Port selected-colour editing through main d42f9da94 and its engine/UI
module layout. Preserve native channel targets, drawing previews, tone
editors and filter behavior in the split modules, with their existing
regressions and public facade exports.
Upstream split smart_cmds.rs into smart_cmds/ and smart_map.rs into
smart_map/: the stored-mesh hook moves to smart_cmds/rendering.rs, the
Puppet Warp import and export arms to smart_map/import.rs and export.rs.
main split atn_cmds.rs into submodules: the Camera Raw Filter mapping moves to
cmd/atn_cmds/steps_image.rs and its event to events.rs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Type Path branch in edit_path caught every `"name":"layer"` edit of a
type-on-a-path layer, including edits of its targeted vector mask: Direct
Selection showed the mask, but a drag moved the Type Path by the mask's
knot indices and left the mask as it was. On main the same command edits
the mask, and path.info / path.get / path.transform still read the mask.
`"name":"layer"` now keeps that meaning: it edits the vector mask when
the layer has one, and the Type Path only when it has none or with
`"typePath": true`, which Direct Selection sends for the Type Path
(PathRef::TypePath). path.list reports the Type Path under its own
`typePath` key, so `layerPath` still names the vector mask, and an edited
Type Path keeps the knot cap type.attachToPath enforces. Both points were
found by CodeRabbit.
The canvas also draws the Type Path with the other tools, as it does a
shape's path: the Pen's Auto Add/Delete edits it, and before this a Pen
click on the baseline added or deleted anchors the user couldn't see.
Co-authored-by: DuBearte <130190089+DuBearte@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
main split widgets.rs into widgets/; the scroll arrows move to widgets/dropdown.rs
and their tests to widgets/tests.rs, reading the offset through clamped_offset()
now that egui 0.37 keeps ScrollArea State::offset private.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The PR read and wrote Halftone Pattern as `Sz ` and `Ptrn` (enum `HlfT`:
`Dt `/`Crc `/`Ln `). The reference application's Filter Gallery uses `HlSz`
for the size and a `ScrT` enum (`ScrD` dot, `ScrC` circle, `ScrL` line), so a
Halftone Pattern saved there still opened as an opaque filter, and one we saved
lost its size and pattern there without a warning. The tests built their fixture
with the same keys as the code, so they passed either way.
Now the keys match, a default Halftone Pattern (HlSz 1, Cntr 5, ScrD) opens as
{size 1, contrast 5, dot} (fails before this commit), and descriptors with the
old keys stay verbatim. The docs no longer claim reopen checks in the reference
application.
Co-authored-by: denstu473 <63150090+denstu473@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
zlib-rs maps level 1 to deflate_quick, which writes static Huffman blocks
only, so every Compression::fast() caller (PDF export and print, brush
presets, PNG Fast) wrote 5-32% bigger streams than miniz's level 1 did.
Level 3 is the lowest level that is never bigger than miniz level 1 on the
docs/images photos, raw and Up-filtered (level 2 was +0.9% on one); it is
0-16% smaller and takes 1.2-3.4 times as long. The new test exports a
noisy-gradient RGB document as PDF and checks the image stream against
miniz level 1 of the same samples: 603007 vs 547315 bytes with fast(),
passes with level 3. The PR's doc no longer says the levels are unchanged.
Main moved the io integration tests into crates/io/tests/it/, so the new
deflate_compatibility.rs moves there and uses crate::common.
Co-authored-by: Rene Leonhardt <65483435+reneleonhardt@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>