duplicate-net-name

Remedy ¶

Rename one of the nets. If the two were meant to be one net, connect them instead, and check whether a connect-by-name path has already merged them somewhere else in the design.

What it means ¶

More than one net in the design states the same non-empty name.

Why engineers want it ¶

On schematic formats connectivity IS naming, since same-named labels join, so this cannot happen there by construction (the solver merges them; see label-alias-conflict for the collapsed-alias hazards). It happens on formats with EXPLICIT net lists (an EDIF netlist can state two (net VCC …) entries) and wherever a tool export or a reader gap splits one logical net into two records. Either way, every consumer that keys on net names, whether diff identity, board joins or reviews, is now quietly wrong.

Impact ¶

A revision diff keyed by name sees one net where the design has two; a schematic-vs-board join attaches findings to the wrong copper.

Scope note ¶

Fires once per claiming net (each is a place to fix). Synthesized stub names (N$, unconnected-(…)) are per-net reader inventions and never collide; empty names carry no claim. A firing on a fresh format’s corpus is ALSO a read-health signal, so check the reader before the design (the pin-net-conflict lesson).

Query structure ¶

select the nets whose exact name is stated by two or more nets.

select N in nets where count(M : M in nets, M.name == N.name) >= 2

Reads: net.names. Tier R.