AI at the Workbench

M07-L02 exercise

Starting checkpoint: CP07. Figures: SS15-03, SS15-04.

In a new CP07 copy, record the good result, seed only the poll comparison >→>=, compile/run the tests, identify the equality failure, restore the comparison, and rerun. Submit the tiny diff and three actual receipts. Do not modify expected values to make the bad version pass. Recovery: return to the preserved copy if the seeded edit expands beyond the intended line.

Record source identity, changed files, exact commands, actual outcomes, and open limitations. Preserve the original exercise copy and review the final diff. Expected results are criteria, not evidence of execution.

Actual guest terminal shows failures at ordinary and rollover equality after the single operator mutation.
SS15-03 · SS15-03. The deliberately premature stale boundary compiles and causes two meaningful assertion failures, with exit 1.
Editor shows the complete controlled boundary experiment and its actual process exits.
SS15-04 · SS15-04. Actual boundary results preserve the passing baseline, two-failure mutation, restored pass and successful build. Injected timestamps are explicitly distinguished from measured device latency.