Conflict: ROADMAP.md: next's FONT1-3 lines with the branch's CJK1 line
Conflict: docs/cjk.md: the branch's rows, with next's aki below and Japanese composer text kept in the leading model row
Conflict: crates/fonts/src/fontdb.rs: both fields (family/style aliases and the ICF cache)
Conflict: crates/compose/src/ruby.rs: the branch's ruby module; its PlacedGlyph (and kenten.rs's) carry next's rendered_char
Conflict: crates/compose/src/lib.rs: modules both; rtl and para_block both; the branch's grid placement (gyoudori, grid references) replaces next's snap, its paragraph-gyoudori span uses next's line_metrics with mark leading and #330's aki below advance; a wrap push moves the line by whole grid lines and its grid end with it; next's trimmed spaces and drop cap, then the branch's ruby and kenten annotation
Conflict: crates/compose/src/shape.rs: next's pieces loop and StyleRuns with the branch's auto tate-chu-yoko split (after marks join their base) and fit, and Upright mode in every shape_segment call; next's label fallback passes the mode on
Conflict: crates/compose/src/tests.rs: both sets of tests appended at the end
Conflict: crates/idml/src/export.rs: next's metadata and text variable writers, then the branch's CJK settings writers
Conflict: crates/idml/src/tests.rs: both tests
Conflict: crates/ui-egui/src/dialogs.rs: the branch's Ruby dialog default section (next's duplicate characterStyleOptions default dropped); next's section arms with the branch's style_cjk::section_ui
Conflict: crates/ui-egui/src/lib.rs: next's missing-fonts test; the branch's side repeated control_bar_shows_its_second_row (next's newer copy stays) and key_names_sent_as_text_are_dropped, whose IME version replaces next's copy in place
Conflict: crates/engine/src/cmd/file.rs: next's file.open params (InDesign .indd/.indt) with the branch's package-parts warnings
Conflict: crates/engine/src/cmd/interchange.rs: next's file.openIdml params (InDesign files) with the branch's warnings list; import_report reads the branch's Imported report, then next's resolve_pdf_crops; open_idml keeps next's .indd conversion and title, then takes import_report's warnings after the conversion's; next's comment on not saving over the .idml or .indd
Conflict: crates/engine/src/cmd/style.rs: both params notes in the paragraph style edit spec (startAt null, kinsoku null); create_para with next's set_json_over, continue_numbering and the branch's explicit_default_kinsoku; edit_para calls both helpers; both tests
Conflict: crates/ui-egui/src/dialogs.rs: next's section arms (advanced to bullets) and the branch's CJK section arm; next's paragraph helpers and paragraph_local before the branch's section_list (cjk flag, scrolling list; doc comment stays removed as on next)
The branch's "Lines:" row in the shared i18n table is dropped: next already has the key (drop caps).
Conflict: crates/engine/src/cmd/style.rs: next's create_para attrs (set_json_over) followed by the branch's continue_numbering
Conflict: crates/ui-egui/src/dialogs.rs: the branch's paragraph section helpers and sections before section_list (whose doc comment stays removed as on next); its tests, without its copy of paragraph_style_options_tabs_section_adds_a_stop_on_its_ruler (next has it from #343)
Conflict: crates/ui-egui/src/i18n.rs, i18n/it.rs, i18n/ja.rs: both blocks; the branch's [Default] rows dropped (next has the key from New Document)
The Tabs commands check the list (TabStop::checked_list) before they write it
and return that list, with the index of the stop they set in it, so the result
matches what the paragraph or style holds.
Conflict: ROADMAP.md: next's IDML interchange line with the branch's sentence on placed PDFs, $ID/ font names and packaged links appended; next's INDD line kept
Conflict: crates/engine/src/tests_idml.rs: next's drop cap test and the branch's two placed PDF tests
Conflict: crates/fonts/src/fontdb.rs: next's plausible/measured_top/typo_ascent and the branch's plain_family; next's canonical_family lookups, with plain_family in place of without_format_suffix in style_traits, face (two) and has_family; the branch's has_exact_family dropped (has_family uses canonical_family)
Conflict: crates/idml/src/import.rs: next's char_attrs (tab-split family and composite font names), falling back to names::font_family_in
Drop cap style through IDML as a Dropcap nested style, spaces after the cap, one-line caps, nested styles after the cap, own nested-style lists clearing an inherited drop cap style (cells and notes too), cleared styles round-tripping, tabs in right-to-left drop caps.
Conflict: ROADMAP.md: next's INDD, Interface language and Show CJK Features lines (the PR's main-merge resolution, already on next)
Conflict: crates/compose/src/lib.rs: next's warichu placement followed by the trimmed trailing spaces
Conflict: crates/compose/src/tests.rs: next's side; the PR's side repeats helpers and tests next already has
Conflict: crates/idml/src/import.rs: next's range_format
Conflict: crates/idml/src/tests.rs: next's side, with the PR's drop_caps_import_and_round_trip and its new drop_cap_fixture and nested-style clearing tests in place of next's older drop cap test
The style path of the Tabs commands writes its list through
ParaAttrs::set_json_over, the same call format_paras makes for
paragraphs, so TabStop::checked_list sees every list a Tabs command
writes.
guide.add takes orientation: both with point: [x, y] and adds a
horizontal guide through y and a vertical guide through x in one edit.
Cmd-drag from the ruler corner uses it.
Cancelling a drag, a preview or an IME composition restores the
document, its selection and now its revision. Revisions come from one
counter, so the numbers a cancelled preview used are never handed out
again and caches keyed by (uid, revision) never show a cancelled
preview.
The Type tool shows an input method's marked text in the story as a
preview of text.insert, underlined (the converting clause twice), with
the candidate window at the caret. The commit is one text.insert and one
undo step; a cancelled composition leaves the document as it was. Keys,
shortcuts, the clipboard and Undo wait while the IME composes; a click,
a tool or document switch, or any journaled command keeps the marked
text as typed. On macOS the app tells the input context to drop marked
text the tool kept.
With a split window, only the pane last clicked takes keyboard events
(both panes saw them, so typing went in twice).
The web build drops only real key names from text events, and nothing
while its text input has the keyboard, so autocompleted words survive.
The control channel's ui.ime, the headless MCP backend's ui.ime and the
MCP ime tool send IME events.
The placed PDF's asset stores its Crop to choice (crop, art, trim, bleed,
media, or the content bounds with visible or all layers) and where that
box sits on the page. The graphic spans the box: the screen, PDF export,
rasterised exports and clipping-path tracing draw only that part of the
page, so a larger frame doesn't reveal the bleed.
IDML export writes the choice as PDFCrop, with GraphicBounds spanning the
box. Import reads PDFCrop back; the engine finds the box on the PDF page
after import (and after relink or update) and maps it onto GraphicBounds,
or gives a graphic without bounds the box's size. Documents saved without
a choice load with the crop box. The Place PDF dialog offers the content
bounds too, and links.list reports the choice.
IDML has no drop cap style attribute. Import reads the nested style list's leading "through 1 Dropcap" entry as the drop cap style; export writes the drop cap style as that entry. The drop cap style counts as that entry when composing, so nested styles after it start after the drop cap.
Preferences > Units & Increments > Origin sets DocSettings.rulerOrigin
(document.preferences {rulerOrigin: spread|page|spine}, IDML
RulerOrigin). Rulers and the X/Y fields (Control, Properties, Info,
transform.set) measure from that origin plus the zero point; with a page
origin the horizontal ruler restarts on each page and an object measures
from the page it is on.
Cmd-drag (Ctrl) from the ruler corner adds a horizontal and a vertical
spread guide through the drop point. The live preview while dragging the
zero point is kept on the pasteboard like the committed value.
Translations for Origin, Spine and Lock/Unlock Zero Point.
The layout-parity files (ROADMAP, compose lib/tests, engine text, fonts fontdb, idml import/tests, canvas) resolve as in the merge of main into feat/layout-parity.
Conflict: ROADMAP.md: as feat/layout-parity's merge
Conflict: crates/compose/src/lib.rs: as feat/layout-parity's merge; this branch's trimmed trailing spaces follow main's warichu::place and come before the drop cap splice
Conflict: crates/compose/src/tests.rs: both sides' tests
Conflict: crates/engine/src/cmd/text.rs: as feat/layout-parity's merge; this branch's TypingFormat/typing_format and main's warichu_cmd
Conflict: crates/fonts/src/fontdb.rs: this branch's styles/traits_of/linked_style over style_traits; style_traits falls back to the family without main's format suffix (exact_style_traits)
Conflict: crates/idml/src/import.rs, crates/idml/src/tests.rs, crates/ui-egui/src/canvas.rs: as feat/layout-parity's merge
Conflict: crates/ui-egui/src/control.rs: main's page_index and automation_contract_tests, then this branch's tests module
Conflict: crates/ui-egui/src/lib.rs: this branch's ClipboardFn and main's Inbox of ImportedFile
Conflict: crates/ui-egui/src/menus.rs: main's Quick Apply tests, then this branch's copy and paste keys test
Conflict: ROADMAP.md: this branch's Text engine, IDML interchange and Right-to-left lines; main's FID2, PDF export and Interface language lines
Conflict: crates/compose/src/lib.rs: list labels as main's number_label/bullet_label/empty-paragraph arms, kept in this branch's list_label form that marks label glyphs; main's layout_line call (ends_para, forced_mid) and warichu::place, then the drop cap splice; rule above uses main's first-line column/end_x span with this branch's rule_color; caret(): drop-cap branch, then main's stacked-row metrics
Conflict: crates/compose/src/tests.rs: both sides' tests
Conflict: crates/engine/src/cmd/text.rs: type.para doc lists the drop cap fields and main's rule fields
Conflict: crates/fonts/src/fontdb.rs: both (plausible/measured_top/typo_ascent and without_format_suffix)
Conflict: crates/idml/src/import.rs: main's table style regions, borders and alternating fill/stroke attrs (this branch's alt_fills helper dropped, unused); underline/strikethrough colour from main with this branch's tint(); frame_options takes the object-style chain's TextFramePreference elements, the frame's column rule colour resolved after it
Conflict: crates/idml/src/tests.rs: main's tests, plus this branch's added tests and helpers
Conflict: crates/ui-egui/src/canvas.rs: main's compose::highlight_quads, then this branch's drop-cap highlight
The zero point is document state (DocSettings.zeroPoint, zeroPointLocked):
an offset from the top-left corner of each spread's pages, so every
spread measures from the same relative place and split views share it.
view.zeroPoint {at, reset} and view.lockZeroPoint {on} set it; moving it
is not an undo step and undo/redo keep it. Values are kept on the
pasteboard, from commands, files and IDML.
Rulers and the X/Y fields (Control, Properties, Info, transform.set)
measure from it. IDML reads and writes Document ZeroPoint.
Dragging from the ruler corner places it where released, with crosshair
lines while dragging; a click that stays in the corner does nothing; a
double-click resets it; the corner's context menu locks it.
style.paragraph.create and style.paragraph.edit store an explicit
startAt: null as Continue from Previous Number, so it overrides a Start At
the style would otherwise inherit.
Export writes a PDFAttribute with the page number and PDFCrop="CropPDF":
the graphic bounds already span the page's crop box, so the frame clips
to the chosen trim, bleed or media box as before. Import reads the page
number back instead of always showing page 1.
A placed PDF page that the PDF does not have (a hostile PageNumber, or
any document holding one) is left out of PDF export with a warning
instead of failing the whole export.
Conflict: ROADMAP.md: kept both entries (New Document, then Numbering & Section Options).
Conflict: crates/ui-egui/src/dialogs.rs: width match keeps "newDocument" => 960.0 and adds "numberingSection" => 560.0, dropping the PR's "newDocument" if is_rtl => 380.0; tests keep next's helper block and add only the PR's numbering_and_section_options_keep_their_actions_in_view.
Conflict: crates/ui-egui/src/i18n.rs: kept both TABLE tail blocks (File › New › Document, then Layout › Numbering & Section Options).
Conflict: crates/ui-egui/src/i18n/it.rs: kept both tail blocks.
layout.section refuses a startNumber outside 1..=99 999, and so does layout.documentSetup for startPage (both go through one start_number check against designcraft_doc::MAX_PAGE_NUMBER). Document::page_number walks continuing sections in a loop with saturating arithmetic, so start numbers loaded from a file can't overflow it, and book pagination saturates its running page count.
Review follow-up. Enter in the Save Document Preset popup created the
document and Escape closed the whole dialog: the dialog footer took
both keys while a popup was open. The dialog now leaves Enter and
Escape to an open popup (read before it draws, as egui closes a popup
on Escape that frame): Enter in the preset name saves the preset,
Escape closes only the popup. Clicking the name field no longer closes
the popup either (menus close on any click; this one only outside), and
the field takes focus when the popup opens.
Saved presets lived only in the UI state, so `file.new {preset: name}`
refused them. They are now engine preferences (saved with prefs.json):
`file.savePreset {name, …the file.new settings}` replaces a preset of
that name and refuses built-in names, `file.deletePreset {name}` removes
one, `file.presets` lists them after the built-in ones (`saved: true`),
and `file.new {preset}` makes documents from them by name. The Saved tab
reads and writes them through those commands.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflict: crates/ui-egui/src/lib.rs: both modules kept, mod new_document before mod object_control
Conflict: crates/ui-egui/src/dialogs.rs: the PR's translated ok_label plus next's findFont arm, translated the same way ("Replace All"); cancel_label is a match: newDocument "Close", findFont "Done", else "Cancel"
Conflict: crates/ui-egui/src/i18n/it.rs: both blocks kept (next's machine-translated and Tabs rows, then the PR's File › New › Document rows)
Conflict: ROADMAP.md: both lines kept (next's IDML1 line, then the Tabs panel line)
Conflict: crates/engine/src/cmd/mod.rs: both spec lists kept (thread::specs, then tabs::specs)
Conflict: crates/ui-egui/src/dialogs.rs: next's delete_paragraph_style and the PR's style_tabs both kept, without the section_list doc comment that next removed; in the tests the PR's copies of draw, ruled_paragraph, undo_steps, the rule tests, command_app, the generic command tests and the text frame options test duplicated next's identical ones and were dropped, keeping only the PR's new Tabs section test
Conflict: crates/ui-egui/src/i18n/it.rs: both blocks kept (next's machine-translated rows, then the Tabs rows)
Conflict: crates/ui-egui/src/i18n/uk.rs: both blocks kept (next's machine-translated rows, then the Tabs rows)
type.tabs.* edit the stops of the selected paragraphs or, with a style
parameter, of a paragraph style. Positions are finite and clamped, a
paragraph holds at most 100 stops, leaders up to 8 characters and the
align-on character is one character. type.tabs.get reports the stops,
indents and the text column they are measured in.
Layout > Numbering & Section Options now opens its own window instead of
the generated command form, on the section of the page in view: Start
Section (always on for page 1), Automatic Page Numbering or Start Page
Numbering at, and a Page Numbering group with Section Prefix, Style,
Section Marker and Include Prefix when Numbering Pages; then Document
Chapter Numbering with its style, Automatic Chapter Numbering, Start
Chapter Numbering at, Same as Previous Document in the Book (only in a
book) and the book's name. OK and Cancel sit beside the options, out of
the scrolling body.
OK runs layout.section and layout.chapterNumbering; a refused value
(a prefix over eight characters or with + or a comma, a start number
below 1) keeps the window open and changes nothing. Opening the window
reads the chapter numbering without running a command, so it doesn't
stop a Data Merge preview. New strings are translated (German, French,
Spanish, Japanese, Chinese, Italian).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflict: ROADMAP.md: next's Right-to-left and Interface language lines (IDML Kashidas, German and Japanese cover every catalog, Ukrainian covers every key) plus the PR's Show CJK Features line
Conflict: crates/ui-egui/src/lib.rs: the PR's cjk_features() before next's opens_as_document (with .indd/.indt and its doc comment)
Preferences > Type > Show CJK Features (cjkFeatures; unset follows the
interface language: on for Japanese, Chinese and Korean) hides the
CJK-only Type menu items (Tate-Chu-Yoko, Ruby, Kenten, Warichu) and the
Ruby Quick Apply entry. The commands, layout and IDML work either way.
A section prefix is at most eight characters and contains no plus or
comma (page references and ranges use them); `layout.section` refuses
others and leaves the section unchanged. Symbols, a footnote style,
no longer number pages. The parameter documentation lists the 01, 001
and 0001 styles, and a test pins the fields the generated dialogs read
from both numbering commands.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A document's chapter number now has a style (1, I, i, A, a, 01, 001,
0001) and a source, as in Numbering & Section Options > Document
Chapter Numbering: Automatic Chapter Numbering (one more than the
previous document in the book; 1 first), Start Chapter Numbering at,
or Same as Previous Document in the Book.
`layout.chapterNumbering {style?, start?, source?}` sets them (undoable)
and, with no parameters, reports them with the open book that holds the
document. The book's Update Numbering (`book.paginate`) resolves chapter
numbers in book order and reports them. Chapter Number variables and
cross-reference chapter numbers use the style.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aiongg