Skip to content

feat(desktop): the assistant looks up the standards of a kind of part before it designs one - #332

Merged
ManuelStaggl merged 2 commits into
mainfrom
feat/standards-lookup
Nov 11, 2026
Merged

ManuelStaggl merged 2 commits into
mainfrom
feat/standards-lookup

Conversation

@ManuelStaggl

Copy link
Copy Markdown
Collaborator

Decided with the product owner today: the remaining kinds of the standards list get no fixed block. Their shape depends on the part; what stays is rules and sizes. Blocks are kept where two parts must match or SolidCraft is tricky.

  • standards.ts: one table with eight topics (sliding lid, clip and clamp, feet, hooks and eyes, cable entry, vents and fans, board mounts, star knob). Each rule says whether its figure comes from a sheet or standard, a range of guides, or our own choice.
  • standardTool.ts: partbuddy_standard {topic} returns a topic's rules; it builds nothing, is allowed by the policy and leaves no line in the work log. The role prompt only names the topics.
  • Unit tests. No contract test: the tool does not touch SolidCraft.

Nothing is printed, and what the assistant builds from these rules is not tested against SolidCraft.

🤖 Generated with Claude Code

ManuelStaggl and others added 2 commits November 11, 2026 18:28
… before it designs one

partbuddy_standard gives the rules and sizes for eight kinds of part
(sliding lid, clamp, feet, hook, cable entry, vent, board mount, star
knob) from one table, instead of a fixed block per kind.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
The gear test read the driver's position a frame before the pause took
hold and then expected that value to be kept (39.2 read, 39.1 kept).

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
@ManuelStaggl
ManuelStaggl merged commit 882fe8b into main Nov 11, 2026
7 checks passed
@ManuelStaggl
ManuelStaggl deleted the feat/standards-lookup branch November 11, 2026 17:43
ManuelStaggl added a commit that referenced this pull request Nov 11, 2026
…its rounding (#333)

Two joint tests of `app.spec.ts` failed now and then in CI (#327, #332).

- Cause: the view's `data-shown` has two decimals, the app keeps a tenth
of the unrounded value. 13.749 shows as 13.75, is kept as 13.7, and the
tests rounded 13.75 to 13.8. Both now compare within 0.06, as the hinge
test above them already does.
- The wait for a standing value added in #332 rested on a wrong guess (a
frame race) and is removed.
- Driving two joints back to 0 gets 30 s instead of a poll's 5 s; on a
loaded runner the second joint was still on its way.

Both tests ran six times each locally.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: ManuelStaggl <17688572+ManuelStaggl@users.noreply.github.com>
Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant