Main split crates/engine/src/prefs.rs into prefs/ and moved the
gpu_startup tests to gpu_startup/tests.rs: read_section now sits beside
read_lenient in prefs/store.rs, and the malformed-sibling assertions
moved with the tests. Doc status lines and revision rows keep both sides.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CodeRabbit on #9775: FontLibrary::finish wrote a failed type.font.favorite to the status bar as plain information; it now sets status_error like the other status writes. The regenerated scorecard counts 12 of 179 preferences that do nothing; ui-parity.md and gaps.md still said 178.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
CodeRabbit on #9775: the encoder picks P5/P6/P7/PFM from the image, not the name, so a .pbm, .pgm or .pfm picked in Save As held other data (an RGB document as P6 in a .pbm). Only the neutral .pnm is offered; File > Open still reads every Netpbm extension.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
CodeRabbit on #9775: newVideoLayerFromFile placed the layer while its content was still empty, so an active artboard always took it and clipped footage centred elsewhere; the first frame is now the content before placement, as in frames_to_layers (footage_off_the_active_artboard_stays_out_of_it fails without it). The merge also left a second check_group_depth ("Duplicate Layers") before main's "Copy Layers" one in layer.copyToDocument, so main's message could never show; it is removed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
CodeRabbit on #9775: paint.stroke with fill:true builds the coverage, its feathered copy, the selection read and the mask over the region, up to the whole canvas (file.new allows 300000 x 300000), with infallible allocations that abort the process when memory runs out. The peak is now reserved through the fallible allocator first, and the selection is read through it; fill_survives_allocation_failure fails without this (the error came from a later blend).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Main's #9849 lets dropdowns grow back after a short list, and its test
`clearing_a_query_that_matches_nothing_grows_the_menu_back` expects a
search with no matches to shrink the font menu. #6526's menu stayed at its
full 534 px: its font list never shrank (auto_shrink y off), and the
vertical separator between the class list and the font list took the row's
whole available height. The list now shrinks, and the line is drawn over
the laid-out row instead, so the menu ends with the class list.
The test now checks that the menu ends with the class list (it can't drop
below half with the class list beside the list), and a new test pins that
the wheel past the end of the list, or over the class list, never moves
the header, also in a window too short for the full menu.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Upstream split canvas.rs and lib.rs into modules and moved the dialogs
under dialogs/. The preview module moves with them
(dialogs/color_picker_preview.rs); its canvas hooks now sit in
canvas/display.rs and canvas/textures.rs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014QvfaMyAX4i4PbKBMnvFj6
Pressing a pixel tool on a type, shape, Smart Object or fill layer opens the
"Rasterize?" prompt. On OK, confirm() ran layer.rasterize.* and then replayed
the stored press as ToolEvent::Down/Up, so the brush, eraser or bucket acted
on a click the user had made only to get past the dialog. Photoshop 2026
converts the layer and does nothing else; the next click paints.
confirm() now only rasterizes and returns {"rasterized"}. The prompt still
records the tool and point in its fields. The control-protocol doc and the
TOOL-213-5 scorecard row describe the new behaviour.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With Show Transform Controls on, dragging a handle opens a Free Transform
session, and while it is open every press goes to the transform tool. A click
outside the box commits it on release and returns true, which swallowed the
event, so Auto-Select never saw that click and the user had to click again.
After the outside click commits and the session is gone, replay the press as a
plain Move tool click at the same point, so Auto-Select, Shift/Cmd and the
empty-spot marquee behave as for any click. A commit handed to a background
job keeps the session open, and then nothing is replayed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main's #9849 bans egui::ComboBox; the preset menu #3558 added now uses
craft_ui::ComboBox, three characters past rustfmt's 160-column limit.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
merged_planes scanned each composite band twice on the calling thread
for alpha, encoded every range into per-channel Vecs one push at a time,
then copied them into the planes. The workers now write their slice of
every plane directly, check alpha as they go, and run a loop
monomorphised per sample type. The 8-bit deinterleave of layer pixels
used one thread per channel with a strided iterator; it now splits the
pixels into one band per core and writes each plane's slice in place.
Same arithmetic, same bytes; a test pins merged_planes against the
previous per-band implementation for RGB, Gray, Lab and CMYK at 8, 16
and 32 bits, with and without the matte.
bench_export .psd, 36 MP, 3 layers, 16 interleaved rounds of 7 reps
(median of medians / min), Ryzen 16 threads, Linux:
709.4 / 413.7 ms -> 540.8 / 318.9 ms (-24 % / -23 %)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A ZIP save encoded the thumbnail and the 1024 px composite preview as PNG
on the calling thread before any tile was compressed. The composite alone
is a ~200 ms single-threaded deflate on a 36 MP document, about 40 % of a
full save. The previews are now encoded on their own threads while the
document is collected and its new objects are compressed on all cores;
the bundle is assembled afterwards in the same order, so the bytes are
unchanged. Directory saves encode them up front as before. wasm stays
sequential.
bench_export, 36 MP, 3 layers, 16 interleaved rounds of 7 reps (median
of medians / min), Ryzen 16 threads, Linux:
.pcraft 517.9 / 471.1 ms -> 319.3 / 287.8 ms (-38 %)
.pcraft with cache 360.1 / 334.1 ms -> 277.5 / 268.3 ms (-23 %)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Real msiexec runs (per-user and all-users, silent and wizard) found two defects
that the metadata fixtures cannot see:
- "Only for me" published ALLUSERS=2, so FindRelatedProducts treated the run as
per-machine and skipped the existing per-user copy. An interactive upgrade
left two registered products and uninstall left files behind. The page now
leaves ALLUSERS unset for a per-user choice.
- Windows Installer resolves Program Files once, for the context it started in,
so choosing "Everyone" in the wizard installed into the user profile with a
machine-wide registration. The wizard now sets the new-install folder itself
after the choice (Program Files, Program Files (x86) for 32-bit, or the
per-user Programs folder). Silent installs are unchanged.
check-install-policy.ps1 now fails if either regresses.
A new layer with an open group active went above the group whenever the
group had any lock (all, pixels, transparency or position). The
reference application puts it inside, on top of the group's layers, even
when the group is locked with Lock All. Locks no longer keep New Layer,
New Group, adjustment and fill layers out. Layers that are edited right
after they are made (Type, Shape, Place's transform box, Paste, the
Gradient tool's live gradient, video layers, Render, Vanishing Point, a
new frame) take Place::AutoEditable and keep the PR's rule: they go
above a locked group, as before.
Stamp Visible with an open group active put the stamp inside the group,
where the group's opacity, mode or mask applied to the composite a
second time. The reference application puts it directly above the
group; with a layer inside the group active it still lands above that
layer, inside the group.
Co-authored-by: SckraT <70698629+SckraT@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#9321's Direct Selection change calls crate::vector_ui::active_type_path, which
main's layout refactor (#9758) made a private path_ops item, so ui-egui did not
build after the batch's main merge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Addresses CodeRabbit review comments on PR #9841:
1. Clear flag on pointer release (primary or secondary),
even when release occurs outside canvas and produces no event.
This prevents the next ordinary canvas click from being incorrectly suppressed.
2. Block tool input (Down, Move, Up events) on outside press before tool handlers run,
not just on click. The press that dismisses the menu is now detected early via
and its gesture is fully consumed (no stroke starts, no drag
moves, no click action).
Files changed:
- crates/ui-egui/src/canvas.rs: Added pointer release flag clearing, added
detection, and gated all tool_event calls with this flag.
After a committed crop the Crop tool shows its untouched default frame
again, and H/P toggled Show Cropped Area and Classic Mode on it, so the
Hand and Pen shortcuts did nothing. Both keys now act on the crop only
while one is in progress (a frame drawn or changed), as in Photoshop.
* refactor(ui): build every dropdown on craft_ui::ComboBox
Repin craft-libs to the revision with craft_ui::ComboBox, whose menu grows back
after its contents shrink while it is open (egui#5225), and ban egui::ComboBox
in clippy.toml.
* fix(type): the font menu grows back after a search
The search field is the ComboBox header and only the family list scrolls, so
the menu follows the list's height: no unfiltered height kept while searching.
* chore(deps): craft-libs 60d60d48 (dropdown fixes)
* chore(deps): craft-libs 6af63481 (dropdown fixes)
* chore(deps): craft-libs 4d7b82f7 (dropdown fixes)
* chore(deps): craft-libs 92d392cf (#387 merged)
Ports the size, leading and tracking preset chevrons onto main's split type
tool (tools/type_tool/panel.rs, next to num_field, which now names its scrub
icon) and fixes what review found:
- The lists differ from the reference app: Size is 6, 8, 9, 10, 11, 12, 14, 16,
18, 24, 30, 36, 48, 60, 72, 150, 300 pt (the PR had 96, no 16/150/300);
Leading is Auto plus the same list (one shared pub(crate) const); Tracking is
-100..-5 | 0 | 5..200 with separators (the PR lacked -5/5 and separators).
- The PR's layout test was red on CI: the chevron ended ~20 px short of the
column. The numeric field now gets the column width minus the chevron's
exact button width, and the test is strict again (the chevron ends at the
panel edge within 0.5 px, the field's suffix slot matches Transform's).
- In mm the labels read "2.1166668 mm" and a pick stored 12.000001 pt. Labels
now have at most two decimals, a picked preset stores its point value
directly, and "selected" uses a relative tolerance.
- The chevrons have accessible names (Font size / Leading / Tracking presets,
translated in every complete catalog); tests find them by name.
New tests: the literal lists (fail on the PR's lists), the strict layout test,
mm labels and exact 12 pt, a preset while editing a span, and presets without
a type layer setting the defaults with no history step.
Co-authored-by: Ni-zav <106657905+Ni-zav@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main moved the engine commands into crates/engine/src/cmd/ and the
dialogs and panels into submodules. The PR's placement calls follow
them, and main's new insert sites (new_adjustment in adjust_params.rs,
Place in file_cmds/open.rs, shapes in vector_cmds/shape.rs) use
layer_place::insert_new like the PR's originals. The New Group dialog
moves to dialogs/new_group_ui.rs; the Layers panel flyout's New Group…
opens it.
Main's Session::precheck now also checks layer.new.group's and
layer.new.groupFromLayers' new params (name, color, mode, opacity,
position), so precheck_agrees_with_execute holds. Main had since added
"New Group…" to every catalog: the PR's duplicate rows are dropped.
Main's newer tests that made a layer with an open group active now
open the group last, as the PR did for the older ones.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Cmd armed the new Distort and Skew pointers in every transform mode, while the
drag did something else: in Rotate mode a Cmd-drag on a corner or edge rotates
but showed the dart or skew arrows; a Skew-mode corner with Cmd showed the dart
but slides along an edge; Distort mode with Cmd drew skew arrows on edges it
refuses. Cmd also hid the curved rotate arrow just outside a corner (the Alias
commit pointer showed) although a Cmd-drag there still rotates.
One helper, handle_indicator, now answers "what does this drag do" with the
same rules as the drag (drag_mods, mode_mods, apply_mode_drag), and both the
cursor and the drawn pointer use it. The pointer is drawn from the canvas's
cursor branch, so a guide, pan or zoom cursor over a handle no longer gets a
second pointer on top, and cursor() no longer hides the pointer while the box
is being applied (the overlay draws nothing then). During a drag the rotate
arrow shows only while the drag rotates (also a Rotate-mode handle drag), not
when a scaling corner strays past another corner. The warp split guide reads
sticky keys like the cursor does.
Co-authored-by: Catur Sura <8650781+catursurapam@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>