Skip to content

StreamVault is a TV-first free IPTV player for Android TV built with Kotlin, Jetpack Compose, Room, Hilt, and Media3.

build(system): unblock baseline profile copy

David Nashashe36d39b04dc9
diff --git a/app/build.gradle.kts b/app/build.gradle.kts
index f9129bc8..ad7d2789 100644
--- a/app/build.gradle.kts
+++ b/app/build.gradle.kts
@@ -325,7 +325,11 @@ abstract class MergeStartupRulesIntoBaselineProfileTask : DefaultTask() {
// tasks omit the dependency to avoid a cycle. Beta/release merges remain explicit even when
// generation and assembly are requested together.
val profileGenerationRequested = gradle.startParameter.taskNames.any { taskName ->
- taskName.substringAfterLast(':') == "generateBaselineProfile"
+ taskName.substringAfterLast(':') in setOf(
+ "generateBaselineProfile",
+ "copyBaselineProfileIntoSrc",
+ "mergeBaselineProfile",
+ )
}
setOf(
"mergeBetaArtProfile",
diff --git a/docs/COMPOSE_REDUCTION_AND_UI_ARCHITECTURE_PLAN.md b/docs/COMPOSE_REDUCTION_AND_UI_ARCHITECTURE_PLAN.md
index 55605c51..29bc4039 100644
--- a/docs/COMPOSE_REDUCTION_AND_UI_ARCHITECTURE_PLAN.md
+++ b/docs/COMPOSE_REDUCTION_AND_UI_ARCHITECTURE_PLAN.md
@@ -999,10 +999,11 @@ paths, plus reversible quick-filter save/reload/remove persistence. The full
Home shelf, open paired benchmark gate, refreshed release-like profiles, and
current app lint baseline failure remain explicitly tracked. System is now
structurally extracted under `:feature:system`; its connected acceptance
- passes 19/19 and paired incremental thresholds pass. The System profile
- refresh, paired clean/warm comparison, physical-device/accessibility gates,
- and authorized production journeys remain open; Playback, Provider, and
- Settings gates remain governed by their separate reports.
+ passes 19/19, paired incremental thresholds pass, and the steady-state
+ clean/warm guardrail passes. The System profile refresh,
+ physical-device/accessibility gates, and authorized production journeys
+ remain open; Playback, Provider, and Settings gates remain governed by their
+ separate reports.
### Phase 6 - Optional Views migrations
diff --git a/docs/COMPOSE_REDUCTION_PHASE5_ROADMAP.md b/docs/COMPOSE_REDUCTION_PHASE5_ROADMAP.md
index a9957dc0..b08c5987 100644
--- a/docs/COMPOSE_REDUCTION_PHASE5_ROADMAP.md
+++ b/docs/COMPOSE_REDUCTION_PHASE5_ROADMAP.md
@@ -148,7 +148,7 @@ System presentation extraction is structurally complete under `:feature:system`:
Welcome, Downloads, Plugins, controller-free graph registration, resources,
contracts, and tests are feature-owned while app adapters and plugin/platform
runtime remain app-owned. Connected acceptance passes 19/19 and paired
-incremental source/test thresholds pass. The supported System profile refresh,
-paired clean/warm comparison, physical-device/touch validation, accessibility
-service traversal, and authorized production-provider/plugin journeys remain
-open; see `COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md`.
+incremental source/test thresholds and the steady-state clean/warm guardrail
+pass. The supported System profile refresh, physical-device/touch validation,
+accessibility service traversal, and authorized production-provider/plugin
+journeys remain open; see `COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md`.
diff --git a/docs/COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md b/docs/COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md
index d3d813b3..2a27e773 100644
--- a/docs/COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md
+++ b/docs/COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md
@@ -12,11 +12,11 @@ retains startup orchestration, shell composition, platform adapters, plugin
discovery/IPC/work/provider/playback infrastructure, and Android entry points.
The extraction is structurally complete and the incremental feature-value
-thresholds pass. The formal clean/warm paired comparison and startup/profile
-refresh remain open because the supported seeded profile workflow stalled in
-its unrelated Live/player critical journey. The module is therefore retained
-as a provisional Phase 5 decision, with the remaining gates explicitly owned
-as follow-up work.
+thresholds pass. The clean/warm build-overhead guardrail also passes in
+steady-state. The general baseline-profile refresh remains open because the
+supported seeded profile workflow stalls in its unrelated Live/player critical
+journey. The module is therefore retained as a provisional Phase 5 decision,
+with the remaining gates explicitly owned as follow-up work.
## Rollback and commit sequence
@@ -203,9 +203,11 @@ only relevant work. Evidence is in
`validation/phase5_system/system-edit-dry-run.txt` and
`system-edit-task-output.txt`.
-Clean debug, warm debug, Beta, and Release guardrails passed. A cache-equivalent
-paired before/after clean/warm duration comparison has not been completed, so
-the ≤10% build-overhead gate is open rather than inferred.
+Clean debug, warm debug, Beta, and Release guardrails passed. The paired
+rollback/current clean run is faster in the current checkout, and steady-state
+warm medians are 11.707s rollback versus 11.955s current, a 2.12% slowdown;
+the ≤10% build-overhead gate passes. Full raw samples and cache/task details
+are in `validation/phase5_system/performance-after.md`.
## Profiles, graph, and remaining work
@@ -223,8 +225,8 @@ run must promote fresh output through the existing workflow. See
graph resolves feature `SystemGraph` ownership and an `AppNavHost` to
`PluginsScreen` path through `AppNavigation`.
-Open follow-ups are the supported profile refresh, paired clean/warm duration
-comparison, physical-device/touch validation, TalkBack service traversal, and
-authorized production-provider/plugin journeys. This report does not close
+Open follow-ups are the supported general profile refresh, physical-device/
+touch validation, TalkBack service traversal, and authorized
+production-provider/plugin journeys. This report does not close
Playback, Provider, Settings, Live, Catalog, accessibility, or repository-wide
performance gates, and does not claim full Phase 5 modernization completion.
diff --git a/docs/COMPOSE_REDUCTION_PHASE5_TRANSITIONAL_DEPENDENCIES.md b/docs/COMPOSE_REDUCTION_PHASE5_TRANSITIONAL_DEPENDENCIES.md
index a2a090b2..3e7e32cc 100644
--- a/docs/COMPOSE_REDUCTION_PHASE5_TRANSITIONAL_DEPENDENCIES.md
+++ b/docs/COMPOSE_REDUCTION_PHASE5_TRANSITIONAL_DEPENDENCIES.md
@@ -258,7 +258,8 @@ Task 9 structural, resource, unit-test, compilation, lint, and assembly gates
are recorded in `validation/phase5_system/task9-structural-verification.md`.
App-wide lint remains a pre-existing baseline-drift failure and is not a
System extraction failure. Task 10 connected acceptance and reviewed goldens
-pass 19/19, and the paired incremental thresholds pass. The full System
+pass 19/19, the paired incremental thresholds pass, and the steady-state
+clean/warm build-overhead guardrail passes. The full System
execution report is `docs/COMPOSE_REDUCTION_PHASE5_SYSTEM_REPORT.md`; the
-supported profile refresh and cache-equivalent clean/warm comparison remain
-open follow-ups.
+supported general profile refresh, physical-device/accessibility gates, and
+authorized production-provider/plugin journeys remain open follow-ups.
diff --git a/validation/phase5_system/performance-after.md b/validation/phase5_system/performance-after.md
index 55c00320..9146aa07 100644
--- a/validation/phase5_system/performance-after.md
+++ b/validation/phase5_system/performance-after.md
@@ -70,6 +70,28 @@ configuration/cache initialization.
- The app lint task remains red on pre-existing app-wide debt and is tracked
separately in the final System report; feature lint passes.
-The source-edit and unit-test thresholds pass. Sibling Kotlin isolation,
-startup/profile status, and the remaining verification matrix are recorded in
-the System report rather than inferred from these two measurements.
+## Paired clean/warm guardrail
+
+The same `clean :app:assembleDebug` and warm `:app:assembleDebug` commands were
+run in a disposable rollback worktree at
+`cddef1526924634819c7fb27d15c8a49b3945a1d` and in the current checkout. The
+rollback worktree received only a copy of the active local SDK configuration;
+the active checkout was not changed.
+
+| Scenario | Rollback | Current | Change | Gate |
+|---|---:|---:|---:|---|
+| clean assemble | 161.907s | 66.502s | 58.95% faster | pass (<=10% slower) |
+| warm repeat 1 | 11.768s | 11.955s | 1.59% slower | pass |
+| warm repeat 2 | 11.707s | 11.935s | 1.95% slower | pass |
+| warm repeat 3 | 11.072s | 12.691s | 14.62% slower | noise sample |
+
+Steady-state warm medians are 11.707s rollback and 11.955s current, a
+2.12% slowdown. The first warm invocation after the clean build included
+task-set configuration-cache setup (22.862s rollback and 26.527s current), so
+it is reported separately and excluded from the steady-state median. The
+clean runs reused the remote build cache with 232 versus 255 actionable tasks;
+the current run executed 113 tasks versus 102 at rollback.
+
+The incremental source/test thresholds, sibling Kotlin isolation, and the
+clean/warm build-overhead guardrail pass. Startup/profile status and the
+remaining device/production journeys are recorded in the System report.
diff --git a/validation/phase5_system/profile-validation.md b/validation/phase5_system/profile-validation.md
index ea4352ba..3a1d399b 100644
--- a/validation/phase5_system/profile-validation.md
+++ b/validation/phase5_system/profile-validation.md
@@ -31,11 +31,30 @@ com/streamvault/feature/system/navigation/SystemGraphKt
com/streamvault/app/system/AppSystemWelcomeAdapter
```
-The subsequent `criticalJourneys` method remained at `1/10` with no failure
-diagnostic for the bounded collection window. It was stopped safely; the
-overall generator process therefore ended with user-interrupt status and did
-not complete the profile install/copy step. This benchmark journey exercises
-seeded Home/Live/player navigation and does not cover the System routes.
+The isolated startup command was:
+
+```text
+./gradlew.bat :benchmark:connectedNonMinifiedReleaseAndroidTest '-Pandroid.testInstrumentationRunnerArguments.class=com.streamvault.benchmark.BaselineProfileGenerator#startup' --no-daemon --console=plain --warning-mode=none
+BUILD SUCCESSFUL in 4m 59s — 1/1 test
+```
+
+The fresh artifact was emitted under
+`benchmark/build/outputs/connected_android_test_additional_output/nonMinifiedRelease/connected/Television_1080p(AVD) - 16/`;
+the timestamped startup profile was 3,240,532 bytes. Its scan returned no old
+app Welcome/Downloads/Plugins or old app graph descriptors and returned
+nonzero `feature/system` descriptors.
+
+The subsequent independent `criticalJourneys` run started 1 test, remained at
+`0/1`, and produced no failure diagnostic during the bounded collection
+window. It was stopped safely. The earlier full generator run showed the same
+behavior as `1/10`. This benchmark journey exercises seeded Home/Live/player
+navigation and does not cover the System routes.
+
+The first direct `:app:copyBaselineProfileIntoSrc --dry-run` exposed a cycle in
+the existing app profile-task wiring. The build guard now treats direct
+`copyBaselineProfileIntoSrc` and `mergeBaselineProfile` requests as profile
+generation requests, and the same dry run completes successfully. An actual
+copy still requires a completed general baseline collection.
The maintained files under `app/src/main/generated/baselineProfiles/` were
not manually edited. They still contain stale pre-extraction System entries,
@@ -48,6 +67,8 @@ source because the supported generator did not finish its complete workflow.
- Baseline-profile refresh is open and must be rerun when the seeded benchmark
journey completes end-to-end.
+- The direct copy-task dependency cycle is fixed and covered by a successful
+ dry run.
- No claim is made that the stale checked-in generated profile is refreshed.
- No physical Android device was available; only the TV emulator was present.
- The System extraction does not alter player composition, stream preparation,