Arctic finite ground: context outcome and remaining corrections

The unchanged native arctic study now has12actual normal desktop/portrait-phone Region/Hex views:current snow-floor control and the complete unique painting at two private placement sizes. All12pan/shared-view reconnect maps are exact. Three of four native Hex foreground RGBA comparisons are exact;the full-size desktop cohort differs at24,610of3,387,488pixels(max channel difference255),so the package remains failed and foreground stability stays open. Direct native/context inspection finds crisp snow/rock detail and coherent cold palette,but a visible rectangular boundary at smaller scale,oversized ice grain at larger scale,and poor contact between exposed blue ice and the landmark white snow apron. Retain the original painting and current live ground;neither finite placement is selected for admission. No image repeats,no source crop or inherited polygon is applied,and no artist file changes. All captured draws validate27client-visible canonical cells;the outside-viewport verified-domain branch has isolated fixture coverage only. Material CPU fallback does not establish broad GPU/state/performance acceptance.

The actual candidate is WOG-A-008379, SHA-256 1c201991c2b28de66fcb9f0ac0d06ffd51fc3c2737e0662f38aa645ce01bd5f5. Pane labels retain the inherited unchanged snow-floor identifier WOG-A-003583; the finite-image SHA and one-instance paint traces identify the study. The two sizes are private world placements, not resized artist files. The original Celestial Crater companion and both Citadel hamlets remain exact.

The inherited metadata polygon was not applied. Its 27.12% native pixel-center coverage therefore remains a question for a future terrain-domain comparison; no active clipping defect is inferred. The complete painting shows useful fine detail but its full rectangle is not suitable as an admitted finite patch. Resolve the visible boundary and place quiet snow under the existing white landmark apron without repeating the image or sacrificing its native detail. GPU/fallback, relevant state, broader scale/contact, landscape-phone and complete performance gates remain open.

The initial attempt retained four controls and four failed painter bindings. The next attempt failed before any capture because the connected probe needed nested effective values as well as flat painter values. The third captured twelve observations but failed PNG identity before preserving detailed rows or a native sheet; those twelve counts are not added as accepted evidence. The final attempt preserves all twelve detailed observations, one native sheet and the strict failed foreground result. It verifies decoded RGBA, not compressed PNG identity; the observed24,610pixel discrepancy is real and remains unattributed. Its legacy foreground width field contains native canvas width; the authored comparison entries explicitly bind it back to the1440pixel desktop viewport through recorded row dimensions.

Fresh connected read-only engine snapshots are exact before/after. All actual draws validated27client-visible cells, with no outside-view fallback used. Separate browser fixtures reject missing or conflicting geography and test the verified outside-view branch. No world command, production save, source permission or artwork approval changed.

Evidence: individual visual decision world-visual-review.json, final normal results normal-v4/results.json, native comparison sheet normal-v4/arctic-finite-ground-native.png, fresh connected probe normal-v4/read-only-effective-probe.json, owned shutdown normal-v4/owned-resources-terminal.json, native preflight preflight.json, first binding diagnosis cell-binding-diagnosis.json, strict domain fixture verified-domain-diagnosis.json and native RGBA proof method foreground-proof-method.json.