amos 9d633e9e67 Batch 15 (20261023-C-hwe-scale-15): 320 HWE-derived repair tasks
One user-space repair task per eligible HWE debug-diag / compile-deploy base
ID: 320 admitted (C 218, C++ 97, build scripts 5), 1 skipped by the content
filter. Each ships a public spec with a check-to-basis table, at least two
defects with single-fix fixtures, a symptom-masking fixture and an alternative
correct implementation. Claude re-checked all 320: structure, the HWE quality
items and trace6 Docker calibration receipts against current digests
(320/320 USABLE). Difficulty is unmeasured. Known follow-up: 212 tasks use 120
ad-hoc family names; the next batch maps them to the canonical list.
The batch's 1,256-line progress log stays in its factory report and the
36-local branch batch15-hwe.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 01:55:56 +08:00

2.5 KiB

config-declared-params author specification

  • Source route: C, original C++17 task inspired by HWE gi-ros2-param-declare-default. source is hwe/gi-ros2-param-declare-default; split key is that base ID. The HWE .c is actually a Python Node stand-in. Its prompt mentions ROS2, but neither its code nor this task uses an ROS2 library. This task tests a deterministic software rule: schema declaration with defaults must precede reads, and documented overrides must remain observable. This is not a claim about a real incident, rclpy behavior, hardware, or ROS2 lifecycle.
  • Initial files: all C++ sources, header, Makefile, README, and contract are newly written. No HWE source is copied. Independent task family config-calibration.
  • Defect 1: startup misses one schema declaration, exposed by BOOT then GET rate. only_fix_declare repairs this while retaining defect 2.
  • Defect 2: valid non-default mode override is acknowledged but not stored, exposed by SET mode manual or safe and later GET/LIST/PATCH/CAS. only_fix_override repairs this while retaining defect 1.
  • Difficulty: UNMEASURED. Calibration does not measure model difficulty. No agent sampling authorized here.

Criterion to public basis

Grader observation Public basis
C++17 build, stdout line protocol, arbitrary stderr diagnostics template/README.md:3-11; template/docs/contract.md:3-5
BOOT declares all four defaults; repeated BOOT resets template/docs/contract.md:9-16,20
Declaration before GET/SET/PATCH/RESET; duplicate and unknown errors template/docs/contract.md:20-24,38
Typed values, allowed ranges, canonicalization, typed comparison template/docs/contract.md:9-16
SET, RESET, LIST, COUNT behavior and failure state template/docs/contract.md:22,26
PATCH order, repeated names, atomic failure and size template/docs/contract.md:24
DESCRIBE, COMPARE, CHANGES, RESTORE template/docs/contract.md:28
REMOVE, CLEAR, VALIDATE, MISSING, DECLARED template/docs/contract.md:30
DECLARE_MANY atomicity, duplicate and errors template/docs/contract.md:32
CAS conditional typed update and DEFAULTS template/docs/contract.md:34
Syntax, whitespace, length, final line template/docs/contract.md:5,20,24,32,36

The private oracle implements these transitions independently in Python and varies sequences and values; it does not call gold for expected output. The public executable only emits observations. Only public behavior is scored. Every fixture builds from the declared candidate files.