Permalink
Browse files

Merge pull request #3695 from Karlson2k/http_header_rfc2616

Fix: process HTTP header according to RFC 2616
  • Loading branch information...
2 parents e20eee9 + e0a06ab commit 80221152faec9ae7f1691ab1d9df5d27a58b6b17 @jmarshallnz jmarshallnz committed Dec 10, 2013
Showing with 101 additions and 37 deletions.
  1. +98 −37 xbmc/utils/HttpHeader.cpp
  2. +3 −0 xbmc/utils/HttpHeader.h
View
135 xbmc/utils/HttpHeader.cpp
@@ -21,6 +21,10 @@
#include "HttpHeader.h"
#include "utils/StringUtils.h"
+// header white space characters according to RFC 2616
+const char* const CHttpHeader::m_whitespaceChars = " \t";
+
+
CHttpHeader::CHttpHeader()
{
m_headerdone = false;
@@ -32,42 +36,68 @@ CHttpHeader::~CHttpHeader()
void CHttpHeader::Parse(const std::string& strData)
{
- if (m_headerdone)
- Clear();
-
size_t pos = 0;
const size_t len = strData.length();
+ const char* const strDataC = strData.c_str();
+
+ // According to RFC 2616 any header line can have continuation on next line, if next line is started from whitespace char
+ // This code at first checks for whitespace char at the begging of the line, and if found, then current line is appended to m_lastHeaderLine
+ // If current line is NOT started from whitespace char, then previously stored (and completed) m_lastHeaderLine is parsed and current line is assigned to m_lastHeaderLine (to be parsed later)
while (pos < len)
{
- const size_t valueStart = strData.find(':', pos);
- const size_t lineEnd = strData.find("\r\n", pos);
+ const size_t lineEnd = strData.find("\x0d\x0a", pos); // use "\x0d\x0a" instead of "\r\n" to be platform independent
if (lineEnd == std::string::npos)
- break;
+ return; // error: expected only complete lines
- if (lineEnd == pos)
- {
- m_headerdone = true;
- break;
- }
- else if (valueStart != std::string::npos && valueStart < lineEnd)
- {
- std::string strParam(strData, pos, valueStart - pos);
- std::string strValue(strData, valueStart + 1, lineEnd - valueStart - 1);
+ if (m_headerdone)
+ Clear(); // clear previous header and process new one
- StringUtils::Trim(strParam);
- StringUtils::ToLower(strParam);
+ if (strDataC[pos] == ' ' || strDataC[pos] == '\t') // same chars as in CHttpHeader::m_whitespaceChars
+ { // line is started from whitespace char: this is continuation of previous line
+ pos = strData.find_first_not_of(m_whitespaceChars);
- StringUtils::Trim(strValue);
+ m_lastHeaderLine.push_back(' '); // replace all whitespace chars at start of the line with single space
+ m_lastHeaderLine.append(strData, pos, lineEnd - pos); // append current line
+ }
+ else
+ { // this line is NOT continuation, this line is new header line
+ if (!m_lastHeaderLine.empty())
+ ParseLine(m_lastHeaderLine); // process previously stored completed line (if any)
- if (!strParam.empty() && !strValue.empty())
- m_params.push_back(HeaderParams::value_type(strParam, strValue));
+ m_lastHeaderLine.assign(strData, pos, lineEnd - pos); // store current line to (possibly) complete later. Will be parsed on next turns.
+
+ if (pos == lineEnd)
+ m_headerdone = true; // current line is bare "\r\n", means end of header; no need to process current m_lastHeaderLine
}
- else if (m_protoLine.empty())
- m_protoLine.assign(strData, pos, lineEnd - pos);
- pos = lineEnd + 2;
+ pos = lineEnd + 2; // '+2' for "\r\n": go to next line (if any)
+ }
+}
+
+bool CHttpHeader::ParseLine(const std::string& headerLine)
+{
+ const size_t valueStart = headerLine.find(':');
+
+ if (valueStart != std::string::npos)
+ {
+ std::string strParam(headerLine, 0, valueStart);
+ std::string strValue(headerLine, valueStart + 1);
+
+ StringUtils::Trim(strParam, m_whitespaceChars);
+ StringUtils::ToLower(strParam);
+
+ StringUtils::Trim(strValue, m_whitespaceChars);
+
+ if (!strParam.empty() && !strValue.empty())
+ m_params.push_back(HeaderParams::value_type(strParam, strValue));
+ else
+ return false;
}
+ else if (m_protoLine.empty())
+ m_protoLine = headerLine;
+
+ return true;
}
void CHttpHeader::AddParam(const std::string& param, const std::string& value, const bool overwrite /*= false*/)
@@ -143,7 +173,10 @@ std::string CHttpHeader::GetMimeType(void) const
{
std::string strValue(GetValueRaw("content-type"));
- return strValue.substr(0, strValue.find(';'));
+ std::string mimeType(strValue, 0, strValue.find(';'));
+ StringUtils::TrimRight(mimeType, m_whitespaceChars);
+
+ return mimeType;
}
std::string CHttpHeader::GetCharset(void) const
@@ -152,25 +185,53 @@ std::string CHttpHeader::GetCharset(void) const
if (strValue.empty())
return strValue;
- const size_t semicolonPos = strValue.find(';');
- if (semicolonPos == std::string::npos)
- return "";
-
StringUtils::ToUpper(strValue);
- size_t posCharset;
- if ((posCharset = strValue.find("; CHARSET=", semicolonPos)) != std::string::npos)
- posCharset += 10;
- else if ((posCharset = strValue.find(";CHARSET=", semicolonPos)) != std::string::npos)
- posCharset += 9;
- else
- return "";
-
- return strValue.substr(posCharset, strValue.find(';', posCharset) - posCharset);
+ const size_t len = strValue.length();
+ const char* const strValueC = strValue.c_str();
+
+ // extract charset value from 'contenttype/contentsubtype;pram1=param1Val ; charset=XXXX\t;param2=param2Val'
+ // most common form: 'text/html; charset=XXXX'
+ // charset value can be in double quotes: 'text/xml; charset="XXX XX"'
+
+ size_t pos = strValue.find(';');
+ while (pos < len)
+ {
+ // move to the next non-whitespace character
+ pos = strValue.find_first_not_of(m_whitespaceChars, pos + 1);
+
+ if (pos != std::string::npos)
+ {
+ if (strValue.compare(pos, 8, "CHARSET=", 8) == 0)
+ {
+ std::string charset(strValue, pos, strValue.find(';', pos)); // intentionally ignoring possible ';' inside quoted string
+ // as we don't support any charset with ';' in name
+ StringUtils::Trim(charset, m_whitespaceChars);
+ if (!charset.empty())
+ {
+ if (charset[0] != '"')
+ return charset;
+ else
+ { // charset contains quoted string (allowed according to RFC 2616)
+ StringUtils::Replace(charset, "\\", ""); // unescape chars, ignoring possible '\"' and '\\'
+ const size_t closingQ = charset.find('"', 1);
+ if (closingQ == std::string::npos)
+ return ""; // no closing quote
+
+ return charset.substr(1, closingQ - 1);
+ }
+ }
+ }
+ pos = strValue.find(';', pos); // find next parameter
+ }
+ }
+
+ return ""; // no charset is detected
}
void CHttpHeader::Clear()
{
m_params.clear();
m_protoLine.clear();
m_headerdone = false;
+ m_lastHeaderLine.clear();
}
View
3 xbmc/utils/HttpHeader.h
@@ -54,9 +54,12 @@ class CHttpHeader
protected:
std::string GetValueRaw(const std::string& strParam) const;
+ bool ParseLine(const std::string& headerLine);
HeaderParams m_params;
std::string m_protoLine;
bool m_headerdone;
+ std::string m_lastHeaderLine;
+ static const char* const m_whitespaceChars;
};

0 comments on commit 8022115

Please sign in to comment.