spec: §IFC-EXPORT-SEED — row 36, IFC export emits an EMPTY file for any resident - #115
Merged
Merged
Conversation
…ny resident
Measured, not inferred: open Duplex (196 elements on screen), call Bonsai.ifc.build(),
get {walls:0, openings:0, rels:0, arrays:0, bytes:592}. A user who picks Export > IFC on a
resident gets an IFC4 header and nothing else. build() branches on GEOM_EXTRUDE_POLY,
GEOM_CUT and GEOM_ARRAY; every ARC-seeded element is a GEOM_INSERT, so the three handled
types match zero ops in a real building.
Spec: emit one IFC product per non-anchor GEOM_INSERT, typed from params.ifc_class, with
geometry as IfcTriangulatedFaceSet over IfcCartesianPointList3D (both confirmed present in
the vendored web-ifc).
The load-bearing design choice, and its reason: REUSE THE RENDERER'S OWN VERTICES rather
than re-derive the world transform. That is the direct lesson of today's §XEDGE-GEOWIRE —
cross_edges.js re-implemented 'world = centre + R*vert', called it 'the same final numbers,
fewer steps', and was wrong for 798 of 934 elements. The folded meshes already carry
world-space positions baked in (measured: matrixWorld identity on all 3,290 meshes, max
delta 0.000e+0 between geometry.boundingBox and Box3.setFromObject), so the export reads
those coordinates verbatim and parity with what the user sees is structural.
Binding constraints recorded: anchors excluded (SampleCastle has 65) since an export is an
audit; no silent box substitution for an unresolvable mesh (the §PRIME LESSON fault); the
three existing op types stay byte-identical.
Witness W-IFC-EXPORT-SEED specified with five claims, E2 asserting the EXACT product count
so a partial export cannot pass. Out-of-scope items named so the slice is not mistaken for
a fully-furnished IFC.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UU4FmnF1izwj4c2tqp6H29
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Spec for bim-ootb #1747 (merged). Written and committed before the code.
Measured defect: open Duplex (196 elements on screen), call
Bonsai.ifc.build(), get{walls:0, openings:0, arrays:0, bytes:592}— an IFC4 header and nothing else.build()handles three op types; every ARC-seeded element is a fourth (GEOM_INSERT).Records the design choice and its reason: reuse the renderer's own vertices rather than re-derive the world transform — the direct lesson of §XEDGE-GEOWIRE the same day, where a re-derivation described as "the same final numbers, fewer steps" was wrong for 798 of 934 elements.
Binding constraints, witness claims (E2 asserts the exact product count so a partial export can't pass), and out-of-scope items all stated up front.
🤖 Generated with Claude Code