Repin craft-pdf, craft-prepress, krilla and the hayro crates to d30598d.
flatten_document writes the X-1a file at PDF 1.4 with GTS_PDFXVersion
PDF/X-1:2003 in the Info dictionary and XMP, so DesignCraft's header
rewrite and incremental update are gone. check_pdfx1a refuses DeviceN
NChannel attributes and warns about an ICC v4 output intent profile in a
file of PDF 1.5 or lower; the export names the built-in Coated FOGRA39
(v4) when it is the intent. validate-pdf lists warnings apart from issues.
Two corrections to the message, both reported by CodeRabbit on #1371.
switched_from being None only means this start did not switch from a
backend that never showed a frame. A backend that failed on an earlier
start is remembered in the record, and choose() then picks the first
candidate that is not in that list -- so calling it the "first choice"
misstates how it was arrived at. The line now says either "nothing has
failed before" or which backends had, so the log says what happened.
The comment claimed that reaching this line means WGPU_BACKEND is
unset. wgpu-types returns an empty mask for an empty or unrecognised
value and begin() only returns early for a non-empty one, so such a
value reaches here with the variable set. The comment says that instead.
Adds .coderabbit.yaml so CodeRabbit no longer reviews or comments on PRs opened by dependabot[bot].
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
standard x1a writes the PDF/X-4 print file, flattens it with
craft_prepress::flatten::flatten_document at the flattener preset's
resolution (high 300, medium 150, low 72 ppi; default 300), then makes it
PDF/X-1a:2003: the PDF 1.4 header, GTS_PDFXVersion and GTS_PDFXConformance
in the Info dictionary and XMP. check_pdfx1a refuses transparency (soft
masks, transparency groups, CA/ca below 1, blend modes) and color outside
DeviceCMYK, DeviceGray, Separation and DeviceN; export fails when the
flattened file still has either.
The X-1a path skips DesignCraft's RGB transparency flattener. Soft drop
shadows are drawn in their swatch's inks through a coverage mask, so
craft-prepress flattens them per ink instead of an RGB image of the
object. validate-pdf takes --standard x4|x1a; the Export PDF dialog gains
the PDF/X-1a:2003 preset and standard and a Transparency Flattener preset.
Found while testing crates/engine/src/cmd/table.rs:1396-1403
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Export PDF, File > Print and Separations Preview share one print
pipeline: print_pdf builds the print PDF of a set of pages (flattener,
rasterized layers and effects, PDF boxes, the writer, the blend space
outside print color), pdf_export adds Export PDF's view options for
file.exportPdf and file.print, and the preview separates print_pdf's
page per spread. Gradient stops on spot swatches resolve through the
same InkSet as fills (inks::spot_output). The hayro patches move to
craft-libs 8fbfea8 with craft-pdf, craft-prepress and krilla.
Each visible spread is exported in memory as Export PDF writes it (its
print settings, the document bleed) and separated with
craft_prepress::plates off the UI thread, at a resolution matched to the
zoom. The canvas draws all inks composited, one plate as film or the ink
limit over gray; Window > Output > Separations Preview lists the plates,
spot plates by name, with each ink's percentage under the pointer.
view.separations takes spot ink names and reports the plates, ui.inspect
shows the preview's sheets, and ink.separations gives coverage and the
areas over an ink limit to automation. The renderer's GCR separation view
is gone; its plate passes stay for the CMYK blend space.
`Fallback::begin` returns before the fallback runs when `WGPU_BACKEND`
is set, logging that wgpu picks the backend. So the line below it is
reached only when the variable is *not* set -- and it said
"(WGPU_BACKEND overrides)", which reads in a log as though the user's
environment had chosen the backend when nothing had.
before: graphics: DirectX 12 (WGPU_BACKEND overrides)
after: graphics: DirectX 12 (first choice; set WGPU_BACKEND to choose another)
The report quotes that line as evidence of which backend a start used,
which is exactly the wrong conclusion to be led to.
Part of #1343.
Repin craft-pdf, craft-prepress and krilla to craft-libs#190 at 8fbfea8.
Plain exports with a CMYK blend space use print::cmyk_blending (CMYK page
and transparency groups, no output intent, no /Trapped); PDF/X-4 keeps
print::settings with the working space as GTS_PDFX output intent, as
InDesign writes an output intent only for PDF/X.
A stop can name a colour swatch (spot, process, tint, [Paper]) next to its
display colour. swatch.create {gradient}, object.gradient, the Gradient panel
and IDML StopColor keep the reference; renames follow it, swatch edits refresh
the stop's colour and a deleted swatch leaves a plain colour. The field is
optional and omitted for plain stops, so saved documents load and save as
before.
file.print builds its PDF with the same pipeline as file.exportPdf
(flattener, placed PDFs with hidden layers and soft effects as images,
print color, view options), accepts Export PDF's settings and returns its
warnings, instead of calling designcraft_pdf::export_pdf with the
hard-edged vector drop shadow.
Groups are pushed with Surface::push_group, so Knockout Group writes
/K true and Isolate Blending /I true, each on its own. File > Export PDF
no longer rasterizes knockout groups.
Overprint Fill, Overprint Stroke, Overprint Gap, overprinting text (ruby,
kenten) and 100% [Black] under Appearance of Black are pushed through
Surface::push_overprint with OPM 1, around the paint they apply to.
PDF/A exports are RGB and leave overprint out with a warning.
Found while testing crates/engine/src/cmd/object.rs:710-711 (also :975)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The Ink Manager settings stay saved by swatch name; PDF export and the
ink commands build a craft_prepress::InkSet from the document's spot
swatches and those settings, and ink.options refuses alias loops. The
hayro crates come from craft-libs at the craft-pdf revision, which
craft-prepress's plate mode needs.
CMYK print exports (a CMYK blend space, and PDF/X-4) take their blending
groups, output intent and /Trapped from write::print::settings instead of
rewriting krilla's bytes. The output intent is the CMYK working space,
Coated FOGRA39 when it has none or its profile can't be written. PDF/X-4
still adds the XMP metadata date and version ID krilla doesn't write.
RGB and PDF/A exports are written as before.
Only path-filtered workflows run on pull requests today, so most PRs land
without a test run. Port of the wordcraft#800 workflow: Linux, standard
GitHub-hosted runner, read-only token, no secrets, craft-fonts pinned to
the release.yml commit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The keymap reserves craft-keymap's system chords for its platform (window switching and closing, the virtual consoles). Binding one is refused with "<chord> is reserved by the system"; an imported set's key on one is left out and the import summary counts it. No default shortcut sits on a reserved chord on any platform (test).
moppeeerandClaude Opus 5.5
aiongg
Hamza
Brandon ThomasandClaude Opus 5.5
Dylan M. Taylor
NetterSilverandClaude Haiku 5.5
con3codeandClaude Opus 5.5
macadminHUandcoderabbitai[bot]
FredAppsandClaude Opus 5.5
Brandon ThomasandClaude Opus 5.5
Daniel