Approved instrumentation rename; source review complete, full acceptance blocked by measured host capacity.
Production changes only helper/category names and matching consumers; interval names, boundaries, fixtures and budgets stay unchanged.
Core: 3824 passed, 55 opt-in skips; fast: 456 passed. Targeted native: 96 passed; targeted UI: 13 passed. Signed Mac app and embedded extension compile passed.
Full native: 1756 reported test bodies, one failed case/two assertions. Full iPad: 17 passed, one failed, one interrupted, 11 never started. Both owned simulators are shut down.
The structured table is partial: full Core plus log-derived iPad outcomes and the native failure. Full native leaf export is unavailable because interrupted finalisation left no Info.plist; the remaining 1755 native bodies are not fabricated as leaf records.
Needs acceptance follow-up
2 failing tests — the native duplicate-submission case and an iPad reading-status hittability case. Finalisation/iPad execution were interrupted under measured host memory/scheduler pressure. Full phone UI and bounded retries are unrun. Draft PR only; no green full-acceptance verdict.
72d0ff8094486b19bf8e779acae855c3fe60d3ae T-2115: record verification outcomes and host capacity blocker e44a4ff58e0c44e7646764b9ad0389127e118ac7 T-2115: document performance category migration 71ba9301abd1b09d31ea87abc548b7a819e9bf3a T-2115: name performance signposts by their current purpose Rename outdated milestone labels to presentation and teaching labels so measurements describe their current purpose. The measured operations stay the same.
PresentationPerformanceSignposts supplies Today and launch; TeachingPerformanceSignposts supplies composed teaching. The independent UI metric literal changes in the same commit. Host contracts check actual emitters and selectors.
OSLog measurement identity changes atomically across producer and consumer while interval lifetimes remain fixed. Negative category/wiring/whitespace mutations prove contract drift detection. Native/UI gates remain necessary; host contracts cannot establish physical sample delivery.
Packages/AsterismCore/Sources/AsterismCore/PresentationPerformanceSignposts.swift
Why it matters. Removes obsolete milestone identities without changing intervals.
What to look at. PresentationPerformanceSignposts and TeachingPerformanceSignposts
Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift
Why it matters. A mismatched category can yield no metric samples.
What to look at. Category literal and the three production consumers
Packages/AsterismCore/Tests/AsterismCoreTests/PerformanceSignpostContractTests.swift
Why it matters. Constant-only checks cannot catch a differently wired metric or emitted interval.
What to look at. Four host source-contract methods
Only namespace identity changes. Existing timing budgets and fixtures remain comparable.
The UI bundle stays independent of Core; host source contracts detect drift.
No device installation or performance measurement is authorised. The owner must confirm actual samples after review.
| Severity | Area | Finding | Resolution |
|---|---|---|---|
| minor | Host source normalizer | Removing whitespace inside quoted literals could hide category drift. | Preserve ordinary/multiline literal text, add a regression, and prove a spaced category mutation fails. Follow-up quality review clean. |
Source: local run at 2026-10-15T03:15:24.474540+00:00 · snapshot 72d0ff8094486b19bf8e779acae855c3fe60d3ae
Baseline: none
Execution: failed (partial results) · JUnit: 2 files · Coverage: none · Baseline: absent
Coverage scope: as the project configures it
Totals: 3841 passed · 2 failed · 56 skipped · 0 errored · 0 flaky
| Suite | Test | Job or artifact | Message |
|---|---|---|---|
| AsterismTests.EntryDetailModelTests | duplicateSubmissionSuppressed | — | Two assertions: updateEntryCallCount and landed edits were 2, expected 1. |
| AsterismUITests.WideLayoutAccessibilityUITests | testTheFiveReadingStatusSegmentsAreOperableInBothWideLayouts() | — | sidebar-library-reading invalid activation point/hittability, UIJourneySupport.swift:133 |
Derived by declaration name, from the diff (no baseline run).
Files that import a changed file on the left, changed files in the centre, files a changed file imports on the right. Snapshot working-tree against base 0e3c91bdba1e0e1017fa6cf2ef0df57dbbd274f6.
18 files: 10 code · 8 docs. Click to expand.
diff --git a/Asterism/Asterism/ViewModels/ComposedTeachingViewModel.swift b/Asterism/Asterism/ViewModels/ComposedTeachingViewModel.swiftindex 64581d37..7f22f83f 100644--- a/Asterism/Asterism/ViewModels/ComposedTeachingViewModel.swift+++ b/Asterism/Asterism/ViewModels/ComposedTeachingViewModel.swift@@ -28,8 +28,8 @@ import OSLog public final class ComposedTeachingViewModel { private static let logger = Logger(subsystem: "me.nore.ig.Asterism", category: "ComposedTeachingViewModel") private static let performanceSignposter = OSSignposter(- subsystem: M3PerformanceSignposts.subsystem,- category: M3PerformanceSignposts.category+ subsystem: TeachingPerformanceSignposts.subsystem,+ category: TeachingPerformanceSignposts.category ) // MARK: - Types
diff --git a/Asterism/Asterism/ViewModels/LaunchMeasurement.swift b/Asterism/Asterism/ViewModels/LaunchMeasurement.swiftindex d1255b34..bebe90a6 100644--- a/Asterism/Asterism/ViewModels/LaunchMeasurement.swift+++ b/Asterism/Asterism/ViewModels/LaunchMeasurement.swift@@ -107,14 +107,14 @@ nonisolated struct LaunchPhaseRecord: Equatable, Sendable { /// A launch being measured: the signpost intervals as they are emitted, and the /// record they are written into. ///-/// One `OSSignposter` on `M2PerformanceSignposts`' subsystem and category, so a+/// One `OSSignposter` on `PresentationPerformanceSignposts`' subsystem and category, so a /// launch sits in the same Instruments track as the `TodayPresentation` read it /// contains. Not gated on the configuration — `Personal` is the build whose /// launch time means anything (design, §Measurement). struct LaunchMeasurement { private static let signposter = OSSignposter(- subsystem: M2PerformanceSignposts.subsystem,- category: M2PerformanceSignposts.category)+ subsystem: PresentationPerformanceSignposts.subsystem,+ category: PresentationPerformanceSignposts.category) private static let logger = Logger(subsystem: "me.nore.ig.Asterism", category: "Launch") private let startedAt: ContinuousClock.Instant
diff --git a/Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift b/Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swiftindex 97222788..1d278b23 100644--- a/Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift+++ b/Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift@@ -17,7 +17,7 @@ import XCTest /// `waitForExistence` timeout on a screen that was never going to appear. final class M4ScaleTodayPerformanceUITests: XCTestCase { private static let subsystem = "me.nore.ig.Asterism"- private static let category = "M2Performance"+ private static let category = "PresentationPerformance" private static let todayPresentation = "TodayPresentation" /// How long the seeded launch is given to reach the home screen. ///
diff --git a/Packages/AsterismCore/Sources/AsterismCore/LibraryRepository+TodayPresentation.swift b/Packages/AsterismCore/Sources/AsterismCore/LibraryRepository+TodayPresentation.swiftindex b4a437c7..63da5dce 100644--- a/Packages/AsterismCore/Sources/AsterismCore/LibraryRepository+TodayPresentation.swift+++ b/Packages/AsterismCore/Sources/AsterismCore/LibraryRepository+TodayPresentation.swift@@ -7,8 +7,8 @@ private let todayPresentationLogger = Logger( category: "TodayPresentation" ) private let recentPerformanceSignposter = OSSignposter(- subsystem: M2PerformanceSignposts.subsystem,- category: M2PerformanceSignposts.category+ subsystem: PresentationPerformanceSignposts.subsystem,+ category: PresentationPerformanceSignposts.category ) extension LibraryRepository {
diff --git a/Packages/AsterismCore/Sources/AsterismCore/M2PerformanceSignposts.swift b/Packages/AsterismCore/Sources/AsterismCore/M2PerformanceSignposts.swiftdeleted file mode 100644index 3efbc8c2..00000000--- a/Packages/AsterismCore/Sources/AsterismCore/M2PerformanceSignposts.swift+++ /dev/null@@ -1,14 +0,0 @@-import Foundation--/// Stable names consumed by XCTest/Instruments physical-device performance hooks.-/// Keep these values source-compatible so historical measurements remain comparable.-public enum M2PerformanceSignposts {- public static let subsystem = "me.nore.ig.Asterism"- public static let category = "M2Performance"- /// Emitted by `LibraryRepository+TodayPresentation`. The M2 teaching- /// signposts that used to sit beside it (`TeachingEditAcknowledgement`,- /// `TeachingFinalPreviewPublication`) were removed with the M4 milestone's- /// teaching surface: nothing emitted them and only a dead device-only UI- /// test named them.- public static let todayPresentation = "TodayPresentation"-}
diff --git a/Packages/AsterismCore/Sources/AsterismCore/M3PerformanceSignposts.swift b/Packages/AsterismCore/Sources/AsterismCore/M3PerformanceSignposts.swiftdeleted file mode 100644index eb3b8fce..00000000--- a/Packages/AsterismCore/Sources/AsterismCore/M3PerformanceSignposts.swift+++ /dev/null@@ -1,17 +0,0 @@-import Foundation--/// Stable names consumed by XCTest/Instruments physical-device performance hooks.-/// Keep these values source-compatible so historical measurements remain comparable.-///-/// Only the subsystem and category survive. `ComposedTeachingViewModel` (which-/// absorbed the deleted `URLTeachingViewModel`) still logs under them, but the-/// `URLTeachingEditAcknowledgement` / `URLTeachingFinalPreviewPublication`-/// interval names went with the URL-teaching surface the M4 milestone removed:-/// nothing emitted them, and their only reader was a device-only UI suite whose-/// accessibility identifiers no longer exist. The composed budgets are measured-/// as `ComposedEditAcknowledgement` / `ComposedFinalPreviewPublication`-/// (Req 8.5).-public enum M3PerformanceSignposts {- public static let subsystem = "me.nore.ig.Asterism"- public static let category = "M3Performance"-}
diff --git a/Packages/AsterismCore/Sources/AsterismCore/PresentationPerformanceSignposts.swift b/Packages/AsterismCore/Sources/AsterismCore/PresentationPerformanceSignposts.swiftnew file mode 100644index 00000000..ea7a4ad0--- /dev/null+++ b/Packages/AsterismCore/Sources/AsterismCore/PresentationPerformanceSignposts.swift@@ -0,0 +1,10 @@+import Foundation++/// Names shared by presentation/launch emitters and physical performance hooks.+/// Keep interval names stable so historical measurements remain comparable.+public enum PresentationPerformanceSignposts {+ public static let subsystem = "me.nore.ig.Asterism"+ public static let category = "PresentationPerformance"+ /// Emitted by `LibraryRepository+TodayPresentation`.+ public static let todayPresentation = "TodayPresentation"+}
diff --git a/Packages/AsterismCore/Sources/AsterismCore/TeachingPerformanceSignposts.swift b/Packages/AsterismCore/Sources/AsterismCore/TeachingPerformanceSignposts.swiftnew file mode 100644index 00000000..4026eb82--- /dev/null+++ b/Packages/AsterismCore/Sources/AsterismCore/TeachingPerformanceSignposts.swift@@ -0,0 +1,9 @@+import Foundation++/// Namespace used by the composed teaching surface's performance intervals.+/// Interval names remain `ComposedEditAcknowledgement` and+/// `ComposedFinalPreviewPublication` (Req 8.5).+public enum TeachingPerformanceSignposts {+ public static let subsystem = "me.nore.ig.Asterism"+ public static let category = "TeachingPerformance"+}
diff --git a/Packages/AsterismCore/Tests/AsterismCoreTests/M4ScaleFixtureTests.swift b/Packages/AsterismCore/Tests/AsterismCoreTests/M4ScaleFixtureTests.swiftindex a460db97..6c575921 100644--- a/Packages/AsterismCore/Tests/AsterismCoreTests/M4ScaleFixtureTests.swift+++ b/Packages/AsterismCore/Tests/AsterismCoreTests/M4ScaleFixtureTests.swift@@ -521,8 +521,10 @@ struct M4ScaleFixtureTests { /// that drift fails cheaply. @Test("Performance signpost names stay compatible with the physical hooks") func signpostNames() {- #expect(M2PerformanceSignposts.subsystem == "me.nore.ig.Asterism")- #expect(M2PerformanceSignposts.category == "M2Performance")- #expect(M2PerformanceSignposts.todayPresentation == "TodayPresentation")+ #expect(PresentationPerformanceSignposts.subsystem == "me.nore.ig.Asterism")+ #expect(PresentationPerformanceSignposts.category == "PresentationPerformance")+ #expect(PresentationPerformanceSignposts.todayPresentation == "TodayPresentation")+ #expect(TeachingPerformanceSignposts.subsystem == "me.nore.ig.Asterism")+ #expect(TeachingPerformanceSignposts.category == "TeachingPerformance") } }
diff --git a/Packages/AsterismCore/Tests/AsterismCoreTests/PerformanceSignpostContractTests.swift b/Packages/AsterismCore/Tests/AsterismCoreTests/PerformanceSignpostContractTests.swiftnew file mode 100644index 00000000..9a86acfa--- /dev/null+++ b/Packages/AsterismCore/Tests/AsterismCoreTests/PerformanceSignpostContractTests.swift@@ -0,0 +1,92 @@+import Foundation+import Testing+import AsterismCore++/// Host-only source contracts: the UI bundle deliberately keeps its metric+/// literals independent of Core. These reads require the checkout, not an+/// installed app or simulator. Runtime interval behavior stays with native/UI tests.+@Suite("Performance emitters and physical measurement select the same intervals")+struct PerformanceSignpostContractTests {+ private static let repositoryRoot: URL = {+ var url = URL(fileURLWithPath: #filePath)+ for _ in 0..<5 { url.deleteLastPathComponent() }+ return url+ }()++ @Test("Today emits the presentation namespace and matching begin/end interval names")+ func todayEmitterMatchesContract() throws {+ let source = try Self.source("Packages/AsterismCore/Sources/AsterismCore/LibraryRepository+TodayPresentation.swift")+ #expect(source.contains("recentPerformanceSignposter=OSSignposter(subsystem:PresentationPerformanceSignposts.subsystem,category:PresentationPerformanceSignposts.category)"))+ #expect(try Self.captures(#"recentPerformanceSignposter\.beginInterval\("([^"]+)""#, in: source)+ == [PresentationPerformanceSignposts.todayPresentation])+ #expect(try Self.captures(#"recentPerformanceSignposter\.endInterval\("([^"]+)""#, in: source)+ == [PresentationPerformanceSignposts.todayPresentation])+ }++ @Test("The physical metric uses the presentation literals as subsystem, category and name")+ func physicalMetricMatchesContract() throws {+ let source = try Self.source("Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift")+ #expect(try Self.captures(#"privatestaticletsubsystem="([^"]+)""#, in: source)+ == [PresentationPerformanceSignposts.subsystem])+ #expect(try Self.captures(#"privatestaticletcategory="([^"]+)""#, in: source)+ == [PresentationPerformanceSignposts.category])+ #expect(try Self.captures(#"privatestaticlettodayPresentation="([^"]+)""#, in: source)+ == [PresentationPerformanceSignposts.todayPresentation])+ #expect(try Self.captures(#"letmetric=XCTOSSignpostMetric\(([^)]+)\)"#, in: source)+ == ["subsystem:Self.subsystem,category:Self.category,name:Self.todayPresentation"])+ }++ @Test("Launch and composed teaching select their intended namespaces")+ func appEmittersMatchContract() throws {+ let launch = try Self.source("Asterism/Asterism/ViewModels/LaunchMeasurement.swift")+ #expect(launch.contains("signposter=OSSignposter(subsystem:PresentationPerformanceSignposts.subsystem,category:PresentationPerformanceSignposts.category)"))+ let teaching = try Self.source("Asterism/Asterism/ViewModels/ComposedTeachingViewModel.swift")+ #expect(teaching.contains("performanceSignposter=OSSignposter(subsystem:TeachingPerformanceSignposts.subsystem,category:TeachingPerformanceSignposts.category)"))+ }++ @Test("Source contracts preserve literal whitespace and ignore comments")+ func sourceNormalizationPreservesLiteralWhitespace() throws {+ let source = #"""+ // private static let category = "PresentationPerformance"+ /* private static let category = "PresentationPerformance" */+ private static let category = "Presentation Performance" // drift+ """#+ let compact = try Self.withoutCommentsOrLayout(source)+ #expect(compact == #"privatestaticletcategory="Presentation Performance""#)+ #expect(try Self.captures(#"privatestaticletcategory="([^"]+)""#, in: compact)+ == ["Presentation Performance"])+ let withMultiline = "let message = \"\"\"\nmessage with spaces\n\"\"\"\nlet metric = XCTOSSignpostMetric(category: Self.category)"+ let compactMultiline = try Self.withoutCommentsOrLayout(withMultiline)+ #expect(compactMultiline.contains("\"\"\"\nmessage with spaces\n\"\"\""))+ #expect(compactMultiline.contains("letmetric=XCTOSSignpostMetric(category:Self.category)"))+ }++ /// Ignore comments and layout outside quoted literals. The literal contents+ /// are the contract: even whitespace changes which signposts are selected.+ /// All inspected calls use ordinary string literals or member references;+ /// this is deliberately not a Swift parser.+ private static func source(_ relativePath: String) throws -> String {+ let source = try String(contentsOf: repositoryRoot.appending(path: relativePath), encoding: .utf8)+ return try withoutCommentsOrLayout(source)+ }++ private static func withoutCommentsOrLayout(_ source: String) throws -> String {+ let expression = try NSRegularExpression(+ pattern: #""""[\s\S]*?"""|"(?:\\.|[^"\\])*"|//[^\n]*|/\*[\s\S]*?\*/|\s+"#)+ let compact = NSMutableString(string: source)+ for match in expression.matches(in: source, range: NSRange(source.startIndex..., in: source)).reversed() {+ // Matching strings first keeps whitespace and comment markers inside+ // a quoted literal intact. Reverse replacement preserves match ranges.+ if !(source as NSString).substring(with: match.range).hasPrefix("\"") {+ compact.replaceCharacters(in: match.range, with: "")+ }+ }+ return compact as String+ }++ private static func captures(_ pattern: String, in source: String) throws -> [String] {+ let expression = try NSRegularExpression(pattern: pattern)+ return expression.matches(in: source, range: NSRange(source.startIndex..., in: source))+ .compactMap { Range($0.range(at: 1), in: source).map { String(source[$0]) } }+ }+}
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex f7b1605e..1ad790a1 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -8,6 +8,12 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/). ### Changed +- **Performance instrumentation (T-2115):** rename the presentation and launch+ signpost category from `M2Performance` to `PresentationPerformance`, and+ composed teaching from `M3Performance` to `TeachingPerformance`. Update saved+ Instruments filters to the new categories; intervals and timing budgets stay+ unchanged.+ - **Manual note character references:** show the work’s ranked cast in new manual notes using the existing share-sheet chips, alias/note cards and expansion. Keep references read-only, preserve authored drafts and URL-free rating-only
diff --git a/specs/OVERVIEW.md b/specs/OVERVIEW.mdindex e944735b..61023c18 100644--- a/specs/OVERVIEW.md+++ b/specs/OVERVIEW.md@@ -2,6 +2,7 @@ | Name | Creation Date | Status | Summary | |------|---------------|--------|---------|+| [Performance Signpost Names](#performance-signpost-names) | 2026-10-15 | In Progress — source implemented; acceptance blocked by host memory/scheduler pressure | T-2115. Replace milestone helper/category names with presentation and teaching names, preserve interval behavior, and guard emitter/metric agreement on the host. Physical sample verification remains owner-run. | | [Cover Image Drop](#cover-image-drop) | 2026-10-11 | Draft Review | T-2433. Drop local cover images through existing crop/draft/Save; no network or schema change. | | [Redundant Chapter Confirmation](#redundant-chapter-confirmation) | 2026-10-11 | Draft Review | T-2436. Keep ordinary share capture automatic and reveal optional chapter alternatives on demand. | | [Export Selected Day's Notes](#export-selected-days-notes) | 2026-10-06 | Complete — scoped checks passed; full phone UI exception recorded; manual Mac check unrun | T2404. Export notes and ratings originally captured on the day selected in Today, with frozen displayed context, existing Markdown blocks and native sharing. |@@ -1038,3 +1039,15 @@ T-2439. All fourteen Core/App implementation tasks are complete and source-revie - [design-review.md](copy-work-records/design-review.md) - [ticket-draft.json](copy-work-records/ticket-draft.json) - [implementation.md](copy-work-records/implementation.md)++## Performance Signpost Names++T-2115 uses descriptive presentation and teaching instrumentation namespaces in place of milestone names.++- [Smolspec](performance-signpost-names/smolspec.md)+- [Tasks](performance-signpost-names/tasks.md)+- [Decision log](performance-signpost-names/decision_log.md)+- [Specification review](performance-signpost-names/review.md)++- [Implementation](performance-signpost-names/implementation.md)+- [Verification and blocked gates](performance-signpost-names/verification-run.md)
diff --git a/specs/performance-signpost-names/decision_log.md b/specs/performance-signpost-names/decision_log.mdnew file mode 100644index 00000000..2b88baf8--- /dev/null+++ b/specs/performance-signpost-names/decision_log.md@@ -0,0 +1,11 @@+# Decision log++## Quick Decisions+| ID | Date | Decision | Rationale |+| --- | --- | --- | --- |+| Q1 | 2026-10-14 | Name the helpers/categories for presentation and teaching. | Current main routes Today and launch through the M2 helper and composed teaching through M3; milestone numbering no longer describes either role. |+| Q2 | 2026-10-14 | Change both helper contracts atomically and retain metric interval names. | T-2115 explicitly asks to consider M3; both changes are mechanical and preserve the existing independent namespaces. |+| Q3 | 2026-10-14 | Leave physical performance confirmation to the owner. | The user forbids phone installation without a further request; AGENTS.md also requires immediate explicit consent for Personal device runs. |+| Q4 | 2026-10-14 | Check the actual emitter intervals and metric argument wiring in a host-only source contract. | Both independent reviews observed that the Today interval helper constant is not consumed by production; pinning that constant alone cannot detect emission drift. |+| Q5 | 2026-10-15 | Reviewed spec and task creation/execution approved. | User said “Both approved. Create the tasks and execute” in Sentinel_61971502333081919f896dbc09a5760d after phone delivery; this authorizes implementation and task execution without another gate. |+| Q6 | 2026-10-15 | Preserve quoted whitespace in source-contract normalization. | Removing whitespace inside strings can hide a metric-category mismatch; the host contract now ignores comments/layout only outside ordinary quoted literals, with a regression and mutation check. |
diff --git a/specs/performance-signpost-names/implementation.md b/specs/performance-signpost-names/implementation.mdnew file mode 100644index 00000000..4de4c026--- /dev/null+++ b/specs/performance-signpost-names/implementation.md@@ -0,0 +1,148 @@+# T-2115 implementation explanation++This assessment compares the implemented source changes with the approved+[smolspec](smolspec.md). The source changes implement the naming and contract+requirements. Acceptance verification is still incomplete; the evidence below+does not establish physical-device sample resolution or a performance result.++## Beginner Level++### What Changed++The app marks the start and end of operations so performance tools can time+them. These markers are called signposts. Their categories previously referred+to old project milestones; they now describe the work being measured:++- `M2Performance` becomes `PresentationPerformance` for Today presentation and launch.+- `M3Performance` becomes `TeachingPerformance` for composed teaching.++### Why It Matters++The names now tell a developer what a measurement covers without requiring+knowledge of milestone history. The physical measurement test has also been+updated, so its category agrees with the app's category. Saved Instruments+filters need the same update, as recorded in the changelog.++### Key Concepts++A category groups related measurements, like a folder. An interval marks a+particular operation's beginning and end. This change renames the folders;+the existing intervals, work performed, fixtures and timing budgets stay intact.+Host contract tests check that the app and measurement test keep using matching+names.++## Intermediate Level++### Changes Overview++`PresentationPerformanceSignposts.swift` replaces the M2 helper and supplies+`LibraryRepository+TodayPresentation.swift` and `LaunchMeasurement.swift`.+`TeachingPerformanceSignposts.swift` replaces the M3 helper and supplies+`ComposedTeachingViewModel.swift`. Both retain subsystem+`me.nore.ig.Asterism`.++`M4ScaleTodayPerformanceUITests.swift` retains its independent literals and now+selects `PresentationPerformance`. `M4ScaleFixtureTests.signpostNames()` pins+both helpers' subsystem/category values and the `TodayPresentation` interval.+The new `PerformanceSignpostContractTests.swift` checks the actual production+source and physical metric argument wiring.++### Implementation Approach++The source contracts read checkout-relative paths using `#filePath`, remove+comments and layout whitespace outside quoted literals, and inspect the resulting source with string+checks and regular expressions. They verify Today uses the presentation helper+and matching begin/end interval literals, launch and teaching select their+intended helpers, and the physical metric uses matching independent literals+through `subsystem: Self.subsystem`, `category: Self.category`, and+`name: Self.todayPresentation`.++The production diff changes helper references and categories without moving+interval calls. `TodayPresentation`, launch's parent/phase intervals, and+`ComposedEditAcknowledgement` / `ComposedFinalPreviewPublication` retain their+names and boundaries. Existing assertions are retained and extended.++### Trade-offs++Keeping the UI bundle independent of Core preserves the existing module+structure, but duplicates three selector values. The host source contracts+make that duplication testable cheaply. These contracts deliberately depend on+the current source shape and require a checkout; they do not replace native+integration, rendered UI checks or a physical XCTest signpost measurement.++## Expert Level++### Technical Deep Dive++The changed category is part of the OSLog/XCTest measurement identity. A+correct interval name under the wrong category can leave the consumer without+samples. Updating production and the independent consumer together addresses+that failure, while the contracts also guard argument miswiring that a simple+constant assertion would miss.++Today's existing `defer` closes its interval on success or throw. Its start+still precedes the shared-lock operation. Launch and composed teaching only+change their signposter helper references. No persistence, synchronization,+workload, timing-budget, runtime gating or interval-lifecycle change appears+in the reviewed diff.++### Architecture Impact++The two-helper design remains intact, with no new dependency. The helpers are+Swift-public for app/Core module access; the approved spec records that no+supported external package consumers are declared. Historical specs retain+their original milestone terminology, while current helper comments and the+changelog describe the new contract.++### Potential Issues++The source checks are intentionally not a Swift parser. They tolerate layout+and line-comment changes, but semantic refactoring of declarations or calls+may require updating the contracts. They do not prove sample delivery,+interval timing, or runtime behavior. External Instruments filters must move+to the new categories. Physical sample confirmation remains an owner-run check+under the device-consent policy.++## Completeness Assessment++### Implemented Requirements++- Presentation and launch use `PresentationPerformance`; teaching uses+ `TeachingPerformance`, all under the unchanged subsystem.+- The physical Today metric selects the same presentation namespace and interval.+- Existing interval names, boundaries, fixtures, budgets and assertions are+ preserved in the reviewed source diff.+- Host contracts cover both categories, the Today begin/end literals, and+ physical metric selector values and argument wiring.+- The changelog records both old-to-new category mappings.++### Verification Evidence++The supplied evidence directory is `../evidence/` relative to the repository+root (outside this checkout). Its `t2115-focused-core.log` records 18 tests in+two suites passing: `M4ScaleFixtureTests` and+`PerformanceSignpostContractTests`.++`t2115-mutation-results.json` records two intentional mutation runs, each+failing as expected with exit 2: a category mismatch and incorrect metric+argument wiring. It records source restoration and exit 0 afterwards.+`t2115-restored-core.log` confirms the restored run passed all 18 tests in the+same two suites. These failures demonstrate drift detection; they are not+outstanding production failures.++### Remaining Verification++Unfiltered `make test-core` reported 3,824 passed and 55 existing opt-in skips, with no failures/errors; `make test-fast` verified and passed 456 declared tests. The corrected four source contracts pass and a spaced category mutation fails as intended, followed by a restored passing run. The selected native launch/teaching suites passed all 96 declarations; focused rendered teaching and both scale-seeding UI checks passed all 13 expected cases with no skips or missing identifiers. The signed Mac app/embedded-extension compile passed. The full native run reported 1,756 test bodies with one failing duplicate-submission case (two assertions), then finalisation stalled and was interrupted. The full iPad attempt reported 17 passed, one failed, one interrupted and 11 never started out of 30. Full phone UI and targeted retries are unrun because measured host memory/scheduler pressure blocked more work. Both owned jobs were drained and both dedicated simulators shut down. Partial Xcode bundles are retained but cannot export without a final Info.plist. No green full-acceptance verdict is claimed; Rune task 2 remains incomplete. See verification-run.md for exact failures and capacity evidence.++Physical measurement and installation are outside implementation scope.+After code review, the owner must authorize and run+`make test-performance-m4-recent` and confirm actual signpost samples resolve.+Host and simulator success cannot close that check.++### Assessment++The implementation addresses the approved source requirements and the earlier+validation gaps around real emitter literals and metric wiring. No new+architectural or compatibility decision requiring escalation is evident.+Source implementation is complete. Acceptance remains blocked by the failed/incomplete/unrun checks recorded above and in verification-run.md; physical confirmation remains a separate owner-run step. This explanation review performed no builds,+device actions or production source edits.
diff --git a/specs/performance-signpost-names/review.md b/specs/performance-signpost-names/review.mdnew file mode 100644index 00000000..dc8aaf66--- /dev/null+++ b/specs/performance-signpost-names/review.md@@ -0,0 +1,26 @@+# T-2115 specification review++Base: origin/main `0e3c91bdba1e0e1017fa6cf2ef0df57dbbd274f6`.+Branch: `T-2115/performance-signpost-names`.++- Read complete live Transit description, comments (none), links (none), and linked PR14 body/comments/reviews. Asterism project UUID: `2403D561-9060-48CC-9AFD-90A35E2B46D6`; ticket UUID: `007C3FCE-7603-43E7-B182-690E2410C2E6`.+- Current main retains the issue. Historical Recent implementation is now Today; current consumers and filenames are enumerated in the smolspec.+- Explanation validation ran from a clean context at beginner, intermediate, and expert levels. Design critic independently reviewed against source. Neither found a blocker or mandatory full-spec escalation.+- All findings were classified as document clarifications: exact emitter/metric contract assertions, Swift-public visibility versus external compatibility, and named native verification suites. Applied in the smolspec. Actual external API users discovered during implementation would require escalation.+- No overlap with App Intent/export files or PR110 code is planned. Only potential shared-file coordination is the existing Today repository file; the requested change there is its signpost helper references.+- No matching generated knowledge note for signposts/performance measurement was found.+- Approval was received after phone delivery: user “Both approved. Create the tasks and execute” (Sentinel_61971502333081919f896dbc09a5760d). Rune tasks were created and sent by the established phone route; implementation is authorised.+- Source changes and four host source contracts are implemented (including quoted-literal normalization regression). Focused Core passed 18 tests in two suites, including the restored run. Category-literal and metric-argument mutations each failed the intended contract; source restoration was verified.+- Four independent pre-push source reviews (reuse, quality/correctness, efficiency, spec/documentation with phase design critique) found no actionable issues.+- Core: 3,824 passed, 55 opt-in skips, zero failures/errors. Fast: 456 passed, no skips. Targeted native: 96 passed. Targeted UI: all 13 expected cases passed. Signed Mac app/embedded extension compile passed.+- Full native: one failed case/two assertions among 1,756 reported test bodies, followed by infrastructure finalisation interruption. Full iPad: 17 passed, one failed, one interrupted, 11 never started. Full phone and bounded retries are unrun due measured host pressure. Both owned simulators/jobs are drained; no green acceptance claim.+- Documentation whitespace check passed. Source and documentation commits are prepared for a draft PR with these limits; physical measurements remain owner-run.+- Production changes only instrumentation type/category names and matching consumers; no live library data has changed. Physical Personal measurement requires owner approval and remains outside execution authorization.++## Approval++The reviewed `performance-signpost-names` spec and implementation execution are approved as recorded in decision Q5. No merge or device installation is authorised. Final runtime results are recorded in verification-run.md.++## Review correction++The parent found that the initial source-contract normalizer erased whitespace inside string literals. It was corrected to preserve quoted contents and recognise existing multiline diagnostics; a regression and spaced-category mutation verify it. Follow-up quality review found no remaining issue. Four independent source-review roles found no actionable production-code findings.
diff --git a/specs/performance-signpost-names/smolspec.md b/specs/performance-signpost-names/smolspec.mdnew file mode 100644index 00000000..4a2c455d--- /dev/null+++ b/specs/performance-signpost-names/smolspec.md@@ -0,0 +1,34 @@+# Performance signpost names++## Overview+T-2115 removes milestone names from live performance instrumentation. Presentation and composed-teaching instrumentation retain their existing intervals and timing, with matching emitter and measurement category names. The ticket's Recent references predate the current Today presentation implementation.++## Requirements+- Production presentation and launch instrumentation MUST use the milestone-independent category `PresentationPerformance` under subsystem `me.nore.ig.Asterism`.+- Composed teaching instrumentation MUST use category `TeachingPerformance` under that same subsystem.+- The physical Today measurement MUST select the presentation category emitted by production.+- Existing interval names, interval boundaries, fixtures, timing budgets, and test assertions MUST remain unchanged.+- Host contract tests MUST pin both instrumentation categories and the `TodayPresentation` interval, verify production begin/end interval literals, and verify that the physical metric uses matching subsystem/category/name selectors.+- Verification MUST distinguish host/simulator checks from physical-device measurement; device measurement remains pending owner execution.++## Implementation Approach+- Rename `Packages/AsterismCore/Sources/AsterismCore/M2PerformanceSignposts.swift` and its type to `PresentationPerformanceSignposts`, with category `PresentationPerformance`.+- Rename `Packages/AsterismCore/Sources/AsterismCore/M3PerformanceSignposts.swift` and its type to `TeachingPerformanceSignposts`, with category `TeachingPerformance`.+- Update the existing consumers in `LibraryRepository+TodayPresentation.swift`, `Asterism/Asterism/ViewModels/LaunchMeasurement.swift`, and `ComposedTeachingViewModel.swift`.+- Update the independent category literal in `Asterism/AsterismUITests/M4ScaleTodayPerformanceUITests.swift` in the same code commit. The UI bundle does not import Core; retain this existing independence.+- Extend `Packages/AsterismCore/Tests/AsterismCoreTests/M4ScaleFixtureTests.swift`'s `signpostNames` contract and add a host-only `PerformanceSignpostContractTests.swift` beside it, using `PlatformSeamTests.swift`'s repository-root/source-reading pattern to check production begin/end calls and the actual `XCTOSSignpostMetric` arguments. A mismatch of either consumer literals or metric argument wiring MUST fail the contract test; demonstrate each with a temporary mutation and restore it.+- Update current explanatory comments; retain historical specs as historical evidence. Record the category transition so external Instruments filters can be updated by the owner.+- Dependencies: Foundation, OSLog, XCTest's signpost metric, and the existing Makefile targets. No new package or runtime dependencies.+- Out of scope: schema/data changes, UI changes, performance optimization, merging, device installation or physical measurement, and PR110.++## Risks and Assumptions+- Risk: a mismatched category silently yields no device metric samples. Mitigation: emitter/consumer changes land atomically, host tests check both sides, and the owner's physical measurement remains an explicit outstanding check.+- Risk: saved Instruments filters using the old categories stop matching. Mitigation: document old-to-new category names; no external consumer is declared in the repository beyond the in-tree UI suite.+- Assumption: these enums are Swift-public for app/Core module access, but the package is bundled only inside this repository, with no supported external API consumers declared. The category/type contract is repository-owned and reversible: all consumers change together and one revert restores it.+- Scope assessment: no user-owned ambiguity, persisted/external compatibility boundary, or contested architecture; the change is incremental and reversible in one code commit using the existing two-helper pattern.+- Targeted checks: Core `M4ScaleFixtureTests` plus the host source contract, `make test-fast`, native `AppLibraryModelLaunchTests`, `LaunchPhaseRecordTests`, `ComposedTeachingViewModelTests`, and `ComposedTeachingPreviewTests`; UI `ComposedSurfaceUITests` plus `M4ScaleTodayPerformanceUITests/testSeededScaleM4ScenarioReachesRecent` and `testSeededScaleM4DuplicateSiteRowsScenarioReachesRecent`. Physical metric methods are intentionally skipped in simulator execution.+- Final gates: unfiltered `make test-core`, `make test-quick` (including signed Mac compile), `make test-ui`, and `make test-ui-ipad`, with isolated caches/simulator and shared build coordination. Record failures and executed test counts; do not claim performance verification from these gates.+- Owner prerequisite after code review: run `make test-performance-m4-recent` with explicit device-run consent and confirm signpost samples resolve.++## Escalation Note+This change was scoped as a smolspec. If implementation reveals ambiguity only the user can resolve, an irreversible boundary (public API, persisted schema, auth path), or a contested architectural choice, stop and escalate to the full spec workflow rather than deciding it inline.
diff --git a/specs/performance-signpost-names/tasks.md b/specs/performance-signpost-names/tasks.mdnew file mode 100644index 00000000..cbeb0e6a--- /dev/null+++ b/specs/performance-signpost-names/tasks.md@@ -0,0 +1,18 @@+---+references:+ - specs/performance-signpost-names/smolspec.md+ - specs/performance-signpost-names/decision_log.md+---+# Performance signpost names++## Implementation and verification++- [x] 1. Presentation and teaching signposts use descriptive names with matching measurement contracts <!-- id:vwwi25e -->+ - Implement the approved smolspec atomically; preserve interval boundaries and behavior.+ - Extend host contracts, prove category and metric wiring mutations fail, restore them, and pass focused Core/native/UI checks.+ - Reference: specs/performance-signpost-names/smolspec.md++- [-] 2. Renamed instrumentation passes final acceptance gates and independent review <!-- id:vwwi25f -->+ - Core and fast passed; targeted native96 and UI13 passed; signedMac app/embeddedextension compile passed. Fullnative1756 bodies had1 unrelatedfailedcase/twoassertions+ - followed by owned finalization interruption. FulliPad partial interrupted under measured host memory/scheduler overload; fullphone unrun. Remaining acceptance and targeted retries blocked until host capacity recovers; no green fullgate claim. Reference: specs/performance-signpost-names/smolspec.md+ - Blocked-by: vwwi25e (Presentation and teaching signposts use descriptive names with matching measurement contracts)
diff --git a/specs/performance-signpost-names/verification-run.md b/specs/performance-signpost-names/verification-run.mdnew file mode 100644index 00000000..f0214806--- /dev/null+++ b/specs/performance-signpost-names/verification-run.md@@ -0,0 +1,48 @@+# T-2115 verification++## Change and selection++Production changes only instrumentation namespaces: Today presentation and launch use PresentationPerformance, composed teaching uses TeachingPerformance. Changed callers are LibraryRepository.todayPresentation, LaunchMeasurement and ComposedTeachingViewModel; no storage/runtime algorithm changes. The physical Today metric repeats the category independently and host contracts inspect actual selector wiring.++Affected runtime adapters are OSLog and XCTest signpost measurement. Development checks retain the real app's launch and teaching integration suites, plus rendered teaching and both scale-seeding consumers. Fixtures remain unchanged.++Expected targeted native suites: AppLibraryModelLaunchTests, LaunchPhaseRecordTests, ComposedTeachingViewModelTests, ComposedTeachingPreviewTests. Use whole Swift Testing suites; verify actual nonzero execution. Targeted UI: all 11 ComposedSurfaceUITests declarations, plus M4ScaleTodayPerformanceUITests/testSeededScaleM4ScenarioReachesRecent and testSeededScaleM4DuplicateSiteRowsScenarioReachesRecent. The native source inventory contains 96 declarations across the selected suites (13 launch, 7 phase record, 71 composed view model, 5 preview); parameterized execution may report additional cases. The two physical metric methods remain skipped in simulator broad runs by their original gates.++Private simulator IDs: phone 04621845-AAC6-4738-B7CB-498A7C6AB69B, iPad 6C7F6D89-2142-4B1D-BC6A-E6351C28C117. Actual heavy commands coordinate through /tmp/asterism-t2056-t2115-build.lock; source/review proceeds independently. Each run has private caches, a fresh result bundle and an external command/exit receipt.++## Results++| Check | Outcome |+| --- | --- |+| Focused Core: M4ScaleFixtureTests + PerformanceSignpostContractTests | Passed: 18 tests, 2 suites; initial and restored runs |+| Physical metric category mutation | Expected failure: contract detected divergent category; restored |+| Physical metric argument wiring mutation | Expected failure: contract detected divergent argument; restored |+| Source restoration | Byte-for-byte verified; restored 18-test cohort passed |+| Four independent source reviews | No actionable findings |+| Contract literal whitespace correction | Four contract tests passed; spaced category mutation failed as intended; restored four passed |+| Documentation whitespace | Passed |+| Unfiltered make test-core | Passed: 3,824 executed, 55 existing opt-in skips, 0 failures/errors (3,879 discovered across Core/Presentation/Intelligence/Constellation); exit 0 |+| make test-fast | Passed: 456 declared identifiers verified and executed, exit 0 |+| Targeted native launch/phase/teaching/preview | Passed: all 96 declared tests executed, 0 failures/skips, exit 0 |+| Targeted teaching and scale-seeding UI | Passed: all 13 expected cases, 0 failures/skips/missing/unexpected identifiers; exit 0 |+| Signed Mac app and embedded extension compile | Passed within make test-quick; no app launch/install |+| Full simulator native make test-quick | 1,756 test bodies completed; one unrelated EntryDetailModelTests.duplicateSubmissionSuppressed failure (two assertions). Xcode finalisation stalled under extreme host load and was interrupted; runner exit 2. Full native gate is not green. |+| Full iPad UI | Incomplete: 17 passed, 1 failed, 1 interrupted, 11 never started of 30 expected; owned run interrupted under confirmed host pressure, exit 2 |+| Full phone UI | Unrun: host capacity blocker; expected inventory 229 cases including two physical-only opt-ins |+| Bounded unrelated-failure retries | Unrun: host pressure did not recover; no repeated broad gates |++Full logs and structured mutation receipts live in the task evidence directory outside the checkout. No physical performance run or device installation is authorised. Host/simulator checks cannot establish physical sample resolution.++## Host contention and interruption++Independent dedicated-simulator runtime overlap was explicitly authorised. During overlapping runs the host load rose into the hundreds (another owner observed over 1,000), despite free-memory percentages around 32–44%. New jobs were paused and only owned work was drained. T-2115's native test bodies had all completed before Xcode finalisation stalled; its exact owned xcodebuild process was interrupted after preserving logs/result data, then its dedicated phone was confirmed shut down. No unrelated process was stopped. The ongoing iPad run continued to progress.++The failing duplicate-submission case lives in unchanged EntryDetailModelTests.swift and exercises unchanged EntryDetailModel.update with MockLibraryProvider, outside the renamed instrumentation callers. A targeted native retry is pending recovered capacity; the original full-gate failure remains part of the result regardless of retry outcome. No cause is attributed solely to host load.++## Final blocked state++Full iPad failure: `WideLayoutAccessibilityUITests/testTheFiveReadingStatusSegmentsAreOperableInBothWideLayouts()` reported invalid activation point/hittability for `sidebar-library-reading` at `UIJourneySupport.swift:133`. The native duplicate-submission and iPad hittability paths are outside this instrumentation rename. No cause is attributed solely to load, and no unrelated source or assertion was changed.++The final owned-job drain snapshot recorded 2,668 processes, 829 runnable, 14,048 threads, load 907/583/210, 31 GiB physical memory used with 115 MiB unused, 7.3 GiB wired, 6.8 GiB compressed, 55% CPU idle and 16 GiB disk available. This is a measured memory/scheduler capacity blocker, not simply CPU utilisation. Both T-2115 simulators were verified Shutdown and both owned test processes were absent. No host settings or unrelated jobs were changed.++Partial `quick.xcresult` and `full-ipad.xcresult` bundles are retained, but interrupted finalisation left no `Info.plist`; export is unavailable. The precise log-derived outcomes are in `quick-outcomes.json` and `full-ipad-outcomes.json` outside the checkout. Native/iPad retries and the full phone gate remain blocked. Rune task 2 and Transit remain in progress. The draft PR must not claim green acceptance.
Do not merge on a green test claim. Resume full phone/iPad acceptance and the bounded unrelated-failure retries only after measured host capacity recovers. No unrelated source fix is bundled.
Update M2Performance to PresentationPerformance and M3Performance to TeachingPerformance in saved Instruments filters.