stampNeedsNewton
stampNeedsNewton(
stamp):boolean
Defined in: packages/compiler/src/stamp-partition.ts:282
Does this stamp have to be re-evaluated inside a Newton loop?
The stamp-level twin of partition.ts’s lawNeedsNewton, and deliberately a second fact
rather than a reuse of the first. A law and its stamps are not the same list — a jack’s
port-engage law lowers to a plain conductance, an ac-source law lowers to two stamps,
an open law lowers to none — and what a solver has to do is decided by what is in the
block, not by what was in the document. Block.linear is read by the runtime to choose
between one pass and an iteration cap, so it has to describe the stamps it will evaluate.
The two mistakes are not symmetric, which is why the table has to be exhaustive. Calling a
linear stamp nonlinear costs a few iterations; calling a nonlinear stamp linear makes
iterate run one pass, evaluate the curve once from a zero start, and emit that as a solved
answer — no held sample, no telemetry, no diagnostic.
Parameters
Section titled “Parameters”Returns
Section titled “Returns”boolean