improve the parsing of form data

examples/file_downloader.cc

@@ -8,13 +8,14 @@
 
 int main(int argc, char* argv[]) {
   if (argc != 3) {
-    std::cout << "usage: file_downloader <url> <path>" << std::endl;
+    std::cout << "Usage: file_downloader <url> <path>" << std::endl;
     std::cout << std::endl;
-    std::cout << "examples:" << std::endl;
-    std::cout << "  $ file_downloader http://httpbin.org/image/jpeg D:/test.jpg"
-              << std::endl;
-    std::cout << "  $ file_downloader https://www.google.com/favicon.ico"
-              << " D:/test.ico" << std::endl;
+    std::cout << "Examples:" << std::endl;
+    std::cout
+        << "  $ ./file_downloader http://httpbin.org/image/jpeg ~/test.jpg"
+        << std::endl;
+    std::cout << "  $ ./file_downloader https://www.google.com/favicon.ico"
+              << " ~/test.ico" << std::endl;
     return 1;
   }
 

examples/form_client.cc

@@ -11,14 +11,19 @@ namespace bfs = boost::filesystem;
 
 int main(int argc, char* argv[]) {
   if (argc < 2) {
-    std::cout << "usage: form_client <upload_dir> [url]" << std::endl;
+    std::cout << "Usage: form_client <upload_dir> [url]" << std::endl;
     std::cout << std::endl;
-    std::cout << "default url: http://httpbin.org/post" << std::endl;
+    std::cout << "Default url: http://httpbin.org/post" << std::endl;
     std::cout << std::endl;
-    std::cout << "examples:" << std::endl;
-    std::cout << "  $ form_client E:/github/webcc/data/upload" << std::endl;
-    std::cout << "  $ form_client E:/github/webcc/data/upload "
+    std::cout << "Examples:" << std::endl;
+    std::cout << "(Post to httpbin.org)" << std::endl;
+    std::cout << "  $ ./form_client path/to/webcc/data/upload" << std::endl;
+    std::cout << "  $ ./form_client path/to/webcc/data/upload "
               << "http://httpbin.org/post" << std::endl;
+    std::cout << "(Post the example 'form_server')" << std::endl;
+    std::cout << "  $ ./form_client path/to/webcc/data/upload "
+                 "http://localhost:8080/upload"
+              << std::endl;
     return 1;
   }
 
@@ -41,10 +46,10 @@ int main(int argc, char* argv[]) {
   webcc::ClientSession session;
 
   try {
-    auto r = session.Send(webcc::RequestBuilder{}.Post(url).
-                          FormFile("file", upload_dir / "remember.txt").
-                          FormData("json", "{}", "application/json")
-                          ());
+    auto r = session.Send(webcc::RequestBuilder{}
+                              .Post(url)
+                              .FormFile("file", upload_dir / "remember.txt")
+                              .FormData("json", "{}", "application/json")());
 
     std::cout << r->status() << std::endl;
 

examples/form_server.cc

@@ -1,5 +1,6 @@
 // A server handling multipart form data.
 
+#include <fstream>
 #include <iostream>
 #include <string>
 
@@ -25,7 +26,13 @@ private:
 
     for (auto& part : request->form_parts()) {
       std::cout << "name: " << part->name() << std::endl;
-      std::cout << "data: " << std::endl << part->data() << std::endl;
+
+      if (part->file_name().empty()) {
+        std::cout << "data: " << part->data() << std::endl;
+      } else {
+        // Save part->data() as binary to file.
+        // ...
+      }
     }
 
     return webcc::ResponseBuilder{}.Created().Body("OK")();
@@ -36,7 +43,7 @@ private:
 
 int main(int argc, char* argv[]) {
   if (argc < 2) {
-    std::cout << "usage: file_upload_server <port>" << std::endl;
+    std::cout << "usage: form_server <port>" << std::endl;
     return 1;
   }
 

webcc/parser.h

@@ -130,6 +130,7 @@ protected:
 
   bool ParseHeaderLine(const std::string& line);
 
+  // Parse the given length of data.
   virtual bool ParseContent(const char* data, std::size_t length);
 
   bool ParseFixedContent(const char* data, std::size_t length);

webcc/request_parser.cc

@@ -50,14 +50,10 @@ bool RequestParser::ParseStartLine(const std::string& line) {
 }
 
 bool RequestParser::ParseContent(const char* data, std::size_t length) {
-  if (chunked_) {
-    return ParseChunkedContent(data, length);
+  if (content_type_.multipart()) {
+    return ParseMultipartContent(data, length);
   } else {
-    if (content_type_.multipart()) {
-      return ParseMultipartContent(data, length);
-    } else {
-      return ParseFixedContent(data, length);
-    }
+    return Parser::ParseContent(data, length);
   }
 }
 
@@ -115,13 +111,18 @@ bool RequestParser::ParseMultipartContent(const char* data,
       std::size_t off = 0;
       std::size_t count = 0;
       bool ended = false;
+
       // TODO: Remember last CRLF position.
-      if (!GetNextBoundaryLine(&off, &count, &ended)) {
-        // Wait until next boundary.
+
+      bool next_boundary_found = GetNextBoundaryLine(&off, &count, &ended);
+
+      if (!next_boundary_found) {
+        part_->AppendData(pending_data_);
+        pending_data_.clear();
         break;
       }
 
-      LOG_INFO("Next boundary found.");
+      // Next boundary found.
 
       // This part has ended.
       if (off > 2) {
@@ -269,7 +270,7 @@ bool RequestParser::IsBoundary(const std::string& str, std::size_t off,
       *end = true;
     }
   }
-  
+
   return strncmp(boundary.c_str(), &str[off + 2], boundary.size()) == 0;
 }
 

webcc/request_parser.h

@@ -26,8 +26,6 @@ private:
   // asks for data streaming.
   bool OnHeadersEnd() override;
 
-  bool Stream() const;
-
   bool ParseStartLine(const std::string& line) override;
 
   // Override to handle multipart form data which is request only.
@@ -51,13 +49,14 @@ private:
   // received. The parsing will stop and fail if no view can be matched.
   ViewMatcher view_matcher_;
 
-  // Form data parsing step.
+  // Form data parsing steps.
   enum Step {
     kStart,
     kBoundaryParsed,
     kHeadersParsed,
     kEnded,
   };
+
   Step step_ = kStart;
 
   // The current form part being parsed.