* craft-dxf: add VectorCraft's ASCII drawing writer
Move the incremental file writer from vectorcraft-cad (crates/cad/src/writer.rs)
into craft_dxf::drawing. It writes the HEADER/TABLES/BLOCKS/OBJECTS skeleton
around a pre-built ENTITIES section, which is the format API VectorCraft's
scene mapper already uses. The existing model::write tag serializer is
unchanged.
Golden files lock the bytes for empty and one-LINE drawings in R12, 2000 and
2018. Part of #791 / #778.
* test(dxf): preserve exact golden bytes on Windows checkout
* fix(dxf): preserve BYBLOCK and use valid image dictionary keys
* test(dxf): clone images before write in dictionary-key test
Drawing::write takes self, so the assertions have to use a clone of
the image list. The rest of the test is unchanged.
---------
Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
Extract the remaining metadata delegation from #1198 after #1331 landed the raster facade changes. Preserve current image-codecs API, pixel transforms, docs and regression tests; omit unrelated resolution reexports and TIFF cleanup.
Original-change: 1d52e89405
A tile's QP map error is reported by its first plane job, so a later tile's
corrupt map no longer wins over an earlier tile's plane error. Equivalence
tests also compare error texts; the corpus test fails when the folder holds
none of the pinned files; docs/cr3.md says when the sensor buffer is allocated.
Part of #1654
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The vertical lifting combined strided columns one at a time; it now lifts
whole rows, per pair of output rows in parallel, with the same integer
equations and edge rules (i64 floor shifts equal the div_euclid it replaces).
The horizontal pass runs its rows in parallel and drops the Result per sample;
overflow is still an error. Differential tests compare both passes with the
previous implementation for every size 1..=17 and every edge combination.
Part of #1654
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every plane of every tile is one parallel task that ends in sensor samples;
the mosaic is then filled per output row pair in parallel, without a modulo,
division, bounds check or Result per sample. The first error in tile and
plane order is still the one reported, and the sensor buffer is allocated
only after every plane has decoded. The CC0 corpus test also pins the four
EOS R5 Mark II files and skips files that are not present.
Part of #1654
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The ten subbands of a plane are separate entropy streams; the first error in
band order is still the one reported. The step is looked up once per band, or
once per QP cell for adaptive gains, instead of once per coefficient.
Part of #1654
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>