=== BEGIN ARTIFACT FEATURE-Streaming.md ===
# FEATURE: Streaming Transformations

| Field       | Value |
|-------------|-------|
| Version     | 20260822 V1 |
| Description | Reconstruct, truncate, and emit jq streaming representations. |
| Depends On  | ARCHITECTURE.md, FEATURE-Generator-Core.md, FEATURE-Input-Streams.md |
| Provides    | tostream, fromstream, truncate_stream |
| Consumes    | ordered generator evaluation, input stream controls |

## Questions

- None.

## Workflow

Streaming filters represent composite JSON values as ordered path/value events. `tostream` emits
the representation, `truncate_stream` removes a specified number of leading path components, and
`fromstream` reconstructs values from the resulting event stream. Implement these filters with
generator ordering, empty-container markers, truncation, and round-trip behavior matching the
manual and corpus.

## Programmatic Acceptance

=== AC streaming-conformance ===
Intent: The streaming implementation passes every selected corpus case covering tostream, fromstream, and truncate_stream.
Suite: scoped
Requires: executable=python3; scope=test

import json
import os
import subprocess
import sys

selector = r"tostream|fromstream|truncate_stream"
result = subprocess.run(
    [sys.executable, "sources/run_conformance.py", "--select", selector, "--json"],
    capture_output=True,
    text=True,
    env={**os.environ, "JQ": f"{os.getcwd()}/jq"},
)
print(result.stdout)
print(result.stderr, file=sys.stderr)
report = json.loads(result.stdout)
summary = report["summary"]
assert sum(summary.values()) > 0
assert summary["fail"] == 0 and summary["error"] == 0
assert result.returncode == 0
=== END AC streaming-conformance ===

=== AC streaming-round-trip ===
Intent: A streamed value can be reconstructed without changing its value.
Suite: scoped
Requires: executable=python3; scope=test

import json
import os
import subprocess
import sys

program = r". as $dot | fromstream($dot | tostream) | . == $dot"
payload = " [0,[1,{\"a\":1},{\"b\":2}]]\n"
result = subprocess.run(
    [f"{os.getcwd()}/jq", "-c", program],
    input=payload,
    capture_output=True,
    text=True,
    env={**os.environ},
)
print(result.stdout)
print(result.stderr, file=sys.stderr)
assert result.returncode == 0
assert json.loads(result.stdout) is True
=== END AC streaming-round-trip ===

## User Acceptance

- None.

## Guardrails

- Streaming filters must preserve event order and must not use third-party runtimes.
- Invalid stream structures must produce jq-compatible runtime failures.
=== END ARTIFACT ===
=== BEGIN ARTIFACT FEATURE-Conformance-Assets.md ===
# FEATURE: Conformance Assets

| Field       | Value |
|-------------|-------|
| Version     | 20260822 V1 |
| Description | Stage and validate the immutable jq conformance assets. |
| Depends On  | ARCHITECTURE.md |
| Provides    | staged conformance corpus, exclusions, runner, scoring script |
| Consumes    | none |

## Questions

- None.

## Workflow

Place the supplied conformance sources under `sources/` without modification. Validate the
harness in process by importing its corpus and exclusion parsers, confirming the pinned corpus
parses completely and every declared exclusion matches a corpus case. This story verifies staging
only and does not claim interpreter behavior.

## Programmatic Acceptance

=== AC conformance-assets ===
Intent: The staged corpus and exclusions parse successfully and remain mutually consistent with the pinned authoritative assets.
Requires: executable=python3; scope=test

import sys

sys.path.insert(0, "sources")
import run_conformance as harness

cases = harness.parse_corpus(harness.CORPUS.read_text(encoding="utf-8"))
excluded = harness.apply_exclusions(
    cases,
    harness.parse_exclusions(harness.EXCLUSIONS),
)
assert len(cases) == 550
assert len(excluded) == 13
=== END AC conformance-assets ===

=== AC conformance-source-integrity ===
Intent: The staged harness exposes the required corpus and exclusion paths and can be imported from the build directory.
Requires: executable=python3; scope=test

from pathlib import Path
import sys

sources = Path("sources")
assert (sources / "jq.test").is_file()
assert (sources / "exclusions.txt").is_file()
assert (sources / "run_conformance.py").is_file()
sys.path.insert(0, str(sources))
import run_conformance
assert run_conformance.CORPUS == sources / "jq.test"
assert run_conformance.EXCLUSIONS == sources / "exclusions.txt"
=== END AC conformance-source-integrity ===

## User Acceptance

- None.

## Guardrails

- Do not modify, regenerate, trim, or substitute any file under `sources/`.
- Module-loader cases remain excluded only through the supplied exclusions file.
=== END ARTIFACT ===
=== BEGIN ARTIFACT FEATURE-Scoped-Conformance.md ===
# FEATURE: Scoped Conformance Verification

| Field       | Value |
|-------------|-------|
| Version     | 20260822 V1 |
| Description | Provide executing, machine-readable conformance gates for implementation slices. |
| Depends On  | ARCHITECTURE.md, FEATURE-Conformance-Assets.md, FEATURE-Streaming.md |
| Provides    | executing scoped conformance gates |
| Consumes    | ./jq -c program execution, staged conformance runner |

## Questions

- None.

## Workflow

Each implementation story uses the supplied runner with a selector matching its owned syntax.
The candidate command is supplied through the inherited `JQ` environment variable. The gate
executes selected cases, parses the JSON report, requires a non-empty selection, requires zero
failures and errors, and checks the runner exit status.

## Programmatic Acceptance

=== AC scoped-conformance ===
Intent: The scoped runner executes a non-empty reduce slice and reports no failures or errors.
Suite: scoped
Requires: executable=python3; scope=test

import json
import os
import subprocess
import sys

selector = r"reduce"
result = subprocess.run(
    [sys.executable, "sources/run_conformance.py", "--select", selector, "--json"],
    capture_output=True,
    text=True,
    env={**os.environ, "JQ": f"{os.getcwd()}/jq"},
)
print(result.stdout)
print(result.stderr, file=sys.stderr)
report = json.loads(result.stdout)
summary = report["summary"]
assert sum(summary.values()) > 0
assert summary["fail"] == 0 and summary["error"] == 0
assert result.returncode == 0
=== END AC scoped-conformance ===

=== AC scoped-environment ===
Intent: The scoped runner receives the candidate through JQ while retaining the inherited execution environment.
Suite: scoped
Requires: executable=python3; scope=test

import json
import os
import subprocess
import sys

selector = r"true"
environment = {**os.environ, "JQ": f"{os.getcwd()}/jq"}
result = subprocess.run(
    [sys.executable, "sources/run_conformance.py", "--select", selector, "--json"],
    capture_output=True,
    text=True,
    env=environment,
)
print(result.stdout)
print(result.stderr, file=sys.stderr)
report = json.loads(result.stdout)
assert report["summary"]["pass"] > 0
assert report["summary"]["fail"] == 0
assert report["summary"]["error"] == 0
assert result.returncode == 0
=== END AC scoped-environment ===

## User Acceptance

- None.

## Guardrails

- Scoped gates must execute cases; enumeration-only or dry-run modes are not acceptance.
- Every runner invocation must extend the inherited environment and set `JQ`.
- Scoped checks must use machine-readable state and must not assert on summary text.
=== END ARTIFACT ===
=== BEGIN ARTIFACT FEATURE-Complete-Conformance.md ===
# FEATURE: Complete Conformance Verification

| Field       | Value |
|-------------|-------|
| Version     | 20260822 V1 |
| Description | Verify the completed interpreter against the full supplied jq conformance corpus. |
| Depends On  | ARCHITECTURE.md, FEATURE-Scoped-Conformance.md |
| Provides    | complete conformance release verification |
| Consumes    | ./jq -c program execution, supplied scoring script |

## Questions

- None.

## Workflow

After every implementation and staging story has completed, run the supplied
`sources/full_test.sh` from the application root. This is the sole whole-corpus verification.
The story succeeds only when the script exits zero; captured output is printed for diagnosis and
is not itself used as an oracle.

## Programmatic Acceptance

=== AC complete-conformance ===
Intent: The completed executable passes the supplied full jq conformance suite.
Suite: full
Requires: executable=sh; scope=test
Requires: executable=python3; scope=test

import os
import subprocess

result = subprocess.run(
    ["sh", "sources/full_test.sh"],
    capture_output=True,
    text=True,
    env={**os.environ},
)
print(result.stdout)
print(result.stderr, file=__import__("sys").stderr)
assert result.returncode == 0
=== END AC complete-conformance ===

## User Acceptance

- None.

## Guardrails

- This is the only story that runs the whole corpus.
- Do not modify the supplied scoring script, runner, corpus, or exclusions.
- Release acceptance is determined by the scoring command's exit status.
=== END ARTIFACT ===