FAILURE MAP
← Case archive

FA-34921 / Notification interfaces / Open access

Notification visual impression qualification: covered-by-overlay from qualifying · case 01

The covered-by-overlay event leaves a notification in reporting instead of occluded.

Verified by executionVariant 1 · 42 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Covered notifications continue qualifying as visually exposed.

VERIFIED REPAIR

Commit the qualifying / covered-by-overlay transition to occluded; preserve the other explicitly stipulated transitions.

Unsuccessful approach: Counting them immediately makes occluded exposure permanent.

Case contract

Notification visual impression qualification is a bounded visual-notification workflow with mutable policy state {'exposure': 0, 'required': 4, 'covered': False, 'reports': 0}. Its default transition relation is {('offscreen', 'enters-viewport'): 'qualifying', ('qualifying', 'exposure-satisfied'): 'reporting', ('qualifying', 'leaves-viewport'): 'offscreen', ('reporting', 'report-success'): 'counted', ('reporting', 'report-failure'): 'report-error', ('report-error', 'retry-report'): 'reporting', ('qualifying', 'covered-by-overlay'): 'occluded', ('occluded', 'overlay-removed'): 'qualifying', ('counted', 'reenter-viewport'): 'counted', ('reporting', 'leaves-viewport'): 'reporting'}; domain inputs can suppress or redirect transitions and update policy fields, as specified in solve. Different-generation events are inert. Batches apply in order. Optional observe returns the selected policy field alongside the phase.

Why this case matters

Visual notification lifecycle ordering can leave an inbox or toast showing the wrong actionable state even when transport succeeds.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(initial, events, generation, payload=None, observe=None):
    table = {('offscreen', 'enters-viewport'): 'qualifying', ('qualifying', 'exposure-satisfied'): 'reporting', ('qualifying', 'leaves-viewport'): 'offscreen', ('reporting', 'report-success'): 'counted', ('reporting', 'report-failure'): 'report-error', ('report-error', 'retry-report'): 'reporting', ('qualifying', 'covered-by-overlay'): 'reporting', ('occluded', 'overlay-removed'): 'qualifying', ('counted', 'reenter-viewport'): 'counted', ('reporting', 'leaves-viewport'): 'reporting'}
    data = json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})) if payload is None else {**json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})), **json.loads(json.dumps(payload))}
    state = initial
    for delivery in events:
        event, event_generation = delivery[:2]
        arg = delivery[2] if len(delivery) > 2 else {}
        if event_generation == generation:
            previous = state
            if event == 'covered-by-overlay': data['covered'] = True
            if event == 'overlay-removed':
                data['covered'] = False
                data['exposure'] = 0
            if event == 'exposure-satisfied':
                if not data['covered']: data['exposure'] += max(0,arg.get('elapsed',data['required']))
                if data['exposure'] < data['required'] or data['covered']: event = 'unknown-event'
            if event == 'leaves-viewport' and state == 'qualifying': data['exposure'] = 0
            state = table.get((state, event), state)
            if state == 'counted' and previous != 'counted': data['reports'] += 1
    return {'phase': state, 'value': data[observe]} if observe else state
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('offscreen / enters-viewport', solve('offscreen', [('enters-viewport', N)], N), 'qualifying')
check('qualifying / exposure-satisfied', solve('qualifying', [('exposure-satisfied', N)], N), 'reporting')
check('qualifying / leaves-viewport', solve('qualifying', [('leaves-viewport', N)], N), 'offscreen')
check('reporting / report-success', solve('reporting', [('report-success', N)], N), 'counted')
check('reporting / report-failure', solve('reporting', [('report-failure', N)], N), 'report-error')
check('report-error / retry-report', solve('report-error', [('retry-report', N)], N), 'reporting')
check('qualifying / covered-by-overlay', solve('qualifying', [('covered-by-overlay', N)], N), 'occluded')
check('occluded / overlay-removed', solve('occluded', [('overlay-removed', N)], N), 'qualifying')
check('counted / reenter-viewport', solve('counted', [('reenter-viewport', N)], N), 'counted')
check('reporting / leaves-viewport', solve('reporting', [('leaves-viewport', N)], N), 'reporting')
check('older notification incarnation', solve('offscreen', [('enters-viewport', N - 1)], N), 'offscreen')
check('future notification incarnation', solve('offscreen', [('enters-viewport', N + 1)], N), 'offscreen')
check('unknown event is inert', solve('offscreen', [('unknown-event', N)], N), 'offscreen')
check('empty delivery batch', solve('offscreen', [], N), 'offscreen')
check('N stale deliveries before current delivery', solve('offscreen', [('enters-viewport', N - 1)] * N + [('enters-viewport', N)], N), 'qualifying')
check('trace enters-viewport then exposure-satisfied', solve('offscreen', [('enters-viewport', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace enters-viewport then leaves-viewport', solve('offscreen', [('enters-viewport', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace enters-viewport then covered-by-overlay', solve('offscreen', [('enters-viewport', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace exposure-satisfied then report-success', solve('qualifying', [('exposure-satisfied', N), ('report-success', N)], N), 'counted')
check('trace exposure-satisfied then report-failure', solve('qualifying', [('exposure-satisfied', N), ('report-failure', N)], N), 'report-error')
check('trace exposure-satisfied then leaves-viewport', solve('qualifying', [('exposure-satisfied', N), ('leaves-viewport', N)], N), 'reporting')
check('trace leaves-viewport then enters-viewport', solve('qualifying', [('leaves-viewport', N), ('enters-viewport', N)], N), 'qualifying')
check('trace report-success then reenter-viewport', solve('reporting', [('report-success', N), ('reenter-viewport', N)], N), 'counted')
check('trace report-failure then retry-report', solve('reporting', [('report-failure', N), ('retry-report', N)], N), 'reporting')
check('trace retry-report then report-success', solve('report-error', [('retry-report', N), ('report-success', N)], N), 'counted')
check('trace retry-report then report-failure', solve('report-error', [('retry-report', N), ('report-failure', N)], N), 'report-error')
check('trace retry-report then leaves-viewport', solve('report-error', [('retry-report', N), ('leaves-viewport', N)], N), 'reporting')
check('trace covered-by-overlay then overlay-removed', solve('qualifying', [('covered-by-overlay', N), ('overlay-removed', N)], N), 'qualifying')
check('trace overlay-removed then exposure-satisfied', solve('occluded', [('overlay-removed', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace overlay-removed then leaves-viewport', solve('occluded', [('overlay-removed', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace overlay-removed then covered-by-overlay', solve('occluded', [('overlay-removed', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace reenter-viewport then reenter-viewport', solve('counted', [('reenter-viewport', N), ('reenter-viewport', N)], N), 'counted')
check('trace leaves-viewport then report-success', solve('reporting', [('leaves-viewport', N), ('report-success', N)], N), 'counted')
check('trace leaves-viewport then report-failure', solve('reporting', [('leaves-viewport', N), ('report-failure', N)], N), 'report-error')
check('trace leaves-viewport then leaves-viewport', solve('reporting', [('leaves-viewport', N), ('leaves-viewport', N)], N), 'reporting')
check('domain state regression 1: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
check('domain state regression 2: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 4})], N, {}, 'exposure'), {'phase': 'reporting', 'value': 4})
check('domain state regression 3: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 9})], N, {'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 4: exposure', solve('occluded', [('overlay-removed', N, {})], N, {'exposure': 3, 'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 5: exposure', solve('qualifying', [('leaves-viewport', N, {})], N, {'exposure': 3}, 'exposure'), {'phase': 'offscreen', 'value': 0})
check('domain state regression 6: reports', solve('counted', [('reenter-viewport', N, {})], N, {'reports': 1}, 'reports'), {'phase': 'counted', 'value': 1})
check('domain state regression 7: exposure', solve('offscreen', [('enters-viewport', N, {}), ('exposure-satisfied', N, {'elapsed': 3}), ('covered-by-overlay', N, {}), ('overlay-removed', N, {}), ('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
offscreen / enters-viewportqualifyingqualifyingPassed
qualifying / exposure-satisfiedreportingreportingPassed
qualifying / leaves-viewportoffscreenoffscreenPassed
reporting / report-successcountedcountedPassed
reporting / report-failurereport-errorreport-errorPassed
report-error / retry-reportreportingreportingPassed
qualifying / covered-by-overlayreportingoccludedFailed
occluded / overlay-removedqualifyingqualifyingPassed
counted / reenter-viewportcountedcountedPassed
reporting / leaves-viewportreportingreportingPassed
older notification incarnationoffscreenoffscreenPassed
future notification incarnationoffscreenoffscreenPassed
unknown event is inertoffscreenoffscreenPassed
empty delivery batchoffscreenoffscreenPassed
N stale deliveries before current deliveryqualifyingqualifyingPassed
trace enters-viewport then exposure-satisfiedreportingreportingPassed
trace enters-viewport then leaves-viewportoffscreenoffscreenPassed
trace enters-viewport then covered-by-overlayreportingoccludedFailed
trace exposure-satisfied then report-successcountedcountedPassed
trace exposure-satisfied then report-failurereport-errorreport-errorPassed
trace exposure-satisfied then leaves-viewportreportingreportingPassed
trace leaves-viewport then enters-viewportqualifyingqualifyingPassed
trace report-success then reenter-viewportcountedcountedPassed
trace report-failure then retry-reportreportingreportingPassed
trace retry-report then report-successcountedcountedPassed
trace retry-report then report-failurereport-errorreport-errorPassed
trace retry-report then leaves-viewportreportingreportingPassed
trace covered-by-overlay then overlay-removedreportingqualifyingFailed
trace overlay-removed then exposure-satisfiedreportingreportingPassed
trace overlay-removed then leaves-viewportoffscreenoffscreenPassed
trace overlay-removed then covered-by-overlayreportingoccludedFailed
trace reenter-viewport then reenter-viewportcountedcountedPassed
trace leaves-viewport then report-successcountedcountedPassed
trace leaves-viewport then report-failurereport-errorreport-errorPassed
trace leaves-viewport then leaves-viewportreportingreportingPassed
domain state regression 1: exposure{'phase': 'qualifying', 'value': 3}{'phase': 'qualifying', 'value': 3}Passed
domain state regression 2: exposure{'phase': 'reporting', 'value': 4}{'phase': 'reporting', 'value': 4}Passed
domain state regression 3: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 4: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 5: exposure{'phase': 'offscreen', 'value': 0}{'phase': 'offscreen', 'value': 0}Passed
domain state regression 6: reports{'phase': 'counted', 'value': 1}{'phase': 'counted', 'value': 1}Passed
domain state regression 7: exposure{'phase': 'reporting', 'value': 3}{'phase': 'qualifying', 'value': 3}Failed

SHA-256 / 6b414bee12c961f0926bb3182b4ceb9cf2f5d8d3fa3d7213998e287237a72bbc

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(initial, events, generation, payload=None, observe=None):
    table = {('offscreen', 'enters-viewport'): 'qualifying', ('qualifying', 'exposure-satisfied'): 'reporting', ('qualifying', 'leaves-viewport'): 'offscreen', ('reporting', 'report-success'): 'counted', ('reporting', 'report-failure'): 'report-error', ('report-error', 'retry-report'): 'reporting', ('qualifying', 'covered-by-overlay'): 'counted', ('occluded', 'overlay-removed'): 'qualifying', ('counted', 'reenter-viewport'): 'counted', ('reporting', 'leaves-viewport'): 'reporting'}
    data = json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})) if payload is None else {**json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})), **json.loads(json.dumps(payload))}
    state = initial
    for delivery in events:
        event, event_generation = delivery[:2]
        arg = delivery[2] if len(delivery) > 2 else {}
        if event_generation == generation:
            previous = state
            if event == 'covered-by-overlay': data['covered'] = True
            if event == 'overlay-removed':
                data['covered'] = False
                data['exposure'] = 0
            if event == 'exposure-satisfied':
                if not data['covered']: data['exposure'] += max(0,arg.get('elapsed',data['required']))
                if data['exposure'] < data['required'] or data['covered']: event = 'unknown-event'
            if event == 'leaves-viewport' and state == 'qualifying': data['exposure'] = 0
            state = table.get((state, event), state)
            if state == 'counted' and previous != 'counted': data['reports'] += 1
    return {'phase': state, 'value': data[observe]} if observe else state
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('offscreen / enters-viewport', solve('offscreen', [('enters-viewport', N)], N), 'qualifying')
check('qualifying / exposure-satisfied', solve('qualifying', [('exposure-satisfied', N)], N), 'reporting')
check('qualifying / leaves-viewport', solve('qualifying', [('leaves-viewport', N)], N), 'offscreen')
check('reporting / report-success', solve('reporting', [('report-success', N)], N), 'counted')
check('reporting / report-failure', solve('reporting', [('report-failure', N)], N), 'report-error')
check('report-error / retry-report', solve('report-error', [('retry-report', N)], N), 'reporting')
check('qualifying / covered-by-overlay', solve('qualifying', [('covered-by-overlay', N)], N), 'occluded')
check('occluded / overlay-removed', solve('occluded', [('overlay-removed', N)], N), 'qualifying')
check('counted / reenter-viewport', solve('counted', [('reenter-viewport', N)], N), 'counted')
check('reporting / leaves-viewport', solve('reporting', [('leaves-viewport', N)], N), 'reporting')
check('older notification incarnation', solve('offscreen', [('enters-viewport', N - 1)], N), 'offscreen')
check('future notification incarnation', solve('offscreen', [('enters-viewport', N + 1)], N), 'offscreen')
check('unknown event is inert', solve('offscreen', [('unknown-event', N)], N), 'offscreen')
check('empty delivery batch', solve('offscreen', [], N), 'offscreen')
check('N stale deliveries before current delivery', solve('offscreen', [('enters-viewport', N - 1)] * N + [('enters-viewport', N)], N), 'qualifying')
check('trace enters-viewport then exposure-satisfied', solve('offscreen', [('enters-viewport', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace enters-viewport then leaves-viewport', solve('offscreen', [('enters-viewport', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace enters-viewport then covered-by-overlay', solve('offscreen', [('enters-viewport', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace exposure-satisfied then report-success', solve('qualifying', [('exposure-satisfied', N), ('report-success', N)], N), 'counted')
check('trace exposure-satisfied then report-failure', solve('qualifying', [('exposure-satisfied', N), ('report-failure', N)], N), 'report-error')
check('trace exposure-satisfied then leaves-viewport', solve('qualifying', [('exposure-satisfied', N), ('leaves-viewport', N)], N), 'reporting')
check('trace leaves-viewport then enters-viewport', solve('qualifying', [('leaves-viewport', N), ('enters-viewport', N)], N), 'qualifying')
check('trace report-success then reenter-viewport', solve('reporting', [('report-success', N), ('reenter-viewport', N)], N), 'counted')
check('trace report-failure then retry-report', solve('reporting', [('report-failure', N), ('retry-report', N)], N), 'reporting')
check('trace retry-report then report-success', solve('report-error', [('retry-report', N), ('report-success', N)], N), 'counted')
check('trace retry-report then report-failure', solve('report-error', [('retry-report', N), ('report-failure', N)], N), 'report-error')
check('trace retry-report then leaves-viewport', solve('report-error', [('retry-report', N), ('leaves-viewport', N)], N), 'reporting')
check('trace covered-by-overlay then overlay-removed', solve('qualifying', [('covered-by-overlay', N), ('overlay-removed', N)], N), 'qualifying')
check('trace overlay-removed then exposure-satisfied', solve('occluded', [('overlay-removed', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace overlay-removed then leaves-viewport', solve('occluded', [('overlay-removed', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace overlay-removed then covered-by-overlay', solve('occluded', [('overlay-removed', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace reenter-viewport then reenter-viewport', solve('counted', [('reenter-viewport', N), ('reenter-viewport', N)], N), 'counted')
check('trace leaves-viewport then report-success', solve('reporting', [('leaves-viewport', N), ('report-success', N)], N), 'counted')
check('trace leaves-viewport then report-failure', solve('reporting', [('leaves-viewport', N), ('report-failure', N)], N), 'report-error')
check('trace leaves-viewport then leaves-viewport', solve('reporting', [('leaves-viewport', N), ('leaves-viewport', N)], N), 'reporting')
check('domain state regression 1: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
check('domain state regression 2: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 4})], N, {}, 'exposure'), {'phase': 'reporting', 'value': 4})
check('domain state regression 3: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 9})], N, {'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 4: exposure', solve('occluded', [('overlay-removed', N, {})], N, {'exposure': 3, 'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 5: exposure', solve('qualifying', [('leaves-viewport', N, {})], N, {'exposure': 3}, 'exposure'), {'phase': 'offscreen', 'value': 0})
check('domain state regression 6: reports', solve('counted', [('reenter-viewport', N, {})], N, {'reports': 1}, 'reports'), {'phase': 'counted', 'value': 1})
check('domain state regression 7: exposure', solve('offscreen', [('enters-viewport', N, {}), ('exposure-satisfied', N, {'elapsed': 3}), ('covered-by-overlay', N, {}), ('overlay-removed', N, {}), ('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
offscreen / enters-viewportqualifyingqualifyingPassed
qualifying / exposure-satisfiedreportingreportingPassed
qualifying / leaves-viewportoffscreenoffscreenPassed
reporting / report-successcountedcountedPassed
reporting / report-failurereport-errorreport-errorPassed
report-error / retry-reportreportingreportingPassed
qualifying / covered-by-overlaycountedoccludedFailed
occluded / overlay-removedqualifyingqualifyingPassed
counted / reenter-viewportcountedcountedPassed
reporting / leaves-viewportreportingreportingPassed
older notification incarnationoffscreenoffscreenPassed
future notification incarnationoffscreenoffscreenPassed
unknown event is inertoffscreenoffscreenPassed
empty delivery batchoffscreenoffscreenPassed
N stale deliveries before current deliveryqualifyingqualifyingPassed
trace enters-viewport then exposure-satisfiedreportingreportingPassed
trace enters-viewport then leaves-viewportoffscreenoffscreenPassed
trace enters-viewport then covered-by-overlaycountedoccludedFailed
trace exposure-satisfied then report-successcountedcountedPassed
trace exposure-satisfied then report-failurereport-errorreport-errorPassed
trace exposure-satisfied then leaves-viewportreportingreportingPassed
trace leaves-viewport then enters-viewportqualifyingqualifyingPassed
trace report-success then reenter-viewportcountedcountedPassed
trace report-failure then retry-reportreportingreportingPassed
trace retry-report then report-successcountedcountedPassed
trace retry-report then report-failurereport-errorreport-errorPassed
trace retry-report then leaves-viewportreportingreportingPassed
trace covered-by-overlay then overlay-removedcountedqualifyingFailed
trace overlay-removed then exposure-satisfiedreportingreportingPassed
trace overlay-removed then leaves-viewportoffscreenoffscreenPassed
trace overlay-removed then covered-by-overlaycountedoccludedFailed
trace reenter-viewport then reenter-viewportcountedcountedPassed
trace leaves-viewport then report-successcountedcountedPassed
trace leaves-viewport then report-failurereport-errorreport-errorPassed
trace leaves-viewport then leaves-viewportreportingreportingPassed
domain state regression 1: exposure{'phase': 'qualifying', 'value': 3}{'phase': 'qualifying', 'value': 3}Passed
domain state regression 2: exposure{'phase': 'reporting', 'value': 4}{'phase': 'reporting', 'value': 4}Passed
domain state regression 3: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 4: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 5: exposure{'phase': 'offscreen', 'value': 0}{'phase': 'offscreen', 'value': 0}Passed
domain state regression 6: reports{'phase': 'counted', 'value': 1}{'phase': 'counted', 'value': 1}Passed
domain state regression 7: exposure{'phase': 'counted', 'value': 3}{'phase': 'qualifying', 'value': 3}Failed

SHA-256 / c1d5ce1a4ce7776a75c054b8200807421275551eba02a43e8298fcd1561ebbd5

3 / The verified repair

Exit 0
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(initial, events, generation, payload=None, observe=None):
    table = {('offscreen', 'enters-viewport'): 'qualifying', ('qualifying', 'exposure-satisfied'): 'reporting', ('qualifying', 'leaves-viewport'): 'offscreen', ('reporting', 'report-success'): 'counted', ('reporting', 'report-failure'): 'report-error', ('report-error', 'retry-report'): 'reporting', ('qualifying', 'covered-by-overlay'): 'occluded', ('occluded', 'overlay-removed'): 'qualifying', ('counted', 'reenter-viewport'): 'counted', ('reporting', 'leaves-viewport'): 'reporting'}
    data = json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})) if payload is None else {**json.loads(json.dumps({'exposure': 0, 'required': 4, 'covered': False, 'reports': 0})), **json.loads(json.dumps(payload))}
    state = initial
    for delivery in events:
        event, event_generation = delivery[:2]
        arg = delivery[2] if len(delivery) > 2 else {}
        if event_generation == generation:
            previous = state
            if event == 'covered-by-overlay': data['covered'] = True
            if event == 'overlay-removed':
                data['covered'] = False
                data['exposure'] = 0
            if event == 'exposure-satisfied':
                if not data['covered']: data['exposure'] += max(0,arg.get('elapsed',data['required']))
                if data['exposure'] < data['required'] or data['covered']: event = 'unknown-event'
            if event == 'leaves-viewport' and state == 'qualifying': data['exposure'] = 0
            state = table.get((state, event), state)
            if state == 'counted' and previous != 'counted': data['reports'] += 1
    return {'phase': state, 'value': data[observe]} if observe else state
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('offscreen / enters-viewport', solve('offscreen', [('enters-viewport', N)], N), 'qualifying')
check('qualifying / exposure-satisfied', solve('qualifying', [('exposure-satisfied', N)], N), 'reporting')
check('qualifying / leaves-viewport', solve('qualifying', [('leaves-viewport', N)], N), 'offscreen')
check('reporting / report-success', solve('reporting', [('report-success', N)], N), 'counted')
check('reporting / report-failure', solve('reporting', [('report-failure', N)], N), 'report-error')
check('report-error / retry-report', solve('report-error', [('retry-report', N)], N), 'reporting')
check('qualifying / covered-by-overlay', solve('qualifying', [('covered-by-overlay', N)], N), 'occluded')
check('occluded / overlay-removed', solve('occluded', [('overlay-removed', N)], N), 'qualifying')
check('counted / reenter-viewport', solve('counted', [('reenter-viewport', N)], N), 'counted')
check('reporting / leaves-viewport', solve('reporting', [('leaves-viewport', N)], N), 'reporting')
check('older notification incarnation', solve('offscreen', [('enters-viewport', N - 1)], N), 'offscreen')
check('future notification incarnation', solve('offscreen', [('enters-viewport', N + 1)], N), 'offscreen')
check('unknown event is inert', solve('offscreen', [('unknown-event', N)], N), 'offscreen')
check('empty delivery batch', solve('offscreen', [], N), 'offscreen')
check('N stale deliveries before current delivery', solve('offscreen', [('enters-viewport', N - 1)] * N + [('enters-viewport', N)], N), 'qualifying')
check('trace enters-viewport then exposure-satisfied', solve('offscreen', [('enters-viewport', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace enters-viewport then leaves-viewport', solve('offscreen', [('enters-viewport', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace enters-viewport then covered-by-overlay', solve('offscreen', [('enters-viewport', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace exposure-satisfied then report-success', solve('qualifying', [('exposure-satisfied', N), ('report-success', N)], N), 'counted')
check('trace exposure-satisfied then report-failure', solve('qualifying', [('exposure-satisfied', N), ('report-failure', N)], N), 'report-error')
check('trace exposure-satisfied then leaves-viewport', solve('qualifying', [('exposure-satisfied', N), ('leaves-viewport', N)], N), 'reporting')
check('trace leaves-viewport then enters-viewport', solve('qualifying', [('leaves-viewport', N), ('enters-viewport', N)], N), 'qualifying')
check('trace report-success then reenter-viewport', solve('reporting', [('report-success', N), ('reenter-viewport', N)], N), 'counted')
check('trace report-failure then retry-report', solve('reporting', [('report-failure', N), ('retry-report', N)], N), 'reporting')
check('trace retry-report then report-success', solve('report-error', [('retry-report', N), ('report-success', N)], N), 'counted')
check('trace retry-report then report-failure', solve('report-error', [('retry-report', N), ('report-failure', N)], N), 'report-error')
check('trace retry-report then leaves-viewport', solve('report-error', [('retry-report', N), ('leaves-viewport', N)], N), 'reporting')
check('trace covered-by-overlay then overlay-removed', solve('qualifying', [('covered-by-overlay', N), ('overlay-removed', N)], N), 'qualifying')
check('trace overlay-removed then exposure-satisfied', solve('occluded', [('overlay-removed', N), ('exposure-satisfied', N)], N), 'reporting')
check('trace overlay-removed then leaves-viewport', solve('occluded', [('overlay-removed', N), ('leaves-viewport', N)], N), 'offscreen')
check('trace overlay-removed then covered-by-overlay', solve('occluded', [('overlay-removed', N), ('covered-by-overlay', N)], N), 'occluded')
check('trace reenter-viewport then reenter-viewport', solve('counted', [('reenter-viewport', N), ('reenter-viewport', N)], N), 'counted')
check('trace leaves-viewport then report-success', solve('reporting', [('leaves-viewport', N), ('report-success', N)], N), 'counted')
check('trace leaves-viewport then report-failure', solve('reporting', [('leaves-viewport', N), ('report-failure', N)], N), 'report-error')
check('trace leaves-viewport then leaves-viewport', solve('reporting', [('leaves-viewport', N), ('leaves-viewport', N)], N), 'reporting')
check('domain state regression 1: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
check('domain state regression 2: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 4})], N, {}, 'exposure'), {'phase': 'reporting', 'value': 4})
check('domain state regression 3: exposure', solve('qualifying', [('exposure-satisfied', N, {'elapsed': 9})], N, {'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 4: exposure', solve('occluded', [('overlay-removed', N, {})], N, {'exposure': 3, 'covered': True}, 'exposure'), {'phase': 'qualifying', 'value': 0})
check('domain state regression 5: exposure', solve('qualifying', [('leaves-viewport', N, {})], N, {'exposure': 3}, 'exposure'), {'phase': 'offscreen', 'value': 0})
check('domain state regression 6: reports', solve('counted', [('reenter-viewport', N, {})], N, {'reports': 1}, 'reports'), {'phase': 'counted', 'value': 1})
check('domain state regression 7: exposure', solve('offscreen', [('enters-viewport', N, {}), ('exposure-satisfied', N, {'elapsed': 3}), ('covered-by-overlay', N, {}), ('overlay-removed', N, {}), ('exposure-satisfied', N, {'elapsed': 3})], N, {}, 'exposure'), {'phase': 'qualifying', 'value': 3})
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
offscreen / enters-viewportqualifyingqualifyingPassed
qualifying / exposure-satisfiedreportingreportingPassed
qualifying / leaves-viewportoffscreenoffscreenPassed
reporting / report-successcountedcountedPassed
reporting / report-failurereport-errorreport-errorPassed
report-error / retry-reportreportingreportingPassed
qualifying / covered-by-overlayoccludedoccludedPassed
occluded / overlay-removedqualifyingqualifyingPassed
counted / reenter-viewportcountedcountedPassed
reporting / leaves-viewportreportingreportingPassed
older notification incarnationoffscreenoffscreenPassed
future notification incarnationoffscreenoffscreenPassed
unknown event is inertoffscreenoffscreenPassed
empty delivery batchoffscreenoffscreenPassed
N stale deliveries before current deliveryqualifyingqualifyingPassed
trace enters-viewport then exposure-satisfiedreportingreportingPassed
trace enters-viewport then leaves-viewportoffscreenoffscreenPassed
trace enters-viewport then covered-by-overlayoccludedoccludedPassed
trace exposure-satisfied then report-successcountedcountedPassed
trace exposure-satisfied then report-failurereport-errorreport-errorPassed
trace exposure-satisfied then leaves-viewportreportingreportingPassed
trace leaves-viewport then enters-viewportqualifyingqualifyingPassed
trace report-success then reenter-viewportcountedcountedPassed
trace report-failure then retry-reportreportingreportingPassed
trace retry-report then report-successcountedcountedPassed
trace retry-report then report-failurereport-errorreport-errorPassed
trace retry-report then leaves-viewportreportingreportingPassed
trace covered-by-overlay then overlay-removedqualifyingqualifyingPassed
trace overlay-removed then exposure-satisfiedreportingreportingPassed
trace overlay-removed then leaves-viewportoffscreenoffscreenPassed
trace overlay-removed then covered-by-overlayoccludedoccludedPassed
trace reenter-viewport then reenter-viewportcountedcountedPassed
trace leaves-viewport then report-successcountedcountedPassed
trace leaves-viewport then report-failurereport-errorreport-errorPassed
trace leaves-viewport then leaves-viewportreportingreportingPassed
domain state regression 1: exposure{'phase': 'qualifying', 'value': 3}{'phase': 'qualifying', 'value': 3}Passed
domain state regression 2: exposure{'phase': 'reporting', 'value': 4}{'phase': 'reporting', 'value': 4}Passed
domain state regression 3: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 4: exposure{'phase': 'qualifying', 'value': 0}{'phase': 'qualifying', 'value': 0}Passed
domain state regression 5: exposure{'phase': 'offscreen', 'value': 0}{'phase': 'offscreen', 'value': 0}Passed
domain state regression 6: reports{'phase': 'counted', 'value': 1}{'phase': 'counted', 'value': 1}Passed
domain state regression 7: exposure{'phase': 'qualifying', 'value': 3}{'phase': 'qualifying', 'value': 3}Passed

SHA-256 / 196bac85febf38fb32abfc36b1d452a00fcff79f6637809b3c8262dd0f5d6002

Verification & scope

Stipulated single-notification state and policy model only. Event delivery and acknowledgements are explicit test inputs. No DOM, accessibility announcements, live timers, networking, actual rendering, or production-platform conformance is simulated. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-29T14:42:36.571952+00:00.

Case digest / 1d2eac266cca2c5e8dbefee5ca9532b8dd27c8bf25caac0037771100d3843e43