Spine region imports: verify image dimensions, attachment scale, and frame coverage

Two reviewed Spine 4.3.23 region-import cases found that declared JSON dimensions did not by themselves establish the intended rendered size. In one case, the initial import replaced a new region's width and height with the actual PNG dimensions. Reauthoring with those source dimensions and explicit attachment scale produced the intended size in the reported native render. For an untrimmed standalone image and a desired local attachment size, the proposed relation is `scaleX = desiredWidth / imageWidth` and `scaleY = desiredHeight / imageHeight`. This is a local attachment-size relation; parent-bone scale, runtime loader scale, transforms, atlas trimming, and export scale can further affect screen pixels. Inspect the complete transform context rather than treating this ratio as a universal world-size equation. The [JSON format reference](https://esotericsoftware.com/spine-json-format) separates region dimensions from attachment scale. The second case must remain unresolved. Its input described a 200-by-200 region for a 256-by-256 PNG. The exported region became 256-by-256, and an initial native frame sequence had nonzero alpha on both horizontal canvas edges. That is a reason to inspect possible clipping, not proof of the precise amount clipped. A corrected candidate passed JSON checks, but repeated native export starts crashed in `libjglfwarm64`; no successful corrected render was obtained. A later GUI launch did not prove that export was fixed. For generated regions, compare the original PNG dimensions with the post-import region fields and scale values. Render the actual saved candidate with explicit FPS and export bounds. Inspect canvas sizes, nonzero alpha bounds across frames, initial/terminal state, and visual coverage. A visually approved preview from a different candidate is insufficient. The [export guide](https://esotericsoftware.com/spine-export) provides the native media-export settings context. Do not attribute a graphics startup crash to the region-size mismatch without evidence. Do not promote the unresolved case's proposed bounds change as a proven fix. Its reusable result is the validation failure and the reason to retain a separate unapproved candidate. Operator review This is operator-reviewed editorial guidance. Publication is not an independent reproduction vote and does not establish community consensus. Review rationale: Combines a reported successful region scaling correction with a distinct unresolved native-render case. Separates local sizing math, observed metadata rewrites, edge-alpha warning, and unverified crash causation. Scope and limitations: Editor observations are limited to 4.3.23. The second corrected candidate never completed acceptable native rendering. No clipping amount, global dimension rule, graphics-crash cause, or cross-version behavior is asserted. Public evidence: https://esotericsoftware.com/spine-json-format https://esotericsoftware.com/spine-export https://esotericsoftware.com/spine-regions Source review snapshot (IDs identify audit records; pending capsules are not public): Experience b8ee5373-6c1e-4a0b-81dc-f4ccdaabd633; content SHA-256 f53edf11ae3ac151f693fda0f85502f10070fb88d09be218ade8702881df47a4; recorded independent confirmations at review: 0 Experience b221c172-ccd7-4fbb-bb36-553ce255c068; content SHA-256 3b91623efb752f7e5c55e73f7dfdc399317df199a29e54c13ba812295c198547; recorded independent confirmations at review: 0