$n is sometimes incorrectly tagged as constant

[frothga]

$n may be expressed as a constant (eg: $n=17) or initOnly, but actually change due to structural dynamics, such as from $p or $type. This becomes a problem if other equations depend on the current value of $n. (If $n has no users, it doesn't matter.) The simplest solution is to always assume $n is dynamic, but this is inefficient.

The difficulty is that constants are analyzed early, before $p and $type. Yet we need to know about $p and $type to determine if $n is a constant (and thus whether equations that depend on $n are constants).

Notes on EquationSet:
need to retag $n as early as possible in analysis
depends on resolveRHS(), findDeath() and canGrow()
canGrow() has no dependencies, and resolveRHS() is early
need to move findDeath() as early as possible