The last four big web-vs-desktop gaps.
XREFs: XATTACH reads the picked drawing's bytes and registers them
name-keyed (the image/PDF-underlay pattern), pre-parses on the web
worker, and the commit merge produces the same xref block flow as
native. Opened drawings resolve xrefs from the registry (trailing-name
match) before cache build; unresolved references produce ONE warning
listing the names and the fix — drop the referenced files on the page
(drops and picker opens now register their bytes) and reopen or
XRELOAD. XRELOAD works off the registry on unsaved drawings too.
Data Link & Extraction: spreadsheet links read picked xlsx/csv bytes
(calamine from slices / CSV from string) through a session registry,
so preview, refresh, and table-insert work on web; extraction takes
external drawings via a multi-file picker parsed on the worker, and
outputs CSV/TXT/.dxex via download_bytes. The disabled 'Add folder'
button explains why.
ARCHIVE/ETRANSMIT: assembles the drawing plus registered images, PDF
underlays, and session plot tables into a ZIP delivered as
{stem}_transmit.zip; skipped dependencies are reported by name.
Save/autosave off the UI thread: the parse worker gains a serialize
action (protocol v5) — the document is prepared once on the main
thread (clone, styles, dimension bake), then the DWG/DXF writer pass,
the dominant cost, runs on the worker; any worker failure falls back
to inline serialization with the identical shared implementation.
The save path also clears the dirty flag only when the tab didn't
change mid-write, and the runtime-fields sidecar now carries
graphic_data, fixing a pre-existing loss of raw DWG round-trip bytes
on web-opened documents.
(cherry picked from commit 2e100634)
Ported onto main (#1717 review): the new ocs_save crate builds on the
pinned codec; the hand-written ZIP writer is replaced by the zip crate
already in the tree; ARCHIVE packs registered xref bytes too and
serializes document_for_save() like a web Save; the web save prepares
document_for_save() (prepare_document_for_save now takes the copy by
value, so no second clone); the wasm resolve_path consults the xref
registry, so every reference lookup — not just resolve — finds dropped
files; XRELOAD keeps main's undo, unload and stale-baseline handling;
native data extraction streams the source drawings one at a time.
A fresh web-vs-desktop audit found the remaining gaps; this closes the
top ten.
Raster images: IMAGE attach read the picked file via image::open on a
name-only path (dead on wasm), and raster references inside opened
drawings decoded the same way. Picked bytes are now registered in a
name-keyed store (the PDF-underlay pattern), decoding runs from memory,
and dropped .png/.jpg/.bmp/.tif files register by name so unresolved
references in opened drawings re-resolve and render (and export to PDF).
Command gaps fixed with the bytes-at-pick / download_bytes pattern:
- WBLOCK now downloads the extracted block instead of failing on fs.
- SAVEALL saves every dirty tab instead of printing a no-op message.
- SCRIPT picks and reads the .scr file through the browser picker.
- CENTERLTYPEFILE picks the .lin/.dnl file and populates from text.
- Custom .ctb tables picked on web now parse, register for the
session, appear in the plot dialog list, and apply (load_named
previously returned builtin-only on wasm).
- LAYTRANS 'Load standards file' parses the picked drawing on the
existing web worker.
- INSERT 'from file' in the block palette parses picked bytes and
feeds the shared import path.
Text fidelity: TTF-named styles rendered as thin stroke glyphs on web
because the font database was empty. Fetched Noto faces are now
registered into it with a family alias table (Arial, Times New Roman,
Cambria, ... -> Noto Sans/Serif), so those styles render filled, shape,
export to PDF as vector fills, and appear in the style picker.
ViewCube labels request semibold again on web.
docs/native-vs-web.md: the hatch and font rows were stale (both work on
web now), the GPU row claimed WebGPU is used where available (the
backend is pinned to OpenGL), and the parallelism row missed the parse
worker — all corrected.
(cherry picked from commit c643a10e)
Ported onto main (#1717 review): conflicts resolved toward main — the
font-file lookup, PDF searchable text, purge service, block palette and
plot-style copy stay as they are on main; the web pieces (font registry
and aliases, session plot styles, SCRIPT/SAVEALL/CENTERLTYPEFILE picks,
block-from-file on the web) are layered on top.
Three reviewers (security, correctness, performance) converged on:
- Download PDF died on the second click: Closure::once_into_js throws
when invoked twice. The handler is now a plain FnMut kept for the
overlay's lifetime.
- Rasterisation was bounded by page count, not pixels: 12 A0 sheets at
150 DPI peaked near 1.8 GB of RGBA plus ~1.7 GB of decoded images,
and a drawing declaring absurd paper sizes made the rasteriser
saturate its u16 dimensions into a ~17 GB allocation that aborts the
tab. Preview now computes a uniform DPI under a 64-Mpx total budget
(floor 24 DPI) from the page tree before allocating, with a per-page
4096-px edge clamp.
- Overlays stacked when previewing twice and their blob URLs were only
revoked via Close. The overlay is now single-instance (id-tagged);
a replacement or Close revokes every URL through a shared registry.
- emit_image resampled before probing the XObject cache, so a title
block placed on 12 layouts ran 12 Lanczos passes to embed once.
The cache is now probed with the target dimensions first.
- Preview downloads are named {stem}_preview.pdf instead of a constant
plot_preview.pdf, and the page label says 'first N of M' when the
preview is capped while the PDF keeps every page.
(cherry picked from commit 2a05479d69)
A plot preview should show the plotted result on screen instead of
producing a file. The web build now does: the plotted PDF bytes
are rasterised with the hayro pipeline already used for PDF underlays
(150 DPI, first 12 pages) and shown in an in-page overlay with
Download-PDF and Close actions.
The previous strategy opened a new tab via window.open from the async
plot task; browsers withhold popups without a live user gesture, and
embedded hosts ship no PDF viewer plugin at all — so the fallback
download fired and preview behaved like an export. Plain PNG <img>
elements need neither permission.
(cherry picked from commit 8722d84910)
The rotation is stored in degrees and turns the bottom-to-top direction
counter-clockwise, measured in pixels; a two-colour gradient blends from
the bottom colour to the top one across the whole viewport.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Restoring a named view gives the model viewport the view's visual style
(the closest render mode), and a 2D wireframe viewport shows no
background, as in the reference.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Color, Gradient and Image at the background prompt open a Background
dialog (None / Solid / Gradient / Image, preview, solid colour, two- or
three-colour gradient with rotation, image file with position and
scale). Apply links the view to a new background object in
ACAD_BACKGROUND; restoring a view gives the model viewport a copy of
it (or none), and saving a view copies the viewport's background.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
new src/scene/block_service.rs (references, delete, rename, purge candidates, name validation, one error type and message). The command-line RENAME, palette Remove, plugin API, control/MCP block_delete and PURGE all use it. Layout, xref, dynamic-block, anonymous and open-in-editor records are protected; PURGE removes an anonymous block only when nothing references it. PROTOCOL CHANGE: control/MCP block_delete refuses a referenced block unless force=true (erases referencing INSERTs in one undo step). New REST route DELETE /api/v1/blocks/{name}?force=true. MCP tool digest re-pinned. Undo labels are Delete block X, Rename block X to Y and Purge.
world = insert_point + Rot * Scale * (P - base_point). Attribute placement, EXPLODE, XCLIP storage, REFEDIT's working copy, the UCS reframe of attributes and the dynamic-block visibility grip ignored the base point and were offset by Rot*Scale*base. They now share insert_transform_with_base. Creation stores a ZERO base and redefine resets it; a no-op redefine moves nothing; the Block Definition dialog prefills the base from the definition. XCLIP filters stored by older builds keep their old displacement.
bake_dimension_blocks copied the dimension's whole EntityCommon, including its extension dictionary handle, into every baked child, so many entities claimed one dictionary and file validation reported an error per entity. Baked children now get no extension dictionary. The same defect is fixed in define_transformed_block and the cross-document paste clone. Each has a test.
Their grip code already works in world space, and the dispatch layer
converted the grips through the OCS a second time, so for a normal of
(0,0,-1) the grips (and grip edits) were mirrored across the Y axis.
Fixes#1683
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A polyline with a thickness but no width is drawn as its extrusion edges
alone, and those edges were forced black (the outline colour meant for the
filled tube of a wide polyline). On a dark background the polyline looked
black and changing its colour did nothing. Only a wide polyline's tube
edges are black now.
Fixes#1682
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#1666 sets group 295 on extrusion history records, which makes the kernel
read the embedded profile and path where they stand and ignore the
matrices. The planar fallback profile (a region-derived LWPOLYLINE) was
stored in its plane's local frame with the frame as matrix, so with the
flag it rebuilt at the origin. Re-express it in its own OCS (normal and
elevation) when the record is built.
Pins the kernel and graph at upstream (opencadkernel#84 merged as f051138)
instead of the ramox81/cadkernel [patch].
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#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 <noreply@anthropic.com>
The informational clippy job fails on two pre-existing literals
(also failing on main, surfaced on this PR's merge): the trim
tangency test's angle list used 3.14 where pi was meant — use the
constant; the point-cloud light-direction fixture in render.rs uses
a hand-tuned 0.318 component that only looks like 1/pi — keep the
tuned value and allow the lint on that statement.
Resolve conflicts from main's rustfmt pass by taking main's versions of
command.rs, style.rs, view/mod.rs, scene/mod.rs, ui/window/mod.rs and
ui/properties.rs and re-applying this branch's Graphic Attributes changes.
Add the Graphic Attributes strings to the new lo-LA catalog.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E1QnME6WJbD5iGFae1eeCb
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.
A lofted surface's region cross sections are recorded as the curve round
each region, as the reference records them; a region recorded as a
region made the reference refuse to open the drawing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The reference writes none; a draft distance on an extruded surface made
it re-extrude the surface with a taper when the profile changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The polyline round a region is written about the region's normal in its
object coordinates with an identity transform, as the reference writes
it; a rebuild from the history placed a polyline given in plane
coordinates with a frame at the wrong place once the region moved. Arcs
of one circle making a whole turn are recorded as that circle.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A swept, extruded or revolved surface made from a region records the
region as the curve round it, as the reference does: a lone circle or
whole ellipse as itself, a loop of lines and arcs as a closed polyline in
the region's plane. A region recorded as a region made the reference
refuse to open the drawing (modeling error 58000).
A region read by an associative action carries the step its edge names
belong to (ACAD_STEPID), so the reference finds the edges again and the
surface follows the region.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The surface extrusion, the extrusion along a path, revolve history and
embedded profiles refused a region whenever it had more than one display
wire, which a region of several edges always has. They now refuse only a
region with holes, and a region's loops for a solid extrusion come from
its modeler edges.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A region keeps one display wire per edge, so a region of several edges
gave no polygon and could not be extruded (solid or surface), and a
circle region was extruded from its display polygon. Profiles of a region
with one boundary loop now come from the region's modeler edges.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A region of several edges kept its display wires one per edge, so the
surface extrusion found no polygon and refused it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A new drawing carries the visual styles the reference's new drawing has
in its visual style dictionary. -VIEW sEttings Visual style offers them
(2dWireframe as "2D Wireframe") and a saved view starts with the 2D
Wireframe style.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A swept, extruded, revolved or lofted surface made from a region now
reads the region through one geometry dependency per edge, each naming
its edge by a persistent subentity id recorded in the drawing's
subentity manager, so the surface follows the region. A lofted surface
keeps one continuity and bulge value per cross section.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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 <noreply@anthropic.com>
A drawing whose plan content is built from blocks reaches thousands of
definitions, and each one took a buffer of its own for its segments and
another for its constants, plus one per placement run. The payloads are a
few hundred bytes, but every allocation costs a page: 14k definitions ask a
device for 1.8 GB to hold 12 MB of geometry. D3D12 runs out, the failed
allocations stay invalid, every later frame writes to them, and the session
dies in the staging belt.
The definitions of a build now share one arena: constants in a single
uniform buffer addressed by dynamic offset, segments in a single geometry
buffer addressed by first_vertex, placements in a single instance buffer
addressed by first_instance, with another arena started whenever one of the
three would outgrow the per-buffer budget. The cache holds the records
rather than the buffers, so a build the device rejected poisons nothing.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
SURFACEASSOCIATIVITY now keeps EXTRUDE, REVOLVE and LOFT surfaces linked to
their sources: the profile (and axis object) or cross sections are kept, the
surface is made again when they change, and the drawing's associative
network carries an action with the extruded, revolved or lofted surface
action body, its path parameters and value parameters as the reference
writes them. Swept, extruded, revolved and lofted surfaces can read
splines (NURBS edge curve, fit-point splines through their fit points)
and ellipses (elliptical edge curve). Associative surfaces of a file
opened here follow their sources too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- 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 <noreply@anthropic.com>
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 <noreply@anthropic.com>
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.
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 <noreply@anthropic.com>
- 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 <noreply@anthropic.com>
- 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 <noreply@anthropic.com>
Afonsop