Java: Fix assert CFG by properly tagging the false successor.

[aschackmull]

Follow-up for #19573 to fix the small issue with asserts.

[Copilot]

Pull Request Overview

This PR adds a dedicated CFG node for failing assert statements and updates tests accordingly.

• Remove stale expected warnings for assert-false paths in Nullness tests
• Update test comment to reflect that the correlated conditions are now handled
• Introduce TAssertThrowNode and AssertThrowNode in the CFG and wire up assert failures

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
java/ql/test/query-tests/Nullness/NullMaybe.expected	Remove three outdated expected-warning lines for xs.
java/ql/test/query-tests/Nullness/C.java	Change comment on xs[0] = 42; from “FP...” to “OK”.
java/ql/lib/semmle/code/java/ControlFlowGraph.qll	Add TAssertThrowNode type and AssertThrowNode class; update CFG rules for assert.

Comments suppressed due to low confidence (1)

java/ql/lib/semmle/code/java/ControlFlowGraph.qll:103

• [nitpick] The new TAssertThrowNode(AssertStmt s) case in the TNode newtype lacks a brief comment. Please add documentation describing its purpose (i.e., modeling a failing assert).

 TExitNode(Callable c) { exists(c.getBody()) } or

[owen-mc]

Great to see a new CFG node being added.

[aschackmull]

Great to see a new CFG node being added.

I've got more: #19885