Bug:
EMF polyline/polygon records carry a ring count as a raw u32.
pts16/pts32 did Vec::with_capacity(count) on it directly - a crafted
4294967295 pre-allocates about 4 billion slots and the allocator aborts.
The existing total > 400_000 guard only checked the sum of ring counts,
never an individual ring's vector.
Reproduce:
1. Craft an EMF whose POLYPOLYLINE/POLYGON count is 0xFFFFFFFF.
2. Attach/import it.
3. Before: "pre-allocated 4294967295 slots" then memory allocation
failed (multi-GB pre-allocation).
4. After: the vector is sized from the bytes actually present in the
record; the parse consumes what the record holds.
Fix:
pts16/pts32 size their capacity as count.min(available_bytes / point_size)
- the record's own remaining length is the ceiling, so allocation is
O(record). The iteration logic is untouched: it stops when the bytes run
out, so output for valid records is byte-identical (the guard is a pure
allocation bound). Tests: the u32::MAX ring no longer pre-allocates;
valid single- and multi-ring records still parse.
Bug:
decode_packed (RLE-packed DIB decoding) guarded the header with a
12-byte check but then read fixed fields at offsets 14-16, 16-20 and
32-36. Any header shorter than 36 bytes that took the non-core path
panicked with an out-of-bounds slice - every input in that range.
Reproduce:
1. Open a BMP whose DIB header takes the non-core path but is shorter
than 36 bytes (a 12-35 byte header claiming packed data).
2. Before: out-of-bounds panic at the fixed-field reads.
3. After: the header is rejected and the image skipped; valid headers
(including a 1x1 DIB) still decode.
Fix:
The non-core branch requires len >= 36 before any of its fixed reads -
every input under 36 bytes panicked before the fix, so no non-panicking
behaviour changes. Valid packed DIBs (tests include a 1x1 image) decode
unchanged.
Bug:
The EMF record reader accepts nSize = 28 (header-only) so legal empty
polylines load, but the sub-arm handling those iTypes also reads a u32
count at offset 28 for records that are not empty - rec.get(28..32) on a
28-byte record is out of bounds, panicking while importing the file.
Reproduce:
1. Craft an EMF with an EMR_POLYLINE (or related iType) record whose
nSize is exactly 28.
2. Attach/import it.
3. Before: out-of-bounds panic reading at offset 28.
4. After: the undersized record is skipped; a full 36-byte record still
renders.
Fix:
The sub-arm that reads at offset 28 now requires rec.len() >= 32 (its
first read), while the shared arm's guard stays at 28 so legal
header-only empty polylines keep loading. Undersized records are skipped
the way other malformed records already are. Tests: iType 7 and 91
records of exactly 28 bytes are skipped, a 36-byte record still renders.
The codec and kernel repositories and crates are renamed opencadcodec and
opencadkernel. Inside OCS they are depended on under short keys, so code
reads `codec::`, `kernel::` and the new `graph::` (opencadgraph, the node
graph engine). File-format and protocol spellings stay as they were: the
ACIS product id `acadrust`, the ACADRUST_* switches and the plugin
version keys `acadrust_version` / `acadrust_source`.
A button left of MCP on the command line opens a translucent node graph
over the viewport, with a palette of object, input, math, logic, list,
text and geometry nodes added by click or drag. An object node owns
drawing entities and its rows are the Properties panel's own rows for
them: the single-object page is extracted from refresh_properties as
entity_property_sections, values use the control protocol's
property_json, and writes go through set_property_value with a property
target override, so the panel, MCP and the graph share one read/write
path. A list reaching an object row makes one entity per element. Each
graph action is one undo step; the canvas pans and zooms.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
OLE2FRAME entities whose compound file carries a metafile presentation
(Excel/Word previews, scanned WMF pages) drew only their border frame —
the old decoder scanned the raw blob for a contiguous BMP, which the
CFB sector layout almost never provides. acadrust now extracts the
picture bytes tagged by kind; the new scene/model/gdi player rasterizes
EMF and WMF record streams on the CPU (pens/brushes, polygons, paths,
DIB blits, clipping, transforms) with text drawn through the existing
system-font glyph cache, and image_model feeds the result into the
raster pipeline. Rasters (BMP/PNG/JPEG/GIF, CF_DIB) decode as before
via the image crate.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Karim Jerbi