xs:integer := 0 if you can pass an empty sequence.
• CG: I think we can discard my proposal to change the signatures.
• NW: So we’ll close this PR and CG will make another?

ACTION QT4CG-149-01: CG to draft a new PR with the common rule note.

• NW will merge the markup changes for the note.

See #2387.

• MK: This was a suggestion from CG. Rather than failing when you have a number that’s Infinity or NaN we output some fallback representation.

MK reviews the PR and the fallback representations.

• RD: So those will be xs:decimal types?
• MK: No, they’re double or float because that’s how you get Infinity.
  • … They’re legitimate JSON numbers, but they’re out of range for IEEE.
• JLO: It’s good that we begin this up again. Last time we talked about out, there were several options. I’m not sure I like the null for NaN.
• MK: The only other alternative is a string. Anything is a bit arbitrary. This seems like the closest approximation to me.
• JLO: What happens if you put these into JSON tools, like jq?
• MK: I haven’t tested any.
• WP: This is just the way the thing gets written out.
• MK: This is the way we serialize the JSON if it contains an Infinity.

Proposal: Accept this PR.

Accepted.

See #2376.

MK introduces the PR.

• MK: The main problem with this is that it overloads an attribute that was originally designed to be a boolean and is becoming more complex.
  • … The @typed data started as a boolean; then it was extended to “strict”, “lax”, and “unspecified”.
  • … We’ve never really covered the case of what typed means if you are matching things that aren’t element nodes.
  • … The relevant section is now extended to cover them.

Proposal: Accept this PR.

Accepted.

See #2373.

• MK: This fixes a bug or inconsistency about whether or not a path starting with “/” triggers conversion to a JNode. I chose not.

Proposal: Accept this PR.

Accepted.

See #2372.

• MK: Issue raised by CG. The rules we have for PIs in HTML are meaningless with HTML 5 which doesn’t recognize processing instructions.
• RD: Does this only apply to the HTML serialization?
• MK: Yes.
• CG: I was a bit surprised to have the new comment; I’d have expected the PI to be omitted.
  • But that’s just one of many ways.
• RD: That depends if you’re using the HTML based parser or the XML based parser. (Some discussion of the HTML DOM and variations in parsers.)
• MK: Leaving it in is purely for diagnostics.
• RD: Or you could have an XHTML document that you load as XML and serialize as HTML5.
• JLO: Why is it a must be maintained as a comment.
• MK: Because on the whole we’re prescriptive unless there’s a good reason not to be.
• JLO: Would it warrant a serialization option to drop them.

(Oh goodness no.)

• WP: As an XSLT person, we can do it all in a pipeline.
  • … What are HTML5 parsers instructed to do with processing instructions?
• RD: The parser parses them into comments.

Some discussion of why we need to do this: because we ought to produce things that are consistent with the HTML5 grammar.

RD displays the relevant part of the WHAT WG specification. Discussion ensues.

ACTION: QT4CG-149-02: MK to update the PR #2372 so that the “?” are part of the comment.

See #2379.

Everyone is happy with NW or MK having to update the generated schema.