fix(solderjumper): default to 2 pins when no pin count is given

[zkasuran]

Closes #784

What

A bare fp.string("solderjumper") (no pin count) silently produced invalid geometry: a pcb_courtyard_rect with width and center.x set to NaN plus a pcb_silkscreen_text with a NaN anchor. solderjumper2 and solderjumper3 were already fine.

The cause is in src/fn/solderjumper.ts: num_pins was read from params with no default, so an omitted count left it undefined and (num_pins - 1) * padSpacing (and everything derived from it) became NaN.

Fix

Default num_pins to 2 when it is omitted, the same way the function already defaults p, pw and ph. This matches the wider convention in this repo where a missing pin count falls back to a sensible default (dip gives 6, ssop gives 8, pinrow gives 6, plus many others). 2 lines up with solderjumper2, the smallest documented variant and the first example in the doc comment. Explicit counts are unchanged.

Tests

• An assertion test: a bare solderjumper now yields 2 pads with finite courtyard geometry and a finite silkscreen text anchor (no NaN). It fails on the current code (0 pads, NaN width) then passes with the fix.
• A visual SVG snapshot (solderjumper_default) rendered from the fixed output, showing the real 2-pin footprint.

Verification

Verified locally before submitting:

• bun test: 523 pass, 0 fail
• bun run build: success (ESM plus dts)
• bunx biome format clean on the changed files
• confirmed the assertion goes red on the pre-fix code then green after the fix. Opened the generated snapshot and confirmed it shows a real footprint

AI disclosure

AI assistance (Claude, Anthropic) was used in developing this change. Verified locally before submitting: full bun test (523 pass, 0 fail), plus bun run build and biome format on the changed files.

[zkasuran]

Added a KiCad parity test up front so this does not need a separate round.

tests/kicad-parity/solderjumper_kicad_parity.test.ts now has a parity/solderjumper-default-pincount case. It renders the bare solderjumper form this PR enables (no pin count, so num_pins falls back to the new default of 2) and compares it against the real KiCad SolderJumper-2_P1.3mm_Open_Pad1.0x1.5mm footprint. Courtyard diff is 0.00%, well under the 0.5 threshold the other parity tests use, so the defaulted output is not just the right pad count, it is geometrically the correct SolderJumper-2.

One honest note on dimensions. footprinter's own solderjumper defaults (2.54mm pitch, 1.5mm square pads) are a generic 0.1 inch jumper that does not track any KiCad standard part. KiCad's SolderJumper-2 is a smaller 1.3mm-pitch bridge pad. So the parity case supplies KiCad's dimensions explicitly (solderjumper_p1.3_pw1_ph1.5), the same way the existing solderjumper2 parity case does. What it proves is that the new default-pin-count path produces the same KiCad-correct geometry the explicit form already did. The pure 2.54mm default is left covered by the finite-geometry assertion plus the solderjumper_default snapshot already in this PR.

Full suite green: 524 pass, 0 fail. bun run build clean, biome format clean on the changed file.