Include thickness and an actual obstacle
Unlike a single point’s swing, a thick leaf occupies a changing rectangular region. Enter its closed-position width along +x and thickness along +y. Locate the pivot relative to the rectangle’s lower-left corner. Obstacle coordinates use the pivot as world origin; keep these two coordinate definitions distinct. Motion is counterclockwise from 0 to the selected angle.
Event-based geometric method
Each door vertex rotates by x′ = x cosθ − y sinθ, y′ = x sinθ + y cosθ. Candidate transition angles solve vertex-to-edge line equations of the form a cosθ + b sinθ = c, in both moving and fixed coordinate systems. Between sorted events, the separating-axis test determines whether the rectangles overlap. Event positions are also checked to retain isolated tangencies. The method does not infer clearance by sampling only a few opening angles.
Independent quarter-turn example
Let the closed door be 2 × 1 in, pivot (0,0), and obstacle x = 1–2 in, y = 0–1 in. Over 0–90°, the rectangles intersect from 0° through 45°, including contact at the obstacle corner (1,1). They separate after 45°. This simple example is not a realistic door-size recommendation.
Numerical and physical limits
The result uses floating-point geometry and displays angles to three decimals. Coordinate comparisons use an allowance of 10⁻¹⁰ times the largest entered coordinate/span magnitude in inches; the result states that allowance. It is a numerical rule, not a manufacturing tolerance. Coordinates are limited to ±1,000,000 in and positive rectangle spans to at least 0.000001 in. Reversed or zero-width obstacle bounds are rejected.
A plan intersection can be irrelevant if objects occupy different heights. Conversely, handles, closers, seals, deformation, and unmodeled obstacles can collide even when these two rectangles do not. Add your actual geometry and appropriate professional clearance analysis; a “no intersection detected” result is not an installation approval. For a point without an obstacle, use door swing geometry.