test(app): add PR0.4 low-end perf profile

.github/workflows/perf-probe-baseline.yml

@@ -11,6 +11,7 @@ on:
       - "packages/app/package.json"
       - "packages/app/playwright.config.ts"
       - "packages/app/script/compare-perf.ts"
+      - "packages/app/script/merge-perf-artifacts.ts"
       - "packages/app/script/e2e-local.ts"
       - "packages/app/src/testing/perf-metrics*"
   workflow_dispatch:
@@ -31,11 +32,13 @@ jobs:
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # actions/checkout@v6
         with:
+          fetch-depth: 0
           path: head
           persist-credentials: false
 
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # actions/checkout@v6
         with:
+          fetch-depth: 0
           path: base
           persist-credentials: false
           ref: ${{ github.event.pull_request.base.sha || github.sha }}
@@ -59,6 +62,8 @@ jobs:
         with:
           path: ${{ github.workspace }}/.playwright-browsers
           key: playwright-${{ runner.os }}-${{ hashFiles('head/packages/app/package.json', 'head/bun.lock', 'base/packages/app/package.json', 'base/bun.lock') }}
+          restore-keys: |
+            playwright-${{ runner.os }}-
 
       - name: Install head dependencies
         working-directory: head
@@ -78,6 +83,56 @@ jobs:
           cp head/packages/app/playwright.config.ts base/packages/app/playwright.config.ts
           cp head/packages/app/src/testing/perf-metrics.ts base/packages/app/src/testing/perf-metrics.ts
 
+      - name: Detect low-end perf scope
+        id: low_end_scope
+        working-directory: head
+        env:
+          EVENT_NAME: ${{ github.event_name }}
+          BASE_SHA: ${{ github.event.pull_request.base.sha || github.event.before }}
+          HEAD_SHA: ${{ github.sha }}
+        run: |
+          set -euo pipefail
+
+          is_low_end_path() {
+            case "$1" in
+              packages/app/src/pages/session/*|packages/app/src/pages/session/**|packages/app/src/pages/session.tsx|packages/ui/src/components/message-part.tsx|packages/ui/src/components/session-turn.tsx|packages/ui/src/components/markdown.tsx|packages/app/e2e/perf/*|packages/app/e2e/perf/**|packages/app/src/testing/perf-metrics*|packages/app/script/compare-perf.ts|packages/app/script/merge-perf-artifacts.ts|.github/workflows/perf-probe-baseline.yml)
+                return 0
+                ;;
+              *)
+                return 1
+                ;;
+            esac
+          }
+
+          if [ "$EVENT_NAME" = "workflow_dispatch" ]; then
+            echo "run_low_end=true" >> "$GITHUB_OUTPUT"
+            exit 0
+          fi
+
+          if [ -z "$BASE_SHA" ] || [ "$BASE_SHA" = "0000000000000000000000000000000000000000" ]; then
+            BASE_SHA="$(git rev-list --max-parents=0 HEAD | tail -n 1)"
+          fi
+
+          run_low_end=false
+          mapfile -t changes < <(git diff --name-status --find-renames --find-copies "$BASE_SHA" "$HEAD_SHA" --)
+          for change in "${changes[@]}"; do
+            IFS=$'\t' read -r status path1 path2 <<< "$change"
+            case "$status" in
+              A*|M*|T*|D*)
+                if is_low_end_path "$path1"; then run_low_end=true; break; fi
+                ;;
+              R*|C*)
+                if is_low_end_path "$path1" || is_low_end_path "$path2"; then run_low_end=true; break; fi
+                ;;
+              *)
+                run_low_end=true
+                break
+                ;;
+            esac
+          done
+
+          echo "run_low_end=$run_low_end" >> "$GITHUB_OUTPUT"
+
       - name: Run perf probe baseline (base)
         env:
           CI: "true"
@@ -92,13 +147,45 @@ jobs:
           PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-head.json
         run: bun --cwd head/packages/app test:e2e:local:perf
 
+      - name: Run low-end perf probe baseline (base)
+        if: steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: base
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-base-low-end.json
+        run: bun --cwd base/packages/app test:e2e:local:perf
+
+      - name: Run low-end perf probe baseline (head)
+        if: steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: head
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-head-low-end.json
+        run: bun --cwd head/packages/app test:e2e:local:perf
+
+      - name: Merge perf profile artifacts (base)
+        run: >
+          bun head/packages/app/script/merge-perf-artifacts.ts
+          --required "${PERF_ARTIFACT_DIR}/perf-base.json"
+          --optional "${PERF_ARTIFACT_DIR}/perf-base-low-end.json"
+          --output "${PERF_ARTIFACT_DIR}/perf-base-combined.json"
+
+      - name: Merge perf profile artifacts (head)
+        run: >
+          bun head/packages/app/script/merge-perf-artifacts.ts
+          --required "${PERF_ARTIFACT_DIR}/perf-head.json"
+          --optional "${PERF_ARTIFACT_DIR}/perf-head-low-end.json"
+          --output "${PERF_ARTIFACT_DIR}/perf-head-combined.json"
+
       - name: Compare base and head perf baselines
         id: compare
         continue-on-error: true
         run: >
           bun head/packages/app/script/compare-perf.ts
-          --base "${PERF_ARTIFACT_DIR}/perf-base.json"
-          --head "${PERF_ARTIFACT_DIR}/perf-head.json"
+          --base "${PERF_ARTIFACT_DIR}/perf-base-combined.json"
+          --head "${PERF_ARTIFACT_DIR}/perf-head-combined.json"
           --output "${PERF_ARTIFACT_DIR}/perf-compare.json"
           --comment-output "${PERF_ARTIFACT_DIR}/perf-comment.md"
 
@@ -118,14 +205,48 @@ jobs:
           PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-head-confirm.json
         run: bun --cwd head/packages/app test:e2e:local:perf
 
+      - name: Confirm low-end perf regression (base)
+        if: steps.compare.outcome == 'failure' && steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: base
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-base-low-end-confirm.json
+        run: bun --cwd base/packages/app test:e2e:local:perf
+
+      - name: Confirm low-end perf regression (head)
+        if: steps.compare.outcome == 'failure' && steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: head
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-head-low-end-confirm.json
+        run: bun --cwd head/packages/app test:e2e:local:perf
+
+      - name: Merge confirmed perf profile artifacts (base)
+        if: steps.compare.outcome == 'failure'
+        run: >
+          bun head/packages/app/script/merge-perf-artifacts.ts
+          --required "${PERF_ARTIFACT_DIR}/perf-base-confirm.json"
+          --optional "${PERF_ARTIFACT_DIR}/perf-base-low-end-confirm.json"
+          --output "${PERF_ARTIFACT_DIR}/perf-base-confirm-combined.json"
+
+      - name: Merge confirmed perf profile artifacts (head)
+        if: steps.compare.outcome == 'failure'
+        run: >
+          bun head/packages/app/script/merge-perf-artifacts.ts
+          --required "${PERF_ARTIFACT_DIR}/perf-head-confirm.json"
+          --optional "${PERF_ARTIFACT_DIR}/perf-head-low-end-confirm.json"
+          --output "${PERF_ARTIFACT_DIR}/perf-head-confirm-combined.json"
+
       - name: Compare confirmed perf baselines
         id: compare_confirmed
         if: steps.compare.outcome == 'failure'
         continue-on-error: true
         run: >
           bun head/packages/app/script/compare-perf.ts
-          --base "${PERF_ARTIFACT_DIR}/perf-base-confirm.json"
-          --head "${PERF_ARTIFACT_DIR}/perf-head-confirm.json"
+          --base "${PERF_ARTIFACT_DIR}/perf-base-confirm-combined.json"
+          --head "${PERF_ARTIFACT_DIR}/perf-head-confirm-combined.json"
           --output "${PERF_ARTIFACT_DIR}/perf-compare-confirm.json"
           --comment-output "${PERF_ARTIFACT_DIR}/perf-comment.md"
 
@@ -188,6 +309,26 @@ jobs:
           PAWWORK_PERF_TRACE: "1"
         run: bun --cwd head/packages/app test:e2e:local:perf
 
+      - name: Capture low-end perf diagnostic trace (base)
+        if: steps.compare.outcome == 'failure' && steps.compare_confirmed.outcome == 'failure' && steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: base
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-base-low-end-trace.json
+          PAWWORK_PERF_TRACE: "1"
+        run: bun --cwd base/packages/app test:e2e:local:perf
+
+      - name: Capture low-end perf diagnostic trace (head)
+        if: steps.compare.outcome == 'failure' && steps.compare_confirmed.outcome == 'failure' && steps.low_end_scope.outputs.run_low_end == 'true'
+        env:
+          CI: "true"
+          PAWWORK_PERF_BRANCH: head
+          PAWWORK_PERF_PROFILE: low-end
+          PAWWORK_PERF_OUTPUT: ${{ github.workspace }}/perf-artifacts/perf-head-low-end-trace.json
+          PAWWORK_PERF_TRACE: "1"
+        run: bun --cwd head/packages/app test:e2e:local:perf
+
       - name: Fail job on comparator regression
         if: steps.compare.outcome == 'failure' && steps.compare_confirmed.outcome == 'failure'
         run: exit 1

packages/app/e2e/perf/perf-probe.spec.ts

@@ -12,9 +12,12 @@ import {
 } from "../selectors"
 import { sessionPath, terminalToggleKey } from "../utils"
 import { installPerfProbe, resetPerfProbe, snapshotPerfProbe, summarizeScenarioRuns } from "./probe"
+import { applyPerfProfile, readPerfProfile, shouldRunScenario, type PerfScenarioName } from "./profiles"
+import { seedTimelineRecomputeSession } from "./timeline-fixture"
 
 const outputPath = process.env.PAWWORK_PERF_OUTPUT ?? path.join(process.cwd(), "e2e", "perf-results", "pr0.1-baseline.json")
 const perfBranch = process.env.PAWWORK_PERF_BRANCH ?? "dev"
+const PERF_PROFILE = readPerfProfile()
 
 const longMarkdown = [
   "# Baseline stream",
@@ -146,6 +149,10 @@ async function scrollTimelineTo(page: Parameters<typeof snapshotPerfProbe>[0], t
   expect(found).toBe(true)
 }
 
+function skipUnlessScenario(name: PerfScenarioName) {
+  test.skip(!shouldRunScenario(PERF_PROFILE, name), `${PERF_PROFILE} profile does not run ${name}`)
+}
+
 test.describe("PR0.1 perf probe baseline", () => {
   test.describe.configure({ mode: "serial" })
 
@@ -155,7 +162,9 @@ test.describe("PR0.1 perf probe baseline", () => {
   })
 
   test("homepage-cold emits a 3-run JSON baseline", async ({ page, project }) => {
+    skipUnlessScenario("homepage-cold")
     await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
     await project.open()
 
     const runs = []
@@ -171,11 +180,13 @@ test.describe("PR0.1 perf probe baseline", () => {
       if (run < 2) await cooldownAfterRun(page)
     }
 
-    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, scenario: "homepage-cold", runs }))
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "homepage-cold", runs }))
   })
 
   test("session-streaming-long emits a 3-run JSON baseline", async ({ page, project, llm }) => {
+    skipUnlessScenario("session-streaming-long")
     await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
     await project.open()
 
     const runs = []
@@ -216,11 +227,13 @@ test.describe("PR0.1 perf probe baseline", () => {
       if (run < 2) await cooldownAfterRun(page)
     }
 
-    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, scenario: "session-streaming-long", runs }))
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "session-streaming-long", runs }))
   })
 
   test("tool-call-expand emits a 3-run JSON baseline", async ({ page, project, llm }) => {
+    skipUnlessScenario("tool-call-expand")
     await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
     await project.open()
 
     const runs = []
@@ -246,11 +259,13 @@ test.describe("PR0.1 perf probe baseline", () => {
       if (run < 2) await cooldownAfterRun(page)
     }
 
-    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, scenario: "tool-call-expand", runs }))
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "tool-call-expand", runs }))
   })
 
   test("terminal-side-panel-open emits a 3-run JSON baseline", async ({ page, project }) => {
+    skipUnlessScenario("terminal-side-panel-open")
     await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
     await project.open()
 
     const runs = []
@@ -273,11 +288,13 @@ test.describe("PR0.1 perf probe baseline", () => {
       })
     }
 
-    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, scenario: "terminal-side-panel-open", runs }))
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "terminal-side-panel-open", runs }))
   })
 
   test("session-scroll-reading emits a 3-run JSON baseline", async ({ page, project }) => {
+    skipUnlessScenario("session-scroll-reading")
     await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
     await project.open()
 
     const runs = []
@@ -312,6 +329,35 @@ test.describe("PR0.1 perf probe baseline", () => {
       })
     }
 
-    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, scenario: "session-scroll-reading", runs }))
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "session-scroll-reading", runs }))
+  })
+
+  test("session-timeline-recompute emits a 3-run low-end JSON baseline", async ({ page, project }) => {
+    skipUnlessScenario("session-timeline-recompute")
+    await installPerfProbe(page)
+    await applyPerfProfile(page, PERF_PROFILE)
+    await project.open()
+
+    const runs = []
+    for (let run = 0; run < 3; run += 1) {
+      await withSession(project.sdk, `perf timeline recompute ${Date.now()}-${run}`, async (session) => {
+        await seedTimelineRecomputeSession(project, session.id)
+        await page.goto(sessionPath(project.directory, session.id))
+        await expect(page.locator(sessionMessageItemSelector).first()).toBeVisible({ timeout: 30_000 })
+        await expect.poll(async () => page.locator(sessionMessageItemSelector).count()).toBeGreaterThanOrEqual(8)
+        await resetPerfProbe(page)
+        await page.locator(scrollViewportSelector).first().hover()
+        for (let index = 0; index < 4; index += 1) {
+          await page.mouse.wheel(0, index % 2 === 0 ? 2400 : -2400)
+          await settleFrames(page, 2)
+          await scrollTimelineTo(page, index % 2 === 0 ? 0 : 1200)
+          await settleFrames(page, 2)
+        }
+        runs.push(await snapshotPerfProbe(page))
+        if (run < 2) await cooldownAfterRun(page)
+      })
+    }
+
+    scenarioResults.push(summarizeScenarioRuns({ branch: perfBranch, profile: PERF_PROFILE, scenario: "session-timeline-recompute", runs }))
   })
 })

packages/app/e2e/perf/probe.ts

@@ -1,5 +1,6 @@
 import type { Page } from "@playwright/test"
-import { aggregatePerfRuns, summarizePerfRun, type PerfRunSummary } from "../../src/testing/perf-metrics"
+import { aggregatePerfRuns, summarizePerfRun, type PerfProfile, type PerfRunSummary } from "../../src/testing/perf-metrics"
+import type { PerfScenarioName } from "./profiles"
 
 type BrowserPerfSample = {
   startedAt: number
@@ -158,6 +159,11 @@ export async function snapshotPerfProbe(page: Page) {
   return summarizePerfRun(sample)
 }
 
-export function summarizeScenarioRuns(input: { branch: string; scenario: string; runs: PerfRunSummary[] }) {
+export function summarizeScenarioRuns(input: {
+  branch: string
+  profile?: PerfProfile
+  scenario: PerfScenarioName
+  runs: PerfRunSummary[]
+}) {
   return aggregatePerfRuns(input)
 }