HTTP server for testing Google Drive.

chrome/browser/chromeos/gdata/test_servers/http_request.cc

+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/chromeos/gdata/test_servers/http_request.h"
+
+#include <algorithm>
+#include <map>
+#include <string>
+#include "base/basictypes.h"
+#include "base/logging.h"
+#include "googleurl/src/gurl.h"
+
+namespace gdata {
+namespace test_servers {
+
+namespace {
+int kRequestSizeLimit = 64 * 1024 * 1024;  // 64 mb.
+}  // namespace
+
+HttpRequestParser::HttpRequestParser() : request_builder_(new HttpRequest()),
+                                         state_(REQUEST_LINE),
+                                         buffer_position_(0),
+                                         crlf_position_(-1),
+                                         crlf_checked_position_(0),
+                                         current_content_length_(0) {
+}
+
+HttpRequestParser::~HttpRequestParser() {
+}
+
+void HttpRequestParser::ProcessChunk(const char *data, int length) {
+  buffer_.append(data, length);
+  if (static_cast<int>(buffer_.length()) + length > kRequestSizeLimit) {
+    // Too big request. Treat it as a syntax error.
+    LOG(ERROR) << "The HTTP request is too large.";
+    state_ = SYNTAX_ERROR;
+    return;
+  }
+  if (crlf_position_ == -1 && state_ != DATA)
+    FindNextCrLf();
+}
+
+HttpRequestParser::STATE HttpRequestParser::ParseRequest() {
+  while (state_ != READY &&
+         state_ != SYNTAX_ERROR &&
+         ShouldParseBuffer()) {
+    std::string token;
+    switch (state_) {
+      case REQUEST_LINE:
+        // Parse a method, eg. GET.
+        token = ShiftToken(IDENTIFIER);
+        if (token.empty()) {
+          state_ = SYNTAX_ERROR;
+          break;
+        }
+        std::transform(token.begin(), token.end(), token.begin(), ::toupper);
+        if (token == "OPTIONS") {
+          request_builder_->method = OPTIONS;
+        } else if (token == "GET") {
+          request_builder_->method = GET;
+        } else if (token == "HEAD") {
+          request_builder_->method = HEAD;
+        } else if (token == "POST") {
+          request_builder_->method = POST;
+        } else if (token == "PUT") {
+          request_builder_->method = PUT;
+        } else if (token == "DELETE") {
+          request_builder_->method = DELETE;
+        } else if (token == "TRACE") {
+          request_builder_->method = TRACE;
+        } else if (token == "CONNECT") {
+          request_builder_->method = CONNECT;
+        } else {
+          request_builder_->method = CUSTOM;
+        }
+
+        // Parse an URL, eg. index.cgi.
+        token = ShiftToken(IDENTIFIER);
+        if (token.empty()) {
+          LOG(ERROR) << "The url token must not be empty.";
+          state_ = SYNTAX_ERROR;
+          break;
+        }
+        request_builder_->raw_url = token;
+
+        // Parse a protocol, eg. HTTP/1.1.
+        // Currently we support only HTTP/1.1, since HTTP/1.0 is quite obsolete.
+        token = ShiftToken(LINE);
+        std::transform(token.begin(), token.end(), token.begin(), ::tolower);
+        if (token.empty() || token != "http/1.1") {
+          LOG(ERROR) << "This protocol is not supported: [" << token << "]";
+          state_ = SYNTAX_ERROR;
+          break;
+        }
+        request_builder_->protocol = token;
+        state_ = HEADER_LINE;
+        break;
+      case HEADER_LINE:
+        // If we have an empty line, then finish parsing headers and validate
+        // the request.
+        if (ShiftCrLf()) {
+          // TODO(mtomasz): Check if these are really case-insensitive in the
+          // rfc.
+          if (request_builder_->headers.find("host") ==
+              request_builder_->headers.end()) {
+            LOG(ERROR) << "Host header is required for HTTP/1.1";
+            state_ = SYNTAX_ERROR;
+            break;
+          } else {
+            // TODO(mtomasz): We add http:// protocol, but how about https://?
+            // Do we want to support it?
+            GURL host = GURL("http://" + request_builder_->headers["host"]);
+            request_builder_->url = host.Resolve(request_builder_->raw_url);
+          }
+          if (request_builder_->headers.find("content-length") !=
+              request_builder_->headers.end()) {
+            current_content_length_ =
+                atoi(request_builder_->headers["content-length"].c_str());
+          }
+          if (current_content_length_ > 0) {
+            state_ = DATA;
+          } else {
+            state_ = READY;
+          }
+          break;
+        }
+        {
+          // Parse a header name, eg. Content-length.
+          std::string header_name = ShiftToken(HEADER_KEY);
+          std::transform(header_name.begin(),
+                         header_name.end(),
+                         header_name.begin(),
+                         ::tolower);
+          if (header_name.empty()) {
+            LOG(ERROR) << "Header key must not be empty.";
+            state_ = SYNTAX_ERROR;
+            break;
+          }
+
+          // Parse a header value, eg. 100 (for Content-Length).
+          token = ShiftToken(LINE);
+          request_builder_->headers[header_name] = token;
+          break;
+        }
+      case DATA:
+        if (current_content_length_ ==
+            static_cast<int>(request_builder_->content.length())) {
+          // Entire data received.
+          state_ = READY;
+        } else {
+          int to_read = current_content_length_ -
+              request_builder_->content.length();
+          if (to_read > static_cast<int>(request_builder_->content.length()) -
+              buffer_position_) {
+            to_read = request_builder_->content.length() - buffer_position_;
+          }
+          request_builder_->content.append(ShiftData(to_read), to_read);
+        }
+        break;
+      case SYNTAX_ERROR:
+        break;
+      default:
+        state_ = SYNTAX_ERROR;
+        break;
+    }
+  }
+  return state_;
+}
+
+scoped_ptr<HttpRequest> HttpRequestParser::GetRequest() {
+  DCHECK(state_ == READY);
+  scoped_ptr<HttpRequest> result = request_builder_.Pass();
+  request_builder_.reset(new HttpRequest());
+
+  // Reset the parser state.
+  state_ = REQUEST_LINE;
+  buffer_position_ = 0;
+  buffer_.clear();
+  crlf_position_ = -1;
+  crlf_checked_position_ = 0;
+  current_content_length_ = 0;
+
+  return result.Pass();
+}
+
+std::string HttpRequestParser::ShiftToken(SHIFT_TOKEN_TYPE token_type) {
+  SkipSpaces();  // Ignore any spaces from the beginning.
+  size_t found_position;
+  int token_start_position = buffer_position_;
+
+  switch (token_type) {
+    case IDENTIFIER:
+      found_position = buffer_.find(' ', buffer_position_);
+      if (found_position != std::string::npos) {
+        buffer_position_ = static_cast<int>(found_position) + 1;
+        return buffer_.substr(token_start_position,
+                              found_position - token_start_position);
+      }
+      break;
+    case HEADER_KEY:
+      found_position = buffer_.find(':', buffer_position_);
+      if (found_position != std::string::npos) {
+        buffer_position_ = static_cast<int>(found_position) + 1;
+        return buffer_.substr(token_start_position,
+                              found_position - token_start_position);
+      }
+      break;
+    case LINE:
+      DCHECK(crlf_position_ != -1) << "CRLF not found, but expected.";
+      buffer_position_ = crlf_position_ + 2;
+      int token_length = crlf_position_ - token_start_position;
+      FindNextCrLf();
+      return buffer_.substr(token_start_position, token_length);
+      break;
+  }
+
+  return "";
+}
+
+bool HttpRequestParser::ShiftCrLf() {
+  if (buffer_position_ == crlf_position_) {
+    buffer_position_ += 2;
+    FindNextCrLf();
+    return true;
+  }
+
+  return false;
+}
+
+const char* HttpRequestParser::ShiftData(int length) {
+  DCHECK(length <= static_cast<int>(request_builder_->content.length()) -
+      buffer_position_) << "Tried to read more than available.";
+  const char *result = buffer_.data() + buffer_position_;
+  buffer_position_ += length;
+
+  return result;
+}
+
+void HttpRequestParser::SkipSpaces() {
+  const char *buffer_data = buffer_.data();
+  int buffer_left = static_cast<int>(buffer_.length()) - buffer_position_;
+
+  // Ignore any spaces from the beginning.
+  while (buffer_left && buffer_data[buffer_position_] == ' ') {
+    buffer_position_++;
+    buffer_left--;
+  }
+}
+
+bool HttpRequestParser::ShouldParseBuffer() {
+  return buffer_position_ < static_cast<int>(buffer_.length()) &&
+      (crlf_position_ != -1 || state_ == DATA);
+}
+
+bool HttpRequestParser::FindNextCrLf() {
+  if (!buffer_.length()) {
+    return false;
+  }
+
+  size_t found_position = buffer_.find("\r\n", crlf_checked_position_);
+  if (found_position == std::string::npos) {
+    crlf_position_ = -1;
+    crlf_checked_position_ = buffer_.length() - 1;
+    return false;
+  }
+
+  crlf_position_ = static_cast<int>(found_position);
+  crlf_checked_position_ = crlf_position_ + 2;
+  return true;
+}
+
+}  // namespace test_servers
+}  // namespace gdata