fix(trace): use pathToFileURL for Windows-safe URLs; scan raw HTML in pageInfo

- Replace `file://${path}` template literals in lib/utils/trace.js
  fileToUrl() and bin/mcp-server.js with pathToFileURL().href so paths
  like `C:\foo\bar` are encoded correctly on Windows.
- captureSnapshot now also returns out.htmlRaw (pre-cleanHtml source) so
  consumers can scan classes that the trash-class filter would strip.
- pageInfo's error-class scan now reads htmlRaw, restoring detection of
  errorClasses containing digits/prefixes like `text-error` or `alert-1`.

Author: DavertMik

bin/mcp-server.js

@@ -14,7 +14,7 @@ import {
   writeTraceMarkdown,
 } from '../lib/utils/trace.js'
 import event from '../lib/event.js'
-import { fileURLToPath } from 'url'
+import { fileURLToPath, pathToFileURL } from 'url'
 import { dirname, resolve as resolvePath } from 'path'
 import path from 'path'
 import { spawn } from 'child_process'
@@ -453,7 +453,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
             text: JSON.stringify({
               status: 'success',
               dir,
-              traceFile: `file://${traceFile}`,
+              traceFile: pathToFileURL(traceFile).href,
               artifacts: artifactsToFileUrls(captured, dir),
             }, null, 2),
           }],
@@ -551,7 +551,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
           error: result.error,
         })
         result.dir = traceDir
-        result.traceFile = `file://${traceFile}`
+        result.traceFile = pathToFileURL(traceFile).href
 
         return { content: [{ type: 'text', text: JSON.stringify(result, null, 2) }] }
       }
@@ -642,10 +642,10 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
               if (t.err) r.error = t.err.message
 
               if (t.artifacts?.aiTrace) {
-                r.traceFile = `file://${t.artifacts.aiTrace}`
+                r.traceFile = pathToFileURL(t.artifacts.aiTrace).href
               }
-              if (t.artifacts?.har) r.har = `file://${t.artifacts.har}`
-              if (t.artifacts?.trace) r.trace = `file://${t.artifacts.trace}`
+              if (t.artifacts?.har) r.har = pathToFileURL(t.artifacts.har).href
+              if (t.artifacts?.trace) r.trace = pathToFileURL(t.artifacts.trace).href
 
               if (!t.artifacts?.aiTrace) {
                 try {
@@ -664,7 +664,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
                       captured,
                       error: r.error,
                     })
-                    r.traceFile = `file://${tracePath}`
+                    r.traceFile = pathToFileURL(tracePath).href
                   }
                 } catch {}
               }

lib/plugin/pageInfo.js

@@ -9,6 +9,7 @@ import { output } from '../index.js'
 import store from '../store.js'
 import { humanizeString, ucfirst } from '../utils.js'
 import { testToFileName } from '../mocha/test.js'
+
 const defaultConfig = {
   errorClasses: ['error', 'warning', 'alert', 'danger'],
   browserLogs: ['error'],
@@ -59,15 +60,21 @@ export default function (config = {}) {
         if (captured.html) {
           const htmlPath = path.join(store.outputDir, captured.html)
           pageState.htmlSnapshot = htmlPath
-          try {
-            const html = fs.readFileSync(htmlPath, 'utf8')
-            const errors = scanForErrorMessages(html, config.errorClasses)
-            if (errors.length) {
-              output.debug('Detected errors in HTML code')
-              errors.forEach(error => output.debug(error))
-              pageState.htmlErrors = errors
-            }
-          } catch {}
+          // Scan raw HTML (pre-cleanHtml) so error classes containing digits
+          // or trash-class prefixes aren't stripped before detection.
+          const htmlForScan = captured.htmlRaw || (() => {
+            try { return fs.readFileSync(htmlPath, 'utf8') } catch { return '' }
+          })()
+          if (htmlForScan) {
+            try {
+              const errors = scanForErrorMessages(htmlForScan, config.errorClasses)
+              if (errors.length) {
+                output.debug('Detected errors in HTML code')
+                errors.forEach(error => output.debug(error))
+                pageState.htmlErrors = errors
+              }
+            } catch {}
+          }
         }
 
         if (captured.aria) {

lib/utils/trace.js

@@ -1,6 +1,7 @@
 import crypto from 'crypto'
 import fs from 'fs'
 import path from 'path'
+import { pathToFileURL } from 'url'
 import Container from '../container.js'
 import { clearString } from '../utils.js'
 import { formatHtml } from '../html.js'
@@ -63,7 +64,7 @@ export function artifactLinks(artifacts, { indent = '  ', consoleCount } = {}) {
 }
 
 export function fileToUrl(dir, basename) {
-  return `file://${path.join(dir, basename)}`
+  return pathToFileURL(path.join(dir, basename)).href
 }
 
 export function writeTraceMarkdown({ dir, title, file, durationMs, commands, captured, error }) {
@@ -193,6 +194,9 @@ export async function captureSnapshot(helper, {
       const file = `${prefix}_page.html`
       fs.writeFileSync(path.join(dir, file), formatted)
       out.html = file
+      // Expose pre-cleanup HTML for consumers that need to inspect classes
+      // stripped by cleanHtml (e.g. pageInfo's error-class scan).
+      out.htmlRaw = html
     } catch {}
   }
 

test/unit/utils/trace_test.js

@@ -161,6 +161,11 @@ describe('lib/utils/trace.js', () => {
     it('joins dir + basename and prefixes with file://', () => {
       expect(fileToUrl('/output/run', 'p.html')).to.equal('file:///output/run/p.html')
     })
+
+    it('encodes spaces and special characters via pathToFileURL', () => {
+      const url = fileToUrl('/output/run dir', 'p age.html')
+      expect(url).to.equal('file:///output/run%20dir/p%20age.html')
+    })
   })
 
   describe('artifactsToFileUrls', () => {
@@ -339,6 +344,16 @@ describe('lib/utils/trace.js', () => {
       expect(html.split('\n').length).to.be.greaterThan(2)
     })
 
+    it('exposes pre-cleanup HTML in out.htmlRaw for class-scanning consumers', async () => {
+      const raw = '<html><body><div class="text-error alert-1">boom</div></body></html>'
+      const helper = fullHelper({ grabSource: sinon.stub().resolves(raw) })
+      const out = await captureSnapshot(helper, { dir, prefix: 'p' })
+      expect(out.htmlRaw).to.equal(raw)
+      const formatted = fs.readFileSync(path.join(dir, 'p_page.html'), 'utf8')
+      expect(formatted).to.not.include('text-error')
+      expect(formatted).to.not.include('alert-1')
+    })
+
     it('normalizes Playwright ConsoleMessage objects to {type, text}', async () => {
       const playwrightLog = {
         type: () => 'error',