fix(schematic): respect schTraceAutoLabelEnabled={false} to suppress port-net auto-labels

[gsdali]

Summary

Closes #2243.

schTraceAutoLabelEnabled is declared as a tri-state prop in the type definitions, but only the true case was wired up at runtime — for opting INTO net labels at complex multi-junction traces (in Trace_doInitialSchematicTraceRender.ts). Setting the prop to false had no observable effect, because the only consumer treats falsy (undefined or false) as "not opted in".

The user-facing pain: named-net connections (every chip pin tied to net.V3_3 / net.GND / etc.) automatically get a schematic_net_label drawn at every endpoint — the canonical "label, not wire" rendering. Users wanted a way to opt out and got none.

Fix

Wire schTraceAutoLabelEnabled === false (explicit, not undefined) into Group_doInitialSchematicTraceRender so that insertNetLabelsForPortsMissingTrace is skipped when the user explicitly opts out. Both call sites (early-return when no routable connections, and post-routing) honour the opt-out.

Tri-state semantics

value	behaviour
undefined (default)	existing behaviour, auto labels emitted
true	opt INTO labels for complex traces (existing, in Trace_doInitialSchematicTraceRender)
false	opt OUT of port-net auto-labels (NEW)

Test plan

• New tests/components/sch/sch-trace-autolabel-disabled.test.tsx:
  • false → 0 schematic_net_label records for a chip on net.V3_3 / net.GND
  • unset (default) → ≥1 schematic_net_label (existing behaviour preserved)
• bun test tests/components/sch: 2/2 pass
• Full-suite delta vs main: same flaky-test profile (1 fail in tests/repros/repro81-panel-unnamed-pcb-group.test.tsx which flakes under full-suite load on main too — passes in isolation; not caused by this change)

What this PR does NOT solve

• The true case behaviour is unchanged from current semantics (only opts into complex-trace labels).
• This PR only suppresses labels at port-missing-trace junctions (the most common user pain). Other auto-label sources (e.g. labels emitted by the trace solver inside applyNetLabelPlacements) are not affected. Reasonable next step is to extend the opt-out to those paths too, but it'd touch more code and is left as follow-up.

Why this scope

The most surprising part of the original report was the prop has zero observable effect when set to false. After this PR, setting false produces an immediately-visible difference (no more port-net auto-labels). Users who want "named nets render as wires, not labels" still need other mechanisms (port-to-port connections via array connections), but at least they have one knob that makes a visible difference.

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
tscircuit-core-benchmarks	Ready	Preview, Comment	May 6, 2026 6:47am