JSON Bug Report:

[mohan-bee]

Solver input JSON attachment

_board#14 _-SchematicTracePipelineSolver-solver-input.json

Notes

No response

[tscircuitbot]

Created or updated #710 for this JSON bug report.

Report id: bug-report-20260724T175257Z

[DPS0340]

The JSON Bug Report issues carry an input attachment but an empty Notes field, so what's actually wrong with each board isn't recorded anywhere. I ran every imported bug report through a set of measurable defect checks so the queue can be triaged without re-deriving each one.

Method

Solve each board with SchematicTracePipelineSolver, then count on the final output (traceCleanupSolver2 traces, netLabelNetLabelCollisionSolver label placements):

• cross — different-net trace segments properly crossing
• coll — different-net segments drawn collinearly (two nets render as one wire)
• inChip — labels ≥50% covered by a chip body, matching CHIP_COLLISION_MIN_OVERLAP_FRACTION
• lblOv — label boxes overlapping each other
• diag — segments that are neither horizontal nor vertical
• failed — routes the line solver gave up on

Results

board             chips traces labels | cross coll inChip lblOv diag failed
20260706T213649Z     35     46    104 |    11    0      0     0    0      0
20260706T220324Z     55     58     85 |     6    0      0     1    0      0
20260721T221026Z     34     51     79 |     5    0      0     2    0      0
20260716T144856Z     16     21     29 |     4    0      1     0    1      0
20260717T022934Z     12     17     24 |     3    0      0     0    1      0
20260707T134549Z     14     22     20 |     3    0      0     1    0      0
20260707T230831Z      7     10     11 |     2    0      4     0    0      0
20260707T092615Z     39     50     53 |     2    1      1     1    0      0
20260708T055430Z      2      6      5 |     2    0      0     0    0      0
20260707T140410Z     17     23     28 |     1    0      0     1    0      0
20260707T020342Z      6      8      3 |     1    0      0     0    0      0
20260707T141421Z     13     18     17 |     0    0      1     1    0      0
20260707T141025Z      4      5      6 |     0    0      1     0    0      0
20260717T031704Z      2      2      1 |     0    0      0     0    0      0
20260708T053736Z      2      1      0 |     0    0      0     0    0      0
20260707T134722Z      5      6      8 |     0    0      0     0    0      0
20260708T095725Z      6      5      9 |     0    0      0     0    0      0
20260717T042845Z      2      5      2 |     0    0      0     0    0      0
20260724T175257Z      4      5      3 |     0    0      0     0    0      0

The part worth acting on

Six boards score zero on every check, and all six issues are still open:

board	issue
20260717T031704Z	#676
20260708T053736Z	#639
20260707T134722Z	#622
20260708T095725Z	#643
20260717T042845Z	#679
20260724T175257Z	#709

Every solver reports solved, no route fails, and none of the geometric defects above are present. Either these were fixed by later work, or the problem is something my checks don't cover — which is exactly what the empty Notes field makes impossible to tell.

Suggestion: for those six, either close them or add a line saying what to look for. Right now they're indistinguishable from the boards that do have real defects, which makes the whole queue harder to work through.

Caveats

• These checks are geometric. A board could be wrong in a way none of them see — wrong net connectivity, a missing label, bad ordering — and score zero here. A zero is "no measurable defect of these kinds", not "correct".
• I've been wrong with checks like these twice today and caught it by inspecting the geometry: a chip-intersection check that flagged traces reaching their own endpoint pins, and a label-overlap check that flagged 1–6% grazes the solver tolerates deliberately. Both are excluded above.

Related PRs measured against this table: #718 (pins crossings), #720 (pins label placement), #719 / #723 (reduce crossings and collinear overlaps), #722 (removes both diagonals).