Integration: resolve with main: .asl joins main's longer File › Open lists (xcf, xcfgz, gz); each language keeps main's new strings and #4049's Styles strings; docs status lines joined.
Integration: resolve with main for area type (#6230)
- Image Size with a new resolution (#7124, canvas_geom::set_resolution) now scales an area type's
path with the text space like a box or a type path, so the same words wrap onto the same lines;
test a_resolution_change_keeps_area_type_wrapping_where_it_was fails without it.
- The Paragraph panel's justify buttons (#7713) stay available on area type, which is paragraph
text in Photoshop; test justify_buttons_work_on_area_type.
- Test files: main's tests plus #6230's; layout.rs imports joined; docs and scorecard regenerated.
Flat and layered TIFFs write and read alpha channels, spot channels and a
Quick Mask as extra samples, with Photoshop's channel resources. This is
the io part of #7415 as it stands, on current main; the next commit keeps
PSD output as it is on main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Zer0C00L01 <12090083+Zer0C00L01@users.noreply.github.com>
Integration: resolve smart_map.rs with main's 27 new smart filters (#6296 and others): both filter lists kept (49 mapped); #5249's enum readers/writers use main's equal enum_name/enum_param instead of its own copies, its test helper is photoshop_fltr_item next to main's photoshop_item, and the unknown-filter test uses Shear, since main now maps Crystallize and #5249 maps Twirl.
Conflicts:
- filter_dialog.rs: preview_document_for_target keeps our proxy scaling
and `p` clone; main's Bitmap rule (the preview drops Output resolution,
#5559) is applied to the same `p`.
- hu/sr/tr/vi catalogs: both sides appended rows at the end; kept both, no
duplicate keys.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A layer keeps its own XMP packet (Layer::xmp), where scripts and embedding
apps store data about the layer under their own namespaces.
- PSD/PSB/layered TIFF: read from and written to shmd > cust > layerXMP
(UTF-16, NUL-terminated, after layerTime), laid out as Photoshop 2026
writes it; an unchanged Photoshop block round-trips byte for byte.
- .pcraft: optional `xmp` on each layer.
- layer.xmpMetadata {layer?, rawData?}: read, set (one undo step) or
remove with null; refuses malformed XML, DTDs, NUL and > 4 MiB.
- Duplicate Layer copies the packet; document.inspect reports hasXmp.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- rafc::zero_stream takes the six-row group count explicitly; a lossless
stream always declared 18 rows, so lossless fixtures of any other height
were rejected as inconsistent. New test decodes 6- and 24-row lossless
and lossy files (fails with the old behaviour).
- The damaged-entropy case keeps the file and the stripe sizes intact and
zeroes bytes inside the first stripe, so it reaches the residual decoder
and must be RawError::Malformed (truncating the file only hit the
"stripe outside the file" check).
- Doc comments on the remaining rafc helpers.
- docs/gaps.md G14: lossy DNG decodes since 2026-10-13; no longer missing.
Opening a RAF shot with lossless (or lossy) compression fell back to the
camera's embedded JPEG ("Fujifilm compressed RAF is not decoded ... opened
the camera's embedded 4416x2944 JPEG preview instead").
Port LightCraft's clean-room decoder for Fujifilm's striped predictive
compression (storytold/lightcraft crates/raw/src/vendor/rafc.rs, MIT OR
Apache-2.0, same organisation), verified there sample for sample against a
reference decoder on 24 files from 13 bodies. New `rafc` module:
- 16-byte `IS` header checked against the raw IFD; stripe size table;
per-stripe quantizer tables for lossy files
- 768-pixel stripes decoded in parallel (`par::map`: rayon natively,
sequential on wasm); X-Trans and Bayer, 12/14/16-bit
- every read bounds-checked: damaged streams return errors, never panic
`raf.rs` sends short strips to it, after checking that header byte 108
(0 uncompressed, 2 lossless, 3 lossy) agrees with the stream and that the
CFA layout is one the stream can encode (else unsupported, preview
fallback as before).
`RafPacking::Compressed { lossy }` in testgen now writes real compressed
streams; tests cover lossless and lossy, X-Trans, RGGB and BGGR, 12/14/16
bits, mode mismatches, unknown layouts, a missing signature, truncated and
random entropy. Verified on an X-T30 lossless file: decodes in 155 ms and
develops to 6240x4160.
PhotoCraft reads GIMP's .xcf (and .xcf.gz) files, read only, through a new
reader in photocraft-io written from the public XCF specification: versions
0 to 26, uncompressed, RLE and zlib tiles, 8/16/32-bit integer and
16/32/64-bit float precisions (linear precisions get a linear profile),
pixel layers at any position, groups by item path with pass-through,
layer masks applied or disabled, opacity, visibility, locks, colour tags,
the 2.10 and legacy blend modes mapped to Photoshop's (modes without an
equivalent open as Normal with a note), the ICC profile, resolution,
guides, the comment, extra channels and the saved selection; indexed
images open as RGB. Text, vector and link layers open as their pixels;
layer effects, paths and floating selections are reported.
Fixtures are written by GIMP 3.2 itself (scripts/xcf_fixtures.py), each
with GIMP's merged rendering as a PNG. In gamma-encoded blend and
composite spaces our composite matches GIMP within 0.5/255; GIMP's
linear-light default, its Clip to backdrop and its clamp-after-alpha
for unbounded modes are documented and measured in docs/xcf.md.
Photoshop's Paragraph panel checkbox, per paragraph style:
- doc: ParagraphStyle.hanging (serde default off; .pcraft keeps it).
- layout: a trailing run of closing marks (. , : ; ! ? and fullwidth
twins, ellipsis) trims out of the aligned width on center/right lines,
so a centered line ending in an ellipsis stays centered; a wrapped row
of nothing but marks steps back onto the line above and hangs past the
box edge (contiguous ranges, merged-row baseline, manual kerning; no
RTL, vertical or tabs).
- PSD: /Hanging bool in the paragraph sheet, right after /LeadingType as
the corpus files show; read defaults off.
- engine: type.setStyle takes "hanging":bool, params doc updated.
- UI: checkbox in the Paragraph section, a Paragraph Styles editor row,
Character dock group grows one row; translations for the new label.
- tests: layout hang runs (trim, merge-back, plain no-op), PSD write and
paragraph-sheet read round trips, engine setStyle, panel harness.
openingRaw is {name, path, demosaic}; ui.demosaic is accepted only while a raw
opens; rawDefaults.demosaic is the method the app's raw opens develop with; Open
re-develops when the chosen method differs from the one the open used.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
docs/scorecard.md regenerated with cargo xtask scorecard (24 of 172); the
Preferences test keeps main's Enhanced Controls rows and the demosaic row;
catalog tails keep both sides.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings in main's Clear History, state ids and History Brush source (#3534, #3536)
through #3535's merge. Headless doc.save keeps main's on_doc and still adds the
save snapshot inside it; DocState has both save_snapshot and source_snapshot.
Two tests from #3536 now account for the automatic first snapshot (id 0): it is a
valid History Brush source, and the snapshot row clicked is the new one. Catalogs:
one line per string (duplicate History Options… lines dropped in hu, tr).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts only in the parity docs: status lines take main's text with the
JPEG Options entry (FILE-215-19) on top; scorecard takes main's rows with
FILE-215-19 done and File compatibility 10 / 12 / 5.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- text::TextLayout::overflows: a box's layout drops the overflowing lines, so
any character past the last line's byte range is dropped content; trailing
breaks that fit gain their own empty lines and never read as overflow.
- type_tool::draw_overlay: boxed text overflows draw a red + port on the box's
bottom edge near the right corner half below it, with the cursor-outline
white rim; hovering it for the tooltip delay paints a card explaining the
fix (theme tokens, tooltip preferences honoured); leaving clears the hover
clock and the delay's tail repaints, so the hint never shows late.
- ui.type_box_overflow (serde skip): automation reads the overflow without a
screen; set each overlay frame, false without a session.
- Tests: text overflows_says_whether_a_box_holds_its_text; ui unit
the_overflow_port_flags_a_box_that_cannot_hold_its_text (true / fitting
false / point false); GPU integration tests/type_overflow_port.rs checks the
red disc on screen and the hover card (skips without an adapter). The tl!
string is translated in all twenty catalogs per the i18n gate.
docs/gaps.md conflicted: kept main's header and G8 text and added the #7962 notes.
The corpus job renders with the runner's fonts. Without the font Photoshop used, a wider
fallback runs further around the circle (far edge 22 px out, IoU 0.21), so the far edge and
overlap are only compared when the font is installed, as photoshop_kerning_oracles does for
Arial. The start edge and vertical extent are still checked with any font.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Japanese printing had no paper of its own: the Print dialog and `file.print`
only offered letter/legal/tabloid/A3–A5/4x6/5x7, so a 年賀状 (New Year card),
はがき, a JIS B5 handout or an L判 photo could not be printed or exported to
PDF at its size. Add the customary papers at their exact metric sizes after the
existing entries, whose order and sizes are unchanged:
はがき 100×148 mm (283.46×419.53 pt), 往復はがき 148×200, JIS B5 182×257
(515.91×728.50), JIS B4 257×364 (728.50×1031.81), A6, L判 89×127
(252.28×360.00), 2L判 127×178, KG判 102×152, 六切 203×254 and 四切 254×305.
Names resolve case-, space- and punctuation-insensitively, and the customary
names are aliases (はがき/ハガキ/葉書/ポストカード, 往復はがき, JIS B5/B5 (JIS),
JIS B4/B4 (JIS), L判/Lサイズ/L-size, 2L判, KG判, 六切/6切, 四切/4切). The Print
dialog lists the sizes with their customary names and metric sizes, and the PDF
page (MediaBox) takes the paper's size.
Tests: `print_cmds::tests::japanese_paper_sizes_and_aliases` pins every size to
0.1 pt, the aliases and the unchanged existing papers;
`japanese_paper_sizes_reach_the_pdf_page` prints on はがき, L判, the はがき alias
and JIS B5 and reads the actual PDF MediaBox (283.46×419.53 and 252.28×360.00 pt).
Docs: docs/printing.md gains a Paper sizes section, docs/target-app-parity.md's
print row and revision history, and ROADMAP.md's progress log and revision
history, dated 2026-10-22; existing revision rows keep their dates.
Keeps main's Clear History, state ids and History Brush source (#3534, #3536 and
the integration fix-up) together with History Options: states carry both their
id and their visibility; non-linear record keeps the pinned brush source; trim_with
keeps the source-first logic. The rollback log now records the index each state
was dropped at (the brush source is never dropped, so it is not always the front)
and the replaced document, so a rollback puts every state back where it was.
Clear History ends the rollback window. Catalogs: one line per string.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Move cached keys and compressed buffers between saves, and compare weak
allocation pointers without temporary strong-reference updates. Keep
payload budgets and archive bytes unchanged.
Allow fallible tile edits after a cached save. Move pixel storage when
only weak references remain; use a fallible copy for shared storage and
restore the original tile if the copy fails.
Cover cache key reuse, edits after saves at every depth, and weak/shared
tile ownership. Name the fixed revisions behind the existing timings.
A marker kept from the opened file (Photoshop's Mtrn/Mt16/Mt32 PSDs, or
the MTrn of its layered TIFFs) was written back verbatim. Saved with an
opaque composite and an alpha channel, the file has a positive layer
count and the marker, so it reopened with that alpha channel taken as
the merged transparency, and the channel was gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>