CoreUtils/HTTPRequest.h

#ifndef __HTTPRequest_h__
#define __HTTPRequest_h__

#include "Parser.h"
#include "LineParser.h"
#include "IMFHeader.h"
#include "CGIFormData.h"
#include "MString.h"
#include <map>

namespace coreutils {

   ///
   /// HTTPRequest parses a full HTTP/1.1 request - request line, headers,
   /// and (for a Content-Length body) either a raw body or, for
   /// application/x-www-form-urlencoded / multipart/form-data, a parsed
   /// CGIFormData - off a shared StreamReader. Resumable across any number
   /// of StreamReader::feed() calls, exactly like every other Parser in
   /// this subsystem: built by chaining LineParser (request line) and
   /// IMFHeader (headers) via checkParse(), the same way CGIFormData itself
   /// chains IMFHeader/IMFBoundary over one shared StreamReader.
   ///
   /// This codebase predates ever parsing raw HTTP off a socket (it has
   /// only ever sat behind a CGI gateway) - IMFRequest/IMFResponse/
   /// IMFMessage are stubs from before this Parser/StreamReader subsystem
   /// existed and are not built on here.
   ///
   /// v1 limitations, deliberately not addressed (flagged, not fixed -
   /// matching the standard already set by other Parser subclasses here for
   /// known-but-out-of-scope gaps):
   ///  - chunked Transfer-Encoding is not supported, only Content-Length.
   ///  - repeated header names (e.g. multiple Cookie: lines) last-wins -
   ///    the same limitation IMFHeader's std::map already has, inherited
   ///    here since HEADERS delegates to IMFHeader directly.
   ///

   class HTTPRequest : public Parser {

    public:

      HTTPRequest(StreamReader &reader);
      virtual ~HTTPRequest();

      ///
      /// Valid once doParse() has returned Status::COMPLETE.
      ///

      MString getMethod();
      MString getPath();
      MString getQueryString();
      MString getVersion();

      ///
      /// Case-insensitive header lookup (HTTP header names are
      /// case-insensitive per RFC 7230, unlike IMFHeader's underlying
      /// std::map, which compares byte-for-byte) - returns an empty
      /// MString if the header wasn't present. Valid once the HEADERS
      /// state has completed (i.e. from BODY parsing onward).
      ///

      MString getHeader(const char *name);

      ///
      /// Every header as parsed, keyed exactly as received (not
      /// case-normalized) - prefer getHeader() for lookups.
      ///

      std::map<MString, MString> &getHeaders();

      ///
      /// Valid once doParse() reports COMPLETE. Populated only when the
      /// request had a body that was neither application/
      /// x-www-form-urlencoded nor multipart/form-data (e.g. a raw JSON
      /// body) - empty otherwise, including when there was no body at all.
      ///

      MString getBody();

      ///
      /// Valid once doParse() reports COMPLETE, when the body's
      /// Content-Type was application/x-www-form-urlencoded or
      /// multipart/form-data - NULL otherwise (including no body at all).
      /// Owned by this HTTPRequest - do not delete.
      ///

      CGIFormData *getFormData();

    protected:

      Status parse() override;

    private:

      enum State { REQUEST_LINE, HEADERS, BODY, DONE } state = REQUEST_LINE;

      MString method, path, queryString, version;
      std::map<MString, MString> headers;
      MString body;
      MString boundary;

      LineParser *lineParser = NULL;
      IMFHeader *headerParser = NULL;
      CGIFormData *formParser = NULL;

   };

}

#endif