boardproof (TotallyNotAHackerTrustMeIAmADolphin) errors

4 recurring problems engineers report in boardproof (TotallyNotAHackerTrustMeIAmADolphin), ranked by impact.

#114 by errors trackedECAD

Recurring boardproof (TotallyNotAHackerTrustMeIAmADolphin) errors, by impact

  1. 1Neither check_connectivity nor ERC reports empty Footprint, missing LCSC or a Datasheet of '~' on placed symbols, so a schematic can look finished and still produce an incomplete BOM. A read-only, hierarchy-wide check would list every symbol missing a footprint, an LCSC number or a datasheet, by reference and uuid. The checks that exist all look at wiring, while the ordering data lives in fields nobody verifies until assembly.ECAD · Agent failure · 1 report74
  2. 2The maintainer of a KiCad agent toolkit is collecting the questions a downstream hardware project's agent had to ask because it could not find the answer. JLCPCB fab limits exist as a Python profile but no tool writes them into the board's .kicad_dru rules, the JLC part tier lookup says 'unknown' for Extended parts, LCSC search fails behind a bot wall with no written fallback, and a reported KiCad 10.0.7 behaviour drops creepage paths whose midpoint falls outside Edge.Cuts.ECAD · Agent failure · 1 report74
  3. 3A peer project asked for one call that checks every part in a project against LCSC: Basic or Extended, stock, and whether the footprint fits the package. Today it is one part at a time, and once it suggested a 0402 part for a 0603 footprint. They want a package and footprint mismatch reported as an error. Same failure as finding 1, seen from the sourcing side.ECAD · Agent failure · 1 report66
  4. 4Renaming X to PWR_X when a PWR_* pattern exists silently puts the net in that class. Python's fnmatch reads [..] as a character class where KiCad's wildcard does not. The board keeps old names until someone runs the update from schematic.ECAD · Agent failure · 1 report66

How it fails

  1. Agent failure4

Sourced from public community reports. Costs and delays on individual entries are modeled estimates, not measurements.