Decode lossless CRX and version 0x100/0x200 C-RAW in portable Rust,
including sensor geometry, timed white balance and black/white levels.
Verify all sensor samples against six independent M50/R100/R8 references.
Any CR3 decode error falls back to the embedded JPEG, as CR3 files did
before this decoder. The full sensor buffer is allocated only after the
first plane decodes, and plane buffers grow row by row. Tile planes
decode in parallel with a windowed bit reader (24 MP: 424 -> 89 ms
lossless, 519 -> 138 ms C-RAW on an M4).
Add guarded CR3 camera-JPEG colour fitting with independent framing
validation, relative white balance and cache invalidation.
Validation: cargo xtask ci.
A region parsed from mwg-rs:Regions is now intersected with the 0..1 frame and dropped when it
misses the photo, so a hostile or stale sidecar (x=1e5, a 1e5-wide box) can't park a box far
outside the image. Lightroom's mwg-rs:Rotation of ±π (Exif orientation 3) is mapped to the upright
frame like the quarter turns; mirrored orientations can't be expressed in MWG and are taken as
written.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CR3 files imported as preview-only with the 1620x1080 PRVW JPEG and no metadata
(no camera, capture time, exposure, lens or GPS): nothing read the container.
- crates/meta/src/cr3.rs: bounded ISO base media box parser for CR3 (Canon uuid
box with CMT1-4 and THMB, tracks with their sample entries, CMP1 coding header
and sample location, XMP uuid box). Every size/offset checked; depth and box
count bounded.
- Metadata: CMT1 (IFD0), CMT2 (Exif) and CMT4 (GPS) merged into one TIFF Exif
block, so CR3 reads like any TIFF-based raw; XMP from its uuid box.
- Preview: the full-size JPEG track (6960x4640 on an EOS R6 Mark III) before
PRVW/THMB; imports report the full image size.
- Tests: synthetic CR3 (boxes, tracks, merged Exif), truncation/corruption sweep,
JPEG track preferred over PRVW with fallback.
Raw (CRX) data is still not decoded; CR3 stays preview-only.
Lightroom Classic stores the box of a photo with Exif orientation 6 / 8 in
the sensor's own frame and records the orientation as mwg-rs:Rotation
(-pi/2 for 6, +pi/2 for 8, 0 for 1) while AppliedToDimensions names the
upright size. Ignoring Rotation put every portrait photo's face box in the
wrong place. Convert such regions into the upright frame, with a synthetic
regression test for -pi/2, +pi/2 and 0.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Adds Metadata::regions: Vec<Region>, populated by parse_xmp from mwg-rs:Regions
(MWG Region Guidelines v2.0) — the format Lightroom, digiKam and Picasa write
for face/pet/focus/barcode regions. Read-only: LightCraft does not write
regions yet (blocked on #92, which would otherwise clobber existing sidecars).
- Canonicalises the mwg-rs/stArea/stDim namespaces so any prefix a writer
picks resolves consistently, reusing the existing generic struct/array XMP
machinery (handles both the element and attribute encodings transparently).
- Area center-coordinates are converted to lightcraft_geom::Rect via
Rect::from_center. Normalized units (the default, and the only unit we've
seen in practice) need no AppliedToDimensions cross-reference; the rare
pixel-unit region is converted when AppliedToDimensions is present, and
dropped (not stored as a nonsense fraction) when it isn't.
- Hostile/malformed regions (missing Area, NaN/absurd/zero/negative extents)
are skipped individually rather than failing the whole list or panicking.
- regions merges like keywords in fill_missing/overlay_user_fields, since it
only ever comes from XMP today.
Not included: nothing yet reads Metadata::regions (no catalog field, no UI,
no command) — this is plumbing for a later PR, with no user-visible effect
and no docs/parity.md row to update.
cargo xtask ci passes (fmt, clippy -D warnings, test, parity, layers, assets,
wasm).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Saving (Cmd+S or auto-write) now merges LightCraft's properties into an
existing sidecar (lightcraft_meta::merge_xmp): owned properties (rating,
label, dc/IPTC/rights fields, lc:*) are replaced or removed, capture time
and GPS only when LightCraft has a value; everything else (crs:,
xmpMM:History, unknown namespaces, packet wrapper) is kept byte for byte.
The result is re-parsed and checked before it is written (temp + fsync +
rename). A sidecar that isn't valid XMP is copied to .xmp.bak-<time>
(never overwriting) before it is replaced; one that can't be read at all
is left alone and the save fails.
Shared stems: when several catalogued files map to one stem sidecar
(IMG.CR3 + IMG.JPG), it belongs to the raw (else the first by name) and
the others write and read <file>.xmp.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GPX 1.0/1.1 reader in lightcraft-meta (timed track points per segment; interpolation only within a
segment, nearest point within maxGap otherwise). Engine command photo.autoTagTracklog (path or gpx
text, camera time-zone offset, maxGap, replace, dryRun; one undo step) and Photo ▸ Auto-Tag from
Tracklog… with a time-zone prompt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Catalog Meta gets copyright_status (unknown / copyrighted / publicDomain), usage_terms and
copyright_url, read and written as XMP Rights Management fields (xmpRights:Marked, UsageTerms,
WebStatement) on import, sidecars and exports (kept by "copyright only"). photo.setMeta,
copy/paste metadata, metadata presets, a Copyright Status smart-album rule and the Info panel.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
lightcraft-meta gains tags.rs: every tag of IFD0 and the EXIF, GPS and interoperability IFDs
as (group, name, value) rows, named per the public TIFF / EXIF / DNG specifications, with
common values spelled out (exposure as 1/250 s, f/2.8, 35 mm, orientation, metering, flash,
compression…) and blobs summarised. photo.allMetadata adds the XMP fields (sidecar, else the
file's packet). Info ▸ All Metadata… opens a searchable dialog grouped by TIFF / EXIF / GPS /
XMP.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Local corrections stored as crs: structures (MaskGroupBasedCorrections and the older
gradient / circular gradient / paint containers) become our masks: linear and radial
gradients, brushes (dab runs → strokes), Subject / Sky, luminance and depth ranges, with
add / subtract / intersect and invert, and all local sliders. lightcraft-meta's XmpData gains
`values`, a structured view of every property (arrays of structs survive); .lrtemplate Lua
tables convert into the same shape.
Applying a preset adds its masks to the photo's own (once; Amount scales them); sidecar masks
replace them and radial masks are fitted to the photo's aspect. Unsupported components
(colour range, other AI selections) are reported as "Mask: <kind>".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Metadata gains alt text, extended description, sublocation, city, state/province, country:
catalog, photo.setMeta, copy/paste metadata, XMP read/write (Iptc4xmpCore AltText /
ExtDescrAccessibility / Location, photoshop City / State / Country), sidecars and export
(place fields dropped with "remove location").
- Info panel: camera card (camera, lens, size · file size, format badge, focal length,
shutter, aperture, ISO), then title, caption, alt text, extended description, copyright,
creator, file name (rename), file path (reveal), captured (edit), place fields, GPS.
- Text fields keep what is typed across frames and save on focus loss (they were rebound to
the stored value every frame). New pencil icon (drawn in code); dates::display_time.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
RAF header/record parsing, raw IFD (0xf000) with 16-bit, 12-bit LSB and
32-bit-word MSB packings found by black-box analysis of CC0 samples,
X-Trans layout (reversed record, anchored at 0,0), crop, black levels,
WB_GRB levels, metadata and preview from the embedded JPEG. Compressed
RAF reports Unsupported. Colour uses the documented neutral fallback.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Classic + BigTIFF, both byte orders, all field types, SubIFDs and Exif/GPS/
Interop pointer IFDs, strips/tiles layout, maker-note offset-base helpers,
IFD writer; hardened against malformed input (proptests). Placeholder
manifests for lightcraft-meta and lightcraft-raw.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Иван Лапочкин
Brandon Thomas
Maurice Rogers
Daniel Voshart
storybored379