✔ valid ASCII supports BOM, CRLF, comments, blank strings, and numeric spellings (1.4816ms)
✔ ArrayBuffer and sliced typed-array inputs are accepted without reading surrounding bytes (0.198ms)
✔ complete file with EOF and invalid numeric coordinates is rejected (0.4722ms)
✔ integer and Boolean typed values cannot contain decimal values or arbitrary words (0.2031ms)
✔ truncation and unmatched section markers are rejected descriptively (0.3695ms)
✔ invalid group codes and null bytes are rejected (0.2246ms)
✔ conservative structural checks accept custom sections and unknown entity types (0.0756ms)
✔ binary DXF is delegated to its binary parser unchanged (0.6762ms)
✔ real corrupt fixture is rejected and every valid local ASCII DXF fixture passes (24.7951ms)
✔ DXF viewport patch installs once, respects 40/41, and preserves model view height (1.6138ms)
✔ direct viewport dimension assignment used by the DWG converter is unchanged (0.1023ms)
✔ actual engineering DXF fixture imports its non-square sheet viewport as 250 × 140 (31.2338ms)
✔ DWG content wins over a mistaken filename extension (0.9207ms)
✔ text DXF allows whitespace, UTF-8 BOM, and comment pairs (0.7878ms)
✔ DXF-looking text inside HTML is not accepted (0.141ms)
✔ binary DXF needs the complete sentinel (0.1253ms)
✔ all legacy DGN signatures reject with local conversion guidance (0.1732ms)
✔ compound containers are not falsely asserted to be confirmed DGN V8 (0.09ms)
✔ typed-array subviews respect their byte offsets (0.0823ms)
✔ raster formats require signatures, not just extensions (0.1408ms)
✔ SVG remains unsupported and does not become a raster route (0.0928ms)
✔ empty, malformed and renamed DGN files fail closed (0.3569ms)
✔ pinned parser fixture reproduces the inactive paper-space loss (0.4992ms)
✔ repair installs once and retains all four E-201 sheet entities (15.2389ms)
✔ duplicate source handles repeated in ENTITIES do not duplicate repaired sheet geometry (13.349ms)
✔ distinct handles are retained even when sheet text geometry is identical (15.3511ms)
✔ per-document deduplication does not suppress matching handles in subsequent documents (16.6114ms)
✔ DXF ID distinguishes coincident-center user viewport from the internal paper viewport (15.2208ms)
✔ directly assigned DWG-style viewport IDs retain the original structural predicate (0.1866ms)
✔ binary DXF filing preserves explicit viewport identity too (6.6659ms)
✔ host viewport recovery passes the actual inactive BTR and does not rerun populated views (5.0973ms)
✔ normalizes CAD separators and dot segments without losing roots (1.1108ms)
✔ host-relative exact resolution is case insensitive and wins over duplicate basenames (0.4109ms)
✔ exact duplicate paths are ambiguous rather than first-match (0.134ms)
✔ moved package uses longest suffix and reports relocation (0.1233ms)
✔ UNC reference can use selected local data without fetching a share (0.104ms)
✔ duplicate relocated basenames are never guessed (0.0969ms)
✔ URL schemes including file and data are blocked (0.1248ms)
✔ missing paths stay missing and empty paths are not basename matches (0.0855ms)
✔ cycle and depth guards are case insensitive and bounded (0.1184ms)
✔ folder drop reads all batches and preserves nested paths (0.7722ms)
✔ folder entry errors and size limits reject instead of returning partial success (0.4821ms)
✔ ordinary multi-file fallback retains webkitRelativePath (0.0773ms)
✔ xref cloning namespaces nested blocks, text, linetypes, and dimensions using real APIs (5.0494ms)
✔ copying an already bound reference does not append duplicate geometry (0.6332ms)
✔ preloaded xref geometry is retained with correct count without reading selected files (0.4577ms)
✔ nested overlay geometry does not propagate, including preloaded cached geometry (0.5832ms)
✔ cycle detection normalizes slashes, case, and dot segments before file reads (0.6621ms)
✔ depth limit reports the omitted reference and never attempts its file read (0.2443ms)
✔ malformed ASCII xref fails before parsing and cannot bind partial reference geometry (1.3935ms)
ℹ tests 50
ℹ suites 0
ℹ pass 50
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 405.3591
