FAILURE MAP
← Case archive

FA-51708 / Raster clipping / Member archive

Scanline polygon span clip: empty span emitted · case 03

Spans clipped to zero width are still emitted.

Member previewVariant 3 · 3 implementations · 13 checks per implementation

Case contract

Each command is [polygon,scissor]. Rows y in the scissor rows [top,bottom) intersected with [0,size) are sampled at y+1/2. Each polygon edge is oriented bottom to top and contributes a crossing when top<=yc<bottom (half-open in y), interpolated exactly. Crossings are sorted and paired even-odd. A pair covers pixels whose centers x+1/2 satisfy xa<=x+1/2<xb, i.e. [ceil(xa-1/2),ceil(xb-1/2)), then clipped to the scissor columns [left,right) intersected with [0,size). Non-empty spans are returned as [row,start,end].

Why this case matters

Raster clipping decides which pixels a primitive may touch; a single wrong clip decision writes outside the target or scissor, or silently drops visible coverage.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixtureActualExpectedOutcome
explicit case 1[[[0, 1, 1], [0, 5, 5], [1, 1, 2], [1, 4, 5], [2, 1, 5], [3, 1, 5], [4, 1, 5]]][[[1, 1, 2], [1, 4, 5], [2, 1, 5], [3, 1, 5], [4, 1, 5]]]Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.

Sign in to the archive ↗