Its translations are part of the combined catalog; this merge records the
branch so the PR closes as merged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Its translations are part of the combined catalog; this merge records the
branch so the PR closes as merged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds Swedish (sv, Svenska) as an interface language. The catalog combines
the open Swedish catalogs of #5985 and #9536: each row takes the best of
their translations, with one Swedish term per concept across the UI and
the reference application's established Swedish terms where they are
unambiguous, and the strings added since are filled in. 4,537 rows, every
measured key; plural rule one|other.
Where the rows come from: 2,710 both catalogs agreed on, 979 from #5985,
655 from #9536, 193 written new; 177 of the taken rows
were edited for consistency.
The tests that used sv-SE as an unsupported locale now use sw-KE and
fi-FI, and sv-SE / sv_FI.UTF-8 are positive cases. macOS lists sv in
CFBundleLocalizations.
Fixes#10020.
Co-authored-by: kler <966132+kler@users.noreply.github.com>
Co-authored-by: davvarn <1687962+davvarn@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The web build has no system fonts and embeds no craft fonts, so an Arabic
(or Japanese, Chinese, Korean, Hindi) interface had no font for its labels:
served fonts were registered only in the document text engine. A served
script fallback now also joins the end of every egui font family, and the
font stack theme::install_fonts rebuilds on a language change includes it
again. The web shell asks for the interface language's script font, not
only the browser languages'.
Review finding on #10006.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts in perf/budgets.toml and docs/scorecard.md: both sides added perf rows (P80 here, P81 and P82 from #9942). Kept all three; scorecard regenerated with cargo xtask scorecard.
egui 0.37 lays text out with the Unicode Bidirectional Algorithm, so the
display-order rewrite #5385 added for egui 0.36 now reverses Arabic and
Hebrew layer names and the in-app title a second time. Names go to egui
as typed; the title line asks for a left-to-right paragraph so
"— PhotoCraft" stays at the right; bidi_text::elide cuts a right-to-left
name at its logical end and egui puts the "…" at the left.
Tests lay the text out with egui and compare glyph positions.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
periodic::Composite said paint.cloneStamp for every tool; Healing Brush, History Brush and Pattern Stamp now go through it too. The command is passed in, and the live stroke passes its own.
* Honor CARGO_BUILD_TARGET in FreeBSD packaging for aarch64 release builds
package.sh took the arch from uname -m. The release VM is an x86_64 guest that
cross-compiles aarch64 with CARGO_BUILD_TARGET set, so an aarch64 build was
named freebsd-x86_64 and read binaries from the host target dir.
Read the arch from CARGO_BUILD_TARGET when it is set, and read binaries from
target/<triple>/release. Skip the man-page example and the --version run when
cross-building, since the aarch64 binaries cannot run on the x86_64 guest.
Unset CARGO_BUILD_TARGET keeps the old behavior.
* Reject unsupported CARGO_BUILD_TARGET and resolve host-tuple in FreeBSD packaging
Accept only x86_64-unknown-freebsd and aarch64-unknown-freebsd. Any other
target exits 2 before any build, so a non-FreeBSD target cannot be packaged
as an x86_64 FreeBSD archive.
Resolve cargo's host-tuple alias to the real host triple before choosing the
binary dir, so the path matches cargo's output.
* Skip the photocraft-cli --version run when CARGO_BUILD_TARGET is set
A target build's binary may not run on the host (aarch64 on the x86_64 VM),
and under set -e the smoke run would fail the release after the archive is
written. The native path (CARGO_BUILD_TARGET unset) still runs it.
Ports the wrap changes into main's split retouch_cmds/ and tools/retouch_ui. Mixer Brush keeps main's new LiveMixer and no longer wraps in this PR; it is left out of pattern_wrap so its commit still matches the live stroke.
A raw opened with a Lightroom or Camera Raw sidecar reported
"3 not supported: CropBottom, CropRight, RawFileName" (#9985): the crop
keys fell through to the unsupported list (CropTop/CropLeft were only
spared because they were 0), and the raw's file name was treated as a
setting.
The XMP reader now resolves the crop from the public crs: schema:
crs:HasCrop="True" with CropTop/CropLeft/CropBottom/CropRight, fractions
of the developed frame in the raw's stored orientation. A missing or
False HasCrop means no crop (the keys are stale and are not reported).
Fractions are clamped to 0..1; an inverted, empty or unreadable crop,
and a straightened one (CropAngle not 0, whose rectangle is measured on
the rotated frame), is not applied and its keys are named as not
supported. RawFileName is bookkeeping.
Camera Raw's settings have no crop stage, so the crop is applied the way
the sidecar's other raw-stage settings are: Open re-develops the raw
(RawTuning::crop) and crops the upright frame after mapping the
fractions through the raw's Orientation, before the Camera Raw step, so
the post-crop vignette centres on the crop. This works with the dialog
and with Open in Camera Raw off; the dialog previews the whole frame.
PhotoCraft's own sidecar keeps the crop when it is rewritten, and Load
Settings in the Camera Raw Filter names a crop as not supported.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two more #9648 review points on the ui.pointer path.
- A tool drag begun before the dialog opened got its release but lost
its moves under the dialog, so a Brush stroke spanning the opening
committed a different path. While a tool drag is open, its moves now
reach the tool too, not only the release.
- The held locator press (__pvPointerDown) stayed set when the primary
release went elsewhere (Space held, another dialog on top, the dialog
closed), so a later hover moved the preview with no button down. Any
primary release now clears it in every open dialog before the modal
handlers run, without consuming the event.
Tests: the spanning Brush stroke keeps the points it got under the
dialog and the preview stays put; a Space-held release and a release on
a Color Picker over the dialog both end the press, so a later move
doesn't centre.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Follow-up to the #9648 review.
- ui.pointer centred the preview on any Move, so a hover (or an event
with no kind, which defaults to a move) moved it. It now centres on a
primary press and on moves while that press is held, remembered across
requests in a private dialog field (__pvPointerDown, stripped from the
params like the other view fields), the way the canvas follows a held
mouse button.
- The locator consumed the release of a tool gesture begun before the
dialog opened, leaving the stroke unfinished. A primary release while
a tool drag is open now reaches the tool.
- center_at narrowed to f32 before clamping, so a huge finite point
(1e40) became infinity and clamp_center put that axis at 0 instead of
the far edge. It clamps in f64 first now.
Tests: a bare move doesn't centre and a press then move does; a Brush
stroke begun before the dialog ends under it as one history step; 1e40
and -1e300 clamp to the far and near edges.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Canon's subsampled lossless JPEG (Y 2x2 or 2x1, Cb/Cr per MCU) with its
own prediction chain, de-sliced into YCbCr planes, bilinear chroma, a
BT.601 inverse and the per-shot sRAW gains from ColorData (version 15),
developed like a three-sample LinearRaw so white balance, the fitted
camera look, tone and the Camera Raw dialog all apply. Verified on EOS
6D Mark II mRAW files; other bodies decode with a warning when their
gain offset is unknown.
Review follow-up for #9879: the engine reads `"color": null` as no colour and uses the `filter` preset, so `complete_params` now completes it the same way instead of leaving the dialog on its neutral colour. Regression case added to `recorded_photo_filter_settings_complete_to_their_preset_colour`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review follow-up for #9879: round-tripping every recorded adjustment
through the engine rewrote values it should keep (a float Photo Filter
colour became hex, a pencil curve's map came back smoothed while its
`smooth` stayed in the stored params, so each confirmation smoothed it
again). `complete_params` now adds just the colour a Photo Filter that
names only its `filter` lacks, and leaves every recorded value and every
other adjustment untouched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
In Photoshop, while a filter dialog with a preview window is open (Smart
Sharpen, Unsharp Mask, Reduce Noise...), the pointer over the document is
a small square and a click centres the dialog's preview on that point.
PhotoCraft ignored canvas clicks under the dialog, so the preview could
only be panned from inside it.
The shared in-dialog preview (filter_preview_ui) gains locator(), the top
dialog when it shows the preview window (not while the image holds the
Blur Gallery pins or Replace Color's armed eyedropper), and center_at(),
which writes the clamped document point into the dialog's existing
private view fields. No command runs, so the document and its history
are untouched.
The canvas's dialog input routes a press or drag on the free canvas
through the view transform to center_at, with any tool, and shows a
small square locator cursor there, like the Color Picker's pipette.
Space, middle-drag and the held zoom keys still pan and zoom. ui.pointer
does the same for agents; before, its events reached the tool under the
dialog and a Brush gesture painted.
Tests: harness tests through the real canvas path (Smart Sharpen click
and drag, Unsharp Mask off-image clamp, rotated view, ui.pointer, and a
click with no dialog still painting), plus a unit test for center_at.
docs/ui-parity.md gets the row and revision entry.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
While a Color Picker opened for another dialog's colour field is open
(Edit > Fill's Color..., Stroke, a filter's or adjustment's colour,
Replace Color's Result, New Document, guide and channel colours,
Preferences), its colour now shows in that field before OK, as in
Photoshop. The dialog's swatch and its own preview follow the picker.
The Crop shield's custom colour previews the same way.
The new color_picker_ui::live_preview module saves the target when the
picker opens. The parent dialog's fields are kept on PhotocraftApp, not
in the picker's fields, which automation can rewrite. On each colour
change it resets the target to that saved state and applies the new
colour, so a dialog that derives settings from the colour never builds
on a colour it was only shown. Cancel, or a picker closed any other
way, restores the target exactly. OK restores it and then sets the
field as before. The document is never touched, so the parent dialog's
OK still makes the single history step, and the parent's own Cmd+Z
history records nothing while a colour is only previewed. Confirming a
dialog while its colour picker is still open is now refused for every
dialog, not only for the guide layout. ui.dialog.set on a picker shows
the colour at once.
Command targets (text colour, gradient stops, artboard, count groups)
and the pasteboard colour are unchanged: open PR #9757 previews text
and gradient stops, the rest are follow-ups. The foreground/background
pickers keep changing the swatches only on OK, as in Photoshop.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review follow-ups for #9879:
- Photo Filter's Filter / Color choice lives in the editor memory, so an
opening that switched it and was cancelled left the next opening on
Filter with a custom colour (the colour button disabled). Each opening
now forgets it once, and the editor derives it from the colour again.
- A smart filter's recorded adjustment settings are completed through
the engine's parser before they fill the dialog, so a recorded Photo
Filter that names only its `filter` shows and saves that preset's
colour instead of the dialog's neutral one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Image › Adjustments › Photo Filter… now opens at the filter, colour,
density and Preserve Luminosity of its last OK, as Photoshop's does,
without holding ⌥. The other adjustment dialogs keep opening at their
neutral settings (⌥ still reopens them at the last ones, #8900).
`adjust_dialog::open` picks the last settings for the kinds listed in
`REMEMBERS_LAST`; the old neutral behaviour is `open_neutral`, which the
smart filter editor now uses so a smart filter's dialog always shows that
filter's own settings.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Main now carries #9031 (Halftone) and #8923; this branch keeps only what it adds on top: the guard for a setting beside a Gallery stack's effect list, its test, and the PSD open warning for patterns that paint nothing.
CodeRabbit: removing the popover's inline eyedropper dropped the only test
of screen sampling from the shape Fill/Stroke controls. The sample now
waits in the Color Picker until OK, then sets the next shape's colours.
docs/ui-parity.md: the status line's preference count matches the
scorecard (12 of 182).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TYsCL5E2GM6hUabP9dmYAv