@reportforge/playwright-pdf 0.28.0 → 0.29.0

package/README.md

@@ -59,6 +59,7 @@ Three templates, one reporter. Every PDF is fully offline: fonts, charts, and sc
 - **Flakiness trend**: top-N flakiest tests table with per-test dot sparkline across stored history runs in the `detailed` template; `flakinessTopN` option (default 5, `0` disables)
 - **Pass-rate trend**: cross-run trend chart from stored history; opt-in `remoteHistory` keeps the trend alive on ephemeral CI runners by storing aggregate numbers only (pass rate, counts, duration, verdict; no test titles or error text, and the branch name never travels in plaintext), with no extra network calls
 - **Run-over-run diff**: every report shows what changed since the previous run on the same branch: new failures, fixed tests, and still-failing tests, and the same summary rides along in Slack, Teams, Discord, and email notifications
+- **Cross-browser comparison**: multi-project runs get a per-test matrix across projects in the `detailed` template — divergent outcomes first (fails-only-in-webkit at a glance), identical tests collapsed to a count; works for Node.js multi-project configs and pytest runs with multiple `--browser` values
 - **Offline failure analysis**: sorts each failure into one of 7 root-cause buckets and renders a root-cause chip under each failed test in the `detailed` PDF (dominant-cause one-liner in `executive`). No network, no AI service: a small embedded classifier with automatic updates; pin it any time. Full details at [reportforge.org/docs/advanced/failure-analysis](https://reportforge.org/docs/advanced/failure-analysis).
 - **On-device training**: teach the classifier your team's failures without any data leaving your machine. Label locally-collected samples (`npx @reportforge/playwright-pdf label-feedback`), then `train-model` trains and cross-validates a personal layer on-device; it activates only when it measurably beats the base model on your own labeled set (`evaluate-model` shows the comparison any time). Commit the model file to your repo via `failureAnalysis.localModelPath` to share it with CI. Training data, the trained model, and evaluation are all local files; the analysis code cannot even name a network primitive (enforced by an automated source audit).
 - `npx @reportforge/playwright-pdf export-feedback`: merges your locally-collected feedback (tokenized + redacted) into one CSV for review or moving between machines. Local file only; there is no upload.
@@ -179,7 +180,7 @@ template = "detailed"
 outputFile = "reports/{date}-{branch}-report.pdf"
 ```
 
-The plugin reuses the Node runtime and Chromium that Playwright for Python already ships, so there is nothing else to install. pytest-xdist and pytest-rerunfailures are supported; failure screenshots from `--screenshot only-on-failure` are embedded. Live runs and step trees are Node.js-only today. Full guide: [reportforge.org/docs/python](https://reportforge.org/docs/python).
+The plugin reuses the Node runtime and Chromium that Playwright for Python already ships, so there is nothing else to install. pytest-xdist and pytest-rerunfailures are supported; failure screenshots from `--screenshot only-on-failure` are embedded. Passing `--browser` more than once splits the report into one project per browser and unlocks the cross-browser comparison section. Live runs and step trees are Node.js-only today. Full guide: [reportforge.org/docs/python](https://reportforge.org/docs/python).
 
 ---
 
@@ -194,7 +195,7 @@ set RF_LICENSE_KEY=RFSU-XXXX-XXXX-XXXX-XXXX
 dotnet test --logger "reportforge;output=reports/run.pdf;template=detailed"
 ```
 
-Full options come from a JSON file with the same camelCase names as this reporter: `--logger "reportforge;optionsFile=rf-options.json"`. The logger reuses the Node runtime Microsoft.Playwright places in your test output and a Playwright-managed Chromium, so there is nothing else to install. Retried tests render as flaky with their attempt history; categories become tags. Live runs and step trees are Node.js-only today. Full guide: [reportforge.org/docs/dotnet](https://reportforge.org/docs/dotnet).
+Full options come from a JSON file with the same camelCase names as this reporter: `--logger "reportforge;optionsFile=rf-options.json"`. The logger reuses the Node runtime Microsoft.Playwright places in your test output and a Playwright-managed Chromium, so there is nothing else to install. Retried tests render as flaky with their attempt history; categories become tags. A `dotnet test` run drives one browser, so the cross-browser comparison section stays hidden here; live runs and step trees are Node.js-only today. Full guide: [reportforge.org/docs/dotnet](https://reportforge.org/docs/dotnet).
 
 ---
 
@@ -255,7 +256,7 @@ All options are the second element of the reporter tuple.
 | `slowTestThreshold` | `number` | `10` | Minimum timed-test count before `SLOW` badges appear in the suite breakdown: below this, every test is trivially the "slowest" so badges stay hidden. On larger runs only genuine outliers (at least 2× the median test duration) are badged, so the badge stays rare. Set to `0` to drop the run-size gate (outliers still required). |
 | `requirementTagPattern` | `string` | `'^@?[A-Z][A-Z0-9]*-\\d+$'` | Regex deciding which tags count as requirement IDs in the Requirements Traceability section (`detailed` template). Matching tags (ticket shapes like `@ODP-5328`, `REQ-001`) get the traceability matrix; everything else (`@regression`, `@sanity`) collapses into a compact tag summary instead of repeating identical rows. Set to `''` to disable the split and list every tag in the matrix. |
 | `templatePath` | `string \| string[]` | n/a | Path to a custom Handlebars (`.hbs`) template file. Takes precedence over `template`. Pass an array to generate one PDF per custom template. See the Custom Templates docs. |
-| `sections` | `object` | per-template | Override which report sections appear, on top of the chosen template's defaults. Flat keys are the baseline for every chosen template; per-template keys (`minimal`\|`detailed`\|`executive`) override per template. 20 keys total. Block toggles: `coverPage`, `analysisOneliner`, `releaseGate`, `summary`, `charts`, `trend`, `requirementsMatrix`, `ciEnvironment`, `suiteBreakdown`, `failureDeepDive`, `failureAnalysis`, `slowTests`, `defectLog`, `briefBand`, `runDiff`. Display modifiers: `passRate`, `fullEnvironment`, `retries`, `fullFailures`, `stackTraces`. `runDiff` (block toggle, on by default in all three templates) shows new failures, fixed tests, and still-failing tests compared with the previous run on the same branch. See the Report Sections docs. |
+| `sections` | `object` | per-template | Override which report sections appear, on top of the chosen template's defaults. Flat keys are the baseline for every chosen template; per-template keys (`minimal`\|`detailed`\|`executive`) override per template. 21 keys total. Block toggles: `coverPage`, `analysisOneliner`, `releaseGate`, `summary`, `charts`, `trend`, `requirementsMatrix`, `ciEnvironment`, `suiteBreakdown`, `failureDeepDive`, `failureAnalysis`, `slowTests`, `defectLog`, `briefBand`, `runDiff`, `browserMatrix`. Display modifiers: `passRate`, `fullEnvironment`, `retries`, `fullFailures`, `stackTraces`. `runDiff` (block toggle, on by default in all three templates) shows new failures, fixed tests, and still-failing tests compared with the previous run on the same branch. `browserMatrix` (block toggle, on by default in `detailed`) compares per-test outcomes across the run's projects and appears only when the run has 2+ projects. See the Report Sections docs. |
 <!-- AUTOGEN:options-table END -->
 
 Full reference: [reportforge.org/docs/configuration](https://reportforge.org/docs/configuration).
@@ -542,6 +543,8 @@ reporter: [['@reportforge/playwright-pdf', {
 | `slowTests` | `SLOW` badge on the duration-ranked slowest tests, shown inline in the breakdown |
 | `defectLog` | `DEF-####` numbered failure table (severity, duration) + opt-in repro detail (`capture`) |
 | `briefBand` | Executive brief band: one plain-language sentence (trend delta + root causes) and a hero pass-rate stat |
+| `runDiff` | Since Last Run: new failures, fixed tests, and still-failing tests vs. the previous run on the same branch |
+| `browserMatrix` | Cross-browser comparison: per-test outcomes side by side across the run's projects — divergent tests first, identical ones collapsed to a count. Appears only when the run has 2+ projects |
 
 #### Display modifiers: tune a section that's on
 
@@ -573,6 +576,8 @@ Omit `sections` entirely and each template renders exactly this (✓ = on):
 | `slowTests` | – | ✓ | ✓ |
 | `defectLog` | – | ✓ | – |
 | `briefBand` | – | – | ✓ |
+| `runDiff` | ✓ | ✓ | ✓ |
+| `browserMatrix` | – | ✓ | – |
 | `passRate` | ✓ | ✓ | ✓ |
 | `fullEnvironment` | – | ✓ | – |
 | `retries` | ✓ | ✓ | – |
@@ -583,7 +588,7 @@ Omit `sections` entirely and each template renders exactly this (✓ = on):
 
 - **`trend` needs `charts`.** The trend line, run-history, and flakiness table live inside the charts block, so `trend: true` does nothing with `charts: false` (it's coerced off). It also needs history data; keep the top-level `showTrend` option on.
 - **Inline failure detail needs `suiteBreakdown`.** `failureDeepDive` (failure detail), `failureAnalysis` (root-cause chip), and `slowTests` (`SLOW` badge) now render *inside* the suite breakdown, so they show nothing when `suiteBreakdown` is off. The chip (`failureAnalysis`) also needs `failureDeepDive`: it lives inside the failure detail, so it's coerced off without it.
-- **Data-gated sections**: `failureAnalysis`, `requirementsMatrix`, `slowTests`, `defectLog` render only when there's matching data (classified failures, tagged tests, a meaningful slow set, failures). Toggling them on with no data shows nothing.
+- **Data-gated sections**: `failureAnalysis`, `requirementsMatrix`, `slowTests`, `defectLog` render only when there's matching data (classified failures, tagged tests, a meaningful slow set, failures). Toggling them on with no data shows nothing. `browserMatrix` is data-gated too: it needs 2+ projects in the run (multiple Playwright projects, or `pytest --browser` given twice), so single-browser runs never show it.
 - **Render-layer only.** `sections` controls what's drawn, never what's collected; the data switches stay separate: `showTrend` (history), `flakinessTopN` (flaky-table size), `failureAnalysis.enabled` (classifier), `includeScreenshots` (images).
 - **Custom templates.** `sections` applies to the built-in templates; a custom `templatePath` controls its own layout (every section available) and ignores `sections`. You'll get a warning if both are set.
 

package/dist/cli/index.js

@@ -6518,6 +6518,7 @@ var init_sections = __esm({
       "defectLog",
       "briefBand",
       "runDiff",
+      "browserMatrix",
       // display modifiers
       "passRate",
       "fullEnvironment",
@@ -6542,6 +6543,7 @@ var init_sections = __esm({
         defectLog: false,
         briefBand: false,
         runDiff: true,
+        browserMatrix: false,
         passRate: true,
         fullEnvironment: false,
         retries: true,
@@ -6564,6 +6566,7 @@ var init_sections = __esm({
         defectLog: true,
         briefBand: false,
         runDiff: true,
+        browserMatrix: true,
         passRate: true,
         fullEnvironment: true,
         retries: true,
@@ -6591,6 +6594,7 @@ var init_sections = __esm({
         defectLog: false,
         briefBand: true,
         runDiff: true,
+        browserMatrix: false,
         passRate: true,
         fullEnvironment: false,
         retries: false,

package/dist/index.d.mts

@@ -1,11 +1,11 @@
 import { Reporter, FullConfig, Suite, TestCase, TestResult, TestStep, FullResult } from '@playwright/test/reporter';
 
 /**
- * Canonical section catalog. 15 block toggles (section present/absent) followed
+ * Canonical section catalog. 16 block toggles (section present/absent) followed
  * by 5 display modifiers (tweak a section that is on). Single source of truth —
  * the Zod schema in src/config/schema.ts builds its flag shape from this array.
  */
-declare const SECTION_KEYS: readonly ["coverPage", "analysisOneliner", "releaseGate", "summary", "charts", "trend", "requirementsMatrix", "ciEnvironment", "suiteBreakdown", "failureDeepDive", "failureAnalysis", "slowTests", "defectLog", "briefBand", "runDiff", "passRate", "fullEnvironment", "retries", "fullFailures", "stackTraces"];
+declare const SECTION_KEYS: readonly ["coverPage", "analysisOneliner", "releaseGate", "summary", "charts", "trend", "requirementsMatrix", "ciEnvironment", "suiteBreakdown", "failureDeepDive", "failureAnalysis", "slowTests", "defectLog", "briefBand", "runDiff", "browserMatrix", "passRate", "fullEnvironment", "retries", "fullFailures", "stackTraces"];
 type SectionKey = (typeof SECTION_KEYS)[number];
 /** A partial map of section flags — used both as the flat baseline and per-template. */
 type SectionFlags = Partial<Record<SectionKey, boolean>>;
@@ -406,18 +406,22 @@ interface ReporterOptions {
      * every chosen template; per-template keys (`minimal` / `detailed` /
      * `executive`) override the baseline for that template only.
      *
-     * 20 keys total: 15 block toggles, 5 display modifiers.
+     * 21 keys total: 16 block toggles, 5 display modifiers.
      *
      * Block toggles: `coverPage`, `analysisOneliner`, `releaseGate`, `summary`,
      * `charts`, `trend`, `requirementsMatrix`, `ciEnvironment`, `suiteBreakdown`,
      * `failureDeepDive`, `failureAnalysis`, `slowTests`, `defectLog`, `briefBand`,
-     * `runDiff`.
+     * `runDiff`, `browserMatrix`.
      * Display modifiers: `passRate`, `fullEnvironment`, `retries`, `fullFailures`,
      * `stackTraces`.
      *
      * `runDiff` shows new failures, fixed tests, and still-failing tests against
      * the previous run on the same branch; on by default in all three templates.
      *
+     * `browserMatrix` compares per-test outcomes across the run's projects
+     * (divergent tests first, identical ones collapsed to a count); on by default
+     * in `detailed` and self-hides when the run has fewer than two projects.
+     *
      * `trend` renders only when `charts` is also on and history data exists (`showTrend` enabled). Sections gated on data
      * (`failureAnalysis`, `requirementsMatrix`, `slowTests`, `defectLog`) appear
      * only when matching data exists — this option never changes data collection.

package/dist/index.d.ts

@@ -1,11 +1,11 @@
 import { Reporter, FullConfig, Suite, TestCase, TestResult, TestStep, FullResult } from '@playwright/test/reporter';
 
 /**
- * Canonical section catalog. 15 block toggles (section present/absent) followed
+ * Canonical section catalog. 16 block toggles (section present/absent) followed
  * by 5 display modifiers (tweak a section that is on). Single source of truth —
  * the Zod schema in src/config/schema.ts builds its flag shape from this array.
  */
-declare const SECTION_KEYS: readonly ["coverPage", "analysisOneliner", "releaseGate", "summary", "charts", "trend", "requirementsMatrix", "ciEnvironment", "suiteBreakdown", "failureDeepDive", "failureAnalysis", "slowTests", "defectLog", "briefBand", "runDiff", "passRate", "fullEnvironment", "retries", "fullFailures", "stackTraces"];
+declare const SECTION_KEYS: readonly ["coverPage", "analysisOneliner", "releaseGate", "summary", "charts", "trend", "requirementsMatrix", "ciEnvironment", "suiteBreakdown", "failureDeepDive", "failureAnalysis", "slowTests", "defectLog", "briefBand", "runDiff", "browserMatrix", "passRate", "fullEnvironment", "retries", "fullFailures", "stackTraces"];
 type SectionKey = (typeof SECTION_KEYS)[number];
 /** A partial map of section flags — used both as the flat baseline and per-template. */
 type SectionFlags = Partial<Record<SectionKey, boolean>>;
@@ -406,18 +406,22 @@ interface ReporterOptions {
      * every chosen template; per-template keys (`minimal` / `detailed` /
      * `executive`) override the baseline for that template only.
      *
-     * 20 keys total: 15 block toggles, 5 display modifiers.
+     * 21 keys total: 16 block toggles, 5 display modifiers.
      *
      * Block toggles: `coverPage`, `analysisOneliner`, `releaseGate`, `summary`,
      * `charts`, `trend`, `requirementsMatrix`, `ciEnvironment`, `suiteBreakdown`,
      * `failureDeepDive`, `failureAnalysis`, `slowTests`, `defectLog`, `briefBand`,
-     * `runDiff`.
+     * `runDiff`, `browserMatrix`.
      * Display modifiers: `passRate`, `fullEnvironment`, `retries`, `fullFailures`,
      * `stackTraces`.
      *
      * `runDiff` shows new failures, fixed tests, and still-failing tests against
      * the previous run on the same branch; on by default in all three templates.
      *
+     * `browserMatrix` compares per-test outcomes across the run's projects
+     * (divergent tests first, identical ones collapsed to a count); on by default
+     * in `detailed` and self-hides when the run has fewer than two projects.
+     *
      * `trend` renders only when `charts` is also on and history data exists (`showTrend` enabled). Sections gated on data
      * (`failureAnalysis`, `requirementsMatrix`, `slowTests`, `defectLog`) appear
      * only when matching data exists — this option never changes data collection.

package/dist/index.js

@@ -10355,6 +10355,7 @@ var SECTION_KEYS = [
   "defectLog",
   "briefBand",
   "runDiff",
+  "browserMatrix",
   // display modifiers
   "passRate",
   "fullEnvironment",
@@ -10379,6 +10380,7 @@ var SECTION_DEFAULTS = {
     defectLog: false,
     briefBand: false,
     runDiff: true,
+    browserMatrix: false,
     passRate: true,
     fullEnvironment: false,
     retries: true,
@@ -10401,6 +10403,7 @@ var SECTION_DEFAULTS = {
     defectLog: true,
     briefBand: false,
     runDiff: true,
+    browserMatrix: true,
     passRate: true,
     fullEnvironment: true,
     retries: true,
@@ -10428,6 +10431,7 @@ var SECTION_DEFAULTS = {
     defectLog: false,
     briefBand: true,
     runDiff: true,
+    browserMatrix: false,
     passRate: true,
     fullEnvironment: false,
     retries: false,
@@ -11296,7 +11300,7 @@ var isTopLevel = (flat, f) => !f.ancestors.some((a) => {
 var testStepDepth = (flat, f) => f.ancestors.reduce((n, a) => n + (flat[a].category === "test.step" ? 1 : 0), 0);
 function flatten(steps) {
   const out = [];
-  const walk = (nodes, depth, ancestors) => {
+  const walk2 = (nodes, depth, ancestors) => {
     for (const s of nodes ?? []) {
       const index = out.length;
       out.push({
@@ -11306,10 +11310,10 @@ function flatten(steps) {
         hasError: s.error != null,
         ancestors
       });
-      if (s.steps?.length) walk(s.steps, depth + 1, [...ancestors, index]);
+      if (s.steps?.length) walk2(s.steps, depth + 1, [...ancestors, index]);
     }
   };
-  walk(steps ?? [], 0, []);
+  walk2(steps ?? [], 0, []);
   return out;
 }
 function materialize(n, isFailing, depth) {
@@ -11917,7 +11921,7 @@ var ShardMerger = class {
   merge(filePaths) {
     if (filePaths.length === 0) throw new Error("shardResults: no valid JSON shard files found");
     const resolved = this.resolveShardPaths(filePaths);
-    const reports = this.loadReports(resolved);
+    const reports = this.loadReports(resolved).map((r) => this.normalizeReport(r));
     const projects = this.mergeProjects(reports);
     const stats = this.mergeStats(reports, projects);
     const failures = sortBySeverity(this.extractFailures(reports));
@@ -11955,6 +11959,49 @@ var ShardMerger = class {
       return report;
     });
   }
+  /**
+   * Two top-level shapes reach the merger. Our language adapters (Python,
+   * .NET) and the engine fixtures emit PROJECT wrapper suites (no `file`,
+   * children are the file suites). Playwright's own --reporter=json emits
+   * FILE suites at top level and hangs the project on each
+   * spec.tests[].projectName instead. Everything downstream (mergeProjects,
+   * extractFailures) assumes the wrapper shape, so real Playwright reports
+   * are pivoted into it here: one wrapper per project, file trees filtered
+   * to that project's tests. An unnamed project ('') renders as 'default',
+   * matching the live reporter path.
+   */
+  normalizeReport(report) {
+    if (report.suites.every((s) => s.file === void 0)) return report;
+    const wrappers = this.collectProjectNames(report).map((name) => ({
+      title: name === "" ? "default" : name,
+      specs: [],
+      suites: report.suites.map((fileSuite) => this.filterSuiteForProject(fileSuite, name)).filter((s) => s !== null)
+    })).filter((w) => w.suites.length > 0);
+    return { ...report, suites: wrappers };
+  }
+  /** Config order first, then any project discovered only on the tests. */
+  collectProjectNames(report) {
+    const names = [];
+    const push = (n) => {
+      if (!names.includes(n)) names.push(n);
+    };
+    for (const p of report.config?.projects ?? []) push(p.name ?? "");
+    const walk2 = (suite) => {
+      for (const spec of suite.specs ?? []) for (const t of spec.tests ?? []) push(t.projectName ?? "");
+      for (const child of suite.suites ?? []) walk2(child);
+    };
+    for (const s of report.suites) walk2(s);
+    return names.length > 0 ? names : [""];
+  }
+  filterSuiteForProject(suite, project) {
+    const specs = (suite.specs ?? []).map((spec) => ({
+      ...spec,
+      tests: (spec.tests ?? []).filter((t) => (t.projectName ?? "") === project)
+    })).filter((spec) => spec.tests.length > 0);
+    const children = (suite.suites ?? []).map((child) => this.filterSuiteForProject(child, project)).filter((c) => c !== null);
+    if (specs.length === 0 && children.length === 0) return null;
+    return { ...suite, specs, suites: children };
+  }
   mergeProjects(reports) {
     const projectMap = /* @__PURE__ */ new Map();
     for (const report of reports) {
@@ -13735,6 +13782,66 @@ function sourceTally(classified) {
   return ["rule", "retry", "local", "base"].filter((s) => counts.has(s)).map((s) => `${counts.get(s)} ${s}`).join(" / ");
 }
 
+// src/collector/BrowserMatrix.ts
+init_cjs_shims();
+var DEFAULT_ROW_CAP = 40;
+function buildBrowserMatrix(projects, cap = DEFAULT_ROW_CAP) {
+  if (projects.length < 2) return void 0;
+  let projectNames = projects.map((p, i) => p.name || `project ${i + 1}`);
+  const byTest = /* @__PURE__ */ new Map();
+  projects.forEach((project, column) => {
+    walk(project.suites, [], (path15, test) => {
+      const key = [...path15, test.title].join(" > ");
+      let statuses = byTest.get(key);
+      if (!statuses) {
+        statuses = projectNames.map(() => null);
+        byTest.set(key, statuses);
+      }
+      statuses[column] = test.status;
+    });
+  });
+  const divergent = [];
+  let identicalCount = 0;
+  const contributing = /* @__PURE__ */ new Set();
+  for (const [title, statuses] of byTest) {
+    const present = statuses.filter((s) => s !== null);
+    if (present.length < 2) continue;
+    statuses.forEach((s, i) => {
+      if (s !== null) contributing.add(i);
+    });
+    if (present.every((s) => s === present[0])) {
+      identicalCount++;
+    } else {
+      divergent.push({ title, statuses });
+    }
+  }
+  if (divergent.length === 0 && identicalCount === 0) return void 0;
+  if (contributing.size < projectNames.length) {
+    const keep = [...contributing].sort((a, b) => a - b);
+    projectNames = keep.map((i) => projectNames[i]);
+    for (const row of divergent) row.statuses = keep.map((i) => row.statuses[i]);
+  }
+  divergent.sort((a, b) => severityRank(b) - severityRank(a) || a.title.localeCompare(b.title));
+  return {
+    projects: projectNames,
+    rows: divergent.slice(0, cap),
+    divergentCount: divergent.length,
+    identicalCount,
+    truncated: divergent.length > cap,
+    hasFailures: divergent.some((r) => severityRank(r) > 0)
+  };
+}
+function severityRank(row) {
+  return row.statuses.some((s) => s === "failed" || s === "timedOut") ? 1 : 0;
+}
+function walk(suites, path15, visit) {
+  for (const suite of suites) {
+    const next = [...path15, suite.title];
+    for (const test of suite.tests) visit(next, test);
+    walk(suite.suites, next, visit);
+  }
+}
+
 // src/redact/redact-report.ts
 init_cjs_shims();
 function redactReportData(data, r) {
@@ -13765,6 +13872,10 @@ function redactReportData(data, r) {
   data.briefSentence = rd(data.briefSentence);
   data.charts.suiteResults.forEach((s) => s.label = r.redactText(s.label));
   data.charts.flakinessTable?.forEach((row) => row.title = r.redactText(row.title));
+  if (data.browserMatrix) {
+    data.browserMatrix.projects = data.browserMatrix.projects.map((p) => r.redactText(p));
+    data.browserMatrix.rows.forEach((row) => row.title = r.redactText(row.title));
+  }
   data.environment.buildUrl = r.redactText(data.environment.buildUrl);
 }
 function redactStep(s, r) {
@@ -14505,7 +14616,8 @@ function buildReportData(collected, licenseInfo, hooks) {
     failures,
     environment,
     charts,
-    runDiff
+    runDiff,
+    browserMatrix: buildBrowserMatrix(projects)
   };
 }
 function resolveProjectNameFromPkg(cwd) {
@@ -14980,7 +15092,7 @@ var MAX_ERR_SNIPPET = 4096;
 var isHook = (c) => c === "hook" || c === "fixture";
 function flatten2(steps) {
   const out = [];
-  const walk = (nodes, depth, ancestors) => {
+  const walk2 = (nodes, depth, ancestors) => {
     for (const s of nodes ?? []) {
       const index = out.length;
       out.push({
@@ -14990,10 +15102,10 @@ function flatten2(steps) {
         hasError: s.error != null,
         ancestors
       });
-      if (s.steps?.length) walk(s.steps, depth + 1, [...ancestors, index]);
+      if (s.steps?.length) walk2(s.steps, depth + 1, [...ancestors, index]);
     }
   };
-  walk(steps ?? [], 0, []);
+  walk2(steps ?? [], 0, []);
   return out;
 }
 var inHookSubtree2 = (flat, f) => isHook(f.category) || f.ancestors.some((a) => isHook(flat[a].category));
@@ -15257,7 +15369,7 @@ var PdfReporter = class {
     if (this.options.redact.enabled) {
       this.redactor = new Redactor(this.options.redact);
     }
-    this.version = true ? "0.28.0" : "0.x";
+    this.version = true ? "0.29.0" : "0.x";
     logger.info(`@reportforge/playwright-pdf v${this.version} initialised`);
     this.licenseClient = new LicenseClient({
       licenseKey: this.options.licenseKey,

package/dist/index.mjs

@@ -10356,6 +10356,7 @@ var SECTION_KEYS = [
   "defectLog",
   "briefBand",
   "runDiff",
+  "browserMatrix",
   // display modifiers
   "passRate",
   "fullEnvironment",
@@ -10380,6 +10381,7 @@ var SECTION_DEFAULTS = {
     defectLog: false,
     briefBand: false,
     runDiff: true,
+    browserMatrix: false,
     passRate: true,
     fullEnvironment: false,
     retries: true,
@@ -10402,6 +10404,7 @@ var SECTION_DEFAULTS = {
     defectLog: true,
     briefBand: false,
     runDiff: true,
+    browserMatrix: true,
     passRate: true,
     fullEnvironment: true,
     retries: true,
@@ -10429,6 +10432,7 @@ var SECTION_DEFAULTS = {
     defectLog: false,
     briefBand: true,
     runDiff: true,
+    browserMatrix: false,
     passRate: true,
     fullEnvironment: false,
     retries: false,
@@ -11297,7 +11301,7 @@ var isTopLevel = (flat, f) => !f.ancestors.some((a) => {
 var testStepDepth = (flat, f) => f.ancestors.reduce((n, a) => n + (flat[a].category === "test.step" ? 1 : 0), 0);
 function flatten(steps) {
   const out = [];
-  const walk = (nodes, depth, ancestors) => {
+  const walk2 = (nodes, depth, ancestors) => {
     for (const s of nodes ?? []) {
       const index = out.length;
       out.push({
@@ -11307,10 +11311,10 @@ function flatten(steps) {
         hasError: s.error != null,
         ancestors
       });
-      if (s.steps?.length) walk(s.steps, depth + 1, [...ancestors, index]);
+      if (s.steps?.length) walk2(s.steps, depth + 1, [...ancestors, index]);
     }
   };
-  walk(steps ?? [], 0, []);
+  walk2(steps ?? [], 0, []);
   return out;
 }
 function materialize(n, isFailing, depth) {
@@ -11918,7 +11922,7 @@ var ShardMerger = class {
   merge(filePaths) {
     if (filePaths.length === 0) throw new Error("shardResults: no valid JSON shard files found");
     const resolved = this.resolveShardPaths(filePaths);
-    const reports = this.loadReports(resolved);
+    const reports = this.loadReports(resolved).map((r) => this.normalizeReport(r));
     const projects = this.mergeProjects(reports);
     const stats = this.mergeStats(reports, projects);
     const failures = sortBySeverity(this.extractFailures(reports));
@@ -11956,6 +11960,49 @@ var ShardMerger = class {
       return report;
     });
   }
+  /**
+   * Two top-level shapes reach the merger. Our language adapters (Python,
+   * .NET) and the engine fixtures emit PROJECT wrapper suites (no `file`,
+   * children are the file suites). Playwright's own --reporter=json emits
+   * FILE suites at top level and hangs the project on each
+   * spec.tests[].projectName instead. Everything downstream (mergeProjects,
+   * extractFailures) assumes the wrapper shape, so real Playwright reports
+   * are pivoted into it here: one wrapper per project, file trees filtered
+   * to that project's tests. An unnamed project ('') renders as 'default',
+   * matching the live reporter path.
+   */
+  normalizeReport(report) {
+    if (report.suites.every((s) => s.file === void 0)) return report;
+    const wrappers = this.collectProjectNames(report).map((name) => ({
+      title: name === "" ? "default" : name,
+      specs: [],
+      suites: report.suites.map((fileSuite) => this.filterSuiteForProject(fileSuite, name)).filter((s) => s !== null)
+    })).filter((w) => w.suites.length > 0);
+    return { ...report, suites: wrappers };
+  }
+  /** Config order first, then any project discovered only on the tests. */
+  collectProjectNames(report) {
+    const names = [];
+    const push = (n) => {
+      if (!names.includes(n)) names.push(n);
+    };
+    for (const p of report.config?.projects ?? []) push(p.name ?? "");
+    const walk2 = (suite) => {
+      for (const spec of suite.specs ?? []) for (const t of spec.tests ?? []) push(t.projectName ?? "");
+      for (const child of suite.suites ?? []) walk2(child);
+    };
+    for (const s of report.suites) walk2(s);
+    return names.length > 0 ? names : [""];
+  }
+  filterSuiteForProject(suite, project) {
+    const specs = (suite.specs ?? []).map((spec) => ({
+      ...spec,
+      tests: (spec.tests ?? []).filter((t) => (t.projectName ?? "") === project)
+    })).filter((spec) => spec.tests.length > 0);
+    const children = (suite.suites ?? []).map((child) => this.filterSuiteForProject(child, project)).filter((c) => c !== null);
+    if (specs.length === 0 && children.length === 0) return null;
+    return { ...suite, specs, suites: children };
+  }
   mergeProjects(reports) {
     const projectMap = /* @__PURE__ */ new Map();
     for (const report of reports) {
@@ -13736,6 +13783,66 @@ function sourceTally(classified) {
   return ["rule", "retry", "local", "base"].filter((s) => counts.has(s)).map((s) => `${counts.get(s)} ${s}`).join(" / ");
 }
 
+// src/collector/BrowserMatrix.ts
+init_esm_shims();
+var DEFAULT_ROW_CAP = 40;
+function buildBrowserMatrix(projects, cap = DEFAULT_ROW_CAP) {
+  if (projects.length < 2) return void 0;
+  let projectNames = projects.map((p, i) => p.name || `project ${i + 1}`);
+  const byTest = /* @__PURE__ */ new Map();
+  projects.forEach((project, column) => {
+    walk(project.suites, [], (path16, test) => {
+      const key = [...path16, test.title].join(" > ");
+      let statuses = byTest.get(key);
+      if (!statuses) {
+        statuses = projectNames.map(() => null);
+        byTest.set(key, statuses);
+      }
+      statuses[column] = test.status;
+    });
+  });
+  const divergent = [];
+  let identicalCount = 0;
+  const contributing = /* @__PURE__ */ new Set();
+  for (const [title, statuses] of byTest) {
+    const present = statuses.filter((s) => s !== null);
+    if (present.length < 2) continue;
+    statuses.forEach((s, i) => {
+      if (s !== null) contributing.add(i);
+    });
+    if (present.every((s) => s === present[0])) {
+      identicalCount++;
+    } else {
+      divergent.push({ title, statuses });
+    }
+  }
+  if (divergent.length === 0 && identicalCount === 0) return void 0;
+  if (contributing.size < projectNames.length) {
+    const keep = [...contributing].sort((a, b) => a - b);
+    projectNames = keep.map((i) => projectNames[i]);
+    for (const row of divergent) row.statuses = keep.map((i) => row.statuses[i]);
+  }
+  divergent.sort((a, b) => severityRank(b) - severityRank(a) || a.title.localeCompare(b.title));
+  return {
+    projects: projectNames,
+    rows: divergent.slice(0, cap),
+    divergentCount: divergent.length,
+    identicalCount,
+    truncated: divergent.length > cap,
+    hasFailures: divergent.some((r) => severityRank(r) > 0)
+  };
+}
+function severityRank(row) {
+  return row.statuses.some((s) => s === "failed" || s === "timedOut") ? 1 : 0;
+}
+function walk(suites, path16, visit) {
+  for (const suite of suites) {
+    const next = [...path16, suite.title];
+    for (const test of suite.tests) visit(next, test);
+    walk(suite.suites, next, visit);
+  }
+}
+
 // src/redact/redact-report.ts
 init_esm_shims();
 function redactReportData(data, r) {
@@ -13766,6 +13873,10 @@ function redactReportData(data, r) {
   data.briefSentence = rd(data.briefSentence);
   data.charts.suiteResults.forEach((s) => s.label = r.redactText(s.label));
   data.charts.flakinessTable?.forEach((row) => row.title = r.redactText(row.title));
+  if (data.browserMatrix) {
+    data.browserMatrix.projects = data.browserMatrix.projects.map((p) => r.redactText(p));
+    data.browserMatrix.rows.forEach((row) => row.title = r.redactText(row.title));
+  }
   data.environment.buildUrl = r.redactText(data.environment.buildUrl);
 }
 function redactStep(s, r) {
@@ -14506,7 +14617,8 @@ function buildReportData(collected, licenseInfo, hooks) {
     failures,
     environment,
     charts,
-    runDiff
+    runDiff,
+    browserMatrix: buildBrowserMatrix(projects)
   };
 }
 function resolveProjectNameFromPkg(cwd) {
@@ -14981,7 +15093,7 @@ var MAX_ERR_SNIPPET = 4096;
 var isHook = (c) => c === "hook" || c === "fixture";
 function flatten2(steps) {
   const out = [];
-  const walk = (nodes, depth, ancestors) => {
+  const walk2 = (nodes, depth, ancestors) => {
     for (const s of nodes ?? []) {
       const index = out.length;
       out.push({
@@ -14991,10 +15103,10 @@ function flatten2(steps) {
         hasError: s.error != null,
         ancestors
       });
-      if (s.steps?.length) walk(s.steps, depth + 1, [...ancestors, index]);
+      if (s.steps?.length) walk2(s.steps, depth + 1, [...ancestors, index]);
     }
   };
-  walk(steps ?? [], 0, []);
+  walk2(steps ?? [], 0, []);
   return out;
 }
 var inHookSubtree2 = (flat, f) => isHook(f.category) || f.ancestors.some((a) => isHook(flat[a].category));
@@ -15258,7 +15370,7 @@ var PdfReporter = class {
     if (this.options.redact.enabled) {
       this.redactor = new Redactor(this.options.redact);
     }
-    this.version = true ? "0.28.0" : "0.x";
+    this.version = true ? "0.29.0" : "0.x";
     logger.info(`@reportforge/playwright-pdf v${this.version} initialised`);
     this.licenseClient = new LicenseClient({
       licenseKey: this.options.licenseKey,

package/dist/templates/layouts/detailed.hbs

@@ -26,6 +26,8 @@
 
 {{#if sections.runDiff}}{{> run-diff variant='full'}}{{/if}}
 
+{{#if sections.browserMatrix}}{{> browser-matrix}}{{/if}}
+
 {{#if sections.requirementsMatrix}}{{> requirements-matrix }}{{/if}}
 
 {{#if sections.ciEnvironment}}{{> ci-environment options=(obj showFullEnvironment=sections.fullEnvironment) }}{{/if}}

package/dist/templates/layouts/executive.hbs

@@ -26,6 +26,8 @@
 
 {{#if sections.charts}}{{> charts options=(obj showCharts=true showTrend=sections.trend heroTrend=true) }}{{/if}}
 
+{{#if sections.browserMatrix}}{{> browser-matrix}}{{/if}}
+
 {{#if sections.requirementsMatrix}}{{> requirements-matrix }}{{/if}}
 
 {{#if sections.defectLog}}{{> defect-log }}{{/if}}

package/dist/templates/layouts/minimal.hbs

@@ -26,6 +26,8 @@
 
 {{#if sections.charts}}{{> charts options=(obj showCharts=true showTrend=sections.trend) }}{{/if}}
 
+{{#if sections.browserMatrix}}{{> browser-matrix}}{{/if}}
+
 {{#if sections.requirementsMatrix}}{{> requirements-matrix }}{{/if}}
 
 {{#if sections.suiteBreakdown}}{{> suite-breakdown options=(obj showRetries=sections.retries showFullFailures=sections.fullFailures showStackTraces=sections.stackTraces) }}{{/if}}

package/dist/templates/partials/browser-matrix.hbs

@@ -0,0 +1,34 @@
+{{#if browserMatrix}}
+<div class="section section--browser-matrix">
+  <h2>Cross-Browser Comparison</h2>
+  {{#if browserMatrix.rows.length}}
+  <p class="browser-matrix__intro">{{browserMatrix.divergentCount}} {{pluralise browserMatrix.divergentCount "test behaves" "tests behave"}} differently across the run's {{browserMatrix.projects.length}} projects.{{#if browserMatrix.hasFailures}} Failures first.{{/if}}</p>
+  <table class="browser-matrix__table">
+    <thead>
+      <tr>
+        <th>Test</th>
+        {{#each browserMatrix.projects}}<th class="browser-matrix__col">{{this}}</th>{{/each}}
+      </tr>
+    </thead>
+    <tbody>
+      {{#each browserMatrix.rows}}
+      <tr class="no-break">
+        <td class="browser-matrix__title">{{title}}</td>
+        {{#each statuses}}
+        <td class="browser-matrix__cell">{{#if this}}{{statusBadge this}}{{else}}<span class="browser-matrix__na">n/a</span>{{/if}}</td>
+        {{/each}}
+      </tr>
+      {{/each}}
+    </tbody>
+  </table>
+  {{#if browserMatrix.truncated}}
+  <div class="browser-matrix__note">+ {{subtract browserMatrix.divergentCount browserMatrix.rows.length}} more divergent {{pluralise (subtract browserMatrix.divergentCount browserMatrix.rows.length) "test" "tests"}} not shown</div>
+  {{/if}}
+  {{#if browserMatrix.identicalCount}}
+  <div class="browser-matrix__note">{{browserMatrix.identicalCount}} {{pluralise browserMatrix.identicalCount "test behaves" "tests behave"}} the same in every project.</div>
+  {{/if}}
+  {{else}}
+  <p class="browser-matrix__intro">All {{browserMatrix.identicalCount}} comparable {{pluralise browserMatrix.identicalCount "test behaves" "tests behave"}} the same across the run's {{browserMatrix.projects.length}} projects.</p>
+  {{/if}}
+</div>
+{{/if}}

package/dist/templates/styles/base.css

@@ -954,3 +954,27 @@ code {
 /* The aggregate failure-analysis cluster section was retired — root causes now
    render as inline per-test chips in the suite breakdown (.root-cause-chip above).
    The .analysis-oneliner one-liner lives further up and is still used by executive. */
+
+/* ── Cross-browser comparison matrix (sections.browserMatrix) ───────────── */
+/* Divergent tests only; identical tests collapse into the note line. Cells
+   reuse the .badge--* status palette so the grid reads like the rest of the
+   report. Lives here (not a template css) because the section is toggle-able
+   across all three layouts. */
+.browser-matrix__intro {
+  font-size: 10px; color: var(--text-3); margin-bottom: 12px;
+}
+.browser-matrix__table th.browser-matrix__col {
+  width: 88px; text-align: center; text-transform: capitalize;
+}
+.browser-matrix__title {
+  font-size: 9.5px; color: var(--text-1); word-break: break-word;
+}
+.browser-matrix__cell { text-align: center; }
+.browser-matrix__cell .badge { min-width: 0; font-size: 7.5px; }
+.browser-matrix__na {
+  font-family: 'JetBrains Mono', monospace;
+  font-size: 8.5px; color: var(--text-3);
+}
+.browser-matrix__note {
+  font-size: 9px; color: var(--text-3); margin-top: 6px;
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@reportforge/playwright-pdf",
-  "version": "0.28.0",
+  "version": "0.29.0",
   "description": "Playwright Test reporter that generates designed PDF reports: minimal, detailed, and executive templates with CI/CD integrations",
   "license": "Elastic-2.0",
   "author": "ReportForge",