Justify (last line left, centred or right) widened only word spaces, so
a Japanese line, which has none, stayed set solid and the right edge of
a justified Japanese paragraph was ragged by up to almost a character
(only Justify All spread the characters). Following JLREQ 3.8, the room
left on a line now goes to the gaps next to a CJK character, shared
with any word spaces:
- not inside a Latin word or a tate-chu-yoko block, nor next to a tab;
- lines without CJK characters are justified as before;
- vertical type too (the same code, in line space).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shippori Mincho (assets/fonts, from #145) moves to storytold/craft-fonts. crates/text/build.rs
embeds the fonts of a craft-fonts checkout when CRAFT_FONTS_DIR is set (web: BIZ UDPGothic
Regular only) as vectorcraft_text::CRAFT_FONTS; unset, it is empty and Japanese falls back to the
installed fonts. Document text gets the Japanese faces after the bundled fonts (Mincho first); the
UI adds them at the end of every egui family (BIZ UDPGothic first), before ui_fonts' system-font
discovery. Release and FreeBSD CI builds check craft-fonts out at a pinned commit; packages carry
the fonts' OFL texts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Opening, placing, pasting or importing a file, the font menus and MCP now find
the installed system fonts (Arial, Times New Roman...) in a fresh session, so a
file renders the same whatever happened earlier, and PDF import keeps the real
family names (Microsoft YaHei, not "Microsoft Ya Hei").
Root cause: FontDb only consulted the system font catalog, which stayed empty
until a CJK fallback or Character panel > Show System Fonts happened to scan
the font folders. Every lookup by family name (face, families, styles,
has_family) now waits for one cataloging scan per session: the app starts it
in the background, other processes run it on first use. The scan reads only
each file's table directory and name table (about 70 ms for 780 faces here,
instead of reading every font file whole), the catalog is indexed by family,
the family list is cached, and an installed style of a loaded family (a
bundled family also installed with more weights) loads when asked for.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shipped code no longer unwraps, expects or panics: every one of the 124
unwrap()/expect()/panic!/unreachable! sites outside tests now handles the
failure (Result/Option, ?, let-else, defaults) and reports errors to the user.
- Lints: workspace clippy denies unwrap_used, expect_used, panic, todo,
unimplemented and unreachable; clippy.toml allows them in tests, and
integration tests, examples and testkit opt out explicitly.
- Safety net (vectorcraft_engine::guard): a panic in a top-level command rolls
the document back and returns EngineError::Internal; a panicking tool is
reset; the MCP server answers with an internal error and keeps serving; the
UI frame and each control request are guarded.
- APIs that hid failures now return them: Rendered::to_png/to_jpeg/to_webp
return Result (JPEG/WebP errors used to write empty files),
FontDb::face returns Option, Region::top returns Option.
- Fuzzing: engine/tests/import_fuzz.rs feeds garbage, hostile and mutated SVG
and PDF through import → render → export (30k cases locally, no panics).
- Docs: AGENTS.md/CLAUDE.md rule, docs/development.md "Robustness" section,
README pointer and a PR template checklist.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
