Skip to content

corpus: mirror rvt-rs #225 — slabs and shading devices exact; full-project surface is 10 fields; spec 1.0.2 - #5

Merged
DrunkOnJava merged 1 commit into
mainfrom
corpus/mirror-rvt-rs-225
Aug 30, 2026
Merged

corpus: mirror rvt-rs #225 — slabs and shading devices exact; full-project surface is 10 fields; spec 1.0.2#5
DrunkOnJava merged 1 commit into
mainfrom
corpus/mirror-rvt-rs-225

Conversation

@DrunkOnJava

@DrunkOnJava DrunkOnJava commented Aug 30, 2026

Copy link
Copy Markdown
Owner

rvt-rs #225 (RE-22) recovers IFCSLAB 80/80 and IFCSHADINGDEVICE 20/20 by
id set (per-instance IFC Export As overrides found in the file; the one
missing slab was an OST_BuildingPad) and slab thickness 79/80 against the
export's extrusion depths. The manifests gain a shading_devices category,
so all three observations change on both artifacts; the full-project
verdict is 10 surface fields / 4 excluded, the element fixture 4 / 9.
SPEC.md mirrors rvt-rs docs/octetproof-spec.md 1.0.2 (worked examples
labelled as dated snapshots). Index re-chained; validators pass.


Note

Low Risk
Documentation and committed corpus/verdict fixtures only; no protocol schema or gate logic changes in this diff.

Overview
Bumps OctetProof to 1.0.2 and refreshes the golden corpus to match rvt-rs slab/shading recovery (RE-22 / #212): 80 IFCSLAB and 20 IFCSHADINGDEVICE with exact export id-sets on the full-project artifact, retiring the old plan-loop slab path and the floor_slab_extrusion_thickness known gap on floors.

On g-2026-0002 (full export), floors becomes known (80 slabs), a new shading_devices row scores 20 devices, entity_counts.IFCSLAB and IFCSHADINGDEVICE join the compared semantic surface (10 fields, 4 excluded), and committed rvt-rs / bridge observations and verdict hashes are updated. On g-2026-0001 (element fixture), the same decoder output is reflected in rvt-rs observations while manifest rows stay decoder_baseline for slabs/shading; exclusions shift from IFCSHADINGDEVICE known_gap to IFCSLAB / IFCSHADINGDEVICE baselines.

SPEC.md adds §19b and clarifies that §6.2/§6.3 examples are dated 2026-08-30 snapshots, not a live mirror of upstream witness files. MANIFEST_INDEX.json is re-chained with new manifest and verdict hashes; both artifacts remain PASS.

Reviewed by Cursor Bugbot for commit 4460f92. Configure here.

…oject surface is 10 fields; spec 1.0.2

rvt-rs #225 (RE-22) recovers IFCSLAB 80/80 and IFCSHADINGDEVICE 20/20 by
id set (per-instance IFC Export As overrides found in the file; the one
missing slab was an OST_BuildingPad) and slab thickness 79/80 against the
export's extrusion depths. The manifests gain a shading_devices category,
so all three observations change on both artifacts; the full-project
verdict is 10 surface fields / 4 excluded, the element fixture 4 / 9.
SPEC.md mirrors rvt-rs docs/octetproof-spec.md 1.0.2 (worked examples
labelled as dated snapshots). Index re-chained; validators pass.
@cursor

cursor Bot commented Aug 30, 2026

Copy link
Copy Markdown

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_1bbe845a-38b5-401a-80a9-5e3daec7453a)

@DrunkOnJava
DrunkOnJava merged commit 8834c78 into main Aug 30, 2026
5 checks passed
@DrunkOnJava
DrunkOnJava deleted the corpus/mirror-rvt-rs-225 branch August 30, 2026 16:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant