From ee302681649df3d3f0398fed560639c1321da636 Mon Sep 17 00:00:00 2001
From: Jon Siwek <jsiwek@ncsa.illinois.edu>
Date: Fri, 23 Sep 2011 10:31:20 -0500
Subject: [PATCH] Teach HTTP parser to derive content length of
 multipart/byteranges bodies.

Addresses #488.
---
 src/HTTP.cc | 45 +++++++++++++++++++++++++++++++++++++++++++++
 src/HTTP.h  |  1 +
 src/MIME.h  |  1 +
 3 files changed, 47 insertions(+)

diff --git a/src/HTTP.cc b/src/HTTP.cc
index 0d154f1873..ee5ee5f1df 100644
--- a/src/HTTP.cc
+++ b/src/HTTP.cc
@@ -5,6 +5,8 @@
 #include <ctype.h>
 #include <math.h>
 #include <stdlib.h>
+#include <string>
+#include <algorithm>
 
 #include "NetVar.h"
 #include "HTTP.h"
@@ -310,6 +312,49 @@ void HTTP_Entity::SubmitHeader(MIME_Header* h)
 			}
 		}
 
+	// Figure out content length for HTTP 206 Partial Content response
+	// that uses multipart/byteranges content-type
+	else if ( strcasecmp_n(h->get_name(), "content-range") == 0 && Parent() &&
+	          Parent()->MIMEContentType() == CONTENT_TYPE_MULTIPART &&
+		      http_message->MyHTTP_Analyzer()->HTTP_ReplyCode() == 206 )
+		{
+		data_chunk_t vt = h->get_value_token();
+		string byte_unit(vt.data, vt.length);
+		vt = h->get_value_after_token();
+		string byte_range(vt.data, vt.length);
+		byte_range.erase(remove(byte_range.begin(), byte_range.end(), ' '),
+		                 byte_range.end());
+
+		if ( byte_unit != "bytes" )
+			http_message->Weird("HTTP_content_range_unknown_byte_unit");
+
+		size_t p = byte_range.find("/");
+		string byte_range_resp_spec = byte_range.substr(0, p);
+		string instance_length = byte_range.substr(p + 1);
+
+		p = byte_range_resp_spec.find("-");
+		string first_byte_pos = byte_range_resp_spec.substr(0, p);
+		string last_byte_pos = byte_range_resp_spec.substr(p + 1);
+
+		if ( DEBUG_http )
+			DEBUG_MSG("Parsed Content-Range: %s %s-%s/%s\n", byte_unit.c_str(),
+		              first_byte_pos.c_str(), last_byte_pos.c_str(),
+		              instance_length.c_str());
+
+		int64_t f, l;
+		atoi_n(first_byte_pos.size(), first_byte_pos.c_str(), 0, 10, f);
+		atoi_n(last_byte_pos.size(), last_byte_pos.c_str(), 0, 10, l);
+		int64_t len = l - f + 1;
+
+		if ( DEBUG_http )
+			DEBUG_MSG("Content-Range length = %"PRId64"\n", len);
+
+		if ( len > 0 )
+			content_length = len;
+		else
+			http_message->Weird("HTTP_non_positive_content_range");
+		}
+
 	else if ( strcasecmp_n(h->get_name(), "transfer-encoding") == 0 )
 		{
 		data_chunk_t vt = h->get_value_token();
diff --git a/src/HTTP.h b/src/HTTP.h
index 6614886e44..373692c0c3 100644
--- a/src/HTTP.h
+++ b/src/HTTP.h
@@ -184,6 +184,7 @@ public:
 			http_event || http_stats) && !FLAGS_use_binpac; }
 
 	int IsConnectionClose()		{ return connection_close; }
+	int HTTP_ReplyCode() const { return reply_code; };
 
 protected:
 	void GenStats();
diff --git a/src/MIME.h b/src/MIME.h
index c8c70cf65b..52d943fb15 100644
--- a/src/MIME.h
+++ b/src/MIME.h
@@ -95,6 +95,7 @@ public:
 	virtual void EndOfData();
 
 	MIME_Entity* Parent() const { return parent; }
+	int MIMEContentType() const { return content_type; }
 	StringVal* ContentType() const { return content_type_str; }
 	StringVal* ContentSubType() const { return content_subtype_str; }
 	int ContentTransferEncoding() const { return content_encoding; }