Review fixes from CodeRabbit on #2749:
- the image file is read through read_capped, so a file that grows after the size check can't exhaust memory
- an image cut by a layout viewport widens the sheet's bounds only by the part inside it, as other primitives do
- uploaded tiles of closed drawings are dropped also when the next list has no images
* fix(render): charge every block entry, index attribute definitions, and walk dashes and outlines in linear time
Array copies skipped ATTDEF entries before charging the expansion budget, each attribute searched its block for a definition, each dash scanned every centreline sample, and outline chaining searched a vertex's segment list from its start on every visit. Each is now charged, indexed or cursored.
* address review: drop wall-clock bound from outline test
* render: a display list past the block-expansion budget says what it left out
* Report viewport truncation only after the expansion budget is exceeded
* fmt: keep ui-egui control.rs rustfmt-clean
A DIMENSION moved by an older AutoCAD keeps its *D block contents where
they were and the move in DXF group 12, and is drawn translated by the
insertion point. Group 12 was never read into Dimension, so the block
rendered at its pre-move position and the extents reached there too.
Read group 12 into a new Dimension::insertion_point, draw the block
translated by it, move its bounds with it in the extents, test the
block's text where it draws, and write group 12 back whenever the block
written is the one read from the file. Regenerated geometry (edits drop
the block link) sits at its definition points, so no offset is kept or
written for it.
Co-authored-by: rahul05ranjan <11793455+rahul05ranjan@users.noreply.github.com>
* Plot: a layout's saved page setup is used by default (#2493)
A layout whose page setup says plot window, scale to fit or a plot scale was plotted 1:1 from
the sheet origin unless the caller passed fit. plot, exportpdf, the PNG device and preview now
read the layout's saved window, scale to fit, scale and centring when the options don't say
(render::paper::saved_plot); explicit options win, and fit: false without a scale is still the
plain 1:1 sheet. Tests: saved_setup_tests in pdf.rs (four), and
preview_uses_the_layouts_saved_page_setup_by_default.
* Parity checklist: regenerated after merging main
* Parity checklist: regenerated after merging main
* Plot: saved page setup in preview and PNG matches the PDF (review on #2586)
Preview honours the saved centre when fitting a window and an explicit scale beats a saved scale to fit (PDF too); PNG places a saved window without fit and clips to the printable area.
* Preview: a saved window clips to the printable area (review on #2586)
The preview drew the placed plot over the margins while the PDF and PNG blank them. The blanking helper now lives in the render crate and both the PNG plot and the preview use it. Test: preview_clips_a_saved_window_to_the_printable_area.
* Preview: pin hostile saved page setups and the default setup (review on #2586)
Degenerate or far-away saved windows and saved scales of 0 or not finite already fall back to the plain plot in the PDF, PNG and preview; tests now pin that, and that a layout with no saved setup plots as before.
* Preview: the printable clip follows the view's fit (review on #2586)
The clip assumed the sheet filled the image; when the height cap applies
View::fit scales and centres it. The printable corners now go through
View::to_screen. Test: preview_clip_follows_the_view_when_the_height_cap_applies.
* docs: leave the generated parity checklist to maintainers
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>
* Hit tests don't lay out dashes: moving the mouse over a block stays smooth (#1285)
Every mouse move hit-tests the block references under the cursor, and
hit_polylines built them with a full-detail display list, one primitive per
dash. A fine pattern on long lines in a block inserted at 0.001 made millions
of dashes per move.
Options::linetypes off draws every line whole and continuous, and
hit_polylines uses it for blocks, dimensions and multileaders. A dashed line in
a block is now hit across its gaps, as one outside a block already was.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* render: with linetypes off, a lightweight polyline is one continuous stroke
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Steffen <camperking@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
An IMAGE whose file exists was drawn as a frame with two diagonals. Now
its pixels are loaded when a drawing opens and drawn on screen and in PNG
output, with rotation, clip boundary, fade and object transparency. A file
that can't be read keeps the crossed-out frame and reports a message.
Ported from the raster commit of #1029 onto the split render and canvas
modules. New here: large images are drawn from a pyramid of levels cut
into tiles uploaded as they come into view, instead of one texture with
pixels skipped, so a large scan stays sharp zoomed in and doesn't shimmer
zoomed out.
Limits are listed in docs/gaps.md (FF10): PDF and SVG output write only
the frame, and on screen images lie under all other objects.
Co-Authored-By: Chanthawat Kiriyadee <enwuft@icloud.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
UPDATE_APP (native builds; craft-update is a native-only dependency) carries our 128 px icon and accent (#3ea6e8, its colour in ArtCraft Launcher). The updater, the self-update
helper and craft-update's startup prompt all get them through crate::UPDATE_APP; the
prompt draws the icon, the accent and "by ArtCraft" (craft-update's default developer).
craft-update is pinned to storytold/craft-libs main at 1065653e7 (#434).
Clippy fixes for Rust 1.99 in untouched code (same logic):
- crates/constraints/src/auto.rs
- crates/engine/src/audit.rs
- crates/engine/src/cmd/inquiry.rs
- crates/engine/src/mtext_edit.rs
- crates/geom/src/nurbs.rs
- crates/io/src/dxf_raw.rs
- crates/io/src/dxf_read/document.rs
- crates/render/src/text.rs
- crates/ui-egui/src/cmdline.rs
- xtask/src/layers.rs
* Dimensions: fill in the measurement when one is created (#2421)
A dimension made in the session kept measurement 0.0 until it was saved and reopened.
add_dim now sets it with the geometry call the DXF/DWG writer uses for group 42.
Block names are still assigned at save time. Test: a_new_dimension_reports_its_measurement.
* Parity checklist: regenerated after merging main
* Dimensions: a jogged radius dimension stores its radius as the measurement (#2421)
draw_jogged_radius set the text but not DimGeometry.measurement, so DIMJOGGED dimensions
kept measurement 0 (and the DXF/DWG group 42 was 0). It now sets the radius.
Test: jogged_radius asserts the stored measurement.
* Parity checklist: regenerated after merging main
* Dimensions: fill in a new dimension's text position when it is created (#2421)
A dimension made in the session kept text_mid (0,0) until it was saved and reopened.
add_dim now sets it from the same dimension_geometry result as the measurement, the
position the writer puts in group 11, unless the user placed the text.
Test: a_new_dimension_reports_its_text_position.
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* fix: keep WIPEOUT closure coherent across grips and outlines
Treat WIPEOUT as a closed polygon with an optional exact terminal closure marker, not as an open polyline. Centralize logical vertices, explicit outlines and closure-preserving vertex edits in the document type. Reuse this contract in grips, geometric primitives, rendering and DXF export without changing the stored schema or writer closure convention.
Add six-version DXF ASCII/binary and DWG regressions for implicit/explicit closure, both windings, all corners, API/interactive edits, undo/redo and repeated saves. Keep open/closed polylines and their coincident endpoints distinct. Cover multiple hot grips, hostile inputs and equivalent mask/frame output.
Validation: cargo xtask ci, all seven stages passed (2571 tests passed, one pre-existing skip). Two unrelated network tests failed transiently in earlier runs; each passed in isolation and the complete final run passed without exclusions. Direct and reopened grip edits produce byte-identical CLI PNG output.
Fixes#2256; related test-only matrix PR #2237 and tracker #2238 B3.
* fix(doc): keep coincident WIPEOUT endpoints separately editable
When an edit makes the endpoints of an implicit boundary coincide, add a separate closing marker rather than reclassifying the edited corner as that marker. Retain the logical vertex count and allow either coincident corner to move independently afterwards.
The document regression fails before the fix. Add API/interactive coverage for both endpoint directions and both closure forms, undo/redo, and subsequent separate edits across all six DXF/DWG versions including binary DXF and save-after-edit. Update the documented editing contract.
Validation: cargo xtask ci passed all seven stages; 2573 tests passed, one pre-existing skip.
Addresses the coincident-endpoint review on PR #2284; related #2256.
---------
Co-authored-by: cloudj <jonbahnes@gmail.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
TrueType and OpenType outlines are defined with the nonzero rule, but glyph
meshes were tessellated even-odd, so any area covered by two contours came
out empty. Arial Unicode (the macOS fallback CADCraft picks for Hangul)
builds 진 (U+C9C4) from a component that already contains the final ㄴ plus
a second copy of that ㄴ; even-odd cancelled the pair and 진 rendered as 지.
Overlapping contours in other fonts lost their overlaps the same way.
Tessellate glyph outlines with FillOptions::non_zero(). Outlines placed in
drawing coordinates without a TextGlyph (complex-script runs) go through the
same tessellator, offset to their first point to keep f32 precision, instead
of the even-odd slab filler.
PNG export (and everything else drawn by raster::render) stroked every
polyline primitive with its own tiny-skia stroke_path call: a new path
and a newly built paint pipeline per primitive. Hatch patterns, linetype
dashes and text strokes produce long runs of short polylines in one
colour, so that per-call setup dominated the render.
Consecutive hairline polylines (zero width, or at most one pixel when
anti-aliased; the default for PNG export) with the same colour, alpha
and width are now collected into one path and stroked once. tiny-skia
draws hairlines segment by segment straight onto the pixmap with the
same per-contour round caps, so the pixels and their blending order are
unchanged. Thicker strokes (filled outlines), fills, points, infinite
lines and polylines with non-finite or far-out screen coordinates break
the run and are drawn exactly as before.
Adds crates/io/examples/export_bench.rs, which exports examples/
apartment.dxf repeated on a grid (10x10: 5,800 entities) and reports
median/min wall time per format.
PNG export, 10x10 grid, 11 interleaved runs (Ryzen, 16 threads, Linux):
before: median 1790.7 ms, min 1703.3 ms
after: median 1037.6 ms, min 926.0 ms (-42.1 %)
A/A noise on the same setup: 1.1 %. Output PNGs are byte-identical
(bench at 10x10 and 20x20, CLI export of apartment.dxf and a 23,200-
entity drawing). New tests compare render() against stroking each
primitive on its own.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- VIEW Save inside an MSPACE viewport stores the viewport's twist; VIEW Restore into an
unlocked viewport turns the viewport to the saved twist (one undo step).
- Undo and redo keep the untiled Model view's twist record in step with the live view.
- A selection window is a crossing window when dragged leftwards on screen, also in a
twisted view (preview and selection alike).
- DVIEW's preview inside a clipped viewport is clipped to the viewport's outline.
- DXF view centres are read relative to the view target (VIEWPORT/VPORT group 17, VIEW
group 12), the origin of display coordinates.
Pure extraction, no behaviour change: same order of side effects, same
early returns and error values. Clears cognitive-complexity findings on
AreaM::input, explode_kind, hatch sweep, the COMPARE overlay paint and
the Files options tab (S3776).
DIMTFILL and DIMTFILLCLR were not modeled, so a dimension's text never had a
background and lines under it stayed visible. Dimension styles now carry both
variables. DIMTFILL 1 hides what is drawn under the text in the drawing's
background color, and DIMTFILL 2 fills the box in DIMTFILLCLR. The box is the
text box widened by DIMGAP on every side. The arc length symbol over the text
and the DIMINSPECT frame are drawn over the background.
The variables work with SETVAR, DIMOVERRIDE and the style fields. DXF stores
them as DIMSTYLE groups 69 and 70 from R2007 and in the style's CADCRAFT
xdata in older files; the reader skips the record's flags, which also use 70.
For DIMTFILL 2, the dimension's *D block and EXPLODE add a SOLID in
DIMTFILLCLR under the text. A dimension drawn from the *D block it was read
with gets a background around each of the block's texts, before the block's
first text or first segment of the text's marks, except for a text that a
SOLID of the block already covers. The background stays after the drawing is
saved and reopened.
The *D block and EXPLODE's MTEXT took the dimension's decoded text, so
%%c arrived as U+2300, which no SHX font has: other programs plotted a
diameter dimension in an SHX style without its sign. They now write
DimGeometry::mtext_contents(): the sign as %%c, which every font draws
(and CADCraft decodes), a % that would run into a code as %%%.
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Use the effective dimension style's decimal separator when formatting
angular labels, preserving the public formatter's decimal-point default.
Add rendering and stored DXF/DWG block regression tests, and update the
file-format progress documentation.
Fixes#1646.
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* Dimensions: text that does not fit follows DIMTMOVE, over the dimension line (with a leader for 1) instead of past the second extension line (fixes#1998)
* Dimensions: tighter check that the line is not carried out to moved text; drop a guard that never changed the outcome (#1998)
A block reference's own linetype scale (DXF group 48) was ignored for the objects in the block:
a DASHED line inside a block inserted at linetype scale 2 kept dashes of scale 1. It now
multiplies their linetype scale, as LTSCALE does for the whole drawing (and as a dimension's
does for its block).
Claude-Session: https://claude.ai/code/session_01VP2GTYMskgGjiMw5GT5NL9
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* security: budget the rendering pipeline against DoS geometry (CWE-400/407/789)
Several render paths let crafted drawing data spin the rendering thread or
allocate without bound:
- Hatch dash sweeping reset its per-span step counter and used emitted geometry
as its only limit, so gap-only families (which emit nothing) could multiply
CPU work. Skip families that cannot draw and charge every dash step against a
shared per-hatch work budget (row 10).
- Hatch polygon deduplication compared loops with an O(n²) cyclic match and
every loop against every kept one. Charge vertex comparisons against a shared
budget and stop deduplicating when spent (row 12).
- Table coverage marked every merge rectangle, so overlapping merges re-marked
covered cells quadratically. Bound marking to a small multiple of the cell
count (row 11).
- Broken-table part layout expanded sparse rows to full width before checking
whether detailed layout is allowed. Only expand when within the cell limit;
otherwise leave cells sparse and let layout draw an outline (row 17).
- Dimension text substitution copied the base for every `<>` placeholder with no
size bound. Estimate the expansion and cap it at a fixed text budget (row 32).
- Linetype range generation cycled all-gap patterns forever (nothing emitted, so
the emit limit never tripped). Return continuous for all-gap patterns and
bound visited pattern elements (row 38).
- Viewport contents built with a fresh expansion counter, so a layout could far
exceed the documented block-expansion limit. Share one budget across the whole
build, seeded with what the enclosing list already spent (row 22).
New tests for the all-gap linetype path; all render tests pass, clippy clean.
Addresses findings rows 10, 11, 12, 17, 22, 32, 38.
* fix: preserve bounded render fallbacks and text output
---------
Co-authored-by: skylartaylor <26588250+skylartaylor@users.noreply.github.com>
* security: bound geometry-kernel degree and work (CWE-407)
Several geometry routines let a crafted spline or region cause quadratic or
worse CPU work:
- B-spline basis evaluation is O(degree²) per sample and tessellation takes
O(degree) samples, so a HATCH/DXF spline edge declaring a very high degree
(DXF allows up to 100 000) turned a tiny spline into billions of iterations.
Add `MAX_DEGREE = 64` and treat higher-degree splines as invalid, so
`eval`/`tessellate` fall back to the control polygon. The render clip boundary
also rejects such degrees up front (rows 18, 28).
- `Spline::derivative` re-ran the O(knots) validity scan on every call, and
arc-length integration calls it once per sample, making a valid spline's
measurement quadratic. Split off `derivative_unchecked` and have integration
(validated once by its public entry points) use it (row 24).
- `infer_knot_param` matched each fit parameter against the whole knot vector
from the start. Both are ascending, so replace it with a linear merge (row 16).
- Fillet seed collection did not stop the enclosing run-pair loops when the
per-call candidate cap was hit, and refined duplicate candidates. Bound total
seeds across pairs and deduplicate before Newton refinement (row 36).
- Region grid intersection counted only distinct pairs against its budget, so a
long segment replicated across many cells re-visited the same pair unbounded.
Charge every candidate-pair visit against a budget (row 41).
New spline degree tests added; all geom and render tests pass, clippy clean.
Addresses findings rows 16, 18, 24, 28, 36, 41.
* fix: preserve spline knot and fillet search behavior
---------
Co-authored-by: skylartaylor <26588250+skylartaylor@users.noreply.github.com>
* fix(render): keep filled highlight dash anchors stable
* outline test: as_chunks for clippy
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>
Found while testing crates/render/src/units.rs:139-160 (format_linear, unit 3 and 4 arms) and feet_inches at :101
Co-authored-by: Claude Haiku 5.5 <noreply@anthropic.com>
Slabs ran between vertex rows and paired the active edges by their
order at the slab's middle, so two edges crossing inside a slab (a
bowtie, two loops crossing) were paired across the crossing and the
trapezoids covered the wrong side of it. A slab with such a crossing is
now cut again at that y and each part paired on its own, under a cut
budget so hostile input stays bounded. Tests check coverage against an
even-odd parity oracle for a bowtie, an off-centre bowtie, two crossing
loops and the plain cases; the three crossing cases fail on main.
Claude-Session: https://claude.ai/code/session_011K1peyrh5MdZ6ttHkVq8Gh
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Paper-space units followed the drawing's MEASUREMENT only, so a layout
whose page setup is in millimetres (plot paper units 1), common in
drawings from AutoCAD with MEASUREMENT 0, was taken as inches: its PDF
drew everything 25.4 times too large, off the page.
PageSetup gains `paper_mm`, read from LAYOUT group 72 = 1; 72 = 0 still
leaves the units to MEASUREMENT, since earlier CADCraft versions wrote 0
for every layout. The layout's sheet, its PDF/raster plots, PREVIEW and
the pens of what its viewports show use the layout's units, and the
writer stores 72 = 1 for layouts in millimetres (it wrote 0 always).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* AutoCAD interop: justified text, DIMPOST, current dimension style, *D block colours
What AutoCAD takes from CADCraft's DWG/DXF as it is:
- Justified TEXT/ATTRIB/ATTDEF (#1050): group 10 is now where the
baseline starts, from the alignment point and the text's width, not
the alignment point itself; AutoCAD drew centred and right-aligned
text from DWG starting at its alignment point. A stored point close to
CADCraft's own (AutoCAD's, exact for its fonts) is kept.
- DIMPOST (#1053) is written even when empty: the DWG bridge defaulted a
missing one to "<>", and AutoCAD then dropped the "R" of radius
dimensions when it regenerated them.
- The current dimension style (#1051): the DXF header now carries its
DIM* variables (AutoCAD draws new and updated dimensions from them);
the DWG bridge points $DIMSTYLE at the style by name (its handle was
renumbered, so the writer fell back to Standard), and reading a DWG
gives the current dimension and text styles their names back.
- *D blocks (#1052): lines, arrows, extension lines and text take the
style's DIMCLRD/DIMCLRE/DIMCLRT instead of ByBlock.
#1054 (group 42 of a rotated linear dimension) was already fixed on
main by #855; acad_interop.rs keeps it covered.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Test: angular dimensions store their angle in radians (group 42)
Three-point (60°, 174°) and two-line (120°) angular dimensions, read back
from the DWG as DXF. Already right on main since #855; this guards it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Justified text made here: Left Middle and Left Bottom write their baseline start too
text_insert kept a stored insertion point that was within 0.1 x an
estimated width + 0.5 x height of CADCraft's baseline start, taking it for
AutoCAD's own. The width was estimated from the alignment point's offset,
which for Left Middle is only h/2 (vertical), so a text made here, whose
stored point is the alignment point, passed and kept it: AutoCAD drew it
half a height high (Left Bottom: a third). Reported with a test by
@lee-cyrille-maskell.
A stored point equal to the alignment point is now always stale. Otherwise
the offset is split along and across the baseline: along it the same slack
as before (a tenth of the text's width plus half its height, as fonts and
their substitutes differ most in width), across it a fifth of the height.
Multiline attribute values keep the distance-only check, since their lines
are placed by the embedded MTEXT.
On 12 drawings made in AutoCAD (1,055 justified texts) the insertion points
written are the same as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Tanish Manku