gen_commands: Generate and shrink model-valid command sequences

In s7contract: Behavioral Contracts and Generative Laws for 'S7'

Generate and shrink model-valid command sequences

Description

Sequence length is sampled from zero through min(size, max). Generation stops early if every command is unavailable. Available commands are selected with equal probability. Generated inputs must satisfy their preconditions.

Usage

gen_commands(initial, commands, max = 10L)

Arguments

initial	Initial model with value semantics.
commands	Non-empty list of descriptors made with new_command().
max	Maximum sequence length.

Details

Shrinking removes contiguous chunks, then shrinks command inputs. Each candidate is replayed against the symbolic model: commands with unsatisfied preconditions or missing output references are removed, together with any commands that depend on them. Retained command IDs never change. Candidate construction is lazy and executes no implementation commands or fixtures.

Value

An S7 generator of lists of steps. Each step contains an integer id, a command name, and its generated input.

s7contract documentation built on Sept. 10, 2026, 1:09 a.m.