The web wire pipeline (fat-instance wire.wgsl) had grown to 17 vertex
attributes as the width-band and taper fields were added; WebGL2 and
WebGPU cap vertex attributes at 16, so pipeline creation failed and
the web viewport drew no lines at all — UI and SDF text still worked,
which is why it looked like "the app stopped working" rather than a
crash (#414). Scalars now ride in packed vec4/vec2 attributes (10
locations total); the instance struct's memory layout is unchanged.
Co-Authored-By: Claude Fable 5 <[email protected]>
GRADIENT silently degraded to a plain hatch: add_hatch rebuilt the
render model from the DXF entity but never encoded the gradient on it,
so the committed fill lost its gradient entirely. The entity now
carries the full gradient record (name, angle, colour stops,
solid-fill flag), and the standard shape set became selectable:
Linear / Cylinder / Spherical / Hemispherical / Curved plus an Invert
toggle, offered as prompt buttons on the GRADIENT command and as a
Gradient type dropdown + Invert box in the properties panel. All nine
standard DXF names round-trip (inverted linear persists by swapping
the colour stops), and the shaders shape each kind distinctly —
mirrored cylinder, eased curve, dome-profiled hemisphere — on both the
native and web pipelines without changing the GPU struct sizes.
Co-Authored-By: Claude Fable 5 <[email protected]>
A wide polyline with per-vertex varying widths rendered at its widest edge
(uniform-width ones were already exact). Draw the true taper instead.
`WireModel.taper_widths` carries a per-point full band width. The wire
instance (native slim + web fat) gains per-endpoint `world_hw_a/b`, and both
wire shaders interpolate `mix(hw_a, hw_b, which_end)` so each segment's band
narrows/widens linearly — depth-tested, and the linetype dash still applies.
Alignment (widths must match the tessellated points) is solved by building
both together: a new `TruckObject::TaperedLines(points, widths)`, filled by
`common::tapered_band_points` (straight = endpoints, arc = 16 samples with the
width lerped along it). to_truck routes a varying-width polyline here —
LwPolyline before the PLINEGEN split, Polyline2D at its return — while a
uniform-width one keeps the constant-band Contour path unchanged. `world_width`
stays the widest edge as the PDF-stroke + zoom-floor fallback.
Blocks fall back to the constant-widest band (`LocalWire` carries no
taper_widths).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A wide polyline's flat band was drawn by the hatch pipeline: a solid fill
with only draw-order depth. It ignored the entity linetype (the width
"crushed" the dash pattern) and showed through 3-D geometry in a tilted view.
Draw the band from the polyline's own centre-line wire instead. A new
`WireModel.world_width` (drawing units) makes the wire vertex shader expand
the quad by `world_half_width / world_per_pixel` (clamped to a hairline), so
the band is sized in world units and tracks zoom. Because it is now a wire:
- the GPU dash shader dashes the band by arc-length — the linetype applies;
- it is depth-tested at the polyline elevation — no draw-order see-through;
- no CPU band triangles or per-dash geometry are generated.
`world_half_width` rides in `WireConst` (native, replacing a pad) and the web
`WireInstance` (loc 14); both wire shaders branch on it. The tessellator sets
`world_width` from the LwPolyline width on the Contour / SegmentedLines arms,
and the model-space fill loop no longer emits a hatch band for any LwPolyline.
Pick still uses the band polygon (`pick_tris`).
Blocks keep the hatch band: `LocalWire` carries no `world_width` (finalize
zeroes it), so a scaled insert still scales its band correctly.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Entering a thickness (DXF 39 extrusion) on a wide polyline now works end
to end:
- The properties Thickness field is a General-group common prop, so it is
routed to apply_common_prop (set_entity_thickness) instead of the
geometry setter, which ignored it — the value was silently dropped and
reverted to 0.
- A wide polyline extrudes its whole band into a 3-D tube (outer + inner
walls with top and bottom caps), not just its centreline.
- The tube is real 3-D geometry, not a flat draw-order hatch: it renders
solid in every view mode with true depth, so nothing shows through the
walls. The flat 2-D band fill is skipped for a thickened polyline.
- A fill whose triangles span depth keeps its real depth (no draw-order
bias), so the extrusion's own edge wires stay in front of the surface
and remain visible instead of only the silhouette.
- The corner and cap edge lines read black, framing the coloured tube.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Rectangular paper viewports now go through the same GPU stencil clip as
non-rectangular ones: viewport_clip_boundary_ndc returns the four rect
corners (a full-target mask) instead of nothing, so the render path has
no rect/non-rect special case — every content viewport carries a clip
boundary and draws with stencil reference 0xFF.
With a single clip path, the old per-wire / -image / -wipeout scissor
mechanism is fully dead and removed: compute_*_scissors, the
*_pixel_scissors buffers, the project_scissor helper, the per-pass
set_scissor_rect blocks, the vp_scissor field threaded through
WireModel / ImageModel / HatchModel and their GPU mirrors (WireGpu /
ImageGpu / WipeoutGpu / HatchWebGpu), from_run's scissor parameter, the
scissor grouping in build_wire_buffers, and the is_arena_eligible gate.
Net -245 lines across 54 files.
Rectangular viewports still clip exactly via their own render rectangle;
the stencil mask is a redundant full-target no-op there, kept only so
the clip code path stays uniform. XCLIP keeps its CPU wire clip.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Paper content viewports with a circle / spline / ellipse clip boundary
drew their model content to the full bounding rectangle instead of the
shape. Each paper content viewport is its own render into its own screen
rectangle, so clipping is done per viewport: the boundary polygon is
projected into that viewport's render-target NDC (paper shape mapped
through the same visible-sub-rect crop as the content, so it stays fixed
to the viewport frame) and stamped into the stencil buffer with an
even-odd Invert, which marks the interior of any shape, convex or not,
without triangulation. Every content pass then draws only where the
stencil is set, so hatch, wire, mesh, 3DFACE, text, image, wipeout and
selection are all clipped uniformly.
Because Invert turns the cleared 0x00 into 0xFF, clipped content draws
with stencil reference 0xFF (unclipped content keeps reference 0). The
shared depth buffer moves to Depth24PlusStencil8 across every pipeline.
Rectangular viewports keep clipping via their own render rectangle and
XCLIP keeps its existing CPU wire clip; neither needs the stencil path.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A gradient hatch was drawn linear from the entity colour to its second
stop, so a red->yellow spherical fill came out as a pale linear wash. Two
fixes:
- Colour: start from the gradient's first stop, not the entity colour.
hatch_model_from_dxf already baked stop-0 in; stop synced_hatch_models
from clobbering it with the entity's resolved colour for gradient fills.
- Shape: spherical / hemispherical / curved gradient names now render as a
radial fill (new shader mode 3, centred on the boundary with radius to the
farthest vertex) instead of linear. Native and web shaders both updated.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A raster image reference now resolves through one memoised path shared by
the loader and the unresolved-reference placeholder:
- http/https URLs are fetched natively (rustls, 8s timeout, 32MB cap); a
local file is read as before. The web build never fetches. The cache is
cleared when a new document opens.
- Online, the fetched image displays; offline or on a broken/missing
reference, the frame + saved path text still shows. One resolver keeps the
two in agreement without fetching twice.
The raster is also clipped to its boundary: the image draws as the
triangulated clip polygon (ear-clipped) with per-vertex UVs instead of a
full quad, so pixels appear only inside the clip. Clip-boundary Y is raster
space (top-down), so it is flipped to the image's Y-up space in both the
outline and the clip triangulation so they align.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
A SOLID's filled quad and a dimension/MultiLeader text background are flat 2-D
overlays that should draw in wireframe and shaded alike, but face3d_gpu
classified fills by `!fill_tris_low.is_empty()` — and a 2-D fill at UTM scale
carries a double-single low residual for precision just like a real 3-D mesh
face. So the arrowhead fill and the "5;8" cyan text background were routed to
the 3-D buffer and only appeared in shaded mode, vanishing back to nothing on
return to wireframe.
WireModel now carries an explicit `fill_is_3d`: true only for a real surface
(PolyfaceMesh / PolygonMesh face, set in the tessellator), false for every flat
overlay fill. face3d_gpu classifies on the flag, so 2-D fills stay in the
always-visible buffer while keeping their low residual (and UTM precision).
Verified on budweiser: the `_BoxFilled` arrow SOLID and the dim '5;8' MTEXT
background both tessellate with fill_is_3d = false.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
An open "C" tube (the € currency symbol's ring) rendered as a full closed
donut: truck builds a torus as a full revolution and can't trim it, and the
wireframe/silhouette arc estimate was wrong.
The arc estimate failed because `collect_face_polygon` reads only the first
loop, and the "widest empty gap = tube body" heuristic breaks once the tube
carries interior holes (where the two horizontal bars punch through it): a hole
splits the body into wide hole-free stretches that masquerade as the opening.
`torus_phi_range` now walks every boundary loop and classifies each: a minor
circle sitting in a constant-φ plane (spread < ~8.6°) is an end cap; a wider
loop is an interference hole. The opening is the gap flanked on both sides by
end caps — a hole never bounds it — so the body is the rest of the turn.
Torus faces now route to the bespoke sampler (like spheres); `tess_torus_face`
clips its φ grid to that arc, so the shaded tube is an open C. `torus_isolines`
and `collect_curved_gens` use the same range for the wireframe rim and the
silhouette. Removes the now-dead truck torus path (`torus_faces`,
`Outward::Torus`, `vadd`).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Extends the curved-face wireframe treatment from cones to spheres and tori, so
a solid built from those faces reads the way a cylinder already did.
ISOLINES: sphere faces get meridian lines across the face's colatitude span at
evenly spaced longitudes; torus faces get minor cross-section circles at evenly
spaced revolution angles. Both clip to the face's own parametric window (from
the boundary polygon via a shared `angular_span` gap-finder), so a partial
sphere or torus arc doesn't draw lines out over the missing part.
Silhouettes (DISPSILH): `CurvedGen` becomes an enum. A sphere silhouettes as
the great circle perpendicular to the view, clipped to the face window; a torus
as the two curves where the tube turns edge-on, traced around the revolution.
The silhouette now uses the camera's view *direction* alone, not the eye
position — a single parallel direction for the whole scene — so the outline
follows the view angle, holds still under pan, and doesn't foreshorten, which
is the orthographic silhouette a CAD wireframe expects.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
ISOLINES show a curved solid from any angle, but its outline still cut across
whichever fixed longitudinal lines happened to face the eye. DISPSILH draws the
true silhouette — the two lines where a cylinder (or cone) turns edge-on to the
view — and it moves with the camera, so it cannot be a baked edge.
Tessellation stores a world-space generator per cone/cylinder face on the mesh
set (axis, radial frame, radius, cone taper, extent), sharing the face-extent
extraction with the ISOLINES collector. Each frame `upload_silhouettes` walks
those generators, solves `du·cosθ + dv·sinθ = -tanα·da` for the two edge-on
angles (`φ ± acos(...)`, `φ ± π/2` for a cylinder), and builds a line list in
the mesh vertex format — double-single, so it stays precise at UTM. It draws
through the existing wireframe / edge pipelines in the wireframe and
shaded-with-edges passes.
Gated on the document's DISPSILH header flag, which defaults off — so the
per-frame pass is a no-op unless the drawing asks for it. Scoped to top-level
solids: a block instance's generators aren't transformed by the INSTANCE pose
yet, so they are cleared on the block path (its isolines and feature edges
still draw).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The uncaptured-error handler from #404 was armed in `Pipeline::new`, which
`MultiPipeline::ensure_len` runs again for every pane added — so the slot was
re-registered on every new pane. Move it to `MultiPipeline::new`, which runs
once per device; if iced ever rebuilds the device it builds a new
`MultiPipeline` too, so the fresh device stays covered — where a `Once` guard
would have left it on wgpu's panicking default.
Throttle the message. The handler slot is per-device and single, so it also
catches iced's own draw errors and is the only thing that can report them — it
has to stay loud. But a validation error normally repeats on every frame, and
an unthrottled print floods stderr at refresh rate and buries the first, most
useful message. Log the 1st, 2nd, 4th, 8th … occurrence instead: the error is
never hidden, the running count shows it is ongoing, and the output stays
finite.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Face3DGpu batched every visible 3DFACE quad and PolyfaceMesh /
PolygonMesh fill triangle into a single vertex buffer. A mesh-heavy
DWG (e.g. a Navisworks import) blows past wgpu's default 256 MB
max_buffer_size once enough layers are thawed; the resulting
uncaptured validation error panics inside iced's main-thread redraw
and aborts the whole process. Split the 3D and 2D fill streams into
per-limit chunks on whole-triangle boundaries — the same scheme
build_mesh_batch already uses (#203) — and register an
on_uncaptured_error handler so any remaining GPU validation error
degrades the frame instead of killing the session.
Co-Authored-By: Claude Fable 5 <[email protected]>
MLINE dashed elements were fully CPU-expanded (apply_along) — one wire per
dash — to keep the parallel lines in phase. Move pure-dash elements onto the
GPU dash shader and end-align every element to the multiline's centre-line so
they share one interior phase.
- wire_distances now accumulates arc-length from the double-single (high+low)
points, so the dash phase stays precise at UTM coordinates. f32-high-only
deltas quantised ~0.1 there and drifted the parallel lines apart — the
original reason MLINE dashes were CPU-side.
- New WireModel.dash_align_end carries a shared begin/end dash length. The
shader uses it as align_end for every parallel element (the interior phase
depends on align_end, not the wire's own length) while align_total stays the
element's own length, so each still ends on a dash at its own endpoint.
- MLINE tessellation derives that length once from the centre-line, routes
pure dash/space/dot elements to the GPU pattern and text-bearing elements to
apply_along.
- apply_along gained an "A"-type mode (shared reference length): a solid
begin/end dash and a phased interior, so its dashes — and embedded text —
line up with the GPU-dashed sibling elements instead of tiling from start.
Pure-dash MLINEs now cost one WireModel + pattern per element instead of N
segments; single (non-MLINE) entities keep the from-start tiling unchanged.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
glam 0.33.2 deprecated Mat4::{orthographic,perspective,look_at}_rh in
favour of the glam::camera::rh::* functions. Switch the 7 call sites in
camera.rs and viewcube.rs to the equivalent camera-module functions
(directx orthographic/perspective = same [0,1] depth range; look_at_mat4
= same view matrix), silencing the deprecation warnings. No behaviour
change.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
DET pipe walls now carry the stroke pattern (dash_from_start), which
tiles from each wall's own start vertex and skips the A-type end
alignment that was force-lighting a solid dash under the end symbol.
Two rendering fixes on top:
- Dedup the host polyline's vertices before offsetting the walls. Some
DGN pipe polylines carry trailing duplicate end vertices; a zero-length
segment gave offset_xy a degenerate normal, leaving that vertex
un-offset and folding the wall down to the centre line. The two walls'
fold segments met at the centre and drew a spurious line linking them.
- Re-key the symbol (end circle / half-ellipse) wires to the host handle
so the whole pipe picks and selects as one entity instead of the
symbols acting as separate phantom entities.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The single-batch arena required a set with no mesh/solid edges, so any drawing
with a 3D solid or mesh (this one: 1693 mesh-edge wires of 12200) disqualified the
WHOLE set — every edit fell back to build_wire_buffers re-uploading ~265 MB
(4.4M instances), stalling ~1s per edit on a shared-memory GPU.
Split the resident set into a regular-2D arena and a mesh-edge arena (the wires
that need the draw loops mesh-edge skip/black treatment), one WireArena each,
passed as &[&WireModel] subsets (no clone). Both patch incrementally; the mesh
arena carries is_3d_mesh_edge so the existing draw loop handles it. A changed
handle absent from an arenas subset is a no-op for it, so each edit touches only
the batch it belongs to. Result: the wire block drops from ~300ms + a 265 MB GPU
upload to ~2ms.
Two correctness points folded in from adversarial review:
- mesh-edge wires are stamped draw_depth 0 (matching WireGpu::from_run) so their
hidden back edges dont peek through the shaded fill; the mesh arena also always
bails on append (all its wires share depth 0, so relocation would reorder
coincident edges).
- buffers are capped at wgpus 256 MB limit — a batch larger than one buffer
returns None so the caller chunks it via the batched path (fixes a create_buffer
panic on the 355 MB regular batch).
Still behind OCS_WIRE_GPU_PATCH (default off). OCS_PERF prints the per-edit wire
timing + outcome.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The in-place-only arena still fell back to a full batched rebuild on add/remove —
re-uploading the entire wire instance buffer, which on a drawing whose wires
expand to millions of segments (~265 MB) stalls ~1s per add/remove on a
shared-memory GPU. Only move/edit avoided it.
Add append/tombstone so add/remove patch the instance buffer too: a new or
segment-count-changed entity bump-allocates a slab at the tail (buffers carry
1.5x headroom); an erased entity is tombstoned (blanked at const slot 0). The
huge instance buffer only grows/blanks by the changed entity.
Three correctness points make add/remove safe (each verified by an adversarial
review, findings folded in):
- draw_depth_map normalises each entity`s z-bias by the block entity count, so a
structural change re-scales every entity`s bias. Keep a CPU mirror of the const
array; on any structural change refresh every slab`s draw_depth and re-upload
the small (~1 MB) const buffer, leaving the instance buffer untouched.
- a tail-appended entity draws last, which only mis-orders wires whose overlap
resolves by submission order rather than the z-bias: transparency, and 3D-solid
fallback edges (Solid3D/Region/Body/Surface are excluded from draw_depth_map, so
their opaque edges share depth 0). `append_unsafe` bails to a full rebuild when
the set contains either.
- an in-place edit also flushes its consts to the GPU (a colour change is not
structural, so the whole-buffer refresh would otherwise skip it).
Still behind OCS_WIRE_GPU_PATCH (default off); every uncertain case returns false
and the caller does the identical full rebuild, so correctness never rides on the
fast path.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
On a geometry edit the resident set gets a fresh content id, so the batched wire
path (build_wire_buffers) re-emits and re-uploads EVERY wire instance — the whole
buffer even when one line moved. Behind the OCS_WIRE_GPU_PATCH flag (default off,
native only), add a persistent per-entity instance arena: one long-lived
VERTEX|COPY_DST instance buffer + its WireConst storage, with a per-entity slab
table. An in-place edit (move / rotate / scale / colour — same segment count)
becomes a queue.write_buffer of just that entity`s slab.
Scope is deliberately IN-PLACE ONLY. draw_depth_map normalises each entity`s
draw-order z-bias by the block`s entity count, so an Add or Remove re-scales EVERY
entity`s z-bias — which a per-slab patch can`t honour — and a segment-count change
relocates the entity. patch() returns false for any of those (and for mesh-edge /
scissored sets), and the caller falls back to the identical full batched rebuild,
so correctness never rides on the fast path. An adversarial review of an earlier
append/tombstone design surfaced exactly these draw-order hazards; restricting to
in-place removes them while keeping the common interactive-edit win.
The Scene hands the changed handles to the arena via model_wire_gpu_patch,
recorded by try_resident_patch for the Model block and matched by content id in
prepare(); scissored paper viewports and mixed 2D/3D sets keep the batched path.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Each pipeline slot owned its own gpu_wires: every paper viewport and every
Model tile rebuilt an identical resident wire set into its own GPU vertex
buffers and its own handle index. With N content viewports drawing one shared
tessellation that is N copies of the (potentially huge) wire buffers in VRAM
plus N batch+index builds on every content change — the issue-131 memory
exposure multiplied by the viewport count.
A wgpu::Buffer is internally ref-counted, so sharing a WireGpu by Arc shares
its GPU allocation. Make Pipeline.gpu_wires an Arc<Vec<WireGpu>> (and
wire_handle_index an Arc<map>), turn upload_wires into build_wire_buffers(&self)
that returns those Arcs, and add MultiPipeline.wire_buffer_cache keyed by
wire_content_id. prepare() builds an entry once on a cache miss and hands Arc
clones to every slot with that id — so N viewports (and, since one MultiPipeline
backs all panes, N panes) upload the wire vertices exactly once between them.
Camera uniforms and scissor stay per-slot, so the draw still issues one call per
viewport with the right view; only the geometry is deduplicated. The shared Vec
is never mutated in place — only reassigned on a content change — so the sharing
is sound.
The cache drops entries no slot still references (Arc::strong_count == 1) once it
passes a small bound, so a live buffer (held by any pane) is never evicted.
Model multi-tile gets the same dedup for free, and two panes showing the same
model now share buffers.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
WebGL2 has no vertex/fragment storage buffers, so the batched storage-buffer
hatch renderer (hatch_gpu) is disabled on wasm and real hatch fills never
rendered on the web build — the pick highlighted the boundary and the command
then appeared to cancel (issue #204).
Add a wasm-only per-hatch renderer (hatch_web_gpu + hatch_web.wgsl) that packs
the variable-length boundary / family / dash arrays into one RGBA32F data
texture read with textureLoad, instead of the fixed-size uniforms of the
wipeout path. This removes the MAX_FAMILIES (16) / MAX_HATCH_BOUNDARY_VERTS
(1024) / MAX_DASHES (128) caps, so every hatch — solid, gradient, and
arbitrarily complex line patterns — renders on the web. The fill algorithm
(in_polygon, per-family line/dash/dot, gradient, LOD, double-single vertex
transform) is identical to the desktop path; only the array ACCESS changes.
Pattern phase is anchored at world_origin with the boundary stored raw, and
scale / gradient-range clamped, to match the desktop batched renderer exactly.
Native (hatch_gpu + hatch.wgsl) and the wipeout mask renderer are untouched;
the new module and pipeline are gated on target_arch = "wasm32".
Also silence the resulting (and pre-existing) wasm-only dead-code warnings for
native-only paths so both the native and web builds compile clean.
Closes#204.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Dashed/dash-dot lines could end in a gap ("empty space at the end"),
especially inside paperspace viewports. The dash pattern was only centred
on each wire, which makes the two ends symmetric but does not force them
onto a solid dash.
Implement AutoCAD "A" (aligned) linetypes: the line begins AND ends on a
solid dash, interior dashes stay nominal, and the two end dashes stretch
symmetrically to absorb the remainder (so parallel lines share an
identical interior). wire_distances computes the end-dash length D and
total length L for a dash-first pattern (D = (L - k*P + a) / 2); the GPU
in_dash walk force-lights the two end regions and phases the interior to
resume right after the start dash. Lines shorter than one pattern period
draw solid.
Only alternating dash-first patterns are aligned; gap/dot-first,
consecutive-dash, NaN-break (plinegen=false) and solid lines keep the
prior behaviour. Native and web (wasm) shaders kept in sync.
Refs #13.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Pipeline slots are addressed by list index, but the renderer drops
off-canvas viewports from the list, so a slot can be reused by a
different viewport across frames — e.g. when the first paper viewport
scrolls off the canvas the second slides into its slot and inherits its
stale, differently frustum-culled GPU buffers. The text upload is gated
on a geometry-epoch change (unaffected by the index shift), so the
survivor rendered the gone viewport's text and its own vanished.
Give each viewport a stable identity, remember which one last used a
slot, and reset all cache keys when the occupant changes so wires, text,
hatches, meshes and Face3D re-upload for the new viewport. No cost in the
steady state — the reset fires only when a viewport enters/leaves view.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Command previews (MOVE, COPY, ROTATE, SCALE, STRETCH, MIRROR, ARRAY,
PASTE, INSERT) build ghost WireModels via WireModel::translated/rotated/
scaled/stretched/mirrored, but those helpers moved only `points`, never
`text_verts`, and the ghosts' glyphs were never uploaded — so text
vanished (or froze at its source) during the preview, same root cause as
the grip-drag case.
Transform text_verts alongside points in all five helpers (shared
map_text_verts, reconstructing + re-splitting the double-single world
position), and gather preview_wires' glyphs into the per-frame
preview-text buffer in the render path. This covers every command
preview automatically. The grip-drag else branch no longer sets
preview_text explicitly (the render gather handles it); the grip-slide
fast path keeps its own scene.preview_text channel (empty preview_wires,
so no double-draw).
MIRROR honours MIRRTEXT: on -> true glyph mirror via mirrored(); off ->
readable relocation via translated().
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Dragging text by a grip hid it until release. The dragged entity is
hidden from the base wire set and shown as a preview overlay, but its
SDF glyph quads (text_verts) were never uploaded for the preview, so
the text vanished mid-drag.
Add a per-frame preview_text GPU buffer (sibling to preview_wires). A
rigid whole-entity move of pure text (TEXT/MTEXT, Square insertion grip)
slides the drag-start glyphs by the drag delta instead of re-shaping the
run every move; reshape grips (MTEXT width), dimensions and wire
entities re-tessellate and feed the same buffer.
Also fold preview_text into render_signature: a pure-text slide leaves
preview_wires empty, so without it the scene-render cache froze the
dragged text at its first frame (grip moved but text stayed, then
jumped to place on release).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The batched path is the sole hatch renderer and has no family cap, yet it
was named hatch_batched_gpu while the old capped per-primitive renderer
kept the hatch_gpu / MAX_FAMILIES name — which reads as if visible hatch
fills are truncated at 16 families (they are not). Swap the names so the
active renderer is hatch_gpu.rs / hatch.wgsl and the legacy one, now used
only for wipeout masks, is wipeout_gpu.rs / wipeout.wgsl. MAX_FAMILIES now
lives only in the wipeout module. Pure rename; no behaviour change.
Refs #323.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Dragging a model-tile pane_grid divider changed each pane's pixel size
every frame, and `ensure_depth_texture` recreated the depth + 4×MSAA +
resolve textures on any size delta. On Windows Firefox (WebGL2 → ANGLE →
D3D11) this per-frame texture realloc hangs the D3D driver — the same
hang class as the native DX12 lock in #120. Linux Firefox (native GL) and
Chromium/Edge tolerate the churn, which is why the freeze is specific to
Windows Firefox.
Round the depth/MSAA/resolve allocation up to a 128px grid and only
recreate when the rounded size changes, so a live drag reallocates at
most once per 128px instead of every frame. Geometry passes still render
at the real size into the [0,0,render] corner of the larger texture, and
the blit UV is scaled by render/alloc so it samples exactly the filled
region — visually a no-op on every platform.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The batched hatch path pre-triangulated every boundary on the CPU
(PR#300) with no cap on the resulting vertex buffer. A hatch-dense
drawing (13k solid-filled hatches) built a ~6.75 GB `hatch_batched.vertex`
buffer and panicked in `Device::create_buffer` (device max 256 MB), after
a long stall spent triangulating on load.
Drop the CPU triangulation and always draw each hatch as its AABB quad
(6 verts) with `poly_test = 1`, letting the shader resolve inside/outside
per fragment against the uploaded boundary — the path the shader always
kept as a fallback. The vertex buffer is now 6 verts/hatch regardless of
boundary complexity, so no drawing can overflow it. The interaction-LOD
flicker PR#300 addressed no longer occurs (OCS_HATCH_LOD defaults off).
Removes the now-unused hatch_tess triangulation module.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
External contribution by sLuCHa. Triangulates each hatch boundary once
(lyon even-odd, cached with the batch build) and rasterizes the triangles
instead of the per-fragment point-in-polygon test, removing the screen-fill
overdraw + boundary raycast that forced the flicker-causing interaction LOD
(now default off behind OCS_HATCH_LOD). Precision preserved (triangles in the
same world_origin-relative frame + RTE path); untessellatable boundaries fall
back to the old AABB-quad + in_polygon path (poly_test=1).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Rearchitect SDF text from a document-wide collector to per-entity
production during tessellation, so each entity owns its glyph quads and
composite/nested text is covered for free (no re-explosion, no per-frame
document walk).
- WireModel gains `text_verts`: glyph quads ride on the wire, so they are
cached by the tess memo, cloned on hit, and transformed by the block
expand loop exactly like `points`.
- Text arm (tessellate.rs) builds the quads when SDF is on and suppresses
the strokes; render gathers text_verts from the viewport wire set
(`gather_text_verts`, cached on wire content id) — the old collector is
removed. Covers model, dimension, table, block-internal and paper-sheet
text uniformly.
- MLEADER text (lays out via layout_mtext, not the Text arm) gets its own
SDF branch driven by the per-run GlyphRun.
- Pick box comes from the rendered glyph quads (true text extent), not
entity_aabb which mis-places the box for MTEXT; composite paths no longer
clobber the empty text wire's tight AABB (dim pick box hugs the text).
- Selection / hover highlight via a text-highlight overlay (upload_text_
highlight, keyed on selection_generation) mirroring the selected-wire
xray overlay; base glyphs stay neutral so a deselect leaves no stale tint.
- Debug: env OCS_TEXT_BOX draws each run's glyph-bounds rectangle.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The batched hatch pass drew a screen-covering AABB quad per hatch and,
for every fragment, ran a point-in-polygon ray-cast over the whole
boundary plus the pattern test, with a discard that disabled early-Z.
Zoomed into a hatch the quad covers the screen, so the shader re-ran
per pixel every frame — the cost the PR #273 interaction-LOD worked
around by suppressing hatches during navigation, which made fills
flicker on and off while zooming.
Triangulate each boundary once (lyon, even-odd — same coverage as the
old in_polygon, islands/holes included) and rasterize those triangles
instead: no screen-filling overdraw, no per-fragment boundary test. The
fragment shader keeps the boundary test only on a per-instance fallback
path (poly_test == 1) for boundaries that fail to tessellate.
Tessellation is cached with the batch build (geometry_epoch) in local
space, so the double-single relative-to-eye precision is preserved.
With the fast path the interaction-LOD is no longer needed; it is gated
behind OCS_HATCH_LOD (default off) as a safety net for now.
Wires the SDF glyph atlas into the render loop end-to-end for top-level
TEXT entities, gated by the OCS_TEXT_SDF env var so the default path is
unchanged. Verified rendering on a real drawing: glyph quads land exactly
over the existing stroke text (position, size, rotation), confirming the
atlas → shader → pipeline → screen path and the RTE / UV math.
- text::text_run_placement — extracted from to_truck (behaviour-preserved)
so the stroke and SDF-quad paths compute an identical run placement.
- sdf_atlas::text_atlas — process-wide GlyphAtlas (Mutex/OnceLock), shared
by the collector (bakes glyphs) and the GPU upload (reads texels).
- Scene::sdf_text_vertices — collects per-glyph TextVertex for the frame's
TEXT entities via layout_glyph_quads + push_glyph_vertices. Empty unless
OCS_TEXT_SDF is set.
- ViewportData.text_verts + Pipeline text_pipeline/atlas/vbuf, upload_text
(rebuilds the R8 atlas texture when it grows), and a text render pass
drawn over the wires.
Still additive over the stroke text; only top-level TEXT is wired (MTEXT,
attributes, and block-internal text are next), and the collector re-runs
per frame (vertex caching by geometry epoch is a follow-up).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Groundwork to render CAD text as per-glyph SDF quads (box-quad snap,
perf, no LOD) instead of stroke-polyline wires. CPU side fully tested;
GPU side compiles but is not yet wired into the render loop.
- scene::text::sdf_atlas — GlyphAtlas: shelf packer + R8 SDF tiles baked
per (font, char). Unified field so one shader draws both kinds: TrueType
= signed distance to the filled outline, LFF stroke = pen band. Uses only
crates already in the tree (ttf-parser/lyon via the existing Glyph).
- scene::text::glyph_quads — layout_glyph_quads: lays a run out as per-glyph
quads using the exact tessellate_text_run transform, so quads land where
the strokes would. Per-char advance (TTF shaping is a follow-up).
- shaders/text.wgsl — SDF fragment shader (smoothstep over fwidth), RTE
vertex path mirroring image.wgsl, per-vertex colour + draw-order bias.
- scene::pipeline::text_gpu — TextVertex + double-single vertex builder,
R8 atlas texture upload, bind group, and the image-cloned pipeline.
13 unit tests across the three CPU stages, all green. Modules carry
#![allow(dead_code)] until the render-loop integration lands.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The batched hatch pass evaluates its fill procedurally per fragment — a
point-in-polygon test over the whole boundary plus the pattern test, with a
`discard` that disables early-Z. When the view sits inside a hatch its quad
covers the screen, so panning / zooming re-runs that full-screen shader every
frame (~40 ms/frame on a hatch-heavy plan) even though the rest of the scene
is cheap — the wire pass and CPU prepare are negligible by comparison.
Suppress the hatch pass while the view is actively panning / zooming /
orbiting, detected from `camera_generation` plus a short settle timer
(`Scene::navigating_lod`), and fold that state into the render signature so
the moment the view settles it renders hatches once and the scene-render
cache holds that full-quality frame. A brief `is_settling` subscription keeps
requesting frames just past the settle point so the hatched frame renders
even when no input event would otherwise fire. Mirrors AutoCAD's adaptive
degradation. The paper sheet keeps its fills.
Every cursor movement over the viewport re-rendered the entire GPU scene
(all wires, hatches, meshes) even when nothing visible changed: iced redraws
the whole window on each input message, and the shader's draw/prepare/render
runs every time. On large drawings one such redraw exceeds the frame budget,
so moving the mouse or panning stutters — and it gets worse the bigger the
drawing.
The pipeline already resolves each frame into a persistent, clip-sized
resolve texture and blits that to the surface. This hashes the inputs that
determine the rendered image (camera uniforms, geometry/selection
generations, wire content id, render-mode flags, clip size, live preview
coordinates) into a per-viewport signature. When a frame's signature matches
the previous one the image is pixel-identical, so `prepare` skips the buffer
uploads + per-frame scissor/LOD recompute and `render` skips passes 1-7 plus
the MSAA resolve, re-blitting the resolve texture instead. A pure cursor move
becomes a single fullscreen blit rather than an O(N) re-rasterization.
The ViewCube is excluded from the signature (it renders in its own always-on
pass) so cube hover still tracks while the scene is cached. Any real change —
edit (geometry_epoch), select/hover (selection_generation), camera move or
render-mode toggle — differs the signature and forces a full render, so
nothing goes stale. Navigation still renders every frame (the view genuinely
changes); this targets the wasted re-render on cursor-only movement.
Mesh (PolyfaceMesh / PolygonMesh) outline edges are drawn through the wire
pipeline, whose shader pulls each wire toward the camera by
draw_depth * DRAW_ORDER_BIAS (0.001) for 2-D draw-order layering. Every
entity gets a draw_depth in [-1, 1], so a mesh's edges were pulled forward
too — and when the mesh's depth extent projects to less than the bias (a
small or distant mesh), its hidden back-face edges peeked through its own
shaded fill. Solid feature edges use the mesh shader, which has no such
bias, which is why only meshes showed it and why adjusting the fill's depth
bias had no effect.
3-D mesh outline edges are real geometry occluded by true depth, so force
their draw_depth to 0 (they are already grouped as `mesh_edge` in
upload_wires). They now occlude exactly like solid edges.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Extends the solid feature-edge behaviour to mesh entities (PolyfaceMesh /
PolygonMesh): their triangulation outline edges now render black when a
filled-with-edges mode is active, and keep the entity colour in the
lines-only modes (wireframe / hidden-line).
A mesh entity emits its face fill and its outline edges as separate
WireModels sharing the entity handle: the fill carries fill_tris plus a
non-empty fill_tris_low (real 3D depth), the edge carries points. The old
is_3d_mesh_edge flag keyed on `!fill_tris.is_empty()`, which tagged the
points-less fill wire rather than the edge wire, so the edges were never
recognised. upload_wires now collects the handles of real 3D fill wires and
flags any points-wire sharing one as the mesh edge (text greek, whose fill
has an empty fill_tris_low, is excluded — matching the face3d test). The
wire draw loop then routes flagged edges through a new black wire pipeline
(fs_black) in filled-with-edges modes.
fs_black lives in wire_indexed.wgsl (the native shader) and wire.wgsl (the
wasm shader).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Solid wireframes rendered every triangle edge of the tessellation, so a
curved solid read as a dense mesh instead of its real outline. Now each ACIS
solid's B-rep face-boundary edges are collected from the SAT `edge` records
(straight edges as segments, ellipse/circle edges sampled along their bounded
arc), body-transformed and split into the double-single pair, and stored on
the MeshLodSet. The block-instance and top-level offset paths transform these
edges alongside the mesh vertices.
The batch builder routes a solid with feature edges into a dedicated LineList
vertex buffer and skips its triangle wireframe; a mesh with no B-rep edges
keeps its triangulation edges. Edge colour follows the display mode: the flat
`fs_edge` shader draws the entity colour in the lines-only modes (wireframe,
hidden-line) and a `fs_edge_black` twin forces black for the edge overlay in
filled render modes.
Verified: the Clinic storefront arch bars now yield ~14-16 clean edge
segments (vs the full triangulation) bounded to the solid's true extent.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Once solids carried a real sub-1 alpha (after the ENC transparency fix),
the single depth-writing mesh pass made a transparent solid occlude the
geometry behind it: it wrote depth at its near surface, so anything farther
failed the LessEqual test and was never drawn — the transparent fill looked
like it erased what was behind it.
The batched mesh builder now routes each solid's triangles into an opaque or
a transparent index stream by its baked colour alpha (< 0.999). The opaque
fills draw first and write depth; the transparent fills draw last through a
new depth-write-disabled pipeline, so they blend over the geometry behind
them instead of culling it. The hidden-line depth prepass still rasterises
both streams so solid surfaces keep occluding hidden edges.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The batch split at a fixed 6M-vertex cap that assumed 40 B/vertex, but
MeshVertex grew to 52 B when the double-single `position_low` field was
added — so a full chunk was 6M x 52 = 312 MB, past wgpu's 256 MB
max_buffer_size, and Device::create_buffer panicked on large models.
Derive the caps from device.limits().max_buffer_size and the real vertex
size, bound both the vertex buffer and the (fatter) wire-index buffer,
and split any single mesh too large for one chunk into triangle-soup
sub-chunks so no batch buffer can exceed the device limit.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
PolyfaceMesh / PolygonMesh / SubDMesh fills and 3DFACE quads were
classified as 2D fills and given the draw-order z-bias, so they were
pulled toward the camera and drew over solids. The tessellator emits a
mesh's edges and its fill as separate WireModels (points-only vs
fill-only), so the old `!points.is_empty()` discriminator misrouted the
fill. Classify a fill as a 3D surface by its double-single low residual
(`fill_tris_low`) instead, and drop the draw-order bias from 3DFACE
quads. 2D annotation fills (text greek, MultiLeader / dimension
backgrounds) leave `fill_tris_low` empty and keep ordering by rank.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Opening / navigating a heavy 3D DWG (e.g. a BIM model with ~10k ACIS solids)
was effectively frozen. Several independent costs, all addressed here:
- Face3D edge/fill buffers were re-uploaded on every camera move (keyed on
camera_generation). They are world-space and selection-independent, so gate
them on (geometry_epoch, fill_mode) instead — a pan no longer re-walks every
wire.
- Solid meshes were drawn one buffer + one draw call per solid (~10k draws/
frame), which strangled the GPU front end. Concatenate every solid's LOD0
into a few large buffers (split only to stay under the 256 MB buffer cap) and
draw the whole set in a handful of calls. Built once per geometry epoch, so
selection / hover never re-pack it.
- Selection / hover highlight is now a small tinted overlay of just the picked
solids, drawn last with an Always-depth pipeline so it shows on top even when
the solid is occluded — instead of re-tinting and re-uploading all meshes.
- Hover / pick hit-testing scanned every triangle of every solid each move
(seconds on a heavy model). Add a 3D-AABB screen-rect broad phase
(MeshLodSet now carries z_aabb) and reuse the renderer's cached expanded mesh
set, so picking is O(solids) cheap projections plus the few solids actually
under the cursor.
- Block-internal hatch hit-testing exploded every INSERT on every hover.
Cache the result per geometry epoch and skip blocks that contain no hatch.
- Curved-surface (cone/sphere/torus) tessellation tolerance is now radius-
relative (CURVE_REL_TOL) so a cylinder's facet count is size-independent and
matches the circle/arc wire tessellation, instead of exploding on large radii.
Also adds docs/tessellation.md mapping every EntityType to its tessellation
path (truck B-rep vs truck curve topology vs direct).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Replace the custom model_tiles tiling with iced 0.14 pane_grid as the layout
source of truth, while paper-space floating viewports are left untouched (they
overlap/free-place, which pane_grid cannot model).
Each Model pane is its own shader widget rendering its camera into the pane
bounds; the renderer Primitive carries a base_slot so several shader widgets
share the type-keyed MultiPipeline storage on disjoint slots. A second pane_grid
of transparent mouse_areas (plus on_resize) layers above the crosshair overlay
to receive input — the overlay's Hidden cursor interaction otherwise levitates
the cursor and starves any layer beneath it. The active pane follows the cursor
(hover-activate), and grid / crosshair / UCS icon render per pane while the
ViewCube and render-mode picker stay on the active pane only.
Split / close / VPORTS presets and the VPort table round-trip now drive
pane_grid (guillotine reconstruction on load). Adds drag-to-swap via a controls
bar handle with a ghost-card preview, visible pane dividers from the pane_grid
split regions, a default model background of RGB(33,40,48), and removes all the
superseded custom-tile code.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Reduce the ViewCube 30% (VIEWCUBE_PX 120→84) and tighten the nav region
(NAV_INSET_F 2.2→2.0) so the home / roll / nudge controls hug the cube in the
corner instead of floating in a large dead area, with sizes scaled to keep them
from overlapping. Bump the home and roll buttons back up slightly, shift the
UCS picker down below the region, and give it a fixed width so it centres
exactly under the cube centre.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
The UCS icon now anchors to the projected UCS origin when it is on-screen
(falling back to the bottom-left corner), and becomes a selectable object:
hover highlights it, a click shows draggable grips, and dragging the origin
grip relocates the UCS while the X/Y tip grips rotate it. Drags snap to
geometry/grid through the existing snapper, and the new UCS is persisted on
release. The icon stays selectable while parked in the corner.
Also make the ViewCube compass (N/E/S/W) world-fixed: the cube body still
reorients with the active UCS, but the cardinals project through a camera-only
rotation and snap in world frame, so the directions no longer spin with the UCS.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Remove the unused `crate::par::prelude` glob import in wire_gpu.rs and
the never-called `viewport_at_paper_point` method; fold its doc into the
screen-space twin that is actually used so no intra-doc link dangles.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Every wire segment is one instance, and each carried the wire's color,
line-weight, dash pattern and draw-depth (~44 B) inline — re-fetched per
segment even though they're constant along the wire. On native, move
those into a per-wire `WireConst` storage buffer (group 1) indexed by a
`wire_id` on the instance. The instance drops from 104 B to 60 B (~42 %)
across every entity's wires, and the shared constants are fetched once
per wire (cached) instead of once per segment — less vertex-attribute
bandwidth and less GPU upload on rebuilds.
WebGL2 has no vertex-stage storage buffers, so the wasm build keeps the
original self-contained fat instance and wire.wgsl; native uses the new
slim instance, wire_indexed.wgsl, and a group-1 bind group built per
batch in WireGpu::from_run / from_batch. Visual output is unchanged
(colors are now full f32 rather than 8-bit, if anything more precise).
Co-Authored-By: Claude Opus 4.8 <[email protected]>
Preview / interim / grip-drag overlay wires were drawn in the main wire
pass with the depth-tested wire_pipeline (depth_compare=LessEqual), so a
live command preview could be hidden behind occluding geometry — a 3D
solid, or 2D geometry drawn in front of it.
Draw them with the existing wire_xray_pipeline instead (depth_compare=
Always, no depth write), the same always-on-top path selection highlights
already use. Previews are now never occluded, for every command.
Co-Authored-By: Claude Opus 4.8 <[email protected]>
