Warn on unused module io ports

[akhilles]

• warn when a module io() is declared but never connected to any realized ports
• add core and CLI regression coverage for forwarded vs genuinely unused module inputs
• add an unreleased changelog entry for the new warning

---

Note

Medium Risk
Adds a new warning emitted during schematic conversion and threads span/call-stack metadata through module parameter tracking, which may change diagnostic output and snapshots across builds.

Overview
Adds a new module.io.unused warning when a module io() value (including interfaces) is ultimately not connected to any realized ports in that module, while avoiding false positives for forwarded nets/partially-used interfaces.

Plumbs io()/config() declaration span + call stack into ParameterMetadata and Diagnostic (new with_call_stack) to produce richer, correctly-attributed warnings, and updates tests/snapshots accordingly.

Fixes stdlib variant APIs by making Crystal.GND and MountingHole.P1 only exist for the relevant package/plating variants, updating downstream tests that previously passed now-nonexistent ports, and documents the changes in CHANGELOG.md.

^{Written by Cursor Bugbot for commit cb2eaac. This will update automatically on new commits. Configure here.}

---

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 3 additional findings.

[github-actions]

Build Performance

Board	Base (median)	Head (median)	Change
demo/DM0001	199ms ±4	197ms ±5	-1.4%
demo/DM0002	176ms ±3	181ms ±4	+2.6%
demo/DM0003	194ms ±3	197ms ±2	+1.3%
arduino/Nano	130ms ±4	138ms ±2	+5.9%
arduino/UNOQ	218ms ±4	232ms ±3	1.07× ±0.02 slower

_{Measured with hyperfine. Times show median ±stddev.}