Added chunked upload support for embedded_test_server

net/test/embedded_test_server/http_request.cc

Index: net/test/embedded_test_server/http_request.cc
diff --git a/net/test/embedded_test_server/http_request.cc b/net/test/embedded_test_server/http_request.cc
index 3156e5134faa29cfe86586bc03d8ecbcef7ea137..7e41adc3ddc24fc8471e257120d22c995f445930 100644
--- a/net/test/embedded_test_server/http_request.cc
+++ b/net/test/embedded_test_server/http_request.cc
@@ -10,6 +10,7 @@
 #include "base/strings/string_util.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_split.h"
+#include "net/http/http_chunked_decoder.h"
 
 namespace net {
 namespace test_server {
@@ -140,6 +141,13 @@ HttpRequestParser::ParseResult HttpRequestParser::ParseHeaders() {
         http_request_->headers["Content-Length"],
         &declared_content_length_);
     DCHECK(success) << "Malformed Content-Length header's value.";
+  } else if (http_request_->headers.count("Transfer-Encoding") > 0) {
+    if (http_request_->headers["Transfer-Encoding"] == "chunked") {
+      http_request_->has_content = true;
+      chunked_decoder_.reset(new HttpChunkedDecoder());
+      state_ = STATE_CONTENT;
+      return WAITING;
+    }
   }
   if (declared_content_length_ == 0) {
     // No content data, so parsing is finished.
@@ -155,6 +163,21 @@ HttpRequestParser::ParseResult HttpRequestParser::ParseHeaders() {
 
 HttpRequestParser::ParseResult HttpRequestParser::ParseContent() {
   const size_t available_bytes = buffer_.size() - buffer_position_;
+  if (chunked_decoder_.get()) {
+    std::string result;
+    result.append(buffer_.data() + buffer_position_, available_bytes);
+    int bytes_written = chunked_decoder_->FilterBuf(
+        &result[0], static_cast<int>(result.size()));

[mmenke, 2015/02/12 18:22:48]

&result[0] -> result.data()

+    buffer_position_ += available_bytes;

[xunjieli, 2015/02/12 16:54:52]

I know why I was having issue with FilterBuf. we should not increment
buffer_position by the bytes_written, which is the number of bytes written after
eliminating CRLF.

But I am not sure if we should increment available_bytes either, since there
might be a case whether FilterBuf does not parse everything available right?

What do you suggest?

Changing buffer_ to std::vector<char> affects many places in this file (since
the parser uses find() method on string), so I think I will keep std::string
result to store the intermediate.

[mmenke, 2015/02/12 18:22:48]

FilterBuf will parse everything you give it, up to the trailer, per my earlier
comment.
Clearing buffer_ and resetting buffer_position_ after each pass, except the
last, where you should keep the last bytes_after_eof bytes (And also reset
buffer_position_).

Also, as you're getting nothing out of using result, which is a non-const
string, instead of buffer_, which is a non-const string...get rid of it.

[xunjieli, 2015/02/12 22:15:45]

I guess one benefit of using result is that we won't be change buffer_ .

+    http_request_->content.append(result.data(), bytes_written);
+    if (chunked_decoder_->reached_eof()) {
+      state_ = STATE_ACCEPTED;
+      return ACCEPTED;
+    }
+    state_ = STATE_CONTENT;
+    return WAITING;
+  }
+
   const size_t fetch_bytes = std::min(
       available_bytes,
       declared_content_length_ - http_request_->content.size());