Replace Machine.invoke with fluent state invocations

[SandroMaglione]

Summary

• Replace Machine.invoke and object invocation configs with state-local invoke: (from) => ... chains for Effects, Streams, timers, reusable logic, and child machines.
• Enforce source-specific typestate: every reachable completion, failure, and element channel is handled exactly once, snapshot handling remains optional, and owner state, protocols, source channels, failures, and services stay inferred through the chain.
• Normalize completed chains into the existing invocation runtime model, preserve state-scoped lifecycle semantics, and migrate tests, fixtures, benchmarks, documentation, and all affected examples.

Changeset

• Added or updated for a library or package-metadata change
• Not required because this PR does not change src/ or package.json

Validation

• pnpm check
• Relevant example checks, when examples changed (platformer, playground, and pokemon)
• Automated type-performance measurement passed
• Automated runtime- and memory-performance measurement passed

[github-actions]

Type performance

Measured with TypeScript 6.0.3 and skipLibCheck=true.

Scenario	Base	PR	Difference
Effect only	55	55	0 (0.0%)
Import effect-machine	55	55	0 (0.0%)
Machine.states (3 states)	3,039	3,039	0 (0.0%)
Machine.make (3 states, 2 events)	10,759	10,759	0 (0.0%)
machine.handle (3 states, 2 transitions)	29,625	28,192	-1,433 (-4.8%)
fluent transition (10 named branches)	107,405	106,006	-1,399 (-1.3%)
fluent invocation (state-dependent Effect)	92,583	92,756	+173 (+0.2%)
machine.handle (depth 24)	196,139	192,588	-3,551 (-1.8%)
machine.handle (wide depth 16)	227,285	220,950	-6,335 (-2.8%)
machine.handle (parallel/history/choice)	130,795	128,389	-2,406 (-1.8%)
machine definition (3 independent implementations)	129,848	127,342	-2,506 (-1.9%)
machine exact input/output/error/services	114,991	113,853	-1,138 (-1.0%)
execution adapter readiness	124,898	118,896	-6,002 (-4.8%)

Marginal instantiations are measured against the matching setup without that API call:

Scenario	Base	PR	Difference
Import effect-machine	0	0	0
Machine.states (3 states)	2,984	2,984	0 (0.0%)
Machine.make (3 states, 2 events)	7,712	7,712	0 (0.0%)
machine.handle (3 states, 2 transitions)	18,866	17,433	-1,433 (-7.6%)
fluent transition (10 named branches)	97,156	95,757	-1,399 (-1.4%)
fluent invocation (state-dependent Effect)	83,318	83,491	+173 (+0.2%)
machine.handle (depth 24)	175,443	171,892	-3,551 (-2.0%)
machine.handle (wide depth 16)	207,796	201,461	-6,335 (-3.0%)
machine.handle (parallel/history/choice)	110,532	108,126	-2,406 (-2.2%)
machine definition (3 independent implementations)	113,879	111,373	-2,506 (-2.2%)
machine exact input/output/error/services	100,480	99,342	-1,138 (-1.1%)
execution adapter readiness	93,457	89,084	-4,373 (-4.7%)

Check times (informational)

Scenario	Base	PR
Effect only	0.03s	0.03s
Import effect-machine	0.03s	0.04s
Machine.states (3 states)	0.12s	0.11s
Machine.make (3 states, 2 events)	0.18s	0.17s
machine.handle (3 states, 2 transitions)	0.29s	0.28s
fluent transition (10 named branches)	0.59s	0.54s
fluent invocation (state-dependent Effect)	0.52s	0.55s
machine.handle (depth 24)	0.88s	0.85s
machine.handle (wide depth 16)	0.91s	0.90s
machine.handle (parallel/history/choice)	0.68s	0.67s
machine definition (3 independent implementations)	0.68s	0.69s
machine exact input/output/error/services	0.64s	0.61s
execution adapter readiness	0.65s	0.60s

Type instantiations are the comparison metric. Check time varies with runner load and is informational only.

[github-actions]

Runtime performance

Median of 5 independent benchmark processes on AMD EPYC 9V74 80-Core Processor with Node v24.19.0.

Pull request baseline

Scenario	Effect Machine
Plan counter transitions	110,421 transitions/s
Drain burst with terminal fence	327,618 increments/s
Drain burst with a change observer	307,787 increments/s
Lookup and send to one child	282,433 increments/s
Start and stop a machine	138,677 machines/s
Start and stop a parent with one child	40,376 families/s
Plan transitions through a compound state	97,521 transitions/s
Plan transitions through parallel regions	80,667 transitions/s
Drain burst through a compound state	278,381 events/s
Drain burst through two parallel regions	284,365 events/s
Drain a compound-state burst with a change observer	263,879 events/s

Process runtime reference points

Scenario	Effect Machine
Start and stop a raw generic process	24,918 processes/s
Start and stop a raw compiled process	64,541 processes/s

Memory profile	Effect Machine
Idle machine	1.8 KiB
Raw generic managed process	13.9 KiB
Raw compiled process	3.2 KiB
Two independent idle machines	3.4 KiB
Idle parent with one child	5.8 KiB
Parent with observed child registry	10.0 KiB
Parent with observed invoked child snapshots	6.3 KiB

Effect Machine change from base

Metric	Base	Base variability	PR	PR variability	Difference
Plan counter transitions	112,746 transitions/s	1.2% MAD	110,421 transitions/s	2.3% MAD	-2.1%
Drain burst with terminal fence	328,057 increments/s	0.3% MAD	327,618 increments/s	0.1% MAD	-0.1%
Drain burst with a change observer	309,133 increments/s	0.9% MAD	307,787 increments/s	0.8% MAD	-0.4%
Lookup and send to one child	283,564 increments/s	0.9% MAD	282,433 increments/s	0.9% MAD	-0.4%
Start and stop a machine	138,102 machines/s	1.3% MAD	138,677 machines/s	0.8% MAD	+0.4%
Start and stop a parent with one child	41,058 families/s	0.4% MAD	40,376 families/s	0.8% MAD	-1.7%
Plan transitions through a compound state	99,569 transitions/s	0.3% MAD	97,521 transitions/s	0.9% MAD	-2.1%
Plan transitions through parallel regions	80,909 transitions/s	1.1% MAD	80,667 transitions/s	1.6% MAD	-0.3%
Drain burst through a compound state	280,461 events/s	1.5% MAD	278,381 events/s	1.1% MAD	-0.7%
Drain burst through two parallel regions	289,416 events/s	0.3% MAD	284,365 events/s	1.3% MAD	-1.7%
Drain a compound-state burst with a change observer	264,516 events/s	0.8% MAD	263,879 events/s	0.5% MAD	-0.2%
Idle machine heap per unit	1.8 KiB	0.2% MAD	1.8 KiB	0.1% MAD	-0.3%
Raw generic managed process heap per unit	13.9 KiB	0.0% MAD	13.9 KiB	0.0% MAD	+0.0%
Raw compiled process heap per unit	3.2 KiB	0.6% MAD	3.2 KiB	0.4% MAD	-0.0%
Two independent idle machines heap per unit	3.4 KiB	0.1% MAD	3.4 KiB	0.1% MAD	+0.5%
Idle parent with one child heap per unit	5.8 KiB	0.0% MAD	5.8 KiB	0.0% MAD	-0.0%
Parent with observed child registry heap per unit	10.0 KiB	0.0% MAD	10.0 KiB	0.0% MAD	+0.0%
Parent with observed invoked child snapshots heap per unit	6.3 KiB	0.1% MAD	6.3 KiB	0.0% MAD	+0.1%

Process runtime reference change from base

Metric	Base	Base variability	PR	PR variability	Difference
Start and stop a raw generic process	25,708 processes/s	1.8% MAD	24,918 processes/s	2.1% MAD	-3.1%
Start and stop a raw compiled process	64,922 processes/s	0.8% MAD	64,541 processes/s	1.2% MAD	-0.6%

Regression guard

No large, noise-adjusted throughput or heap regressions detected.

Versions and interpretation

• Effect Machine: 0.17.0

Higher throughput is better; lower heap is better. Variability is the median absolute deviation across independent processes, relative to their median. Small differences on shared GitHub-hosted hardware remain informational; the required guard rejects only large changes beyond the measured noise allowance.