* refactor: replace wildcard imports with explicit ones (S2208, batch_03)
* refactor: replace wildcard imports with explicit ones (S2208, batch_03)
* Format imports so the CI fmt gate passes
* refactor: replace wildcard imports with explicit ones (S2208, batch_00)
* refactor: replace wildcard imports with explicit ones (S2208, batch_00)
Explicit lists now resolve the names each file uses, and the test
module keeps its glob so it still sees the parent's items.
Picks up the startup prompt without the glyphs a font may lack (storytold/craft-libs#343, #344): the close mark is painted and the version pill is ASCII, so apps that panic on a missing glyph no longer take the prompt down.
Part of tracking issue storytold/craft-libs#319
Picks up the update prompt fix that draws its close mark and version pill
without glyphs a font may lack, so the prompt no longer takes the app down
when a bundled face has no ✕ or →.
Part of tracking issue storytold/craft-libs#319🤖 Generated with Codebuff
Co-authored-by: Codebuff <noreply@codebuff.com>
Picks up the fix that keeps the close request pending while the app holds unsaved work, the executable-name check in the Windows update helper, and the zip 9 / dirs-free build.
Part of tracking issue storytold/craft-libs#319
* feat: startup update prompt modal and direct self-updating
Ref: storytold/craft-libs#319
* build: pin craft-update to craft-libs main (b9ba47f) and refresh the lock
* refactor(updates): drive the flow with craft-update's Updater
Results are read independent of the prompt, a dropped worker is an error, failures return to the prompt, and no update starts over unsaved work. A manual check during the startup check is honored.
craft-ui and craft-docpath were pinned at intermediate commits of the egui
0.37 port (craft-libs#268), and the PDF stack at 6e7f285. a0916808 is main
after #268 merged and contains all three.
* CLI --save, convert and sample write through a staged file so a failed write keeps the previous drawing (#912)
* cli: write through the engine's atomic_write instead of a second copy
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Read dynamic block element tags from the entity's own xdata (#942)
* dynblock test: an untagged object at index 1 is keyed 1_000_001
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Keep associative-array parameters when an xref is loaded (#943)
* xref load test: initialise Ctx's new bytes counter
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
AutoCAD stores the wipeout frame setting in the drawing's WIPEOUTVARIABLES object (group 70:
0 = frames hidden). CADCraft ignored it and always showed the frames (the default of 1), so
masks such as the diamond inside a numbered circle were outlined. The flag is now read into
the WIPEOUTFRAME setting.
Co-authored-by: simanov <asimfali@mail.ru>
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/units.rs:333-340 (parse_dms)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* ui: accept IME composition and commit in the command line (#1761)
The command line is a custom buffer with keyboard focus staying on the
drawing, so no egui TextEdit ever owned the IME: egui never published an
IME output, winit never called set_ime_allowed(true), and a Windows IME
(Microsoft Pinyin, Japanese, Korean) could not even switch on. Even when
IME events arrived they were ignored, so in-place single-line TEXT edits
(and every other text prompt) accepted ASCII only.
- Publish an IME output (command-line rect and caret) while a prompt asks
for text, enabling the OS IME; release it when the prompt no longer
collects text, so ASCII typing and shortcuts keep their behaviour.
- Handle Event::Ime: Preedit is drawn as an underlined composition after
the buffer but never saved; Commit inserts the committed string once.
Newline/Return commits are ignored like egui's own text fields, and
Enter/Escape/submit clear any composition.
Tests: a headless command-line test asserts the IME is enabled for a
TEXTEDIT prompt, the preedit stays out of the buffer, the commit lands
exactly once and the text entity changes; a kittest double-click test
edits TEXT in place with IME preedit and commit.
* cmdline IME: build the IMEOutput with egui 0.37's constructor and a context generation
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* PLOTWINDOW: pick both corners even with ORTHO on
PLOTWINDOW's opposite-corner prompt used Prompt::base, so the canvas bent
the picked point onto the nearest axis with ORTHOMODE on: the second corner
landed above/beside the first, the window had no width or height, and
set_plot_window refused it. The prompt now uses Prompt::corner, the flag
ZOOM Window, RECTANG and ARRAYRECT already use to keep a rectangle's
opposite corner free of ortho. The first corner sets the base as before, so
the rubber band and direct distance entry still work.
Regression test drives the interactive machine through snap_cursor with
ORTHOMODE on, picks two non-axis-aligned corners and checks the stored
window keeps both extents; an ordinary LINE prompt still follows ortho.
Fixes#1641
* docs: document the functions touched by #1641
Found while testing crates/engine/src/lisp/builtins.rs:798-800 (glob_inner depth limit; depth is incremented per '*' at line 808)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Found while testing crates/engine/src/units.rs:124-127 (wrap_full_turn, used at line 106 for AUNITS 3 and at lines 105 and 120)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
CECOLOR now reads "BYLAYER", "BYBLOCK", the index as text ("1" after
"red") or "RGB:r,g,b" for a true colour, from getvar, LISP and the
SETVAR ? list, where it was the ACI number (256, 0, 1, or the nearest
index of a true colour). The header still keeps the index and
CECOLOR_RGB, so DXF output and new objects are unchanged.
SETVAR CECOLOR takes "RGB:r,g,b" too, so a value saved with getvar
can be set back, true colour included.
Found while testing crates/geom/src/curve.rs:568-581 (Polyline::centroid); compare area() at 533-538
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Found while testing crates/dxf/src/lib.rs:511-524 (escape_unicode) and 226-258 (unescape_unicode, applied in parse_ascii_with_limit at 278)
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
An egui modal showed a frame late (its first frame only measured it) and
then faded in, dim included, over 0.2 s. These alerts answer a key press or
menu choice, so they now skip the fade, and their first show reruns the pass
with the measured size, which draws them in the first frame presented.
* macOS menu choices and their shortcuts run at once
A menu choice is no window event, so the app ran it only at its next frame
for another reason: File ▸ Close (Cmd+W) on an edited drawing showed
"Save changes?" and dimmed the floating palettes a second or two late. The
menu's event handler now wakes the app.
* cargo fmt after merging main
The egui 0.37 migration moved craft-pdf, krilla, hayro-syntax and
hayro-jbig2 to 7895f5e, 206 commits older, though none of them depends on
egui. 6e7f285 keeps krilla patch 13 (non-finite link and page-box geometry
refused), craft-libs#282 (URI links only for http(s) and mailto) and
craft-libs#255 (CCITT and JPEG 2000 decode bounds). The embedded ICC
profiles are identical at both revisions.
Dylan M. Taylor