[siracozmen01]

Problem

No test asks m2Joint_GetLocalAxisA for the axis of a joint that has none, or a tuning reader for a tuning the joint's kind does not carry. Their headers promise zero, and the tuning readers one misuse, but a change to either refusal would pass every suite (#357).

Change

test/test_joints.c gains TestJointReadersCheckTheKind:

• With one revolute, one filter and one distance joint, each of these reads zero:
  • the revolute joint's axis (m2Joint_GetLocalAxisA);
  • the filter joint's spring (m2Joint_GetSpringHertz);
  • the distance joint's constraint stiffness (m2Joint_GetConstraintHertz);
  • the revolute joint's weld spring (m2WeldJoint_GetLinearSpringHertz).
• Each counts one misuse.
• As the near miss, the revolute joint's own spring, set to 3 Hz, reads back and counts nothing.

A test-only change.

Verification

• test_joints passes.
• Mutants, each built into a scratch library and run against test_joints:

Mutant	Fails
the axis reader's kind check is skipped	the axis read, the misuse count and the near miss
a refused tuning read returns 1	the three tuning reads
the tuning slot checks only that the joint is live	the misuse count and the near miss

• tools/check.sh changed passes. The changed file compiles clean with gcc-14 and clang-20 under the strict flags.
• test_golden prints the 13 lines of test/hashes.txt unchanged in the scalar (linux-clang-20-asan) and AVX2 (linux-clang-20) presets.

Closes #357

[siracozmen01]

Review at 4a5662a: no change needed. I read the code; I did not run it. Checks were still pending when I read them.

Each read matches what include/maul2d/joint.h documents for a joint of the wrong kind:

• m2Joint_GetSpringHertz on a filter joint (joint.h:818-827). The filter joint is not in its list (distance, wheel, revolute, motor, mouse). It returns 0 and counts one misuse.
• m2Joint_GetConstraintHertz on a distance joint (796-805). The distance joint is not in its list (revolute, prismatic, gear, pulley, ratchet). It returns 0 and counts one misuse.
• m2WeldJoint_GetLinearSpringHertz on a revolute joint (840-849). It returns 0 and counts one misuse.
• m2Joint_GetLocalAxisA on a revolute joint. Its own comment (786-794) promises only zero and says nothing about a misuse. The count the test asserts comes from the file comment at joint.h:7-9: a call made for one joint type "does the same" on another type, meaning it changes nothing, returns zero and counts one misuse. So misuse + 4 is what the header promises as a whole. The function comment could say the misuse as the tuning readers' comments do, but that is not needed for this change.
• The near miss is valid. The revolute joint is in m2Joint_GetSpringHertz's list. The baseline is taken after m2Joint_SetSpringHertz, so the setter cannot shift the count. The final check also asserts the count did not move.

The body has Problem, Change and Verification. No line starts at column 0 with a word( or type(scope):.