jpmschweitzer/settled-reach
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

data(assets): first production environment assets — four spike GLBs promoted (T-1204) #214

Closed
jpmschweitzer wants to merge 0 commits from furniture-promotion into main
pull from: furniture-promotion
merge into: :main
Conversation 1 Commits - Files Changed -
jpmschweitzer commented 2026-07-26 12:59:23 +02:00
Owner
Copy Link
Copy Source

Summary

Batch-7 visual track: the first real environment assets enter the game per the furniture-props brief (T-1049/PR #213).

  • table_baroque, chair_modernist, desk_scifi (furniture, with mask sidecars) and lion_statue (props, non-tintable) promoted into client/assets/models/ under the category-first convention — byte-identical to the spike sources (MD5-verified), spikes/ untouched, vw_beetle excluded per the Q-067 deferral.
  • Postprocess verified already-applied by Blender node-graph inspection (mat_primary, roughness=1.0/specular=0.0 at the shader-input level) — not re-run.
  • All four registered in manifest.json at 'planned'; models.md catalog rows + README counts updated.
  • Discovery documented (conventions.md §2 + glb-gen SKILL.md): Godot externalizes each GLB's embedded texture as a loose <model>_Image_0.png referenced by binary UID even with embedded_image_handling=3 — a mandatory, previously-undocumented commit-with-the-model sidecar; deleting it breaks loading with an error no text-grep can trace.

For the review round (araminta)

footprint_tiles ruling needed: all four spike GLBs are Blender-normalized to 1m max dimension, so real-world footprint is not derivable from geometry. The brief's [2,1] example turns out to be illustrative, not derivable. This PR sets [1,1] for all four as a flagged judgment call — rule whether normalized spike assets default [1,1] or whether promotion requires an author-supplied real-world scale, and whether the brief gets amended.

Test plan

  • In-engine proof: clean headless import; ResourceLoader.load + instantiate verified for all four
  • Pre-push gate green (full gdUnit)
  • No runtime rendering path consumes these yet (Phase-5 concern) — no eyeball owed; the load proof is the acceptance
## Summary Batch-7 visual track: the first real environment assets enter the game per the furniture-props brief (T-1049/PR #213). - **table_baroque, chair_modernist, desk_scifi** (furniture, with mask sidecars) and **lion_statue** (props, non-tintable) promoted into client/assets/models/ under the category-first convention — byte-identical to the spike sources (MD5-verified), spikes/ untouched, vw_beetle excluded per the Q-067 deferral. - Postprocess verified already-applied by Blender node-graph inspection (mat_primary, roughness=1.0/specular=0.0 at the shader-input level) — not re-run. - All four registered in manifest.json at 'planned'; models.md catalog rows + README counts updated. - **Discovery documented** (conventions.md §2 + glb-gen SKILL.md): Godot externalizes each GLB's embedded texture as a loose `<model>_Image_0.png` referenced by binary UID even with embedded_image_handling=3 — a mandatory, previously-undocumented commit-with-the-model sidecar; deleting it breaks loading with an error no text-grep can trace. ## For the review round (araminta) **footprint_tiles ruling needed:** all four spike GLBs are Blender-normalized to 1m max dimension, so real-world footprint is not derivable from geometry. The brief's [2,1] example turns out to be illustrative, not derivable. This PR sets [1,1] for all four as a flagged judgment call — rule whether normalized spike assets default [1,1] or whether promotion requires an author-supplied real-world scale, and whether the brief gets amended. ## Test plan - In-engine proof: clean headless import; ResourceLoader.load + instantiate verified for all four - Pre-push gate green (full gdUnit) - No runtime rendering path consumes these yet (Phase-5 concern) — no eyeball owed; the load proof is the acceptance
jpmschweitzer added 1 commit 2026-07-26 12:59:24 +02:00
data(assets): promote the four spike GLBs — first production environment assets (T-1204) 2f87c1a65f 
table_baroque, chair_modernist, desk_scifi (furniture, with mask
sidecars) and lion_statue (props, non-tintable) enter
client/assets/models/ under the category-first convention, byte-
identical to the spike sources (MD5-verified; spikes/ untouched,
vw_beetle excluded per the Q-067 deferral). Postprocess verified
already-applied by Blender node-graph inspection (mat_primary,
roughness=1.0/specular=0.0 shader inputs — the convenience properties
read stale defaults, the graph is authoritative), so not re-run. All
four registered in manifest.json at 'planned' with footprint_tiles
[1,1] (a flagged judgment call — normalized spike geometry erases
real-world scale; the brief ruling goes to araminta in review).
In-engine proof: clean headless import + ResourceLoader load +
instantiate for all four. Discovery documented in conventions.md §2 +
glb-gen SKILL.md: Godot externalizes the GLB's embedded texture as a
<model>_Image_0.png sidecar referenced by binary UID — it must be
committed or loading breaks with an undiscoverable dependency error.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer commented 2026-07-26 13:03:26 +02:00
Author
Owner
Copy Link
Copy Source

Review: furniture-promotion -> main (type: visual/assets)

Reviewers: Hoshe (QA) + Araminta (art direction — the ruling this PR requested is hers). Hoshe APPROVE; Araminta REQUEST_CHANGES (one doc amendment).

Hoshe (QA): APPROVE

No defects. Independently re-verified: MD5 byte-identity of all 4 GLBs + 3 masks against the spike sources; spikes/ untouched; vw_beetle and lion_statue_mask.png correctly absent; manifest.json valid and schema-exact with correct mask presence/absence and the _-prefix skip contract intact; models.md rows and README count exact; the conventions/SKILL sidecar docs mutually consistent; the mat_primary/roughness/specular claim corroborated against postprocess_glb.py source; live re-verification — clean headless import + ResourceLoader load/instantiate of chair_modernist and lion_statue, script deleted, processes clean.

Araminta (art direction): REQUEST_CHANGES

THE RULING (on the PR's own question): [1,1] STANDS for all four items — each justified by real-world class (café-scale table, chair, compact desk, statue), none misrepresented at 1 tile = 1m (D-243). The judgment call was sound.

General rule (binding going forward): a normalized promotion never infers footprint from geometry — normalization erases real-world scale. footprint_tiles is an AUTHORED value: [1,1] only when the object's real-world class is plausibly single-tile; ordinarily-multi-tile classes require the promoting author to state an intended footprint explicitly. The manifest stays the field of record; the authoring step is what was missing.

# File Issue
1 briefs/furniture-props.md §3 + conventions.md §4 The rule above must be codified BEFORE merge (her exact wording supplied) — the texture-heavy wall stories start next and must inherit a written rule, not tribal knowledge from this PR. Applied lead-side in the fix commit (her wording verbatim; conventions.md §4 gains the authored-not-derived sentence with a pointer to the brief's full rule; a pre-existing "Single Sunday-tile" typo in the same paragraph fixed to "Single-tile" in passing).

Her confirmations on record: category-first renames all correct (desk_scifi's direction verified against the spike source name); mask handling per §3 exact (lion_statue ships NO mask file, not an empty one); the extracted-texture subsection is correctly placed in §2 and stays mechanics-not-taste, no wording changes; D-257 conformance verified at the node-graph level and none of the four is glazing-class; the models.md row honestly carries the spike's known "mask too aggressive" issue rather than silently promoting it as clean; the manifest schema generalizes structurally to the wall stories — the transferable lesson ("don't let a pipeline artifact stand in for an authored value") is exactly why the rule is now written.

Verdict: CHANGES REQUESTED -> amendment applied; merging after gate.

# Review: furniture-promotion -> main (type: visual/assets) Reviewers: Hoshe (QA) + Araminta (art direction — the ruling this PR requested is hers). Hoshe APPROVE; Araminta REQUEST_CHANGES (one doc amendment). ## Hoshe (QA): APPROVE No defects. Independently re-verified: MD5 byte-identity of all 4 GLBs + 3 masks against the spike sources; spikes/ untouched; vw_beetle and lion_statue_mask.png correctly absent; manifest.json valid and schema-exact with correct mask presence/absence and the _-prefix skip contract intact; models.md rows and README count exact; the conventions/SKILL sidecar docs mutually consistent; the mat_primary/roughness/specular claim corroborated against postprocess_glb.py source; live re-verification — clean headless import + ResourceLoader load/instantiate of chair_modernist and lion_statue, script deleted, processes clean. ## Araminta (art direction): REQUEST_CHANGES **THE RULING (on the PR's own question): `[1,1]` STANDS for all four items** — each justified by real-world class (café-scale table, chair, compact desk, statue), none misrepresented at 1 tile = 1m (D-243). The judgment call was sound. **General rule (binding going forward):** a normalized promotion never infers footprint from geometry — normalization erases real-world scale. `footprint_tiles` is an AUTHORED value: `[1,1]` only when the object's real-world class is plausibly single-tile; ordinarily-multi-tile classes require the promoting author to state an intended footprint explicitly. The manifest stays the field of record; the authoring step is what was missing. | # | File | Issue | |---|------|-------| | 1 | briefs/furniture-props.md §3 + conventions.md §4 | The rule above must be codified BEFORE merge (her exact wording supplied) — the texture-heavy wall stories start next and must inherit a written rule, not tribal knowledge from this PR. **Applied lead-side in the fix commit** (her wording verbatim; conventions.md §4 gains the authored-not-derived sentence with a pointer to the brief's full rule; a pre-existing "Single Sunday-tile" typo in the same paragraph fixed to "Single-tile" in passing). | Her confirmations on record: category-first renames all correct (desk_scifi's direction verified against the spike source name); mask handling per §3 exact (lion_statue ships NO mask file, not an empty one); the extracted-texture subsection is correctly placed in §2 and stays mechanics-not-taste, no wording changes; D-257 conformance verified at the node-graph level and none of the four is glazing-class; the models.md row honestly carries the spike's known "mask too aggressive" issue rather than silently promoting it as clean; the manifest schema generalizes structurally to the wall stories — the transferable lesson ("don't let a pipeline artifact stand in for an authored value") is exactly why the rule is now written. ## Verdict: CHANGES REQUESTED -> amendment applied; merging after gate.
jpmschweitzer added 1 commit 2026-07-26 13:04:07 +02:00
docs(assets): codify the authored-footprint rule (araminta ruling, PR #214) 82d44cf483 
footprint_tiles is an authored value, never derived from normalized
mesh geometry — [1,1] defaults only for plausibly-single-tile classes;
ordinarily-multi-tile classes require an explicitly stated footprint at
promotion. Her exact wording in furniture-props.md §3; conventions.md
§4 gains the authored-not-derived sentence pointing at the full rule
(and a pre-existing 'Single Sunday-tile' typo fixed in passing). The
[1,1] values shipped for the four promoted items stand as ruled.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer closed this pull request 2026-07-26 13:13:21 +02:00

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
jpmschweitzer (Jeroen Schweitzer)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: jpmschweitzer/settled-reach#214