|
1 | 1 | MODEL: GPT-5.3-codex |
2 | | -REASONING: medium |
| 2 | +REASONING: high |
3 | 3 |
|
4 | 4 | TASK: |
5 | | -Apply PR 11.42. |
6 | | - |
7 | | -Remove `buildDefaultPayload` from: |
8 | | -- 3D Asset Viewer |
9 | | -- 3D Camera Path Editor |
10 | | -- 3D JSON Payload Normalizer |
11 | | - |
12 | | -This cleanup supports the explicit sample JSON/input contract from PR 11.41. |
13 | | - |
14 | | -Rules: |
15 | | -- Do not fabricate payloads. |
16 | | -- Do not auto-load hidden defaults. |
17 | | -- Do not create fallback sample geometry/camera/path data. |
18 | | -- When no payload/input exists, show a safe empty state explaining explicit JSON/input is required. |
19 | | -- Preserve behavior when explicit payload/input is provided. |
20 | | -- Keep changes limited to the listed tools and direct shared helpers only if required. |
21 | | - |
22 | | -Search terms: |
23 | | -- buildDefaultPayload |
24 | | -- 3d asset viewer |
25 | | -- 3d camera path editor |
26 | | -- 3d json payload normalizer |
27 | | -- payload normalizer |
| 5 | +Apply PR 11.43. |
| 6 | + |
| 7 | +Use PR 11.41 audit output as the source of truth: |
| 8 | +docs/dev/reports/PR_11_41_sample_json_ownership_audit.md |
| 9 | + |
| 10 | +Goal: |
| 11 | +Perform the first targeted cleanup batch for deferred sample JSON ownership findings. |
| 12 | + |
| 13 | +Important: |
| 14 | +Do NOT run the full samples smoke test by default. |
| 15 | +It takes about 20 minutes. |
| 16 | + |
| 17 | +Testing policy: |
| 18 | +- Prefer targeted sample-specific tests only. |
| 19 | +- Run syntax checks for changed JS files. |
| 20 | +- Run targeted smoke only for samples changed by this PR. |
| 21 | +- Run full samples smoke only if this PR changes shared sample launch infrastructure or a broad shared loader. |
| 22 | + |
| 23 | +Cleanup policy: |
| 24 | +For deferred PR 11.41 items: |
| 25 | +- KEEP + WIRE if JSON belongs to current executable sample and can be visibly used. |
| 26 | +- MOVE / REHOME if JSON clearly belongs to a different sample/tool. |
| 27 | +- DELETE only if stale/obsolete and coverage remains elsewhere. |
| 28 | +- CREATE / UPDATE CORRECT SAMPLE if coverage would otherwise be lost. |
| 29 | +- DEFER if ownership is uncertain. |
| 30 | + |
| 31 | +Sample 1902: |
| 32 | +- remains exempt |
| 33 | +- do not apply single-tool ownership cleanup to 1902 |
| 34 | + |
| 35 | +Batch selection: |
| 36 | +- Choose the smallest safe batch from the deferred 11.41 list. |
| 37 | +- Prefer items with clear ownership and low blast radius. |
| 38 | +- Do not attempt to fix all deferred items at once. |
28 | 39 |
|
29 | 40 | Do NOT: |
30 | | -- change unrelated tools |
31 | | -- change sample 1902 workspace payload except as needed for correct explicit-input messaging |
| 41 | +- run full samples test unless justified |
| 42 | +- add hidden fallback/default data |
| 43 | +- create decorative JSON |
| 44 | +- change unrelated tool logic |
32 | 45 | - undo SVG Asset Studio rename |
33 | 46 | - touch start_of_day folders |
34 | 47 |
|
35 | 48 | Validation: |
36 | | -Run targeted syntax checks for changed JS files. |
37 | 49 | Run: |
38 | | -node ./tests/runtime/LaunchSmokeAllEntries.test.mjs --samples --tools |
| 50 | +- node --check for changed JS files |
| 51 | +- targeted sample/tool smoke for changed samples only |
39 | 52 |
|
40 | | -If full smoke is too broad, run relevant targeted smoke and document why. |
| 53 | +If a full test is required, explain why in the report before running it. |
41 | 54 |
|
42 | 55 | Reports: |
43 | 56 | Write: |
44 | | -docs/dev/reports/PR_11_42_validation.txt |
| 57 | +docs/dev/reports/PR_11_43_cleanup_batch_1.md |
| 58 | +docs/dev/reports/PR_11_43_validation.txt |
45 | 59 |
|
46 | 60 | Report must include: |
47 | | -- files changed |
48 | | -- each removed buildDefaultPayload location |
49 | | -- new no-input behavior for each tool |
50 | | -- validation command results |
| 61 | +- source PR 11.41 findings used |
| 62 | +- resolved JSON items |
| 63 | +- action taken for each |
| 64 | +- coverage preserved statement |
| 65 | +- deferred items remaining |
| 66 | +- exact targeted tests run |
| 67 | +- whether full suite was skipped and why |
0 commit comments