Adding tests for related log importer PR here: https://github.com/piwik/piwik-log-analytics/pull/94

tests/PHPUnit/Fixtures/ManySitesImportedLogs.php

@@ -192,7 +192,7 @@ class ManySitesImportedLogs extends Fixture
      * Logs a couple visits for the site we created and two new sites that do not
      * exist yet. Visits are from Aug 12, 13 & 14 of 2012.
      */
-    public function logVisitsWithDynamicResolver()
+    public function logVisitsWithDynamicResolver($maxPayloadSize = 1)
     {
         $logFile = PIWIK_INCLUDE_PATH . '/tests/resources/access-logs/fake_logs_dynamic.log'; # log file
 
@@ -201,8 +201,8 @@ class ManySitesImportedLogs extends Fixture
         $opts = array('--add-sites-new-hosts'       => false,
                       '--enable-testmode'           => false,
                       '--recorders'                 => '1',
-                      '--recorder-max-payload-size' => '1');
-        self::executeLogImporter($logFile, $opts);
+                      '--recorder-max-payload-size' => $maxPayloadSize);
+        return implode("\n", self::executeLogImporter($logFile, $opts));
     }
 
     /**

tests/PHPUnit/System/ImportLogsTest.php

@@ -8,11 +8,14 @@
 namespace Piwik\Tests\System;
 
 use Piwik\Access;
+use Piwik\Common;
 use Piwik\Plugins\SitesManager\API;
 use Piwik\Tests\Framework\Fixture;
 use Piwik\Tests\Framework\TestCase\SystemTestCase;
 use Piwik\Tests\Fixtures\ManySitesImportedLogs;
 use Piwik\Tests\Framework\TestingEnvironmentVariables;
+use Piwik\Tracker\Request;
+use Piwik\Tracker\RequestSet;
 
 /**
  * Tests the log importer.
@@ -103,10 +106,15 @@ class ImportLogsTest extends SystemTestCase
     /**
      * NOTE: This test must be last since the new sites that get added are added in
      *       random order.
+     * NOTE: This test combines two tests in order to avoid executing the log importer another time.
+     *       If the log importer were refactored, the invalid requests test could be a unit test in
+     *       python.
      */
-    public function testDynamicResolverSitesCreated()
+    public function test_LogImporter_CreatesSitesWhenDynamicResolverUsed_AndReportsOnInvalidRequests()
     {
-        self::$fixture->logVisitsWithDynamicResolver();
+        $this->simulateInvalidTrackerRequest();
+
+        $output = self::$fixture->logVisitsWithDynamicResolver($maxPayloadSize = 3);
 
         // reload access so new sites are viewable
         Access::getInstance()->setSuperUserAccess(true);
@@ -120,6 +128,10 @@ class ImportLogsTest extends SystemTestCase
 
         $whateverDotCom = API::getInstance()->getSitesIdFromSiteUrl('http://whatever.com');
         $this->assertEquals(1, count($whateverDotCom));
+
+        // make sure invalid requests are reported correctly
+        $this->assertContains('The Piwik tracker identified 2 invalid requests on lines: 10, 11', $output);
+        $this->assertContains("The following lines were not tracked by Piwik, either due to a malformed tracker request or error in the tracker:\n\n10, 11", $output);
     }
 
     public function test_LogImporter_RetriesWhenServerFails()
@@ -163,23 +175,53 @@ class ImportLogsTest extends SystemTestCase
     {
         $testingEnvironment = new TestingEnvironmentVariables();
         $testingEnvironment->_triggerTrackerFailure = null;
+        $testingEnvironment->_triggerInvalidRequests = null;
         $testingEnvironment->save();
     }
 
+    private function simulateInvalidTrackerRequest()
+    {
+        $testEnvironment = new TestingEnvironmentVariables();
+        $testEnvironment->_triggerInvalidRequests = true;
+        $testEnvironment->save();
+    }
+
     public static function provideContainerConfigBeforeClass()
     {
         $result = array();
+        $observers = array();
 
         $testingEnvironment = new TestingEnvironmentVariables();
         if ($testingEnvironment->_triggerTrackerFailure) {
-            $result['observers.global'] = \DI\add(array(
-                array('Tracker.newHandler', function () {
-                    @http_response_code(500);
+            $observers[] = array('Tracker.newHandler', function () {
+                @http_response_code(500);
 
-                    throw new \Exception("injected exception");
-                })
-            ));
+                throw new \Exception("injected exception");
+            });
         }
+
+        if ($testingEnvironment->_triggerInvalidRequests) {
+            // we trigger an invalid request by checking for triggerInvalid=1 in a request, and if found replacing the
+            // request w/ a request that has an nonexistent idsite
+            $observers[] = array('Tracker.initRequestSet', function (RequestSet $requestSet) {
+                $requests = $requestSet->getRequests();
+                foreach ($requests as $index => $request) {
+                    $url = $request->getParam('url');
+                    if (strpos($url, 'triggerInvalid=1') !== false) {
+                        $newParams = $request->getParams();
+                        $newParams['idsite'] = 1000;
+
+                        $requests[$index] = new Request($newParams);
+                    }
+                }
+                $requestSet->setRequests($requests);
+            });
+        }
+
+        if (!empty($observers)) {
+            $result['observers.global'] = \DI\add($observers);
+        }
+
         return $result;
     }
 }

tests/resources/access-logs/fake_logs_dynamic.log

@@ -8,8 +8,8 @@ whatever.com 72.44.32.10 - - [12/Aug/2012:15:49:48 +0200] "GET /translations/ HT
 piwik.net 175.41.192.09 - - [12/Aug/2012:22:56:45 +0200] "GET /docs/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (X11; Linux i686; rv:6.0) Gecko/20100101 Firefox/6.0"
 piwik.net 175.41.192.09 - - [12/Aug/2012:23:00:42 +0200] "GET /docs/manage-users/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_0) AppleWebKit/536.3 (KHTML, like Gecko) Chrome/19.0.1063.0 Safari/536.3"
 whatever.com 79.125.00.21 - - [13/Aug/2012:20:03:40 +0200] "GET /newsletter/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; Trident/5.0)"
-whatever.com 175.41.192.34 - - [13/Aug/2012:21:59:50 +0200] "GET /faq/how-to/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; Trident/5.0)"
-anothersite.com 175.41.192.34 - - [13/Aug/2012:22:01:17 +0200] "GET /faq/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (X11; U; Linux x86_64; fr-FR) AppleWebKit/534.7 (KHTML, like Gecko) Epiphany/2.30.6 Safari/534.7"
+whatever.com 175.41.192.34 - - [13/Aug/2012:21:59:50 +0200] "GET /faq/how-to/?triggerInvalid=1 HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; Trident/5.0)"
+anothersite.com 175.41.192.34 - - [13/Aug/2012:22:01:17 +0200] "GET /faq/?triggerInvalid=1 HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (X11; U; Linux x86_64; fr-FR) AppleWebKit/534.7 (KHTML, like Gecko) Epiphany/2.30.6 Safari/534.7"
 anothersite.com 177.71.128.21 - - [13/Aug/2012:22:21:03 +0200] "GET /docs/manage-websites/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (compatible; MSIE 10.6; Windows NT 6.1; Trident/5.0; InfoPath.2; SLCC1; .NET CLR 3.0.4506.2152; .NET CLR 3.5.30729; .NET CLR 2.0.50727) 3gpp-gba UNTRUSTED/1.0"
 anothersite.com 177.71.128.21 - - [13/Aug/2012:22:21:28 +0200] "GET /intranet-analytics/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (X11; U; Linux x86_64; fr-FR) AppleWebKit/534.7 (KHTML, like Gecko) Epiphany/2.30.6 Safari/534.7"
 whatever.com 177.71.128.21 - - [13/Aug/2012:22:22:08 +0200] "GET /blog/2012/08/survey-your-opinion-matters/ HTTP/1.1" 200 3574 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/536.6 (KHTML, like Gecko) Chrome/20.0.1092.0 Safari/536.6"