Swift: add new TypeValueExpr to CFG

[redsun82]

We do see errors about missing types from type reprs, which is not ideal. However I propose to postpone looking at that when type value expressions get out from being experimental.

Also, changed qltest.sh to be more liberal about the location of special option comments in the test source code.

[Copilot]

Pull Request Overview

This PR extends CFG support for the new experimental TypeValueExpr, adds ValueGenerics tests in both AST and CFG, and relaxes how extractor options/env markers are picked up in qltest.sh.

• Relax sed usage in qltest.sh to capture extractor options and env from anywhere in the source file
• Update control‐flow and AST tests (and their expected outputs) to account for TypeValueExpr, next(isolation:), and ValueGenerics support
• Introduce a TypeValueTree class stub in ControlFlowGraphImpl.qll for the new expression kind

Reviewed Changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
swift/tools/qltest.sh	Allow //codeql-extractor-*- markers on any line (not just line 1)
swift/ql/test/library-tests/controlflow/graph/cfg.swift	Reorder member‐ref tests and add ValueGenerics snippet
swift/ql/test/library-tests/controlflow/graph/Cfg.expected	Update expected CFG output for TypeValueExpr, next(isolation:), and ValueGenerics
swift/ql/test/library-tests/ast/cfg.swift	Mirror CFG changes in the AST test and add ValueGenerics snippet
swift/ql/test/library-tests/ast/PrintAst.expected	Adjust variable indices and add missing‐type entries for ValueGenerics
swift/ql/test/library-tests/ast/Missing.expected	Add missing-type entries for TypeValueExpr in ValueGenerics
swift/ql/test/library-tests/ast/Errors.expected	Add corresponding UnspecifiedElement errors
swift/ql/lib/codeql/swift/controlflow/internal/ControlFlowGraphImpl.qll	Add TypeValueTree stub to handle TypeValueExpr in the CFG builder

Comments suppressed due to low confidence (1)

swift/ql/lib/codeql/swift/controlflow/internal/ControlFlowGraphImpl.qll:1887

• [nitpick] The new TypeValueTree stub is declared but has no integration logic or registered use. Ensure it’s hooked into the tree‐builder dispatch so TypeValueExpr nodes actually produce this tree in the CFG.

private class TypeValueTree extends AstLeafTree {

[jketema]

LGTM. I assume we generally want to do this for every new expression and/or statement that is introduced?

[redsun82]

LGTM. I assume we generally want to do this for every new expression and/or statement that is introduced?

yes, I entirely forgot to mention it on your PR, and it only occurred again to me when rereading the ticket about it that was mentioning CFG 😅