Commit Graph

1394 Commits

Author SHA1 Message Date
sascha2101 07ef94e102 Hole: drill only the shown bodies when no target is picked (#7520) (#7562)
* claim: #7520

* Hole: drill only the shown bodies when no target is picked (#7520)

A Hole that names no bodies drilled every body in its path, hidden ones
too, and its dialog preview replaced the hidden bodies with drilled ones.
Like a cut (#3961), it now takes the shown bodies of the active component
as its participants while some are hidden, stored so showing the others
later changes nothing. With nothing hidden, or everything, it still drills
every body.

Its face points are named on the bodies it drills first, so a hidden
body's face in the same plane no longer takes the picked face's name.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FN5eEMrqWWXqi65FcK8GXq

* Hole: drill the picked body though it is hidden (#7520)

Review on #7562: with the picked face's body hidden (picked through Select
Other, or hidden while the dialog is open), the shown-bodies default left it
out, so the hole drilled only shown bodies further along, or nothing, without
a message. The body whose face holds the hole's position now joins the shown
ones, unless a shown body has a face there too (a face shared by a shown and
a hidden body is the shown one's, as a pick in the view takes it).

The test also covers editing a hole made with nothing hidden while a body it
drills is hidden: timeline.redefine stores participants only when the edit
passes them, so the hole keeps its empty scope and the hidden body its bore.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EMbqRqYimvY26QzcvrojSC

* Hole: test that a hole made with a body hidden replays the same (#7520)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* Hole: follow the face-name lookup now returning a Result (#7520)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* Hole: tests follow main's Vec3 and world_state (#7520)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* Hole: format (#7520)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

* Hole: with every body hidden, drill them all as before (#7520)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

---------

Co-authored-by: sascha2101 <221609842+sascha2101@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 19:19:18 +00:00
austin62jake 8f4d7959df Agents and scripts: Parameter Sweep describes its exact parameters and reply (#10977) (#10978)
* claim: #10977

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Parameter Sweep: input and output schemas (#10977)

inspect.parameter_sweep gets a typed contract and reply types in
schema/parameter_sweep.rs and leaves unconverted.txt. Each key reads as
the handler reads it, with its message; the reply is built from
ParameterSweep and its item types, with the same JSON as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Parameter Sweep schema: a missing parameter reaches the handler's message

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* parameter_sweep schema: do not publish a null default on the required parameter

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 19:00:58 +00:00
SS-EN 727bb23c7e Section Analysis: a body standing on another keeps its own cap, also in CPU pictures (#4622)
Where one body stands on another (a post on a plate), the post's bottom,
seen through the cut, and the plate's top are the same plane. In the CPU
rasterizer (headless pictures without a GPU) the plate's top won that
depth tie and showed through the post's cut. The cut now wins a tie by a
few f32 steps (DEPTH_EPS).

The viewport already draws this right since #6363 and #6728; a GPU
regression test now covers it, both creation orders, orthographic and
perspective, with both depth formats.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:52:15 +00:00
semmelrinde ef92f35b16 Sketch drag: a step solved ahead that solves becomes the sketch, one step per core (#7648) (#7649)
Past a failed step a scripted drag solves the next steps at once and takes
those that fail in a row as tried; the first that solved was then solved
again by the walk. It now becomes the sketch from its solved copy, so no
step is solved twice. A batch holds one step per core (at most 8), and on
one core or in the browser build, where a batch would run its steps one
after another, the walk solves the steps itself. The walk takes the same
steps with the same outcomes, so results and counted solves don't change.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:51:32 +00:00
PaddleStroke cfc4908291 Sheet metal: returns on a notched tray now close at the convex corner (#8823) (#9714)
* claim: #8823

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sheet metal: returns on a notched tray close at the corner away from the notch (#8823)

Closing two returns at a corner was refused when either return already lay
over another flange in the flat pattern, even far from that corner. On a tray
with a notch, the walls on either side of the notch's inner corner both unfold
into the notch, so the return on one of them lies over the other there; the
return's opposite convex corner was then left open ("closed, the flanges would
overlap in the flat pattern") and the two returns ran through each other.

A closure is now refused only when it makes the two returns overlap or adds to
what either already shares with another flange in the flat.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:41:07 +00:00
froglemonade2 1432131e95 English: command labels in title case, other labels in sentence case (#9085) (#10734)
* English: command labels in title case, other labels in sentence case (#9085)

The rest of #9085 after #9194 (the plural and the spelling) and #9907
(the Preferences page names). Command labels are in title case (422 of
430 were already); the seven in sentence case follow, and Fix/UnFix
reads Fix/Unfix. The UI text skill states the rule. The user docs name
the Preferences footer's button by its current label, Restore page
defaults (#10295).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Navigation bar menu test: the echo names the command by its own label

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* UI text skill: state the case rule in one short sentence, under the 3600 limit

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Localization doc: the case rule in full, with examples

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Localization doc: the example button is Restore page defaults now

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:39:52 +00:00
takuto-NA 91ecea8cfd Properties: account for sheet perforations (#6025) (#6028)
* claim: #6025

* fix(engine): account for sheet perforations in physical properties

* fix(inspect): retain kernel topology counts for sheet bodies
2026-11-04 18:30:31 +00:00
austin62jake 80889762ce Agents and scripts: Configure, Configuration Column, Row and Rule describe their exact parameters and reply (#11895) (#11896)
* claim: #11895

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Configuration table: input and output schemas for Configure, Configuration Column, Row and Rule (#11895)

describe_command now gives the exact parameters and reply of the four
commands that change a design's configuration table, and the warn and
strict parameter checks use them. The handlers read their parameters
from the raw JSON, so the new contracts in schema/config.rs read each
key as its handler does, with its message: lists of names as
params::string_list (whose json_brief and MAX_STRING_LIST become
pub(crate)), exactly one column target, faces picked or a face column's
key, a row's cells, and rules (whose schema describes them; the handler
keeps its own check). What only the design can answer stays the
handler's. All four reply with ConfigTableReply, as config.table does.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 18:19:35 +00:00
rugbymauri a3f293c987 Error messages: an out-of-range value names the field that is wrong (#9625) (#9684)
* claim: #9625

Signed-off-by: Maurizio Monticelli <maurizio@whatwedo.ch>

* Name the value at fault when a hole, sheet rule, joint or extrude is refused (#9625)

Several refusals were right but named a check the input had passed, so
the user changed the wrong field:

- solid.hole: an unknown `fit` was reported as an unknown bolt size; it
  now says the fit must be close, normal or loose. A countersink angle
  out of range was reported as a countersink too narrow; the angle and
  the width are now checked apart. A `diameter` number out of range was
  reported as missing; it now gives the range, as other parameters do.
- Sheet rules: a thickness of 10000 mm or more said "must be positive";
  it now gives both limits. The K-factor and the bend radius are named
  on their own, and the message no longer implies a radius of 0 is
  refused.
- joint.drive: a number out of range said "numbers or expressions"; it
  now gives the range.
- Extrude: a distance past the kernel's limit said `end = 1000000000`;
  it now says the extrusion is too long and gives the limit.

Fixes #9625

Signed-off-by: Maurizio Monticelli <maurizio@whatwedo.ch>

* Refuse a too-long extrusion before the kernel and match the countersink floor (#9625)

The kernel is paused, and its `finite()` check also guards a revolve's
angle and a cylinder's and a torus's sizes, so its message must not say
"extrusion". crates/kernel is back as it was; the doc evaluator now
refuses an extrude (plain, tapered, or redefined from the Edit Extrude
dialog) whose end lies 10000000 mm or more from the sketch plane,
mirroring the kernel's limit, and says so. A cylinder of radius 2e7
keeps the kernel's own message.

The countersink check refused 0.1° while the message said "more than
0°". The floor stays (the cone grows as 1 / tan of the half angle) and
the message now names it: more than 0.115°.

Signed-off-by: Maurizio Monticelli <maurizio@whatwedo.ch>

---------

Signed-off-by: Maurizio Monticelli <maurizio@whatwedo.ch>
2026-11-04 18:18:14 +00:00
austin62jake aba00427fa Render shots: input and output schemas for Duplicate, Rename and Remove (#11450) (#11451)
The handler's strict parameter type moves to schema/render_shots.rs as
ShotParams, with descriptions and JsonSchema, and becomes the contract of all
three commands; the handler reads the same type. It builds ShotChange instead
of a json! literal, and takes its operation as a ShotOperation enum.

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 18:02:53 +00:00
alfonsotames cccf8c1ece Snap points: a flat face offers its corners, edge middles and every hole's centre (#11476) (#11477)
* Snap points: a flat face offers its corners, edge middles and every hole's centre (#11476)

A flat face now offers its whole outline as snap points, read from the
exact faces and edges: every corner, the middle of every straight edge
and the centre of every round edge, holes included, as well as the centre
of its area. Corners on a closed circle's seam, or where the kernel split
one circle into pieces, are left out.

snap_points_from gives each point with where it came from (SnapSource):
the face it was picked on, the face's outward normal and a point inside
its edge, so a tool can keep the point by name (joint.create's edge,
corner and on) and turn a frame to the face. SnapPoint itself does not
change, so code that builds one keeps compiling.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Snap points: a round face's end centres say which way their end faces (#11476)

The centres at a cylinder's or cone's two ends now carry a normal in their
SnapSource: the axis pointing out of that end, which is the normal of the
face a hole opens onto or of a pin's end. A joint origin taken from one can
then lie on that face, as the centres a flat face offers already do.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Snap points: a hole centred on its face keeps its own point and rim (#11476)

Snap points were one point when they lay at one place, whatever their kind,
so a hole centred on its face was lost behind the face's centre, and with it
the rim its point keeps. Now only points of one kind are merged: the face's
centre and the hole's centre both stay, and only the hole's has its rim.

SnapSource::normal's doc now says it is also the axis out of a cylinder's or
cone's end.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Snap points: a washer's two rims keep a centre each (#11476)

Centres at one place are one point only when they are on one circle: the
inner and outer rims of a washer face now keep a centre each, with its own
rim, while the pieces the kernel split one circle into are still one point.

SnapSource::normal's doc now says that at a round face's end it is the
axis out of that end of the face, which at a blind hole's bottom points
into the material below, not out of it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Snap points: rims are told apart by their radii's size, not their place (#11476)

Two centres at one place were one circle when their rims' radii differed by
less than the place's tolerance, which grows with the distance from the
origin: 60 m out, rims of 4 and 4.05 counted as one and one lost its point.
The radii are now compared with a tolerance from the radii themselves.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Snap points: nearby points stay apart far from the origin (#11476)

Points of one kind at one place were merged within a tolerance of 1e-6 of
their distance from the origin, 0.06 mm at 60 m, so two small holes there
whose centres are 0.05 apart lost one centre. Places now use 1e-9 of that
distance (0.06 µm at 60 m): well above the rounding of coordinates and of a
circle fitted to an imported edge, so a point found twice is still one, and
far below any feature. The radii keep their own tolerance.

The merge is its own function, so its test can give it points directly.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Alfonso Tamés <mufufu@Mufufu-MacBook-Pro.local>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:02:44 +00:00
Dahak Builds dcbedbd805 Don't blindly reopen a design that stopped the last session (#5064) (#10443)
A design that crashes or hangs the app while it is being reopened was
reopened blindly every launch, bricking the app each time. A launch that
reopens designs now marks them in the recovery folder; the marker is
cleared as each design finishes opening. A launch that finds the marker
left behind skips those designs once and says which one the last session
stopped on, so the app starts usable and the loop breaks.

Co-authored-by: Dahak Builds <dahakbuilds@users.noreply.github.com>
2026-11-04 17:16:15 +00:00
Mateusz Sych 6488a540c9 Windows: a lost graphics device restarts the app instead of crashing (#10240) (#11109)
* claim: #10240

* Lost GPU: a loss met by a queue write restarts the app instead of crashing (#10240)

The per-frame probe (#6528) finds a device the driver reset before the
frame's uploads, but the driver resets a device at any moment: a reset
that lands after the probe is still met first by a queue write or the
present, and wgpu-core unwraps the driver's Lost there
(PendingWrites::activate) instead of losing the device. That panic
ended the app and wrote a crash report.

- egui-wgpu: the painter runs each frame under DeviceLost::survive. A
  panic the frame meets is taken for the loss once the probe then finds
  the device lost (the probe's path loses the device and runs its lost
  callback); the painter draws nothing more and the app restarts as for
  any lost device, keeping unsaved changes for recovery. Any other panic
  carries on unwinding.
- Crash reports: a panic inside wgpu-core that unwraps Lost is the
  device loss, not a bug, and is not reported. A command that meets it
  fails; the next frame's probe restarts the app.

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-11-04 17:15:58 +00:00
froglemonade2 3e70ad9876 View navigator: right-click to choose Axes, Cube or a Custom body (#10266)
* View navigator: a custom body as the navigator, and a right-click menu to switch types

A right-click anywhere on the view navigator opens a menu with its three types: Axes, Cube and
Custom. Custom opens the Custom View Navigator panel (a feature panel with OK and Cancel): pick a
body in the view, its texture (Themed: the cube's colours; Modeled: the body's own appearance and
the view's edge colour), and whether the cube's rotation arrows and Home button go around it. The
navigator previews the panel's choices while it is open; Cancel leaves it as it was.

The custom navigator draws the body at the cube's size through a small depth-buffered raster, so
hidden faces stay hidden; hovering a face highlights it, a click looks straight at it (and frames
the model with navigator_fit on), a drag orbits. Until a body is chosen it draws the cube.

The body is kept as a packed snapshot (16-bit, base64) with the preferences and in settings
profiles, so it stays across runs and designs and travels with an exported profile. Bodies over
10,000 triangles, and long or flat ones (thinnest width under a fifth of the widest), are refused.

New command view.navigator (style, body, texture, arrows, home) does the same without opening
anything; new preferences navigator_texture, navigator_arrows and navigator_home, and a Custom
choice for navigator_style. Defaults are unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: a hovered face in Modeled takes the view's selection colours

In the Modeled texture the face under the pointer is filled and outlined the way the view selects
a face of the model (sel_face, sel_edge), not in the cube's hover blue and accent, which Themed
keeps.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: a hovered face in Modeled lights up as the view's mouse-over does

The face under the pointer keeps its own colour, a little lighter (hover_face_lift) and still
shaded, with its edges in the view's mouse-over edge colour (hover_edge), as the 3D view shows a
face under the pointer; not the selection colours. Themed keeps the cube's hover blue and accent.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: rotation arrows and Home on by default

navigator_arrows and navigator_home start on, so a new custom navigator has the cube's arrows and
Home button around it; the Custom View Navigator panel turns either off.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: the strings in every complete language, and test fixes

- The new messages in Arabic, Hebrew and Vietnamese (complete catalogs, like Japanese and Chinese).
- The navigator's raster goes through one colour helper, listed with its reason among the files
  allowed colour literals (the design's appearance shaded into pixels).
- nav_settings joins the requested-icon manifest (the view.navigator command's icon).
- A visible edge piece is kept only when the stretch seen with it, across the edge's segments, is
  at least two samples long: a hidden edge no longer shows a sliver at a silhouette corner.
- The navigator layout test treats Custom without a body as the cube it draws; the panel and
  size tests no longer assume an empty undo list, the style before opening, or a coarse sphere.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: profiles carry its texture, arrows and Home

Settings profiles keep a fixed set of preference keys (#9907), which the custom navigator's newer
settings are not part of. They travel in the profile's own `view_navigator` entry, next to its
body (`view_gizmo`): captured, checked (a value this version can't read is left out) and applied,
leaving the fixed key set unchanged. The cube's arrow inversion (#10286) applies to the custom
navigator's arrows too, through the shared arrows().

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: follow main's Vec3 move and fallible world_state

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: list view.navigator in the rig's EXEMPT, refuse a stored body that can't save again, right-click without a widget over the navigator

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: keep the body unless picked again, save only a profile that reads back, no navigator menu through a popup

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: refuse a stored body with a triangle past its points

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: a stored body meets the triangle and aspect limits a picked one does

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Custom view navigator: a body at the 5:1 limit stays inside it once packed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 17:12:27 +00:00
Daniel 4d738cd210 Sketch: an ellipse is drawn with its axes, and D measures its half-axes (#9291) (#9292)
* Sketch: an ellipse comes with its axes, and D measures its half-axes

An ellipse could only be sized by its minor radius, drawn like a circle's
radius, and there was no way to give its full axes. sketch.ellipse now also
adds the major and minor axes as construction lines through the centre.
The major axis runs from the major end to the point opposite it; the minor
one is perpendicular to it, ends on the ellipse, and both are bisected by
the centre. That adds six coordinates and six equations, so the ellipse
keeps its five degrees of freedom, and it needs no new constraint kind and
no format change.

With the axes in place every ellipse size is an ordinary dimension. D on an
axis line gives its length, the whole axis. D on the outline gives a half-
axis, the distance from the centre to an axis end, chosen the way a line
chooses horizontal or vertical but in the ellipse's own frame: text pulled
across the major axis measures along it (the major half-axis), text pulled
out past an end of it measures across it (the minor one). The end on the
text's side is used. Near the boundary between the two, a dimension being
placed keeps the half-axis it had (sketch.dimension's previous_axis, which
the Dimension tool sends), so the preview does not flip. A half-axis and
its whole axis determine each other, so the second becomes a driven
dimension through the existing redundancy path.

Deleting the ellipse leaves its axes, as deleting a Center Rectangle's
edges leaves its construction diagonals. Ellipses without axes, from
existing designs, keep the minor radius as before, and type radius still
gives it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Guide: an ellipse's axes and half-axis dimensions

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: test that an ellipse's half-axis preview keeps its axis near the boundary

The Dimension tool sends the half-axis its last preview chose, so near the
boundary between the major and the minor half-axis the preview keeps it.
This covers that path from the tool, from both sides of the boundary.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's axes select and drag with it, as a rectangle's diagonals do

A double-click on a Center Rectangle's edge or diagonal selects its edges
and both diagonals, and dragging a diagonal moves the rectangle without
changing it, because its lines share corner points. An ellipse's axes are
held by constraints instead, so a double-click selected only the ellipse
(and Delete then left the axes behind), and dragging the major axis moved
the ellipse but stretched its minor half-axis.

Sketch::ellipse_axes now finds an ellipse's construction axes (the
detection sketch.dimension already used, moved into the sketch crate).
Sketch::chain treats an ellipse and its two axes as joined, so a
double-click on any of them selects all three. Dragging either axis moves
the ellipse and both axes as one, keeping both half-axes. Dragging the
outline still sets the minor half-axis.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a corner test reads its line's id instead of assuming l1

sketch.ellipse now adds two axis lines, which take the first line ids, so
the extend test reads the id of the line it draws from the reply.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's axes cut a trim, as a Center Rectangle's diagonals do

Trim, extend and break take every other curve as a cutter, construction
lines included, so an ellipse's construction axes stop them as a Center
Rectangle's diagonals do. A test now states it for both shapes, and the
guide says it. The freeform cutter test keeps the axes it used to delete
and asserts the piece a trim removes: from where the major axis crosses
the circle to where the ellipse does.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: test dragging an ellipse's outline with its axes present

Dragging the outline sets the minor half-axis, and the minor axis line,
held on the ellipse at one end, follows it; the centre and the major
half-axis stay.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: other dimension types on an ellipse with axes keep the minor radius

Only auto, major and minor give a half-axis on an ellipse drawn with its
axes. Any other type (radius, diameter, horizontal and the rest) gives
what it gives on an ellipse without axes, as the parameter doc says,
instead of quietly becoming a half-axis chosen by the text. The arm for
a missing text_at is dropped: with no text the rule already picks the
minor half-axis.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: test an ellipse's half-axis dead zone from the major side too

The tests only showed a previous minor half-axis being kept near the
boundary; the point they used on the major side is major by geometry
alone. A point inside the band where geometry says minor now shows a
previous major half-axis kept, in the engine and in the Dimension
tool's preview.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a placed ellipse half-axis is the one its preview showed

On an ellipse already sized, a half-axis is added as a driven dimension
on the click, and that command left out the half-axis the preview kept
near the boundary, so the click could place the other one. It now
passes it as the preview does.

The remembered half-axis also comes only from a draft whose text was
placed (one without text is the minor half-axis by default and biased
the first placement), is kept per sketch as well as per picks, and is
forgotten once a dimension is placed or the picks are cleared.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: say how an ellipse's axes are recognised and what that costs

The axes are read from the constraints, with no mark in the design, so
the doc comment says which lines count, that deleting the midpoint or
point-on-curve constraint leaves an ellipse without axes, and that each
lookup scans curves times constraints.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse with one axis deleted still picks its half-axis by the text

With one construction axis deleted, D on the ellipse fell back to the minor radius wherever the
text was pulled, because the half-axis choice needed both axes. It now reads what is left of
them and chooses as before: the major half-axis goes to the major end (also from the far side,
when the point opposite is gone with its line), the minor one to its axis end, or as the minor
radius when the minor axis is the one deleted. An ellipse with neither axis keeps the minor
radius.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the one-axis half-axis test checks what is measured to

The test for an ellipse with one axis deleted checked only the reply's axis and value, which
both ends of an axis give alike. It now checks the dimension's kind and end: the axis end on
the text's side, the major end from either side once the point opposite is gone, a radius once
the minor axis is. With the major axis deleted and no text the reply names the minor axis and
measures to its end, where it used to be a radius with no axis named.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's minor radius is drawn as its half-axis, without R

A radius on an ellipse is the minor half-axis, and it was drawn as a circle's radius: one
leader from the centre towards the text, marked R, which measures nothing a reader can see.
It is now drawn as the distance from the centre to the minor axis end on the text's side,
the way the half-axis is drawn while the ellipse has its axes, and labelled with the length
alone. This is what D gives once the minor axis line is deleted, and on an ellipse with no
axes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's half-axis keeps a radius's stored text

The half-axis frame read the stored text from the middle of the half-axis, so a radius placed
on an ellipse in an existing design would show its text somewhere else. The text is stored from
the centre again, as a radius's always was, and the side drawn to follows it. An unplaced one
sits where a linear dimension to the minor axis end puts its text, and stacked texts step out
from the centre as a radius's do. A view test reads the painted label, so the drawing cannot
go back to R unnoticed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse moved, turned or scaled takes its axes along

Move, Sketch Transform and Scale moved only the curves picked, so an ellipse moved without its
axes left them where they were: its centre no longer halved them, the minor end left the
ellipse, and the sketch stayed unsolved. A Center Rectangle's diagonals share its corners and
always go along; an ellipse's axes share only the major end. The axes now go with the ellipse,
as they do when it is dragged. Copy is unchanged: it copies no constraints, for a rectangle
either.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a preference draws ellipses without their axes, as before

Drawing an ellipse with its construction axes changes what the Ellipse tool makes out of the
box, so the old ellipse stays one setting away: Preferences › Sketch › Draw ellipses with their
axes, on by default. Off, the tool draws the ellipse alone. sketch.ellipse takes axes: false
for the same, and keeps drawing the axes when it is not given.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: sketch.ellipse refuses an axes value that is not true or false

A string such as "false" was read as not given, so the axes were drawn; it is now refused as
the text commands refuse a non-boolean bold or italic. The move test also covers an ellipse
whose minor axis was deleted: the major axis left goes along.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: D picks a half-axis on an ellipse without axes too

An ellipse with no axis lines (both deleted, a mirrored copy, one from an older design) gave the
minor radius wherever the text was pulled, so its major half-axis could not be dimensioned on
it. The half-axis is now chosen by where the text is on every ellipse: the major one to the
ellipse's own major end, the minor one as its minor radius. With no text it is the minor
radius, as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: with one half-axis dimensioned, D on the ellipse gives the other

D on an ellipse chose the half-axis by where the text was pulled even when that half-axis was
already held by a dimension, so the second D usually guessed the one already there and was
refused as over-constrained. With a driving dimension already on one half-axis (to an axis
end, the minor radius, or the whole axis along its line or between its ends), D with no type
now gives the other, in either order and with or without the axis lines. With both held it
chooses as before; an explicit major or minor is kept.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the held half-axis rule is documented and its cases tested

The Dimension command's parameters and the guide now say that with one half-axis held, D on
the ellipse gives the other. Tests cover the whole axis held as a distance between its ends
and as its line's length with the other axis deleted, and check that a mirrored ellipse has
no axes. The placement test undoes its first dimension, so its second half still tests where
the text is rather than the held half-axis.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: Transform previews an ellipse with its axes

OK in Sketch Transform moves, turns and scales an ellipse's axes with it, but the window drew
where OK would leave only the picked curves, so the axes stayed put in the preview until OK. An
ellipse picked alone is now drawn with the axis lines it has.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the typed-radius ellipse test counts the ellipse's axes

Main's #10282 test drew an ellipse with typed radii and expected one curve; with the
ellipse_axes preference on by default the tool also draws the two axes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's remaining axis still drags and selects it

With one axis deleted, dragging the other axis did not hold the ellipse (with the minor axis
left, the drag squashed it), and a double-click on the ellipse or that axis selected only the
one curve. The drag and Sketch::chain now read whichever axes are left through
ellipse_axis_lines, as Transform and Move already do.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: drop the unreachable third arm of the held half-axis check

A distance from the centre to the major end or the point opposite it is already matched by the first arm, so the third arm and its major_end closure never ran.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: D on an ellipse's axis line is the whole axis wherever the value is pulled

An axis line took the free line's rule, so on a turned ellipse a value pulled beside or above the box the axis spans gave its horizontal or vertical distance. An ellipse's axis is a construction axis, so auto now always gives its length, which the held half-axis check already counts.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's axis is never a side of a Fillet or Chamfer corner

A line drawn from an ellipse's major end shares that point with the major axis, so Fillet or Chamfer there took the two as a corner (two lines skip the construction filter) and rounded the line into the axis. The axes are left out of a corner, picked by its point or as a and b.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: test the move reply, extend and the one-axis Transform preview for ellipse axes

The reply of sketch.move, sketch.transform and sketch.scale lists the axis lines, Extend stops at an axis, and Sketch Transform previews an ellipse with the one axis left after the other was deleted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: with Draw ellipses with their axes off, D on an ellipse is its minor radius as before

The preference brought back the old drawing only; D on an ellipse had changed with no way back. sketch.dimension takes half_axes (default true, refused when not true or false); false makes auto on an ellipse the minor radius wherever the text is. The Dimension tool and the context menu's Sketch Dimension pass it from the preference, and with it off a radius on an ellipse is drawn and labelled R as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: read an ellipse's axes in one pass over the constraints

ellipse_axis_lines scanned every constraint for every curve, so 500 ellipses with their axes (1500 curves, 2000 constraints) took 1.6 s to read and Move on all of them 2.0 s. It now collects the lines the centre bisects first: 3 ms and 0.36 s.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the dimension and corner parameters say how an ellipse's axis lines take part

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Preferences: Draw ellipses with their axes says D gives the minor radius when off

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: ask only the ellipses centred on a line's midpoint whether it is their axis

is_ellipse_axis read every ellipse's axes to answer for one line. An axis is bisected by its ellipse's centre, so it now asks only the ellipses centred on a point that bisects the line. The answer is the same.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a horizontal or vertical dimension across a level major axis holds it for D on the ellipse

half_axes_held counted a distance, a length and a radius, but not a horizontal or vertical dimension across the major axis, which is its length when the axis lies that way. D on the outline could then take the held major half-axis and be refused as determined; it now gives the minor one. Across a slanted axis such a dimension is only a projection and does not count.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Dimension: a draft on an ellipse is drawn with the axes preference it was made with

The frame read the preference as it is now, while the dimension added is the draft's.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a horizontal or vertical dimension across an upright minor axis holds it too

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: drop the unused Sketch::ellipse_axes; ellipse_axis_lines says how the axes are read

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse's own axes do not cut it, so Trim takes a lone ellipse whole and stops at real crossings

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: Fillet and Chamfer offer no corner at an ellipse's axis end, as the engine takes none

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a deleted ellipse leaves its centre on its axes, as a Center Rectangle's centre stays

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: ellipse axis docs say what they do, and the corner picker tests a curve's ends before its axis role

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the trim-at-a-crossing test checks that part of the ellipse is left

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: an ellipse axis moved, turned or scaled on its own takes its ellipse along, as a drag does

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: a selection dragged by an ellipse axis takes the ellipse along, as the axis alone does

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Docs: the Sketch preferences page lists Draw ellipses with their axes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: the ellipse axes help text says D gives the minor radius, drawn and labelled R as before, wherever the value is pulled

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 16:54:46 +00:00
austin62jake 2cd54b6af9 Agents and scripts: the standard part commands describe their exact parameters and reply (#11003) (#11004)
* claim: #11003

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Standard parts: input and output schemas (#11003)

parts.library, parts.insert, parts.fastener and parts.resize get typed
contracts and reply types in schema/parts.rs and leave unconverted.txt.
Each key reads as the handler reads it, with its message; the replies
are built from PartsLibrary, InsertedPart and ResizedPart, with the same
JSON as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Standard parts schema: Vec3 comes from solidcore_record

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Standard parts schema: hole picks must start with a list

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* parts.insert schema: the first hole pick must be a point, not any list

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 15:30:52 +00:00
N3MI-DG c9c063620a Hole: Bridged option for counterbores (#11103) (#11104)
* claim: #11103

* Hole: Bridge option for counterbores (#11103)

A counterbore hole takes an optional `bridge` length: two 3D-printing
layers that thick below the counterbore's floor, a slot along X across
the counterbore and then a square, both 0.02 mm wider than the hole, so
a printer bridges the step instead of sagging into it. Empty (the
default) keeps a plain counterbore.

The clearance keeps the bore, cut straight through both layers as for a
plain counterbore, clear of their sides: at exactly the hole's width it
ran tangent along them and broke through. The command and the geometry
refuse a layer that is not a positive length, layers that don't fit in
the hole below the counterbore, and a bridge on any other hole type.

The Hole dialog shows a Bridge field below Drill Point for counterbores.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Hole: translate the Bridge label for the complete catalogs (#11103)

Arabic, Hebrew, Japanese, Vietnamese and Chinese carry every English
message, so `cargo xtask i18n --check` needs the new label in each.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Hole: bridged counterbores refuse layers past a through hole's exit, too narrow a counterbore and a malformed bridge (#11103)

- Through all, the second bridge layer must end inside a body (exact point classification),
  short of the exit face; the tool's reach is no measure of the material.
- A bridged counterbore must be at least 0.04 mm wider than the hole, so the slot's ends have
  room on its wall (no square root of a negative number reaches the extrude).
- `solid.hole` refuses a `bridge` that is given but not a length expression (`true`, `[]`, `""`)
  instead of making a plain counterbore; `null` still leaves it out.
- Loop2 and Seg2 now come from solidcore_record (main moved them).
- The "nothing saved" check looks for the `"bridge"` key: `format_pending` names
  `bridged-counterbore` in every design.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Unit defaults test: solid.hole has the bridge field too

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 15:24:15 +00:00
austin62jake bec762fb2b Agents and scripts: the render library browsing commands describe their exact parameters and reply (#11443) (#11444)
* claim: #11443

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Render library: input and output schemas for List, Describe, Targets and Preview (#11443)

Describe, Targets and Preview read their strict parameter types from
schema/render_library.rs, which become their contracts. List keeps its two
readers (the physical profiles' takes availability: null, the other
catalogs' doesn't); its contract takes what either takes. The handlers build
LibraryPage, LibraryResource, LibraryTargets and LibraryPreview instead of
json! literals; a catalog resource is read into ResourceSummary through a
small Listed reader. The physical profiles' replies keep their JSON, which the
schemas describe.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Render library schema: the list says which availability each catalog takes

render.library.list has two readers: the physical profiles' takes
availability all, usable or null; the other catalogs' (and a call
without kind, which lists surfaces) take all, usable, specification or
blocked, but not null. The schema now says so with an anyOf on kind.
The type still takes what either reader takes: the readers check the
query and the page before availability, so refusing it in the type
would change which message a call with two mistakes gets in strict.
A new test checks that the schema takes ten calls exactly when the
handler does.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 14:50:44 +00:00
leValdemar19 826fce68eb Assembly: Move Occurrence says when the part's joints put it back (#11723) (#11791)
* claim: #11723

* Assembly: Move Occurrence says when the part's joints put it back (#11723)

A part that rigid joints hold to something grounded was moved, then put
back by the joint solve after the command, while occurrence.move still
replied with the moved transform. occurrence.move now solves the joints
before it replies. When they put the part back, the reply gives where the
part is, adds reverted: true and a warning that names the joint holding it
and what to change instead. A free chain of parts follows the move as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Move Occurrence: the warning names the joint for a part in a repeated assembly too (#11723)

A part inside an assembly placed more than once has no path of its own, so
the reverted-move warning asked for one instead of naming the joint. It now
looks at the placement the caller named, or else the first one.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Move Occurrence: list the part's joints when no single joint holds it back (#11723)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 14:39:17 +00:00
bherbruck fabeaa329a Desktop app: closing a design stops its kernel work, and designs build the same on any kernel (#11771)
* Desktop: closing a design stops its kernel work; same result on any kernel

A design's jobs now run under a stop token that closing or replacing the
design fires: an operation under it refuses to start, and one running ends
at solidcore's next check. The kernel it ran on is not used again. Calls
that answer from a body itself always finish, since a stopped call can leave
what it borrows half-changed, and what fills a body's shared cells is not
stopped.

Every kernel call starts clean (the same random rays, no stations from
earlier calls), so a design builds the same on any kernel of the pool, down
to its B-rep records. The pool is bounded: at most 8 kernels wait between
calls, and at 256 running calls a new call fails rather than grow or wait.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Stop tokens: a copy never stops its design; a stopped job can't decide a refusal

A copy replacing its own design stopped the token it shares with the design
it came from, so that design's later work would be refused. Only the design
itself stops it now. Which face a refusal names no longer depends on whether
the job that found it had been stopped: the refusal is kept with the body,
and a stopped job's probes were all refused, which named the wrong face.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 12:35:39 +00:00
austin62jake 1cbfb889d5 Agents and scripts: Compare Bodies describes its exact parameters and reply (#10994) (#10995)
* claim: #10994

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Compare Bodies: input and output schemas (#10994)

inspect.compare gets a typed contract and reply types in
schema/compare.rs and leaves unconverted.txt. Each key reads as the
handler reads it, with its message; `color` checks each colour with the
handler's own color_param, now pub(crate) and re-exported from cmd. The
reply is built from Comparison and its parts, with the same JSON.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Compare: drop the json import the typed reply no longer uses

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Compare schema: colour strings carry the hex pattern

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 06:45:10 +00:00
Colin Faulkingham 954232cef1 AI client setup: preserve configuration files when a write fails (#11363) (#11410)
* claim: #11363

* AI client setup: replace configuration files atomically (#11363)
2026-11-04 06:27:37 +00:00
austin62jake 0fc4cd2747 Agents and scripts: the render stage commands (Ground, Backdrop, Clear) describe their exact parameters and reply (#11455) (#11456)
* claim: #11455

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Render stage: input and output schemas for Ground, Backdrop and Clear (#11455)

The three handlers' strict parameter types move to schema/render_stage.rs as
GroundParams, BackdropParams and ClearParams, with descriptions and
JsonSchema, and become the contracts; the handlers read the same types. The
handlers build the shared StageSaved reply instead of json! literals, and the
saved setup is serialized with an error instead of json!'s panic path.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 06:11:09 +00:00
leValdemar19 6d83f95fba Isolate a body: other components are hidden as components, in one step (#7896) (#7951)
* claim: #7896

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Isolate a body: other components are hidden as components, in one step (#7896)

Isolating a body hid every other body one by one: browser.visibility with `items` writes only
hidden_bodies, so every other component's Browser row kept an open eye and normal text, and
its eye could not bring it back (the first click hid an already invisible component).

browser.visibility takes `isolate: true` with `items` (bodies), as it does with `paths`. The
picked bodies' own placements stay open, their other bodies (and the root's) are hidden one by
one, and every other placement is hidden as a placement. Isolate in the Browser and the view
runs that one command, so it is also one undo step instead of two.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Isolate a body: a hidden Bodies folder or group it is in shows again (#7896)

Isolating a body inside a hidden Bodies folder or group left that folder or
group hidden, so the picked body stayed hidden. Isolate now shows the
folders and groups (with the groups they are in) that hold the picked
bodies; their other bodies are hidden one by one as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Isolate a body: look for hidden folders only when some are hidden, and find body owners once (#7896)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:06:23 +00:00
semmelrinde ed3510a91f Typed schemas for the design properties commands (part of #2829) (#10920)
document.metadata, document.metadata.set and document.metadata.remove
get input and output schemas: contracts in schema/metadata.rs, checked
in front of the unchanged handlers, and reply types the handlers now
build their replies from. Their lines leave unconverted.txt.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:03:21 +00:00
semmelrinde d25f8198fa Pieces of a cut keep their names when the design is evaluated again (#11658)
A cut that leaves a body in pieces names them after their faces
(`order_lumps`): the piece whose least face name sorts first takes the
first name, and so on. It moved the names but left the pieces where the
kernel put them. The session then pinned the feature's body names in
that state order, and the next full evaluation (opening the design, or
editing a parameter upstream) handed them out again in a different
order: two pieces swapped names, and a later feature that named one met
the other. Found by a local model fuzzer.

The pieces now also take the family's places in name order, so the
names pinned are the order the next evaluation hands them out in. A
design saved before keeps the names a fresh evaluation gave it.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:59:42 +00:00
semmelrinde 03b5248dc1 Sketch: Line and Construction Line refuse a second point next to the first (#11595)
A through-point 0.0000001 mm from the start made an infinite line whose
direction nothing held, and a Line that short had both ends connected to
the origin, a line of length zero. Both now refuse points closer than the
distance at which a new point is connected to an existing one.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:49:05 +00:00
Colin Faulkingham f9dd77e383 Configurations: reject invalid suppression values when saving rows (#9088) (#11571)
* claim: #9088

* Configurations: reject invalid suppression values before saving rows (#9088)
2026-11-04 05:39:04 +00:00
kaltenbrunnermarkus 8a74336713 Tests: the autosave recovery test no longer fails at random on a busy machine (#11815)
* claim: Recovery: the save test waits for the autosave worker before it checks

* Recovery: the save test waits for the autosave worker before it checks

saving_drops_the_entry_and_a_clean_close_leaves_nothing marked the
session saved right after the tick that started an autosave. Under load
the worker was still writing, so the next tick found the write running,
took nothing up and dropped nothing; the file landed afterwards and the
check that it is gone failed (seen in CI on #9021).

The test now ticks, bounded, until the autosave has caught up
(Autosaver::caught_up), as the app's own frames would, before it marks
the session saved. A small helper in the test module does this. The
app's behaviour is unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:22:20 +00:00
Colin Faulkingham 5e4c52a7cb Sketch Transform: keep the applied rectangle at the previewed rotation and size (#11617) (#11620)
* claim: #11617

* Sketch Transform: preserve dimensioned geometry on apply
2026-11-04 05:21:27 +00:00
SilentWolf75 3a23d21795 Lip: clearer guidance when the face is not a shelled rim (#10437) (#10871)
* claim: #10437

Co-authored-by: Cursor <cursoragent@cursor.com>

* Lip: say Shell first when the face is not a rim (#10437)

A solid box top left OK disabled with a vague shelled-body hint. Name Shell and the rim-around-an-opening requirement in the refusal, the dialog tip, and the command help.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Lip: keep the dialog hint as it was, match the new refusal text in the test

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 04:42:08 +00:00
austin62jake e199f9aeff Agents and scripts: Match Surface describes its exact parameters and reply (#10981) (#10982)
* claim: #10981

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Match Surface: input and output schemas (#10981)

surface.match gets a typed contract in schema/surface_match.rs and the
shared FeatureResult as its output, and leaves unconverted.txt. Each key
reads as the handler reads it, with its message; like solid.scale it
declares `name` and `participants` itself, since a match makes no body.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Follow main's kernel merge: world_state and the shared continuity reader

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Surface match schema: Vec3 comes from solidcore_record now

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 04:14:11 +00:00
austin62jake e06d6e36d8 Agents and scripts: the material card checks describe their exact parameters and reply (#11036) (#11037)
* claim: #11036

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Physical material property commands: input and output schemas (#11036)

material.properties.validate and material.properties.directional get
typed contracts and reply types in schema/material_properties.rs and
leave unconverted.txt. The card stays the handler's to check; the
direction has the handler's messages. The replies mirror the document's
Validation and DirectionalModulus, destructured in full.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 04:08:15 +00:00
Kai Evans 9d1ad9f0b9 Feature editor: OK accepts the joint placements a roll back re-solves (#11094) (#11095)
* Feature editor: a roll back's re-solved placements are not a changed design

Rolling the marker solves the joints again and stores where that puts each occurrence, so the
editor's pinned original differed from its source there and OK was refused with "the original
editor model does not match its source design". The check now takes the placements as the roll
leaves them.

* Feature editor: also accept the ratio and lift of coupled joints that rolling the marker works out

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 03:56:46 +00:00
austin62jake d39a2d28db Agents and scripts: the clearance check describes its exact parameters and reply (#11051) (#11052)
* claim: #11051

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Clearance: input and output schemas (#11051)

inspect.clearance gets a typed contract and reply types in
schema/clearance.rs and leaves unconverted.txt. Each key reads as the
handler reads it, with its message; MAX_CLEARANCE and MAX_NEAREST move
to the schema module so both read one limit. The reply is built from
Clearance and its item types, with the same JSON as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 03:44:59 +00:00
Tyler Chiriboga 23c2244307 Inspect: center of mass reports the same volume as measure (#6173) (#11262)
* claim: #6173

* Inspect: center of mass reports the same volume as measure (#6173)

inspect.center_of_mass measured each body's display mesh, whose flat
facets cut curved faces short, so its volume and centre differed from
inspect.measure's (a sphere of radius 10: 4183.10 vs 4188.49 mm3). It now
takes both from the same cached kernel measurement Measure uses, and like
Measure it leaves out bodies that enclose no volume, counting them in
open_bodies_skipped.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Centre of mass tests: read the skipped list that main now returns

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Tyler Chiriboga <18270301+Tyler-Chiriboga@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 03:13:56 +00:00
bherbruck 481687fff3 Modeled threads: the thread runs down to a flat face with no uncut block, so a nut can bottom out (#11782)
* Modeled threads: run down to a flat shoulder so a nut can bottom out

An external thread running the whole face stopped with its groove clear of
a shoulder, leaving up to a turn uncut above it. The groove now runs on past
the plane of the face's end and is trimmed there, so its root reaches the
shoulder at every angle and nothing past the plane is cut.

A shoulder rising back along the face keeps the square stop, as do internal
threads, tapers, runouts and lengths stopping short. Where the kernel refuses
the trimmed cut, the groove also keeps the square stop and the feature warns.

Closes #11610

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Kernel ladder: record the thread cases that now reach their shoulder

Six cases gain the faces of the trimmed last turn; every other case is
unchanged apart from its timing.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Thread at a shoulder: review fixes, and a test of the fallback

The shoulder check starts just off the face, as the free-end check does, so
a slope rising from the face counts. A groove the trim leaves nothing of now
falls back to the square stop too. A test forces the fallback and reads its
warning. The ladder's order and report go back to main's, changed only for
the cases this moved.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Thread at a shoulder: check all the radii the groove cuts

The check for a shoulder rising along the face looked a fixed distance out,
which falls short of the groove on an undersized shaft. It now covers the
groove's reach. A test reads how each kind of shoulder is classified.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Thread at a shoulder: test the trim that leaves nothing of the groove

The test now reaches the path where the half-space keeps nothing, which is
the one that must fall back to the square stop.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Thread at a shoulder: say so when the shoulder can't be checked

If the kernel can't make the rising-shoulder check, the groove keeps the square
stop as before, and the feature now warns that its last turn is left uncut.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 02:43:36 +00:00
jyviko f74c6605c4 Fillet: Mitre or Swept corners where two rounds meet (#11498)
* SMLib wrapper: a swept or mitred turn for constant rounds (Turn)

Brep::blend_dropping takes a Turn: Mitre (the default, as before) or Swept,
which sets solidcore's FilletSession::set_swept_turns before the corners are
made. A swept turn solidcore refuses at CreateFilletCorners comes back as
Error::TurnRefused with its reason and place (NoRoom at a concave side edge,
else CornerUnsupported). Swept is for Blend::Round only.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Kernel: fillet rounds turn in a mitre or swept (FilletTurn)

fillet_removing, fillet_max, fillet_max_each and fillet_each take a
FilletTurn and hand it to SMLib, through the reblend of earlier rounds at a
later corner too. A swept corner solidcore does not build is refused as
invalid input with its reason and place: no smaller radius builds it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Design: a fillet saves its turn, mitre (default) or swept

FeatureKind::Fillet gets `turn` (FilletTurn: mitre, the default and not
saved, or swept), one pending format file. Evaluation hands it to the kernel,
pattern copies included, and refuses a swept turn with a style, fit, corner
or setbacks the kernel's swept corner does not take (SWEPT_TAKES).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* solid.fillet: the turn parameter, and the command's input and output schemas

`turn: mitre|swept` on solid.fillet, refused when unknown or with options a
swept corner does not take. solid.fillet now has its contract (FilletParams)
and reply type (FilletResult), registered with checked!, and leaves
schema/unconverted.txt.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Fillet dialog: a Turns choice, Mitre or Swept

A Constant radius fillet's Turns row sends `turn: swept`; choosing a narrowed
radius or a curvature corner puts it back to Mitre, and Swept puts those back.
Its tooltip says what each does where two rounds turn round a corner edge.
New strings in a pending locale file (en, de, ja, ko, zh); glossary entry.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Fillet turn: the schema tests' fixes and the complete catalogs' strings

The no-contract example in the schema tests is solid.chamfer now that
solid.fillet has one; the reply test checks the mid-radii form against the
contract (solidcore refuses that law's corner yet, #430) instead of building
it; Arabic, Hebrew and Vietnamese, complete catalogs, get the new strings.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Fillet schema test: variable and spline forms checked against the contract

solidcore fails a variable round on a box edge at this pin (status 1001, as
variable_radius_on_a_box_edge), so the reply test checks those forms against
the contract instead of building them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Fillet turn: the swept rig cases pass; the MCP no-contract example is solid.chamfer

cov-fillet-turn-swept-boss-base-r1 and
cov-fillet-turn-swept-pocket-sharp-corner-refused-r1 pass now that
solid.fillet takes `turn`: off the rig's known-failing list. The MCP bridge
test's example of a command without a contract was solid.fillet, which has
one now.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

---------

Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 20:36:00 -05:00
jyviko e12afd11cc Ellipse and spline extrudes make one wall; coplanar faces never lose their names to a budget (#11687)
* Planar sources: decide coplanar faces by support group, no budgets that drop sources

map_planar_sources kept two budgets, `metadata` (2 MB) and `identity_bytes`
(16 MB), charged by each face's source descriptor and by every extra
candidate on its plane. When one ran out, faces silently lost their source
and with it their persistent name: a 300-hole plate whose top a comb of
grooves cuts into 62 strips lost the cap source from strip ~47 on.

The budgets bounded the real cost: each result face looked at every
candidate on its plane, faces x candidates when many faces are coplanar.
The candidates are now grouped by support once (`same_support` against
each group's first plane, through the existing PlaneIndex), each group's
chosen source, ambiguity and inner flag decided once, and the faces
without a role together with the competing tool faces grouped the same
way. Each result face then looks up its plane's few groups. The decision
is the old one: blocked by an unknown or competing support, else the
first candidate on the plane unless another candidate there has a
different plane, source or role.

Work: grouping looks at each candidate's index window once per group
(each candidate is assigned on its group's first look), and each face
looks at the group representatives near its plane, so the planes looked
at grow with faces + candidates. A counter (`looked`) checks it: n
pieces of one cap and n faces on its plane take exactly 2n looks (n*n
before), 3n with n blocking supports.

Tests: the holed plate keeps the cap source on all 62 strips and the
bottom (fails on the old code: face 352 lost its source); the existing
planar source tests pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Extrude: an ellipse or a spline sketch curve sweeps one wall

An extruded ellipse had four side faces (`F2:side:e1#1` to `#4`), one
per conic quarter, and a spline one per span: the profile holds such a
curve as several exact segments and SMLib got each as its own curve.

The Extrude feature's builder (`extrude_whole_sides`) now takes a
`one_curve` test from the caller. Each run of free-form segments of one
degree that it says are one curve goes to SMLib as one B-spline: the
Béziers' control points in a row, knots repeated the degree times at
each joint, each span as long as keeps the speed through the joint the
same (so the curve is C1 there and SMLib makes one face over it, as it
would not over a C0 parametrisation), weights scaled to meet. Exact: the
same curve, reparametrised. A loop starts at a run's start, so no run
wraps round it. The wall's history lists every piece, and its
construction source is its first piece's (as joined collinear lines).

The document says which segments are one curve: ExactSketchCurves now
also holds each free-form sketch curve's exact segments
(`find_free_form`: a segment that is one of them, either way round), and
the profile roles use it, so the wall is named `side:<curve>` from its
history rather than by the curve nearest its mesh centre.

Saved designs rebuild with these walls renamed (`F2:side:e1#k` ->
`F2:side:e1`); accepted, no alias kept.

Tests: kernel `pieces_of_one_curve_extrude_into_one_wall` (an ellipse in
four conics is 3 faces, its volume pi a b h to 1e-6, the wall generated by
all four pieces; a 3-span spline closed by lines is one wall; the same
solid as a face per piece); doc
`an_extruded_ellipse_or_spline_is_one_wall_named_by_its_curve` (one
`F2:side:e1`, no piece names, the same names on a second build and after
reopening).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Tests: a tray's far rim name is checked on the exact B-rep face

The tray test took the point `sheet:...:rim.far` finds as right when it
lay within 1e-3 mm of the strip z = 30, 199 <= y <= 200: a slack sized to
the mesh. It now asserts on the face the name finds, on exact geometry:
the point's distance to that face's surface (FaceSurface::nearest) is
within rounding at the part's scale (1e-12 x its size), and the point is
inside the face's boundary, counted by crossings of its record edges
(each asserted to be a line, the rim being planar).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Tests: delete tests/naming/, the records no test reads since the naming corpus stopped comparing against them

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Viewport test: smooth wall joins drawn from eight G2 spline walls

The PR extrudes each spline or ellipse sketch curve as one wall, so the
offset-ellipse ring the test built has no joins between faces any more.
The test now extrudes a circle drawn as eight cubic splines: eight walls
meeting with one curvature along eight upright edges (24 edges in all).
It still checks that those joins are left out of the shaded view, drawn
in the wireframe and selectable as edges.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

* Kernel ladder: 009 and 012 record the one-wall extrude (an extruded spline or ellipse sweeps one wall)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ

---------

Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 21:27:14 +00:00
semmelrinde bcac191847 Typed schemas for Change Design Units (part of #2829) (#10934)
document.units gets input and output schemas: `units` (mm, cm, m, in or
ft, with the handler's messages for a missing or unknown unit) and its
reply, `units` and `kept`, which the handler now builds from the reply
type. Its line leaves unconverted.txt.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 17:13:37 +00:00
bherbruck c019e2555c Threads: a 12 mm hole now takes an M12 thread, and a modelled M12 builds in it (#11651)
* Threads: a 12 mm hole takes M12, and a modelled thread narrows a bore that is too wide (#11609)

Automatic sizing of a hole now takes the size whose nominal diameter the hole has (a 12 mm hole
is M12, not M14, the size whose tap drill is 12 mm); a hole near no nominal diameter is still
matched by tap drill. A modelled internal thread in a hole as wide as the thread or wider no
longer refuses with "too wide": the bore is filled in to the thread's minor diameter over the
threaded length and the groove cut there, and the feature's warning says so. A size that does
not fit the hole names the size that does.

A saved thread with no designation keeps the old rule through a pending format step, so an
existing design rebuilds with the size it had.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Threads: keep an old automatic size through edits; kernel-rig cases expect the narrowed bore (#11609)

Editing a saved automatic thread keeps `old_auto_size`, so it does not change size. The five
kernel-rig cases that expected a modelled thread in a hole as wide as the thread to be refused
now expect the narrowed bore, with the volume worked out analytically in each case. The thread
dialog's reopen pattern names the new field.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Thread dialog test: a 5 mm hole is suggested M5, its nominal size (#11609)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 17:09:27 +00:00
semmelrinde 2741df3629 Typed schemas for Appearance, Appearance Library and List Appearances (part of #2829) (#10976)
* Typed schemas for Appearance, Appearance Library and List Appearances (part of #2829)

appearance.assign, appearance.library and appearance.list get contracts
that read what their handlers read, with the handlers' messages, and
reply types the handlers build their replies from. Their lines leave
unconverted.txt.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Appearance library: the opacity is already a number

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 16:52:09 +00:00
Mateusz Sych bfb32cd6e1 Align: move bodies by frames that follow sketch and model geometry (#6257) (#7123)
* claim: #6257

* Align: move bodies by frames that follow sketch and model geometry (#6257)

solid.align takes from_frame and to_frame: a joint origin on the model
(a face's or round end's centre, an edge's middle, a corner, found again
by name) or a sketch point or a sketch circle's, arc's or ellipse's
centre. Both are resolved on every evaluation, so the bodies follow
their target when it changes. An optional angle turns the bodies about
the target's direction after the move.

The Align dialog takes sketch points and curves as From and To, offers
the Joint Origin snap and an Angle row.

* Align frames: box the joint origin (clippy large_enum_variant)

* Align frames: every moved body on the frame's component, placed once

* Align frames: a sketch id's length in characters

* Align frames: the dialog test's sketch curve hit carries its point

* Align frames: the dialog test takes Vec2 from solidcore-record

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-11-03 15:47:52 +00:00
Nathan Dane fbd30b0bd0 Panels: inspect and change layouts through commands (#11304) (#11454)
* claim: #11218

* Panels: add shared layout rules and retained content runtime

* claim: #11304

* Panels: inspect and change layouts through shared commands (#11304)

* Panels: account for layout commands in geometry coverage
2026-11-03 13:55:50 +00:00
jyviko dc38349cbf Layer split stage 1: the B-rep record, native body file and Vec/Plane come from solidcore-record (#11496)
* Kernel: the B-rep record and the native body file come from solidcore-record

SolidCraft's portable B-rep record (`solidcraft_geom::brep`: brep.rs, topology, history,
envelope, part, curve2) and the native body file (`solidcraft_kernel::native`: the tagged binary
codec and the container) move to solidcore's `solidcore-record` crate: geometry as data, with
structural validation and no maths (solidcore's docs/layers.md, stage 1). Every caller uses
`solidcore_record::brep` and `solidcore_record::native`; the old copies are deleted.

What computes from coordinates stays with the kernel, in geom's `brep`:
- `brep::check`: the record's geometric checks (orthonormal frames, non-degenerate planes,
  unit revolution axes, non-singular placements). `check::validate` and
  `check::validate_built` are the record's validation and these together, and every caller of
  `Brep::validate` uses them, so what is accepted is unchanged;
- `brep::label`: a surface's natural ranges (a cone's apex), `curve_interval`,
  `edge_curve_interval` and `form_turns`;
- `brep::pcurve::eval`: a pcurve's point;
- `geometry::check`, `placement::moved` and `split_edges` become functions on the record.

The kernel's native module keeps the conversion between a body and its saved record
(metadata.rs) and checks the decoded record's geometry. `native/analytic.rs`, which no module
declared, is deleted.

solidcore-record is pinned to its rev on solidcore's `layer-record` branch.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NvoxKMbMFSQU8tzRtoJhtT

* Geometry: Vec2, Vec3, Plane and the profile types come from solidcore-record

The plain data of vectors and points, sketch planes and planar profiles (`Seg2`, `Loop2`,
`Region2`) moves to `solidcore-record`, with its constants and component-wise operators only
(owner decision, solidcore layer split). Their geometry stays in geom as traits:
`Vec2Geom`, `Vec3Geom`, `PlaneGeom`, `Seg2Geom`, `Loop2Geom`, `Region2Geom`, `Seg2ArcFit`,
`Loop2ArcFit` and `Region2ArcFit`, in scope with `use solidcraft_geom::prelude::*`. Every caller
names the types from `solidcore_record`; geom no longer defines or re-exports them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NvoxKMbMFSQU8tzRtoJhtT

* Kernel: pin solidcore-record to solidcore main f5e8e9a29

solidcore-record is merged on solidcore main (tangency-labs/solidcore#2519); its crate there is
identical to the rev pinned before but for its clippy configuration. The lock is regenerated
against the merged rev.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NvoxKMbMFSQU8tzRtoJhtT

---------

Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-03 09:04:26 -05:00
jyviko 01abf8af17 Engine test: the later round at a mitred corner, as solidcore 8ddce9016 builds it
At solidcore 8ddce9016 (pinned) the r2, r2, r3 corner at (20, 0, 20) is one patch: 10 faces,
7928.743041456134 mm³, the r3 round and the patch the second fillet's. The test is pinned to
that and leaves the known-failing list. The fork's setback corner (15 faces, 7928.142563938)
waits for solidcore#2310; solidcore#2382, not merged, moves the case to 7927.724282512234.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
2026-11-02 21:47:28 -05:00
jyviko ac3e6e0105 A panic solidcore's guard catches reaches the engine as EngineError::Internal
smlib's run() catches a solidcore panic and replaces the kernel (kept: the app does not die
natively), returning Error::Internal; the kernel turned it into KernelError::Failed and the
engine into Other, so hostile_params_never_panic and the fuzzer no longer saw kernel panics.
Now the kernel's wrapper errors go through KernelError::smlib, which keeps Internal; a body the
wrapper could not take because of a panic is Internal too. A feature whose evaluation hit it
records the result as internal (not cached: the kernel was replaced), and the command that
made it, or any command whose kernel call returns it, fails with EngineError::Internal under
its own id. Crash reports mark a panic inside a solidcore call recovered (kernel::guarded), and
the crash test's kernel mode panics inside a solidcore call.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
2026-11-02 21:39:02 -05:00
jyviko d9310b2389 Crash test: a panic inside the kernel is the command's internal error
The kernel crash test's command must fail with EngineError::Internal naming
diagnostics.crash_test, as a panic in the command itself does: the panic detectors
(hostile_params_never_panic, the fuzzer) look for Internal. Today it is EngineError::Other.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
2026-11-02 21:39:01 -05:00
jyviko 440fdaed13 Kernel work runs on threads with solidcore's 8 MiB stack
solidcore's algorithms recurse (subdivision, intersection, filleting) and its README asks for
8 MiB, a native main thread's stack; Rust starts other threads with 2 MiB, and only the web
build set more. kernel::KERNEL_STACK is 8 MiB and kernel::kernel_thread a thread builder with it.
The kernel's parallel map runs on a pool of its own with that stack (on a worker of a pool
already, as in map_with_local_pools, it stays on that pool as before), the one-thread mesh pools
and map_on_threads' helpers have it, and so do the engine's interference worker and the
evaluator's plan-mesh thread. A test recurses 5 MiB deep on each.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
2026-11-02 21:38:50 -05:00
jyviko 3131b933c7 Engine blend-fit test: solidcore is the only kernel, so the refusal always names its edge
The cherry-picked test asked kernel_route which kernel ran; that choice is gone.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5M41G8Nua7HvuFPZkmJ
2026-11-02 20:38:25 -05:00