@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 189 draft minutes #minutes-11-17
Draft minutes published.
Issue #3068 closed #closed-3068
`item` axis: document order
Issue #3069 closed #closed-3069
3068 `item` axis: document order
Issue #3066 closed #closed-3066
Node constructors: JNodes
Issue #3067 closed #closed-3067
3066 Node constructors: JNodes
Issue #3050 closed #closed-3050
Static inferencing with `xs:error`
Issue #3054 closed #closed-3054
3050 Clarifications regarding errors and type inferences
Issue #3049 closed #closed-3049
Make $err:map into a record
Issue #3061 closed #closed-3061
3049 err:map becomes a labeled record
Issue #3059 closed #closed-3059
`fn:codepoints-to-string`, `fn:char`: use same error code
Issue #3060 closed #closed-3060
3059 fn:char: FOCH0001 instead of FOCH0005
Issue #2997 closed #closed-2997
JNodes: fewer jvalue() calls?
Issue #3029 closed #closed-3029
Method invocation syntax: might `$rectangle/size()` work?
Issue #3079 closed #closed-3079
Proposal to classify `%fixed` as a non-delimiting terminal symbol
Issue #3080 closed #closed-3080
3079 Split %fixed into two terminal symbols
Pull request #3096 created #created-3096
3066B [XSLT] Replace JNodes by their jvalue when constructing elements
Fixes the XSLT part of issue #3066.
Issue #3095 created #created-3095
`MethodCall`: placeholder restriction is stated in prose but reported as `XPST0003`
The grammar for MethodCall allows argument placeholders, because it reuses PositionalArgumentList:
MethodCall ::= PostfixExpr "=?>" NCName PositionalArgumentList
PositionalArgumentList ::= "(" PositionalArguments? ")"
PositionalArguments ::= (Argument ++ ",")
Argument ::= ExprSingle | ArgumentPlaceholder
The prose then forbids them:
The argument list in a method call must not include an argument placeholder; that is, the call must not be a partial function application [err:XPST0003].
But $x =?> f(?) is a valid instance of the grammar, and XPST0003 is only defined for grammar violations. This is the same kind of problem as in #3038. The test case method-920 covers this case.
Proposal: exclude placeholders in the grammar and delete the sentence:
MethodCall ::= PostfixExpr "=?>" NCName MethodArgumentList
MethodArgumentList ::= "(" (ExprSingle ** ",") ")"
Alternatively, XPST0017 could be extended to cover this case, or even a new error code could be introduced.
Issue #3078 closed #closed-3078
3077 Function assertions in reserved namespaces
Issue #3094 created #created-3094
Function assertions: match annotations
Function assertions and annotations look identical but are unrelated. The following query reads like “is f private?”, but the assertion is either rejected or ignored:
declare %private function f() {};
f#0 instance of %private function(*)
Proposal: A function assertion %N matches a function item if the item has an annotation N (with the same values, if arguments are given). Assertions defined in a specification, such as %updating` in the XQuery Update Facility, keep their own semantics. XQST0045 is no longer raised for function assertions.
Pull request #3093 created #created-3093
3092 Optional types for options with default ()
Closes #3092
Issue #3092 created #created-3092
Options with default `()` should have optional type
In most functions, an option with default () has an optional type (xs:string?, xs:QName*, …): match in fn:collection, encoding in fn:unparsed-text, etc. Examples:
fn:collation:versionandstrengthdefault to(), but their types (xs:string,enum(…)) don’t allow it.fn:collection:formisxs:string, whereasfn:uri-collectiondeclares the same option asxs:string?.fn:parse-html:encodingisxs:stringwithout default,fn:unparsed-textandfn:unparsed-text-linesusexs:string?.
Pull request #3091 created #created-3091
3082 function-name() of partial applications
Closes #3082
Pull request #3090 created #created-3090
2846 file:write, file:append: defaults, notes
I started off with the intent to make the file write functions more liberal, and I ended up making them adhere closer to fn:serialize… I believe that we should define fn:serialize more rigorously, instead of introducing too many rules that target file output, so that in most cases…
$data -> file:write($file, ., $options)
$data -> serialize(., $options) -> file:write-text($file, .)
…will remain equivalent (except for byte-level issues, such as encoding and BOM treatment).
Closes #2846
Pull request #3089 created #created-3089
3045 fn:serialize: defaults for omitted argument
I have tried to resolve ambiguities in the current fn:serialize rules (and I hope that the rules are still 3.1-compliant).
Closes #3045
Pull request #3088 created #created-3088
3085 Editorial rewrite of function equivalence
Fix #3085
Pull request #3087 created #created-3087
3051 fn:type-of: records
As proposed, I have droped the record() wrapper in the fn:type-of output.
- I believe that the output will be self-explanatory enough.
- It will improve consistency, unless we introduce wrappers for other types, too.
- The inclusion of the namespace URIs should prevent users from misinterpreting the output.
Closes #3051
Pull request #3086 created #created-3086
3052 `fn:update`: refinement of actions
- editorial revisions
- actions renamed (
replace-value,insert-before,insert-after) - explicit
deleteaction
Closes #3052 Tests (updated & reformatted): qt4cg/qt4tests#516
Issue #3085 created #created-3085
Function equivalence revisited
We have made a brave attempt at defining function equivalence, but in the light of the test cases, I don't think the current definition always provides a clear answer. I propose to make another attempt at improving it.
QT4 CG meeting 189 draft agenda #agenda-11-17
Draft agenda published.
Issue #3084 created #created-3084
Forwards compatibility of options parameters
I'm having to modify some load-xquery-module tests to specify trusted=true, because the dynamically-loaded module in turn does static imports of further modules. Fortunately the queries still run under 3.1, because in 3.1, processors are required to ignore unknown options.
This makes me question whether it was really wise to make 4.0 processors reject unknown options: it may well be that the option is given a meaning in a future version of the specification.
One way around this would be the XSLT mechanism of forwards compatibility mode: a version N processor executing a query labelled with version N+x should accept and ignore unknown options.
Alternatively, revert this change; make the presence of an unknown option a warning condition rather than an error.
See 6413 more statuses in yearly archives.