Python: Handle guards being compared to boolean literals

[yoff]

Inspired by https://github.com/github/codeql/pull/21288/changes#diff-502cb70dbb74ee46cda83c2b3c626120ae120bcd444e146be09bc46599cf981eR194-R215

While we wait for adoption of the shared guards library (which will handle much more than this), this PR is an easy addition to the guards handling in Python.

When a guard such as isSafe(x) is defined, we now also automatically handle isSafe(x) == true and isSafe(x) != false (as well as is and is not).

[Copilot]

Pull request overview

This PR extends Python’s guard-node handling in the dataflow/tainttracking library so that guard calls (e.g., is_safe(x)) are recognized even when compared against boolean literals using ==, !=, is, and is not.

Changes:

• Extend guard-node recursion to treat comparisons against boolean literals as guard nodes with appropriate polarity.
• Add taint-tracking tests covering ==/!= and is/is not comparisons to True/False, and update expected results.
• Add a change note documenting the enhancement.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
python/ql/test/library-tests/dataflow/tainttracking/customSanitizer/test_logical.py	Adds a new test function covering boolean-literal comparisons for guards.
python/ql/test/library-tests/dataflow/tainttracking/customSanitizer/InlineTaintTest.expected	Updates expected sanitizer/guard nodes for the new test cases.
python/ql/lib/semmle/python/dataflow/new/internal/DataFlowPublic.qll	Extends guard-node logic to recurse through boolean-literal comparisons.
python/ql/lib/change-notes/2026-02-08-guards-compared-to-boolean-literals.md	Adds a release note for the new guard-comparison handling.