@qt4cg statuses

This page displays recent status updates about the QT4CG project.

The are also captured in an RSS feed.

By year: 2026, 2025, 2024, 2023, 2022, 2021, 2020

QT4 CG meeting 151 draft minutes #minutes-02-03

03 Feb at 17:45:00 GMT

Draft minutes published.

Issue #2407 closed #closed-2407

03 Feb at 17:36:27 GMT

`fn:type-of`: function vs fn

Issue #2409 closed #closed-2409

03 Feb at 17:36:26 GMT

2407 Change function to fn in type-of output

Issue #2398 closed #closed-2398

03 Feb at 17:34:24 GMT

fn:highest documentation in F&O spec not up to date?

Issue #2410 closed #closed-2410

03 Feb at 17:34:23 GMT

2398 Fix fn:highest to match fn:lowest

Issue #2406 closed #closed-2406

03 Feb at 17:32:17 GMT

Rounding dates/times and durations

Issue #2416 closed #closed-2416

03 Feb at 17:32:16 GMT

2406 Add fn:parts-of-dateTime and fn:build-dateTime functions

Issue #2365 closed #closed-2365

03 Feb at 17:30:14 GMT

Record types: extensible and non-extensible pairs

Issue #1484 closed #closed-1484

03 Feb at 17:30:13 GMT

Functions that expect a record type should make it extensible

Issue #2413 closed #closed-2413

03 Feb at 17:30:12 GMT

2365 Drop extensible record types

Issue #2428 closed #closed-2428

03 Feb at 17:28:08 GMT

2422 Drop XSLT section on embedded stylesheet modules

Issue #2421 closed #closed-2421

03 Feb at 17:26:06 GMT

XSLT edge case incompatibility with simplified stylesheet

Issue #2423 closed #closed-2423

03 Feb at 17:26:05 GMT

2421 document XSLT incompatibility with simplified stylesheets

Issue #2292 closed #closed-2292

03 Feb at 17:23:48 GMT

The XSLT document() function

Issue #2419 closed #closed-2419

03 Feb at 17:23:47 GMT

2292 XSLT document() function: options parameter

Issue #2397 closed #closed-2397

03 Feb at 17:12:27 GMT

Additions for "Functions Defined in XSLT" section in F&O spec

Issue #2411 closed #closed-2411

03 Feb at 17:12:26 GMT

2397 add to F&O list of functions defined in XSLT

Issue #2396 closed #closed-2396

03 Feb at 17:09:58 GMT

Missing "New in 4.0" labels for functions in F&O Spec

Issue #2395 closed #closed-2395

03 Feb at 17:09:58 GMT

The new fn:regex-groups function is not labelled "New in 4.0"

Issue #2412 closed #closed-2412

03 Feb at 17:09:56 GMT

2395 2396 Add missing "new in 4.0" entries

Issue #2438 closed #closed-2438

03 Feb at 17:09:16 GMT

Michaelhkay 2403 enhancements to fos xsd

Pull request #2438 created #created-2438

03 Feb at 17:08:48 GMT
Michaelhkay 2403 enhancements to fos xsd

This is MK's PR with the compiled form of the schema added.

The CG agreed to merge this PR at meeting 151.

Close #2404 Close #2403 Close #2177

Issue #2426 closed #closed-2426

03 Feb at 17:07:47 GMT

2408 editorial omnibus

Issue #2429 closed #closed-2429

03 Feb at 17:04:31 GMT

Feature/2026 01 28 draft review

Issue #1962 closed #closed-1962

03 Feb at 17:02:49 GMT

fn:map-to-element

Issue #2053 closed #closed-2053

03 Feb at 17:02:32 GMT

Add fn:collection-available

Issue #2430 closed #closed-2430

03 Feb at 17:00:27 GMT

Updates to schema for xslt

Issue #2437 created #created-2437

03 Feb at 13:15:31 GMT
SimpleNodeTest: TypeTest → RegularItemType?

With the current grammar…

AxisStep         ::=  (AbbreviatedStep | FullStep) Predicate*
AbbreviatedStep  ::=  ".." | ("@" NodeTest) | SimpleNodeTest
FullStep         ::=  Axis NodeTest
NodeTest         ::=  UnionNodeTest | SimpleNodeTest
SimpleNodeTest   ::=  TypeTest | Selector
TypeTest         ::=  NodeKindTest | JNodeType

…type tests in axis steps are limited to node() and its subtypes as well as jnode(). The test cases include tests for additional types like gnode() or array(*).

Maybe TypeTest should be replaced by the RegularItemType:

RegularItemType  ::=  AnyItemTest | NodeKindTest | GNodeType | JNodeType | MapType | ArrayType | RecordType | EnumerationType

Issue #2436 created #created-2436

03 Feb at 12:04:25 GMT
`jnode` type: arguments (spec vs. tests)

The current spec defines the following grammar for the jnode type:

JNodeType  ::=  "jnode" "(" (("*" | NCName | Constant) ("," ("*" | SequenceType))?)? ")"
Constant   ::=   StringLiteral | ("-"? NumericLiteral) | QNameLiteral | ("true" "(" ")") | ("false" "(" ")")

As far as I can judge, none of the current test cases seems to use this syntax. Instead, the tests I found expect a single sequence type argument, for example fn-jtree-006:

<test-case name="fn-jtree-006">
  <description> JNode applied to an array - type of result</description>
  <created by="Michael Kay" on="2025-06-16"/>
  <test>fn:jtree([1,2,3]) instance of jnode(array(xs:integer))</test>
  <result>
    <assert-true/>
  </result>
</test-case>

Is it the test suite or the spec that needs to be updated?

If this is still subject to discussion, my preference would be to disallow constants in the jnode syntax:

  • It is not clear to me how fn:jtree({ 'a': 1, 'b': 2 }) instance of jnode(a)` can be interpreted.
  • Instance checks for linear hierarchies are generelly more intuitive.
  • With the presence of get(), the jnode constants should be redundant.
  • We can only use atomic items in the node tests for which literals exists.

Issue #2435 created #created-2435

03 Feb at 11:14:19 GMT
Incorrect namespace prefixes in EXPath Binary example

In EXPath Binary Module 4.0 some example functions have been redefined in a different namespace prefix (which was bin:) to avoid suggesting they were part of the supported library.

However in 2.2 Example – reading and writing variable length ASN.1 integers, the definitions of asn:int-octets() and asn:encode-ASN-integer() still contains references to the original prefix definition bin:int-octets() which should be asn:int-octets()

See 5002 more statuses in yearly archives.