#1681 places a dimension's picture block by its insertion point, rotation
and scale. Edits (move/rotate/scale/mirror, copy, paste) transform the
picture's sub-entities in world space, and the codec's dimension transform
moves the insertion point too, so a rotated or copied dimension would get
the placement applied twice. Fold the placement into the subs before the
edit and leave the dimension at the identity placement afterwards.
The picture insert no longer takes the dimension's extrusion: OCS bakes
picture sub-entities in world coordinates, so only the stored insertion
point goes through the OCS. A zero insertion-scale component counts as 1.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
style_new/copy/delete funneled through a load_style_bufs no-op for Table/MLine, leaving stale editor buffers pointed at the fresh style; both arms now route to the existing loaders. Table cell text-style typos persisted as dangling names; unknown names are now rejected with command-line feedback, keeping the buffer text for correction. Silent numeric parse-drops deliberately untouched (codebase-wide keep-old-valid norm). Tests: table_new_reloads_cell_buffers plus table_cell_unknown_text_style_is_rejected incl. retention assert (RED each verified by temporary revert). Full suite: 2048 passed / 0 failed / 25 ignored, zero warnings.
Tests previously ran on Ubuntu only, so Windows- and macOS-specific breakage went unseen. Add a 3-OS test matrix and a separate lint job for rustfmt and clippy. Lint is continue-on-error because the tree is not yet fmt-clean or clippy-clean.
Also add BEDIT characterization tests that pin current session, undo and refusal behavior ahead of the editor refactor. Locked tests guard behavior that must survive; documenting tests record questionable behavior expected to change.
The edge-stack closure duplicated dock_panel_visible arm-for-arm; it now delegates (tab is provably the active tab at that site). Ribbon.layer_names was written but never read outside one test; deleted, test reads via layer_infos. Faces and combo-refresh sub-items investigated and dropped (already unified upstream; no live stale path). Tests: dock visibility parity pins incl. Start-tab. Full suite: 2046 passed / 0 failed / 25 ignored, zero warnings.
XrefManagerSelect/RightClick/RowOp/PathPick/FindReplace carried positional display indices that rescans could retarget. New XrefRowId{key,saved_path} (the identity refresh uses) in all 5 variants, resolved via row_index at dispatch with graceful no-op; panel selection stays index-based. display.index remains for view-time styling/lookup only. Tests: xref_rowop_unknown_id_is_noop (RED first); existing dialog/simulator/click tests migrated to ids. Full suite: 2044 passed / 0 failed / 25 ignored.
AnnoObjectScaleToggle had three silent no-op paths (no target, locked layer, unknown scale); each now pushes command-line info mirroring the open-path guards. Xref empty-state hardcoded grey now uses muted_style. DD_BOUNDS eviction and attr-clamp changes investigated and deliberately dropped (no reachable defect in either). Tests: anno_scale_toggle_tests cover all three paths (RED: silence). Full suite: 2032 passed / 0 failed / 25 ignored, zero warnings.
editor_layer_row and view_editor took Vec<String>, cloning N strings per row per frame (times two build-closure runs). Both take Arc<[String]> now; a plain borrow cannot cross the twice-run closure (E0373), hence the Arc. Cow empty-filter short-circuit for both search queries. Test: view_editor_accepts_shared_linetypes (RED first). Full suite: 2029 passed / 0 failed / 25 ignored, zero warnings.
VP-freeze columns used the global-freeze glyph; shift-ranges swept filtered-out rows into bulk ops. Glyph swapped to layer_vp_freeze (ribbon precedent); new visible_indices helper shared by the view loop and the shift arm. Test: shift_select_skips_filtered_out_rows (RED: hidden row in range). Full suite: 2028 passed / 0 failed / 25 ignored.
A region's loop area sums its pieces with Curve::chain_area, the
composable share: enclosed_area closes an open polyline by its chord.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The pill number used the first-content fallback while the label and sync badge used explicit-only resolution, showing e.g. a 1:50 number beside a 1:1 label with no badge. Label and badge now resolve through target_viewport_handle like the number; render-path handle fn, set-paths and Model branches untouched. Tests: scale_pill_sources_agree_without_targeted_viewport (RED: live probe Some(0.02)/1:1/None) plus Model-unchanged pin. Full suite: 2027 passed / 0 failed / 25 ignored.
props_asym_scale was app-global keyed by raw handle values, but fresh documents restart numbering at seed 1, leaking one tab's unchecked Uniform box into another tab reusing the value. The set moves to DocumentTab; all uses tab-scoped. Also removes a dead destructure left in overlay.rs. Test: asym_scale_opt_out_does_not_leak_across_tabs with handle-reuse premise assert (RED first). Full suite: 2025 passed / 0 failed / 25 ignored, zero warnings.
draw_axes projected with raw project_point3, so a behind-camera axis endpoint wrapped to a finite mirrored point (proven: clip.w=-100 projected to screen centre). Extracted shared project_point with the eye-plane guard (grid_segments delegates verbatim) and axes now skip when an end is unprojectable plus clip_seg. Legacy canvas offset preserved pixel-identically; remaining raw project_point3 sites elsewhere are out of scope. Tests: project_point_rejects_behind_eye plus axes_segments geometry pins. Full suite: 2024 passed / 0 failed / 25 ignored.
LayerPanel.editing and color_picker_row were positional indices, so a sort/sync/filter between start and commit retargeted renames and swatches. Both are Option<String> now; commit resolves fresh with graceful no-op on deleted layers. Existing rename tests migrated. Tests: layer_rename_commit_follows_name_across_resort (RED first) plus deleted/empty no-op and picker-resort pins. Full suite: 2019 passed / 0 failed / 25 ignored.
rename_style_storage and rename_symbol rewrote entities and table rows but left dimtxsty stale, so the next Apply resolved a NULL handle. The follow now lives in both shared functions; the plugin workaround is removed as redundant. Tests: text_rename_follows_dimtxsty drives the manager path (RED first); pre-existing plugin rename test stays green via the shared path. Full suite: 2015 passed / 0 failed / 25 ignored.
StyleKey::TableStyle only updated the ribbon mirror, so inserts kept using the header value and resync re-seeded from the mirror. The arm now writes current_table_style_name plus dirty when the style exists, and sync_ribbon_styles seeds from the document like Text/Dim. Tests: tablestyle_ribbon_pick_writes_header_and_dirties plus pick_flows_into_table_insert covering pick-to-dialog (both RED first). Full suite: 1996 passed / 0 failed / 25 ignored.
A dimension carries its graphics in a block of its own, drawn in that
block's frame, and the entity records where the frame sits. The renderer
inserted it at the origin instead, so every dimension whose file carries a
non-zero insertion point drew far from the geometry it measures — the whole
picture offset by that point. Dimensions inside block definitions are where
this shows: a drawing of plan blocks scattered their dimension graphics
thousands of units away, over empty sheet.
The picture now follows the dimension's insertion point, rotation, scale and
extrusion, the way any block reference is placed.
Co-Authored-By: Claude Opus 5 <[email protected]>
- Pins: codec 8189f83, kernel a82f17d, graph d3558bb (all manifests and
the python plugin lock).
- MASSPROP reports volume, centroid, inertia, products, radii of
gyration and principal moments of 3D solids and bodies from their
B-rep (exact where analytic, else from a fine mesh).
- OFFSET of ellipses and splines through the kernel's curve offset.
- FILLET between circles, ellipses and splines through the kernel's
curve-curve fillet, trimming arcs and open splines to the tangents.
The kernel bump brings the DWG ACIS audit fixes: slices, booleans and
press/pull on cones, slanted cylinder sections, survey coordinates,
spline walls and touching parts.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- refresh_associative_sweeps runs on every edit and scanned the xdata of
every entity; it now returns at once when no associative swept surface
exists (its application is not registered) and looks at surfaces only.
- An arc's end angle was raised by TAU in a loop, which never ends on a
huge angle from a damaged file; the sweep is taken with rem_euclid.
- The sweep DELOBJ test expected surface sweeps to delete their sources,
which #1663 changed on purpose: an associative surface keeps them.
- codec c9ffcac (opencadcodec #103) adds an explicit viewport-off bit: the
viewport on state is read with is_on() and set with turn_on()/turn_off()
everywhere, so a viewport the file marks off shows off and can be turned
back on from Properties.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The reference cannot open a drawing whose extrusion history has a taper
and leaves the profile frame to be computed (ErrorStatus 53). Extrusion
and press-pull records now set group 295, as the reference writes them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
SWEEP: base point from the path; kernel sweep validity fixes
The fork pin the PR carried ([patch] to ramox81/cadkernel 30b3810) is
resolved to upstream instead: kernel 99b6b48 (opencadkernel #78-#83),
codec c9ffcac (opencadcodec #102, #103), graph 7f6aae7, in every manifest.
The Properties palette reads the layer names when it is built, and a
layer rename only refreshed the layer panel and the ribbon, so the
palette's layer list kept the old name beside the new one. The layer
panel refresh now rebuilds the palette whenever the set of layer names
changed: rename, new, delete and purge all pass through it. A layer
toggle that leaves the names alone does not rebuild it, as the rebuild is
costly for a large selection.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Idle and object selection drew the same cursor, crosshair arms with the
pickbox, so the cursor did not tell a pick from a free move. While a
command gathers a selection or waits for an object pick the arms now go
and the pickbox stays: idle keeps arms and box, point input arms and dot.
Fixes#1661
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On a non-Retina screen macOS draws the interface at 1x, which reads small
on a Full HD monitor, and there was no way to enlarge it. Options >
Display now offers an interface scale of 75–200 % in 25 % steps, applied
through the window's scale factor on top of the display's own and kept
with the user settings. It is a list rather than a slider because the
dialog rescales under the pointer as the value changes.
Fixes#1655
Co-Authored-By: Claude Opus 5.5 <[email protected]>
PURGE repeats until a pass removes nothing. A drawing whose complex
linetypes use shape files carries unnamed STYLE records for them; the
table holds those under keys the empty name does not reach, so every pass
listed them again and removed nothing, and the loop never ended.
- Unnamed and shape-file styles, and styles a complex linetype reaches by
handle, are never listed for purging.
- In-use checks compare names case-insensitively, as the tables do, so a
layer, style or linetype referenced as WALL is not purged as Wall.
- Counts are what the tables actually removed, and a pass that removes
nothing ends the loop.
Fixes#1662
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Swept surfaces made with SURFACEASSOCIATIVITY on keep their profile and
path and are written as a swept-surface action of the drawing's
associative network, reading the profile and path and their scale and
twist expressions. They follow edits of their sources.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Add Section 19 benchmarks covering camera pan viewport uniform updates,
spatial hover candidate search, and viewport data assembly scaling
- Add navigation telemetry probes in src/perf.rs (active only when
PERF=1) to track frame interval, batched input moves, and layout-draw gap
underlay_resolution_stale() was requesting continuous redraw frames at
display refresh rate even when no raster underlays were present in the
drawing. Cache has_underlays conservatively on Scene and early-return
false when no underlays exist, eliminating the idle tick leak.
A single profile's base point comes from the path it is swept along: a
path leaving the profile's plane from inside the profile sweeps it where
it stands. Kernel 30b3810: pcurves in the edge's parameterisation (valid
spline-faced sweeps), corner crossings fitted to the modeller's
resolution, run faces through stations evenly spaced along the run.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Language gains lo-LA, labelled ລາວ in every catalog, so the new
translation can be chosen.
- Two Lao keys under stale names now use the catalog's names, and the
touchpad option added since gets its Lao label.
- The web build gets a Lao font (Noto Sans Lao, generated like Thai) and
routes Lao and Thai drawing text to their fonts; Thai had a font but
was never routed to it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Outside macOS a touchpad's two-finger drag arrives as wheel scrolls that
nothing tells from a wheel. A new option (Options, off by default) makes
every scroll pan and Ctrl+scroll — what a pinch sends there — zoom.
Fixes#1560
Co-Authored-By: Claude Opus 5.5 <[email protected]>
select takes window [x1, y1, x2, y2] in world XY, with crossing to also
take what the box crosses. It runs the viewport's own box test, now one
helper (box_selection_handles) that the drag and click-move-click
gestures share; the drag gesture applies the selection filter as the
click one already did.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The closed combo is at least 250 px wide (about 25 characters of name)
and its dropdown 300 px.
Fixes#1648
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The text was placed by its alignment rather than its attachment point,
so text on the left of the landing ran over the leader. It now attaches
by the attachment point, and putting the text on a side (on creation or
by a grip) turns left- or right-justified text to read away from the
leader.
Fixes#1646
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Circle and ellipse shaders scaled the pattern length by the linetype
scale (PSLTSCALE in a viewport, annotation scale in model) but not the
dash and gap elements, so a dashed arc broke into dots.
Fixes#1644
Co-Authored-By: Claude Opus 5.5 <[email protected]>
COPY and PASTE kept the source's reactors on each copy, so a copied
dimension still pointed at the source's dimension association (and
boundary or group members at the source hatch or group). Another
application reports such a drawing as damaged and, after recovery,
moves the dimensions. A copy's reactors now follow only targets copied
with it; group copies give their members the new group's reactor; and a
drawing opened with such stray dimension reactors has them dropped.
Fixes#1645Fixes#1649
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The paste kernel added every entity to model space. It now takes the
space a new entity goes to: the block being edited, a layout's paper
space when no viewport is active, or model space.
Fixes#1652
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Hakan Seven