Route the remaining absolute-coordinate CPU projection through the
relative-to-eye path so it no longer multiplies by a view_proj matrix
carrying a ~1e7 translation (which cancels catastrophically in f32 after
world_offset removal):
- camera pick_on_plane / pick_on_target_plane delegate to
unproject_on_plane_f64 (rotation-only inverse + f64 eye)
- snapper (world_to_screen), hit_test (click_hit / mesh / hatch), the
scene mesh/solid hit methods, grid + UCS axes, and polar_constrain_near
all take view_rot + eye and project relative-to-eye
- compute_grid_step measures the per-unit scale via the rotation-only
matrix (no eye term needed)
view_proj remains only for conservative CPU frustum-cull / scissor and the
screen-space viewcube / UCS-icon directions. NOTE: SnapResult.world and
the command-point chain are still Vec3 (f32) — a ~0.5 m quantization at
extreme UTM zoom remains; the command layer goes f64 next.
Co-Authored-By: Claude Opus 4.7 <[email protected]>
Route the on-screen chrome that projects absolute world coordinates
through the f64 relative-to-eye helpers so it stays glued/jitter-free at
UTM scale after the world_offset removal: grips (grips_to_screen) and
grip hit-testing (find_hit_grip) now project from the f64 GripDef.world
via Camera::project_f64; the model grid and its UCS axes carry the camera
eye (DVec3) + rotation-only matrix and a f64 origin, projecting each line
relative to the eye. Snap-marker / OST-point / compute_grid_step still use
the f32 path (single points / spacing — negligible) and can follow.
Co-Authored-By: Claude Opus 4.7 <[email protected]>
Delete the world_offset field from Scene and DerivedCaches and the
load-time assignment; every reader now uses a zero origin. Geometry is
absolute and the double-single relative-to-eye pipeline keeps it precise,
so the coarse origin is gone. world_offset_from_centers still derives
local_extent_max (camera/cull span) but returns no offset. The remaining
`world_offset: [f64;3]` function parameters are now always zero — dead
plumbing left for a later mechanical sweep; behaviour is unchanged.
Co-Authored-By: Claude Opus 4.7 <[email protected]>
3D solid / surface meshes were the last single-f32 path. truck's mesh
tessellation now returns a paired verts_low residual (TruckTessResult::Mesh,
polygon_to_result), MeshModel/MeshLodSet carry it, the model-space offset
and block-instance transforms reconstruct the f64 position and re-split it
after subtracting world_offset, MeshVertex gains position_low (location 3),
and mesh.wgsl reconstructs the eye-relative position. Native SAT / ACIS /
OBJ mesh paths default to a zero residual (no regression — they were never
sub-f32 precise).
With this every world-space pipeline (wire, hatch, image, face3d, mesh)
is full double-single relative-to-eye; large-coordinate jitter is gone.
Co-Authored-By: Claude Opus 4.7 <[email protected]>
Table / PolyfaceMesh / greek-text / mleader-background fills go through
the face3d pipeline; their vertices were single f32 and quantized at UTM
scale. WireModel now carries a fill_tris_low residual (filled by the
tessellate Lines path), Face3DVertex gains position_low (location 3), and
face3d.wgsl reconstructs the eye-relative position from the pair. 3DFACE
quads built from key_vertices use a zero residual (no low source).
Co-Authored-By: Claude Opus 4.7 <[email protected]>
Foundation for double-single positions: WireModel now carries an
index-paired f32 residual buffer (`points_low`) next to the existing
`points` high half. Empty means "all-zero residual", so every existing
constructor (interactive draw, previews) is unchanged in behaviour;
tessellation from CAD f64 will fill it in the next steps so vertex
precision survives at UTM-scale coordinates.
Co-Authored-By: Claude Opus 4.7 <[email protected]>
Foundation for relative-to-eye rendering: the pivot the camera orbits
must stay precise at large offset-relative coordinates so the derived eye
(which the RTE path subtracts per vertex) is exact. `Camera::target` is
now DVec3 and `eye_f64()` exposes the full-precision eye; `view_proj`
still casts to f32 at the matrix, so behaviour is unchanged for now.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
PASTECLIP recreated missing block definitions, tessellated pasted solids,
and refreshed the layer panel after a paste; PASTEORIG and PASTEBLOCK did
not. A block pasted through either of those rendered empty because its
definition was never recreated in the target drawing.
Both paths now call merge_clipboard_blocks (so referenced/nested block
definitions exist) and populate_meshes_from_document (so pasted ACIS
solids tessellate); PASTEORIG also refreshes the layer panel.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
New annotations now place their text and reading geometry along the
active UCS X axis (identity = world), so they read square to the user's
coordinate system instead of world-horizontal:
- TEXT / MTEXT: rotation defaults to the UCS X angle
(DocumentTab::ucs_rotation_angle, model space only).
- Linear dimensions: measure along the UCS axis the span is nearest, and
bake the UCS X angle as the text rotation so the value reads square
even under a style that forces horizontal text. DIMCONTINUE / DIMBASELINE
derive their axis from the base dim's rotation (any angle), not a
world H/V test.
- LEADER: the MText annotation rotates to the UCS, and the hookline runs
along the leader's horizontal_direction (set to UCS X), used by the
renderer instead of world horizontal.
- MLEADER: context.text_direction = UCS X drives both the text rotation
and, in the renderer, the text side and dogleg/landing direction.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Commands that construct geometry with implicit axis directions ignored
the UCS, so a rectangle drawn under a rotated UCS came out world-aligned.
Give commands the UCS and build square to it:
- CadCommand::set_ucs(Mat4) (UCS→wire affine, default no-op); the host
pushes the current UCS via App::push_ucs_to_cmd before each point /
preview / text dispatch (DocumentTab::ucs_wire_affine, identity outside
model space). New axis-aligned commands just override set_ucs.
- RECT (two-corner) and RECT_CEN (center) place their derived corners on
the UCS axes (ucs_box_corners / ucs_box_around_center).
- ARRAYRECT row/column offsets run along the UCS axes.
RECT_ROT, POLYGON, and ELLIPSE already take their orientation from picked
points, so they need no change.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The dynamic-input boxes displayed deltas/angle in the UCS, but
dyn_resolve_point still resolved the typed (and live) values on the world
axes, so committing a typed coordinate ignored the UCS. It now rotates
the cursor delta into the UCS, interprets the typed cartesian/polar
values in that frame, and maps the relative result back to world via
UcsXform::vec_to_wcs — display and commit now agree. Absolute first-point
fields (no base) stay world; identity UCS is unchanged.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Two leftovers from the UCS grid work:
- The coloured X/Y/Z axis lines (draw_axes) were drawn at the world
origin along world axes; they now run through the UCS origin along the
UCS axes, matching the grid lines.
- Grid snap rounded the cursor on the world XYZ lattice; it now rounds in
the UCS grid frame (to-UCS, round, back-to-world) so it lands on the
grid the user sees. The UCS origin (wire space) + rotation reach the
snapper via DocumentTab::ucs_grid_basis; identity outside model space.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
OTRACK alignment rays (0°/90° and polar steps, plus the polar pencil from
the last point) were built along world X/Y. They now rotate with the
active UCS via a Mat4 passed into otrack_snap (scene.viewcube_ucs_mat() —
the UCS in model space, identity in paper), so tracking lines and the
intersection lock align to the user's coordinate system like ortho,
polar, and the grid. sync_ucs_to_scene now reuses UcsXform::rotation_mat.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Clicking a ViewCube face rotated the gaze into the UCS frame but derived
the up vector from world X/Y/Z, so the snapped view landed world-aligned
(most visibly Top, where the UCS shares world Z and only the in-plane
roll differs). snap_to_direction now takes the UCS basis and picks its up
from the UCS axes, so a face snap lands square to the user's coordinate
system. Identity UCS reproduces the world-aligned snap.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The on-canvas dynamic-input boxes measured relative X/Y deltas and the
polar angle in world axes. They now read in the active UCS plane (via
UcsXform::vec_to_ucs on the offset-invariant delta), matching the UCS
ortho/polar constraints and coordinate readout. Identity UCS reproduces
the world-frame values; absolute first-point fields are unchanged.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The ViewCube showed world Front/Top/Right. It now composes the camera
rotation with the active UCS (scene.viewcube_ucs, synced from the tab's
active_ucs), so its faces follow the user's coordinate system in model
space. Render uniform, hover/click hit-test, and the click→snap direction
all go through scene.viewcube_ucs_mat() so they stay in lock-step;
identity outside model space leaves paper-viewport navigation unchanged.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The grid ruled along world X/Y/Z through the world origin. It now rules
along the active UCS axes through the UCS origin (passed in render/wire
space), so it lines up with the user's coordinate system. Paper space
stays plain WCS; identity UCS reproduces the old world-aligned grid.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Ortho and polar drawing constraints snapped to world X/Y; under an active
UCS they now snap to the UCS axes, routed through the shared UcsXform
(identity = unchanged world-XY behaviour). Applied at all three input
sites (grip drag, command point, command preview). Typed coordinates
already flow through the converter via ucs_to_wcs / ucs_rotate_vec.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Introduce one bridge between WCS (how geometry and the file are stored)
and the active UCS (the coordinate system the user works in), so every
UCS-aware system routes through it instead of re-deriving axis math.
- UcsXform (app/helpers.rs): the converter — to_wcs/to_ucs (points),
vec_to_wcs/vec_to_ucs (directions), axes(), is_identity(). Orthonormal
axes, so the inverse is the transpose. The old ucs_to_wcs / ucs_rotate_vec
free fns now delegate to it (one implementation).
- DocumentTab::ucs_xform() builds it from the tab's active_ucs (None = WCS).
- DocumentTab::adopt_active_ucs_from_header() loads the document's saved
current UCS (header model-space UCS) into active_ucs on open, so the file's
coordinate system is live immediately; called from both load paths.
- Status-bar coordinate readout reports the cursor in the active UCS.
- UCS icon tripod projects the active UCS axes, so it rotates to the UCS
instead of always showing world X/Y/Z.
First slice of routing the app's coordinate systems through the converter;
input/picking, grid, ViewCube, and the remaining readouts follow.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Model-space tiled VPORTs and paper-space floating VIEWPORTs each carried
their own copy of the saved-view→camera math, so a fix to one drifted
from the other (the recurring grid/twist regressions). Route both, plus
the View-table and sheet-viewport paths, through one decoder and fix the
bugs that the duplication hid.
Decoder unification:
- New camera_from_view(direction, target, center, height, twist,
world_offset): the single saved-view→Camera decoder. camera_from_vport
(tiled), camera_for_viewport (floating, + auto-fit), and the View-entry
/ sheet-viewport apply paths all delegate to it. Tiled vs floating now
differ only in field source and the floating auto-fit fallback.
Twist round-trips on layout switch and file save:
- camera_from_view applies roll = -twist; the write path was dropping it,
so model→paper→model (and reload) reset the view to raw orientation.
- New Camera::roll() recovers the camera roll; sync_camera_to_document,
write_camera_view_entry, and vport_from_camera now store -roll as the
view twist on the *Active VPort, the OpenCADStudio_Camera_Model View
entry, the per-tile VPorts, and the paper sheet viewport.
Rotation-aware frustum cull:
- New view_cull_aabb(): the entity-cull rectangle is built from the
camera's right/up basis (so a twisted/rotated view's rotated rectangle
is enclosed) instead of world X/Y axes; returns None for tilted views
to disable the flat-XY cull rather than wrongly hide geometry. Applied
to the model-tile and paper-viewport cull sites.
Paper viewport scale:
- The floating-viewport auto-fit overlap test summed view_target +
view_center raw; under a twist view_center is a DCS offset, so the sum
landed far from the real WCS centre and wrongly tripped the auto-fit,
rendering content at the wrong zoom. It now tests the decoded
(twist-folded, wire-space) target. The CPU projection path derives its
scale from that same camera, locking hit-test/snap to the GPU render.
Grid enumeration:
- ViewportInstance gains grid_on; the two grid-view collectors collapse
into one grid_views() derived from the same active_viewports() list the
renderer uses, so the grid overlay can no longer drift from the views
on screen.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Gate target-specific code so neither build warns about the other's items:
native-only helpers (view, window icon, file dialogs, file association, plugin
host/guard/registry, self-update) are cfg'd to non-wasm; the wasm-only font
fallback (script_of / cjk_lang / request) is cfg'd to wasm. Drop an unused mut,
allow the few genuinely-conditional dead items (mesh_click_hit, GPU layout
helpers, parity stubs), and gate the embedded PlotSettings-unrelated dead
constants. Both targets now build clean (only the unavoidable nom/quick-xml
future-incompat dependency notice remains).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Grid (and grid-snap) state was shared through the global show_grid flag, so
toggling it in model space also turned it on in paper space, and a floating
viewport's grid painted across the whole paper — and beyond the 3-D view when
zoomed in.
- Store grid/snap per view: a model tile in model space, the active floating
viewport or the layout's sheet viewport in paper space, round-tripped via the
viewport status flags.
- adopt_view_display on layout switch and viewport enter/exit so the live
toggle follows the view you are in.
- Render one grid per view (paper_viewport_grid_views): the sheet over the
paper, each floating viewport clipped to its screen rectangle.
- Clamp the grid clip to the widget so an overflowing viewport rectangle never
paints grid outside the 3-D view.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Builds on the acadrust viewport/PlotSettings round-trip fixes:
- Write the sheet viewport's view position into view_center (DCS) with
view_target at the origin, matching AutoCAD; the old view_target form shifted
the layout.
- PAGESETUP now writes paper size, rotation, origin and units onto the Layout's
embedded PlotSettings (not just a side object), and bumps geometry, so an edit
shows on the sheet immediately and survives a save.
- Draw the printable-area guide (paper inset by plot margins, rotation-aware) as
a dashed rectangle, matching AutoCAD's layout view.
- Plot/PDF export falls back to the Layout's embedded PlotSettings via
effective_plot_settings(), so a loaded file's rotation/origin/scale apply.
Bumps acadrust to pick up the viewport-status and PlotSettings round-trip fixes.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
File/view/layer/style/undo-redo and similar one-shot commands dispatch a
single action instead of installing an interactive CadCommand, so they had no
module to register from and never appeared in autocomplete. Add a consolidated
CommandRegistration covering all 132 user-typeable one-shots (NEW, OPEN, SAVE,
SAVEAS, PLOT, PURGE, UNDO, REDO, ZOOM, LAYER, XREF, …). Internal dispatch
tokens the user never types (REFEDIT_BEGIN, REFCLOSE_SAVE, REFCLOSE_DISCARD)
stay excluded.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Merge PR #130 (KarimJerbi) and resolve review findings:
- Conflict (settings.rs): keep current persisted-settings layout, append
texteditmode field.
- Conflict (command/mod.rs): keep both new CmdResult variants — main's
live-entity ops and the PR's text-edit suspend/resume.
- Bug: invalid TEXTEDITMODE value no longer aborts via the Measurement
arm; it now re-prompts with the error and stays active.
- Dedup: single parse_texteditmode() helper replaces the value parsing
duplicated across commands.rs, settings.rs and textedit.rs.
- Strip trailing whitespace.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
In-place block edit now reads more like a dedicated editor:
- Fade everything except the edited entities toward the background (wires,
block instances and 3D solid meshes) so the block stands out while the
surrounding drawing stays visible for context.
- Don't leave a command active on enter — the user edits the block's geometry
freely (move, grips, draw, erase) and finishes explicitly.
- Show the right-edge vertical toolbar with clickable Save (REFCLOSE_SAVE) and
Discard (REFCLOSE_DISCARD) buttons while a session is active, so the edit
can be ended without typing.
The fade clears on save/discard, restoring full brightness.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Double-clicking a block reference (its wire or its solid body) now enters
in-place block edit (REFEDIT), so its geometry can be modified and the change
reflects in every instance on REFCLOSE.
Also drop REFEDIT's "non-uniform scale not supported" restriction: the
session now carries the INSERT's full forward/inverse affine (OCS, rotation
and non-uniform / mirrored scale) and applies it via the entity transform,
instead of a uniform scale + rotate + translate sequence. Mirrored blocks
(e.g. x_scale = -1) are editable now.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Clicking a block whose body is a solid tried top-level solid meshes first
and block-internal ones only as a fallback, so a stray solid behind the
block always won and the block couldn't be picked. Merge both into one
depth-sorted hit-test (solid_click_hit) keyed by the solid's own handle or
the parent INSERT, so the front-most body wins regardless of kind.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Box/lasso only tested top-level solid meshes, so a block whose visible body
is a solid (its mesh lives in the per-INSERT instanced cache, and it has no
wires to catch) was never picked up by a window/crossing/lasso. Add
block_mesh_box_hit / block_mesh_poly_hit that test each INSERT's instanced
block solid meshes and resolve a hit to the parent INSERT, and wire them
into the selection paths.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Box, window, crossing and lasso selection only tested wires and hatches, so
3D solids (which render as meshes) could never be selected that way. Add
mesh_box_hit / mesh_poly_hit that project the solid mesh and test it against
the box / lasso (window = all vertices inside, crossing = any vertex inside
or the region sitting inside the solid), and wire them into the selection
paths.
Also re-tessellate solids after a move/rotate/scale/mirror/copy: the entity
moved but its cached mesh didn't, so the solid stayed drawn at its old
position. Rebuild the mesh caches when a transformed/copied handle is an
ACIS volume.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The preview wires are tessellated in the target drawing's offset-relative
frame, but the centroid driving the preview translation was the source
drawing's, so the ghost drifted by the world_offset difference. Shift the
centroid into the target frame to match the committed paste. Zero within one
drawing.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Solids were missing from the entity transform dispatch, so move / rotate /
scale / mirror were silently no-ops on a 3D solid (only its wires shifted)
and a pasted solid stayed at its original location. Add a Transformable impl
for Solid3D/Region/Body/Surface (delegating to acadrust, which now composes
the move into the ACIS body placement) and register them in the dispatcher.
Also make cross-drawing paste land at the cursor: the clipboard centroid is
measured in the source drawing's offset-relative frame, so correct the paste
translation by the source/target world_offset difference (zero within one
drawing). Bumps acadrust for the solid-geometry transform fixes.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Cross-drawing paste recreated layers/linetypes/styles only for the
top-level selection, so objects inside a pasted block landed on layers the
target drawing didn't have. Scan the entities inside captured block
definitions for their dependencies too. Also refresh the layer panel after
a paste so the brought-in layers show in the manager and the dropdown.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
ACIS solids bake their colour into the mesh, so a property-panel or ribbon
colour (or layer) change left the solid showing its old colour — only the
wires, which resolve colour at render time, updated. Run recolor_meshes
(now covering block-definition meshes too) when property targets are
invalidated, and route the ribbon colour change through the same path.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Paste added the entities but never rebuilt the solid mesh caches, which are
only populated by populate_meshes_from_document (file open, undo, a few
commands). A pasted 3D solid — top-level or inside a recreated block
definition — therefore had no mesh and stayed invisible. Rebuild the mesh
caches after the paste so the solids show up.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Add an Invert Selection entry to the selection context menu (below
Select Similar): it replaces the current selection with every other
selectable object in the active layout. Candidates come from the visible
wire set, so objects on off/frozen layers stay excluded.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Copying an INSERT and pasting it into another drawing rendered nothing:
the clipboard snapshotted referenced layers/linetypes/styles but not the
block definition itself, so the pasted reference pointed at a block the
target drawing didn't have.
Capture every block an copied INSERT references — walking nested INSERTs
transitively, skipping model/paper space and xref blocks — and store the
definition's base point plus its owned entities. On paste, recreate any
of those blocks the target drawing lacks via a new define_block_raw
(entities kept in block-local coordinates), before adding the pasted
references. Same-document pastes are unaffected since the blocks already
exist.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
JOIN previously only merged collinear lines or co-circular arcs into a
single Line/Arc and rejected everything else as "not co-linear" — so the
segments of a broken polyline, which meet at vertices but run at angles,
could never rejoin. It also flattened endpoints to the first segment's Z,
silently discarding height.
Rewrite the geometry pass to chain segments wherever their endpoints
touch, regardless of angle: a collinear straight run still collapses to
one Line, a planar chain becomes an LwPolyline (arcs carried as bulges),
and a chain with varying Z becomes a Polyline3D, preserving height.
Open/closed is detected from the endpoints; a selection that isn't a
single connected chain, or holds an unsupported entity / tilted arc, is
refused with a clearer message.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The rollover hit-test now sweeps wires, hatches, block-internal hatches
and shaded solid bodies, which is O(N) per cursor frame and was freezing
the crosshair on large drawings. Each idle move now clears the live
highlight and arms a dwell timer; HoverDwellTick runs the pick once the
cursor has been still for HOVER_DWELL_MS (120 ms), using the active
tile's bounds so the deferred pick matches the entity actually under
the cursor in tiled viewports.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A buggy external plugin that panicked anywhere — during load (version,
register, manifest), during ribbon build / dispatch, or inside an
interactive command (prompt, on_point, on_enter, on_object_pick) — took
the host down with it (#134). Run every cdylib entry point under
catch_unwind so the host stays up: the plugin's call is treated as
inert (load fails, dispatch returns "didn't handle it", interactive
hooks resolve to Cancel) and a "Plugin <hook> panicked: …" line lands
in the command line on the next tick. (#145)
Scope is Rust panics across the cdylib boundary, not genuine UB —
real isolation (segfault-proof) still requires out-of-process plugins.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Ctrl+V was globally bound to the entity paste command, so it never pasted
into an open text editor, and even when routed there iced's clipboard read
is a no-op on the web. Route Ctrl+V to the focused editor and, on the web,
read the system clipboard via the browser's async Clipboard API (the
keypress is a user gesture, so the read is permitted):
- MText editor: insert at the caret (web + native).
- TEXT field: append on the web; native lets the text_input paste itself.
- No editor open: falls through to the entity paste command as before.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The web build has no system-font access, so CAD text in non-Latin scripts
(Cyrillic, Greek, CJK, …) rendered as nothing — build_fallback was a hard
None on wasm (#141). Instead, ship per-script Noto subsets under web/fonts
(one alphabet per file, SIL OFL 1.1) and fetch them lazily over HTTP the
first time a drawing uses that script, then outline glyphs with ttf-parser.
- web_font: Script enum + script_of (char → script font), a per-script
store with lazy fetch, and a pending queue drained by the app loop.
- CJK is split by language (chinese/japanese/korean). Han ideographs share
code points but differ by language, so the shared block is routed by the
document's $DWGCODEPAGE (932→JP, 949→KR, GB/936→CN); kana is always
Japanese, Hangul always Korean. Re-tessellates when the language changes.
- build_fallback (wasm) outlines from the fetched subset; clear_fallback_cache
lets glyphs that missed while a font was in flight reappear on load.
- PollWebFonts subscription (web only) + WebFontLoaded message drive fetches.
- Trunk copy-dir serves web/fonts; nothing is bundled into the native binary
(desktop keeps using system fonts).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The clap CLI work made boot() read crate::cli, but cli is cfg'd out on
wasm (the web build takes no args), so the wasm target failed to compile.
boot() is only used by the native iced::daemon run(); the web build boots
via boot_web(). Gate boot() with cfg(not(wasm32)) to match.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Replace the ad-hoc `--serve` arg sniff with a clap-parsed CLI:
- positional <file> to open at startup (also the OS file-association path)
- --new, --read-only, --backend, --safe-mode/--no-gpu
- --serve/--port (headless JSON automation, unchanged behaviour)
- --export IN OUT (headless format convert via the automation load/save infra)
- --script FILE (run command lines at startup)
- --log/RUST_LOG, plus --version/--help for free
GUI-only options pass to app::boot through a OnceLock config; read-only
gates the save paths. The web build skips parsing (no args).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A close signal can arrive while the app is backgrounded, leaving the
in-window unsaved-changes prompt unseen behind other windows. Funnel the
prompt opener through gain_focus + a Critical user-attention request so
the window raises (Windows/macOS/X11) or at least flags for attention
(Wayland via xdg_activation). Every unsaved-dialog caller shares this path.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Ribbon layer dropdown rows were a single button, so a click on the
visible/freeze/lock icon fell through to the make-active handler instead
of toggling that state. Split the row: each icon is its own toggle
button, only the swatch+label selects the active layer. Toggle handlers
now resync the ribbon so the icon updates while the dropdown stays open.
Layer Manager columns are now sortable: clicking a header sorts by that
column, clicking again flips direction (SVG arrow indicator). Sort is
reapplied on every panel sync so it survives edits, preserves the
selection by name, and keeps the ribbon dropdown order in step.
Fixes#133
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Solids (meshes) now tint on rollover and selection: hover washes orange,
selection washes blue, with selection winning when both apply. Meshes
re-upload their tint when the highlight set changes, and the rollover pick
now also hits shaded solid bodies and hatches (mirroring the click-select
order: wire → hatch → block hatch → solid).
Wires get the same scheme: the selection xray overlay recolours selected
entities blue and the hovered entity orange (previously both blue). Drops
the now-unused highlight_handles union helper and its cache.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
3D solid meshes inside block definitions were drawn once at block-local
coordinates without the INSERT transform, so a block authored in mm and
placed into a metre drawing at a 0.001 INSERT scale rendered ~1000x too
large. Split solid meshes into top-level (layout-owned, offset and drawn
flat) and block-definition (kept block-local) sets, and emit a transformed
instance of each block-def mesh per model-space INSERT, accumulating nested
transforms. Layout blocks are identified from the Layout objects'
block_record handles, since BlockRecord::is_layout() also flags ordinary
blocks in some files. (#123)
Co-Authored-By: Claude Opus 4.8 <[email protected]>
When drawing with Ortho (or Polar) on, a point that snapped to an object
(e.g. an Endpoint) was still pulled onto the ortho axis, so you couldn't
close a shape exactly on an existing vertex. Skip the ortho/polar
constraint while an object snap is locked, in both the preview and commit
paths. Grid snap is positional rather than an object snap, so it still
combines with ortho/polar. (#132)
Co-Authored-By: Claude Opus 4.8 <[email protected]>
