Squashes the working tree that had accumulated on cloud-agent. Two strands of work, kept in one commit at the maintainer's request. Crash and stability (this session): - Subscribe to MetricKit so crash, hang, CPU, and disk-write diagnostics are recorded on device. Keyboard-extension memory kills are the most likely way a user loses the keyboard mid-sentence and iOS does not classify them as crashes, so they never reach App Store Connect. - Upload those reports to POST /v1/analytics/diagnostics from internal builds only (local Debug and TestFlight), reusing the analytics installation identifier and skipping entirely when analytics are off. App Store builds keep everything on device. - Shed the Chinese engine, English lexicon, and in-flight pipeline work at the 40 MiB and 48 MiB budget thresholds instead of waiting for a system memory warning that jetsam often preempts. - Install audio taps with the input node's live format through one shared adaptive resampler, so an audio-route change mid-recording no longer raises an uncatchable format-mismatch exception. Covers the dictation preview and the Mac recorder, which still used the pattern the Flow capture path had already moved away from. - Harden the librime bridge: never pass a nil string across a nonnull boundary when a user dictionary holds malformed UTF-8, null-check the engine API vector, and distrust a candidate count without its array. - Guard the expanded candidate grid against a list that shrinks while the collection view is still laying out against the previous count. - Resolve localized text without the App Group so the screen shown when that container is missing renders its message instead of crashing on launch, which is the one situation it exists to handle. - Gate CI on the App Group and Keychain wiring that the keyboard extension hard-depends on, with a self-test proving the check fails. Clipboard, reply, and semantics (pre-existing working tree): - Unify the reply center, refresh keyboard AI features, and add native event and reminder exporters. - Expand clipboard semantic routing with authored corpora and a product-blind holdout, plus the scripts that build and validate them. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
153 lines
6.2 KiB
Swift
153 lines
6.2 KiB
Swift
// KeyboardExtensionMemoryBudgetTests.swift
|
|
// OSGKeyboardTests
|
|
|
|
@testable import OSGKeyboardShared
|
|
import XCTest
|
|
|
|
final class KeyboardExtensionMemoryBudgetTests: XCTestCase {
|
|
func testMemoryLevelsUseDocumentedBoundaries() {
|
|
XCTAssertEqual(
|
|
KeyboardExtensionMemoryBudget.level(forPhysFootprintMB: 35.9),
|
|
.normal
|
|
)
|
|
XCTAssertEqual(
|
|
KeyboardExtensionMemoryBudget.level(forPhysFootprintMB: 36),
|
|
.warning
|
|
)
|
|
XCTAssertEqual(
|
|
KeyboardExtensionMemoryBudget.level(forPhysFootprintMB: 40),
|
|
.high
|
|
)
|
|
XCTAssertEqual(
|
|
KeyboardExtensionMemoryBudget.level(forPhysFootprintMB: 48),
|
|
.critical
|
|
)
|
|
}
|
|
|
|
func testUnavailableFootprintIsNotMisclassifiedAsSafe() {
|
|
XCTAssertEqual(
|
|
KeyboardExtensionMemoryBudget.level(forPhysFootprintMB: -1),
|
|
.unavailable
|
|
)
|
|
}
|
|
}
|
|
|
|
/// Covers the *acting* half of the memory guard. The budget thresholds above
|
|
/// were always correct; what was missing was anything that used them. A
|
|
/// keyboard extension gets no reliable system memory warning before jetsam,
|
|
/// so these thresholds are the only chance to shed load before the keyboard
|
|
/// is killed mid-sentence.
|
|
@MainActor
|
|
final class KeyboardExtensionMemoryReliefTests: XCTestCase {
|
|
override func tearDown() {
|
|
KeyboardExtensionMemoryTelemetry.reliefHandler = nil
|
|
KeyboardExtensionMemoryTelemetry.stopSampling()
|
|
super.tearDown()
|
|
}
|
|
|
|
private func armedLevels(
|
|
accepting: Bool = true
|
|
) -> (record: () -> [KeyboardExtensionMemoryBudget.Level], reset: () -> Void) {
|
|
let box = LevelBox()
|
|
KeyboardExtensionMemoryTelemetry.begin(context: "test")
|
|
KeyboardExtensionMemoryTelemetry.stopSampling()
|
|
KeyboardExtensionMemoryTelemetry.reliefHandler = { level in
|
|
box.levels.append(level)
|
|
return accepting
|
|
}
|
|
return ({ box.levels }, { box.levels.removeAll() })
|
|
}
|
|
|
|
@MainActor
|
|
private final class LevelBox {
|
|
var levels: [KeyboardExtensionMemoryBudget.Level] = []
|
|
}
|
|
|
|
func testNormalAndWarningFootprintsDoNotShed() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(20)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(37)
|
|
XCTAssertEqual(levels(), [])
|
|
}
|
|
|
|
func testHighFootprintRequestsSoftRelief() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(41)
|
|
XCTAssertEqual(levels(), [.high])
|
|
}
|
|
|
|
func testSustainedPressureDoesNotReshedAtTheSameLevel() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
// Ten seconds of polling above the threshold. A timed cooldown would
|
|
// have shed three more times by now; hysteresis sheds once.
|
|
for _ in 0..<10 {
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
}
|
|
XCTAssertEqual(levels(), [.critical])
|
|
}
|
|
|
|
func testPressureMustRecedeBelowWarningBeforeSheddingAgain() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
// Landing back in the warning band is where a shed leaves the keyboard;
|
|
// treating that as recovery would re-arm on the very next sample.
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(37)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
XCTAssertEqual(levels(), [.critical])
|
|
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(22)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
XCTAssertEqual(levels(), [.critical, .critical])
|
|
}
|
|
|
|
func testDeclinedReliefStaysArmedForTheNextPoll() {
|
|
let (levels, _) = armedLevels(accepting: false)
|
|
// The host declines while the user is mid-composition. Recording it as
|
|
// a completed shed would leave the keyboard at 48 MiB, having released
|
|
// nothing, until pressure receded on its own.
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
XCTAssertEqual(levels(), [.critical, .critical, .critical])
|
|
}
|
|
|
|
func testRepeatedHighFootprintShedsOnce() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(41)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(42)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(43)
|
|
// One shed per level per episode: re-tearing down the typing surface on
|
|
// every 1 s poll would make the keyboard unusable, not safer.
|
|
XCTAssertEqual(levels(), [.high])
|
|
}
|
|
|
|
func testEscalationToCriticalShedsEvenAfterASoftShed() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(41)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
// The hard shed must not wait out a soft one — 49 MiB is ~11 MiB from
|
|
// the observed jetsam boundary.
|
|
XCTAssertEqual(levels(), [.high, .critical])
|
|
}
|
|
|
|
func testCriticalDoesNotDownshiftBackToHigh() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(41)
|
|
XCTAssertEqual(levels(), [.critical])
|
|
}
|
|
|
|
func testNewSessionCanShedAgain() {
|
|
let (levels, _) = armedLevels()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(49)
|
|
XCTAssertEqual(levels(), [.critical])
|
|
// `begin` is called once per extension process; a reused process must
|
|
// not inherit the previous presentation's armed state.
|
|
KeyboardExtensionMemoryTelemetry.begin(context: "test-2")
|
|
KeyboardExtensionMemoryTelemetry.stopSampling()
|
|
KeyboardExtensionMemoryTelemetry.simulateFootprintForTesting(41)
|
|
XCTAssertEqual(levels(), [.critical, .high])
|
|
}
|
|
}
|