CLE-10-01 deep review — September 29, 2026

The study guide maps all 36 numbered objectives, grouped 6/5/5/4/4/5/4/3 across the eight blocks. Current CLE-10-01 remains the July 24, 2025 syllabus. This same-context AI review retains human review as pending.

Source evidence

The full canonical exam page and certification catalog, plus applicable entry-level policies, were read via public web fallback after direct timeouts. Manual comparison supports retaining the objective/status snapshots; the automated objective monitor did not succeed. Weights remain 13.25/13.25/13.25/13.25/16.5/16.5/7/7 and item counts 4/4/4/4/5/5/2/2. The provider's 30-point maximum and differing item values do not justify equal points per item. CLE-specific TestNow delivery and seven-day failed-retake policy are distinct from associate/professional Pearson rules. Current booking UI was not inspected.

There are 23 registered sources, with direct health outcomes {'error': 5, 'ok': 16, 'blocked': 2}. The public WG14 N1570 PDF contains 701 pages; selected C11 clauses were read through the web reader after the local download failed. No complete standard audit or local PDF capture is claimed. Selected SEI rules and four complete Microsoft runtime pages support the repairs. Microsoft’s C index describes a C89-based organization with extensions. C11 is an explicit practice choice, not an inferred exam requirement. Public source reading is bounded individually in the operation record.

Teaching and execution

All original teaching and 30 readiness answers were reviewed and improved. The allocation fragment and unbounded int-sum example became complete programs. New explanations cover numeric scanf overflow versus assignment count, whole-line versus prefix conversion, ordinary C character constants, cast timing and representability, floor/ceil/round directions, grouping versus sequencing, row-level array bounds, one-past arithmetic, allocation size/lifetime and byte length versus string capacity.

Eight exact public C blocks compiled and ran through Compiler Explorer. GCC 14.2 and Clang 19.1 each passed 72 core checks and 118 scenario checks; an additional GCC harness passed nine checked_sum boundary checks. Twelve final build/run receipts use -std=c11 -Wall -Wextra -pedantic -lm, with no final compiler warnings. An initial pass correctly exposed ambiguous-looking logical expressions as warnings; parentheses and named operands made intent explicit. A text-case expected byte count was corrected from 13 to 14 after recounting its spaces and tab; the program's output was correct.

The line reader rejects over-capacity and embedded-NUL input while consuming that line, distinguishes empty input from EOF and accepts a final unterminated line. Decimal conversion checks full suffix/range and preserves output on failure. Inventory checks zero/eight items, one/eight-billion-cent totals, malformed/missing input and release paths. Its allocation-failure seam returns null without exhausting memory. Text checks exact fit, absent terminator within capacity, unchanged destination on rejection and whitespace word transitions. The menu checks empty/single/reverse/sorted/duplicate sequences, bounds, invalid commands and full capacity. These original programs use explicit fictional contracts and still require callers to provide valid storage.

The scenario harness consolidates identical shared helpers, removes only the entry mains and feeds original functions through temporary files inside the remote sandbox. It does not replace the parser/sorter with a simulation. No local C compiler was installed or used, no deliberately undefined program ran, and no sanitizer or interactive debugger session occurred. The live-partial catalog value describes remote original-code execution. Eight broader learner activities remain proposed.

Learning resources and remaining limits

OpenEDG's complete public landing lists 42 hours, seven/week and modules 0–5. Beginner/Basics versus Intermediate copy and pointer/array shorthand are qualified. SEI's development index warns of incomplete/error-containing pages; selected examples have placeholders or malformed snippets. The original code uses actual error branches and explicit contracts instead of copying those samples. The sizeof shorthand is qualified for VLA cases, and the displayed snprintf snippet is not treated as a general compilable truncation check. Broader exploit histories and analyzer claims are not added exam objectives.

Paid OReilly/Udemy contents were inaccessible. Cisco/YouTube returned shells, so partner runtime and prior video duration remain unverified. No lesson, paid question, transcript or playback review is claimed. Author study estimates are labeled. The operation record preserves mappings, source boundaries, exact code hashes and remote execution receipts. Full unit, repository, catalog and strict generated-site gates are recorded after completion. The outcome is reviewed-with-blockers, including current account/booking details and pending human review. No account, purchase, exam attempt or external message occurred.