fix: memory leak caused by storing base64 encoded files recieved by CDP Network.requestWillBeSent (#22460)

* filtering out data urls from networkRequestWillBeSent CDP requests

* fix: memory leak caused by storing base64 encoded files recieved by CDP Network.requestWillBeSent

* Update packages/server/lib/browsers/cdp_automation.ts

grammar :)

Co-authored-by: Sam Tsai <samtsai@gmail.com>

* Update packages/server/lib/browsers/cdp_automation.ts

grammar :)

Co-authored-by: Emily Rohrbough <emilyrohrbough@users.noreply.github.com>

* added unit tests

Co-authored-by: Matt Henkes <mjhenkes@gmail.com>
Co-authored-by: Sam Tsai <samtsai@gmail.com>
Co-authored-by: Emily Rohrbough <emilyrohrbough@users.noreply.github.com>

packages/server/lib/browsers/cdp_automation.ts

@@ -202,6 +202,15 @@ export class CdpAutomation {
     // in Firefox, the hash is incorrectly included in the URL: https://bugzilla.mozilla.org/show_bug.cgi?id=1715366
     if (url.includes('#')) url = url.slice(0, url.indexOf('#'))
 
+    // Filter out "data:" urls from being cached - fixes: https://github.com/cypress-io/cypress/issues/17853
+    // Chrome sends `Network.requestWillBeSent` events with data urls which won't actually be fetched
+    // Example data url: "data:font/woff;base64,<base64 encoded string>"
+    if (url.startsWith('data:')) {
+      debugVerbose('skipping `data:` url %s', url)
+
+      return
+    }
+
     // Firefox: https://searchfox.org/mozilla-central/rev/98a9257ca2847fad9a19631ac76199474516b31e/remote/cdp/domains/parent/Network.jsm#397
     // Firefox lacks support for urlFragment and initiator, two nice-to-haves
     const browserPreRequest: BrowserPreRequest = {

packages/server/test/unit/browsers/cdp_automation_spec.ts

@@ -71,15 +71,103 @@ context('lib/browsers/cdp_automation', () => {
       this.sendDebuggerCommand = sinon.stub()
       this.onFn = sinon.stub()
       this.sendCloseTargetCommand = sinon.stub()
+      this.automation = {
+        onBrowserPreRequest: sinon.stub(),
+        onRequestEvent: sinon.stub(),
+      }
 
-      this.automation = await CdpAutomation.create(this.sendDebuggerCommand, this.onFn, this.sendCloseTargetCommand, null, false)
+      this.cdpAutomation = await CdpAutomation.create(this.sendDebuggerCommand, this.onFn, this.sendCloseTargetCommand, this.automation, false)
 
       this.sendDebuggerCommand
       .throws(new Error('not stubbed'))
       .withArgs('Browser.getVersion')
       .resolves()
 
-      this.onRequest = this.automation.onRequest
+      this.onRequest = this.cdpAutomation.onRequest
+    })
+
+    describe('.onNetworkRequestWillBeSent', function () {
+      it('triggers onBrowserPreRequest', function () {
+        const browserPreRequest = {
+          requestId: '0',
+          type: 'other',
+          request: {
+            method: 'GET',
+            url: 'https://www.google.com',
+            headers: {},
+          },
+        }
+
+        this.onFn
+        .withArgs('Network.requestWillBeSent')
+        .yield(browserPreRequest)
+
+        expect(this.automation.onBrowserPreRequest).to.have.been.calledWith({
+          requestId: browserPreRequest.requestId,
+          method: browserPreRequest.request.method,
+          url: browserPreRequest.request.url,
+          headers: browserPreRequest.request.headers,
+          resourceType: browserPreRequest.type,
+          originalResourceType: browserPreRequest.type,
+        })
+      })
+
+      it('removes # from a url', function () {
+        const browserPreRequest = {
+          requestId: '0',
+          type: 'other',
+          request: {
+            method: 'GET',
+            url: 'https://www.google.com/foo#',
+            headers: {},
+          },
+        }
+
+        this.onFn
+        .withArgs('Network.requestWillBeSent')
+        .yield(browserPreRequest)
+
+        expect(this.automation.onBrowserPreRequest).to.have.been.calledWith({
+          requestId: browserPreRequest.requestId,
+          method: browserPreRequest.request.method,
+          url: 'https://www.google.com/foo', // we only care about the url
+          headers: browserPreRequest.request.headers,
+          resourceType: browserPreRequest.type,
+          originalResourceType: browserPreRequest.type,
+        })
+      })
+
+      it('ignore events with data urls', function () {
+        this.onFn
+        .withArgs('Network.requestWillBeSent')
+        .yield({ request: { url: 'data:font;base64' } })
+
+        expect(this.automation.onBrowserPreRequest).to.not.be.called
+      })
+    })
+
+    describe('.onResponseReceived', function () {
+      it('triggers onRequestEvent', function () {
+        const browserResponseReceived = {
+          requestId: '0',
+          response: {
+            status: 200,
+            headers: {},
+          },
+        }
+
+        this.onFn
+        .withArgs('Network.responseReceived')
+        .yield(browserResponseReceived)
+
+        expect(this.automation.onRequestEvent).to.have.been.calledWith(
+          'response:received', {
+            requestId: browserResponseReceived.requestId,
+            status: browserResponseReceived.response.status,
+            headers: browserResponseReceived.response.headers,
+          },
+        )
+      })
     })
 
     describe('get:cookies', () => {