Unify state construction and forward fresh root input

[SandroMaglione]

Summary

Root initial constructors now receive machine input directly, so one-off values can belong to the initial child without being retained at root. Root transitions require fresh input and reuse those constructors; root data updates retain the current child configuration.

Replace event initial with target. Named selectors and history fallbacks now take construction objects with data, a per-node decoded flag, states, and declared owner update values. Preserve source-local parallel retention, explicit choice routing, complete history fallbacks, and inspectable branch destinations. Remove the unused fluent-builder runtime and migrate examples, tests, and guides.

Changeset

• Added or updated for a library or package-metadata change
• Not required because this PR does not change a publishable package

Minor changeset with direct migration examples; the fixed package group releases together.

Validation

• pnpm check
• Automated type-performance measurement passed or was not required
• Automated runtime- and memory-performance measurement passed or was not required

Local type budgets and runtime/memory smoke measurements passed. New coverage checks root-input requirements, nested-input exclusion, subtree and owner types, root reentry, validation failure before commit, explicit parallel choices, and generic-versus-indexed execution. Paired CI reports will be reviewed before merge.

[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.state (3 children)	3,521	3,521	0 (0.0%)
Machine.make (3 states, 2 events)	9,491	9,488	-3 (-0.0%)
machine.handle (3 states, 2 transitions)	49,816	49,981	+165 (+0.3%)
declared transition (10 named branches)	66,784	63,038	-3,746 (-5.6%)
atomic construction, guards, and reentry	91,682	88,529	-3,153 (-3.4%)
registered invocation (input-taking Effect)	49,109	49,785	+676 (+1.4%)
machine.handle (depth 24)	197,583	197,380	-203 (-0.1%)
machine.handle (wide depth 16)	221,254	221,123	-131 (-0.1%)
machine.handle (parallel/history/choice)	101,203	95,104	-6,099 (-6.0%)
machine definition (3 independent implementations)	102,696	100,566	-2,130 (-2.1%)
machine exact input/output/error/services	88,468	85,901	-2,567 (-2.9%)
execution adapter readiness	75,683	73,376	-2,307 (-3.0%)

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

Scenario	Base	PR	Difference
Import effect-machine	0	0	0
Machine.state (3 children)	3,466	3,466	0 (0.0%)
Machine.make (3 states, 2 events)	5,962	5,959	-3 (-0.1%)
machine.handle (3 states, 2 transitions)	40,325	40,493	+168 (+0.4%)
declared transition (10 named branches)	57,673	53,930	-3,743 (-6.5%)
atomic construction, guards, and reentry	75,870	72,720	-3,150 (-4.2%)
registered invocation (input-taking Effect)	41,409	42,088	+679 (+1.6%)
machine.handle (depth 24)	179,729	179,529	-200 (-0.1%)
machine.handle (wide depth 16)	204,178	204,050	-128 (-0.1%)
machine.handle (parallel/history/choice)	88,160	82,064	-6,096 (-6.9%)
machine definition (3 independent implementations)	91,471	89,344	-2,127 (-2.3%)
machine exact input/output/error/services	75,730	73,166	-2,564 (-3.4%)
execution adapter readiness	23,335	23,179	-156 (-0.7%)

Check times (informational)

Scenario	Base	PR
Effect only	0.02s	0.02s
Import effect-machine	0.02s	0.02s
Machine.state (3 children)	0.08s	0.08s
Machine.make (3 states, 2 events)	0.11s	0.11s
machine.handle (3 states, 2 transitions)	0.25s	0.25s
declared transition (10 named branches)	0.31s	0.32s
atomic construction, guards, and reentry	0.38s	0.38s
registered invocation (input-taking Effect)	0.25s	0.25s
machine.handle (depth 24)	0.47s	0.47s
machine.handle (wide depth 16)	0.51s	0.51s
machine.handle (parallel/history/choice)	0.39s	0.38s
machine definition (3 independent implementations)	0.41s	0.41s
machine exact input/output/error/services	0.39s	0.38s
execution adapter readiness	0.33s	0.33s

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 7763 64-Core Processor with Node v24.20.0.

Pull request baseline

Scenario	Effect Machine
Plan counter transitions	107,716 transitions/s
Drain burst with terminal fence	368,728 increments/s
Drain burst with a change observer	346,391 increments/s
Lookup and send to one child	303,366 increments/s
Start and stop a machine	111,272 machines/s
Start and stop a parent with one child	23,748 families/s
Plan transitions through a compound state	96,117 transitions/s
Plan transitions through parallel regions	75,815 transitions/s
Drain burst through a compound state	317,683 events/s
Drain burst through two parallel regions	293,693 events/s
Drain a compound-state burst with a change observer	300,407 events/s

Process runtime reference points

Scenario	Effect Machine
Start and stop a raw generic process	15,410 processes/s
Start and stop a raw compiled process	60,680 processes/s

Memory profile	Effect Machine
Idle machine	1.9 KiB
Raw generic managed process	13.9 KiB
Raw compiled process	3.2 KiB
Two independent idle machines	3.5 KiB
Idle parent with one child	6.0 KiB
Parent with observed child registry	10.0 KiB
Parent with observed invoked child snapshots	6.5 KiB

Effect Machine change from base

Metric	Base	Base variability	PR	PR variability	Difference
Plan counter transitions	109,232 transitions/s	1.9% MAD	107,716 transitions/s	1.7% MAD	-1.4%
Drain burst with terminal fence	367,987 increments/s	1.7% MAD	368,728 increments/s	0.6% MAD	+0.2%
Drain burst with a change observer	345,881 increments/s	0.6% MAD	346,391 increments/s	0.7% MAD	+0.1%
Lookup and send to one child	299,191 increments/s	1.0% MAD	303,366 increments/s	0.6% MAD	+1.4%
Start and stop a machine	112,905 machines/s	0.4% MAD	111,272 machines/s	1.5% MAD	-1.4%
Start and stop a parent with one child	23,425 families/s	2.2% MAD	23,748 families/s	1.5% MAD	+1.4%
Plan transitions through a compound state	92,861 transitions/s	1.9% MAD	96,117 transitions/s	1.5% MAD	+3.5%
Plan transitions through parallel regions	74,803 transitions/s	1.8% MAD	75,815 transitions/s	2.9% MAD	+1.4%
Drain burst through a compound state	291,246 events/s	0.8% MAD	317,683 events/s	1.7% MAD	+9.1%
Drain burst through two parallel regions	272,997 events/s	0.5% MAD	293,693 events/s	1.2% MAD	+7.6%
Drain a compound-state burst with a change observer	276,060 events/s	0.3% MAD	300,407 events/s	0.9% MAD	+8.8%
Idle machine heap per unit	1.9 KiB	0.1% MAD	1.9 KiB	0.6% MAD	+0.2%
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	0.3% MAD	3.2 KiB	0.5% MAD	-1.2%
Two independent idle machines heap per unit	3.6 KiB	0.6% MAD	3.5 KiB	0.0% MAD	-0.4%
Idle parent with one child heap per unit	5.9 KiB	0.0% MAD	6.0 KiB	0.0% MAD	+0.1%
Parent with observed child registry heap per unit	10.0 KiB	0.0% MAD	10.0 KiB	0.0% MAD	+0.1%
Parent with observed invoked child snapshots heap per unit	6.5 KiB	0.1% MAD	6.5 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	15,281 processes/s	0.5% MAD	15,410 processes/s	0.5% MAD	+0.8%
Start and stop a raw compiled process	61,084 processes/s	0.4% MAD	60,680 processes/s	0.5% MAD	-0.7%

Regression guard

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

Versions and interpretation

• Effect Machine: 0.35.0

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.