Pressure module: three comparisons and a placement counterexample
Author: Signal & Ship. BoardPort comparison executed 17 September 2026.

Why this case exists
A report with 3 PASS / 0 DIFF / 0 NOT CHECKED can coexist with a changed footprint position. BoardPort compares eight pad/net mappings, five footprint references and defined outline/routed-segment properties for this design. It does not compare footprint coordinates or run DRC.

Inputs and results
All three comparisons use the same source.eagle-9.6.2.brd in ../pressure-module/.
A: destination.kicad-10.0.6.kicad_pcb in ../pressure-module/ -> 3 / 0 / 0.
B: edited-one-net.kicad_pcb in ../pressure-module/ -> 2 / 1 / 0. PAD2 pad 1 was deliberately changed N$3 -> N$4.
C: moved-pad2-1mm.kicad_pcb in this folder -> 3 / 0 / 0. Only PAD2's outer X position changed 1.143 -> 2.143 mm. Y stays -1.016 mm; tracks and pad/net assignments remain byte-identical. Reversing this one-byte edit restores A exactly. All report audit findings for A and C are identical.

Reproduce C
1. Create a new folder. Save the archived A file, LICENSE.txt and reproduce-placement.py there.
2. Run: python3 reproduce-placement.py
   This helper checks A's SHA-256, retains the original, and refuses to overwrite an existing moved-pad2-1mm.kicad_pcb. It does not execute KiCad or BoardPort.
3. Compare output SHA-256 with manifest.json. macOS: shasum -a 256 moved-pad2-1mm.kicad_pcb. PowerShell: Get-FileHash moved-pad2-1mm.kicad_pcb -Algorithm SHA256.
4. Read all three raw report-*.json files. Full comparison recomputation requires the paid BoardPort product; use the same original EAGLE file with each KiCad version. The free compatibility checker tests readability, not these comparisons. Downloading, inspecting, and reproducing the edit is free.

Interpretation and next step
For B, locate PAD2 / 1 and establish the intended net from the original. Even when the comparison passes, inspect footprint placement and routing in KiCad with the intended project/rules; refill applicable zones, run DRC, investigate actual findings, and review fabrication outputs. Do not change a net or weaken a rule solely to make a report pass.

Provenance and limits
Original: UCLA Networked & Embedded Systems Laboratory, locomote commit 1e199a91d4daabcabbf20aa5a1c3c0f1c5beb415, BSD-3-Clause. Preserve LICENSE.txt on redistribution. Original authors do not endorse BoardPort. See ../pressure-module/provenance.json and toolchain.json for the archived KiCad 10.0.6 import recorded 2026-08-31T21:42:40Z.
The archived pcbnew generator header is retained in C to isolate one byte; it does not mean KiCad produced C or was run again. The explicit filename, manifest and this notice identify an authored test copy, not an importer defect.
Only the unchanged BoardPort comparison was run on 17 September. No KiCad import, GUI loading, DRC, ERC, schematic parity, Gerber, manufacture or electrical test was performed. No particular DRC error is predicted. This single controlled case does not measure accuracy, customer benefit, migration success or AI citation probability. Engine/report hashes and runtime are in manifest.json; raw reports record product/engine/rule versions and input hashes.

Method reference: https://dev-docs.kicad.org/en/file-formats/sexpr-intro/#_position_identifier
Guide: https://boardport.signalandship.com/eagle-kicad-migration-check#comparison-boundary