Add randomly chunked HTTP parsing stress test, fix bugs

src/HttpParser.h

@@ -58,6 +58,10 @@ struct HttpRequest {
             headers->valueLength = std::max<int>(0, headers->valueLength - 9);
         }
 
+        if (cursor == nullptr) {
+            return 0;
+        }
+
         return cursor - data;
     }
 
@@ -98,7 +102,7 @@ struct HttpRequest {
 };
 
 class HttpParser {
-private:
+public://private:
     std::string fallback;
     int remainingStreamingBytes = 0;
 
@@ -134,6 +138,8 @@ public:
                 int emittable = std::min(remainingStreamingBytes, length);
                 dataHandler(user, std::string_view(data, emittable));
                 remainingStreamingBytes -= emittable;
+
+                data += emittable; // denna var buggen?
                 length -= emittable;
 
                 ret += emittable;
@@ -177,9 +183,39 @@ public:
             fallback.reserve(maxCopyDistance + 32); // padding should be same as libus
             fallback.append(data, maxCopyDistance);
 
+            // helst ska denna inte emitta någon data alls, vi gör det efteråt!
             if (int consumed = fenceAndConsumePostPadded<true>(fallback.data(), fallback.length(), user, &req, requestHandler, dataHandler); consumed) {
+
+                // I guess?
+                fallback.clear();
+
                 data += consumed - had;
                 length -= consumed - had;
+
+                // ska vi inte tömma fallback här?
+
+
+                // exakt samma if-sats som ovan!
+                if (remainingStreamingBytes) {
+                    // at this point we reset the timeout timer, we are streaming and we got a chunk
+
+                    if (remainingStreamingBytes >= length) {
+                        dataHandler(user, std::string_view(data, length));
+                        remainingStreamingBytes -= length;
+                        // no change to the socket here! we read all data in the buffer, return
+                        return;
+                    } else {
+                        dataHandler(user, std::string_view(data, remainingStreamingBytes));
+
+                        data += remainingStreamingBytes;
+                        length -= remainingStreamingBytes;
+
+                        remainingStreamingBytes = 0;
+
+                        // okay we are done with that, let's parse some more
+                    }
+                }
+
             } else {
                 if (fallback.length() == MAX_FALLBACK_SIZE) {
                     // here we failed to parse any header in the 4kb we were given!
@@ -191,6 +227,7 @@ public:
         }
 
         int consumed = fenceAndConsumePostPadded<false>(data, length, user, &req, requestHandler, dataHandler);
+
         data += consumed;
         length -= consumed;
 
@@ -198,8 +235,8 @@ public:
             if (length < MAX_FALLBACK_SIZE) {
                 fallback.append(data, length);
             } else {
-                // invalid http!
-                std::cout << "invalid http! fuck off!" << std::endl;
+                std::cout << "tail is invalid http!" << std::endl;
+                exit(-1);
             }
         }
 

src/Tests.h

@@ -5,20 +5,72 @@
 
 #include <chrono>
 
+// todo: random test of chunked http parsing of randomly generated requests
 void testHttpParserPerformance() {
 
-    char data[] = "GET /hello.htm HTTP/1.1\r\n"
+    char headers[] = "GET /hello.htm HTTP/1.1\r\n"
             "User-Agent: Mozilla/4.0 (compatible; MSIE5.01; Windows NT)\r\n"
             "Host: www.tutorialspoint.com\r\n"
             "Accept-Language: en-us\r\n"
             "Accept-Encoding: gzip, deflate\r\n"
-            "Connection: Keep-Alive\r\n\r\n ";
+            "Connection: Keep-Alive\r\n"
+            "Content-length: 1048576\r\n\r\n";
 
+    const int requestLength = sizeof(headers) - 1 + 1048576;
+    char *request = (char *) malloc(requestLength + 32);
+    memset(request, 0, requestLength);
+    memcpy(request, headers, sizeof(headers) - 1);
+
+    char *data = (char *) malloc(requestLength * 10);
+    int length = requestLength * 10;
+    //int currentOffset = 0;
+
+
+    int maxChunkSize = 10000;
+    char *paddedBuffer = (char *) malloc(maxChunkSize + 32);
+
+    // dela upp dessa 10 i 5 segment
     HttpParser httpParser;
-
     int validRequests = 0;
+    size_t dataBytes = 0;
 
-    auto start = std::chrono::high_resolution_clock::now();
+    for (int i = 0; i < 10; i++) {
+        memcpy(data + requestLength * i, request, requestLength);
+    }
+
+    for (int j = 0; j < 1000; j++) {
+        for (int currentOffset = 0; currentOffset != length; ) {
+
+            int chunkSize = rand() % 10000;
+            if (currentOffset + chunkSize > length) {
+                chunkSize = length - currentOffset;
+            }
+
+            // kpiera skiten
+            memcpy(paddedBuffer, data + currentOffset, chunkSize);
+
+            httpParser.consumePostPadded(/*data + currentOffset*/ paddedBuffer, chunkSize, nullptr, [&validRequests](void *user, HttpRequest *req) {
+
+                validRequests++;
+                std::cout << "validRequests: " << validRequests << std::endl;
+
+
+            }, [&dataBytes](void *, std::string_view data) {
+                //std::cout << "data bytes: " << data.length() << std::endl;
+                dataBytes += data.length();
+            }, [](void *) {
+
+                std::cout << "Error!" << std::endl;
+                return;
+            });
+
+            currentOffset += chunkSize;
+        }
+    }
+
+    std::cout << "Data bytes: " << dataBytes << std::endl;
+
+    /*auto start = std::chrono::high_resolution_clock::now();
     for (int i = 0; i < 10000000; i++) {
         httpParser.consumePostPadded(data, sizeof(data) - 2, nullptr, [&validRequests](void *user, HttpRequest *req) {
             validRequests++;
@@ -28,9 +80,9 @@ void testHttpParserPerformance() {
 
         });
     }
-    auto stop = std::chrono::high_resolution_clock::now();
+    auto stop = std::chrono::high_resolution_clock::now();*/
 
-    std::cout << "Parsed " << validRequests << " in " << std::chrono::duration_cast<std::chrono::milliseconds>(stop - start).count() << "ms" << std::endl;
+    //std::cout << "Parsed " << validRequests << " in " << std::chrono::duration_cast<std::chrono::milliseconds>(stop - start).count() << "ms" << std::endl;
 }
 
 #endif // TESTS_H