X-Git-Url: http://git.tdb.fi/?p=libs%2Fdatafile.git;a=blobdiff_plain;f=source%2Fparser.cpp;h=82ac0069b95d4f22e3d0de1ab0c0b66b04c76423;hp=8fccec2f94c208d93f966ec88f4279eeb133b081;hb=27630d44298cb67e075c166f4421288cc8ca117e;hpb=c4930d8d15a5a248ca921e0ed3f9bca8aa18b322

diff --git a/source/parser.cpp b/source/parser.cpp
index 8fccec2..82ac006 100644
--- a/source/parser.cpp
+++ b/source/parser.cpp
@@ -6,10 +6,10 @@ Distributed under the LGPL
 */
 #include <cctype>
 #include <sstream>
-#include "error.h"
+#include "binaryparser.h"
 #include "parser.h"
 #include "statement.h"
-#include "token.h"
+#include "textparser.h"
 
 using namespace std;
 
@@ -19,9 +19,15 @@ namespace DataFile {
 Parser::Parser(istream &i, const string &s):
 	in(i),
 	src(s),
-	good(true)
+	good(true),
+	mode(new TextParser(in, src))
 { }
 
+Parser::~Parser()
+{
+	delete mode;
+}
+
 Statement Parser::parse()
 {
 	if(!good)
@@ -29,351 +35,28 @@ Statement Parser::parse()
 
 	try
 	{
-		return parse_(0);
-	}
-	catch(const Exception &e)
-	{
-		good=false;
-		throw;
-	}
-}
-
-Statement Parser::parse_(const Token *t)
-{
-	Statement result;
-	bool      sub=false;
-	bool      finish=false;
-
-	while(in)
-	{
-		Token token;
-		if(t)
-		{
-			token=*t;
-			t=0;
-		}
-		else
-			token=parse_token();
-
-		if(result.keyword.empty())
-		{
-			if(token.str.empty())
-				break;
-			else if(token.type!=Token::IDENTIFIER)
-				throw ParseError(get_location()+": Syntax error at token '"+token.str+"' (expected an identifier)", src, in.get_line_number());
-			result.keyword=token.str;
-			result.valid=true;
-			result.source=src;
-			result.line=in.get_line_number();
-		}
-		else if(sub)
+		while(1)
 		{
-			if(token.str=="}")
+			Statement st=mode->parse();
+			if(st.keyword=="__bin")
 			{
-				sub=false;
-				finish=true;
+				delete mode;
+				mode=new BinaryParser(in, src);
 			}
-			else
+			else if(st.keyword=="__text")
 			{
-				Statement ss=parse_(&token);
-				result.sub.push_back(ss);
+				delete mode;
+				mode=new TextParser(in, src);
 			}
-		}
-		else if(finish)
-		{
-			if(token.str!=";")
-				throw ParseError(get_location()+": Syntax error at token '"+token.str+"' (Expected a ';')", src, in.get_line_number());
-			break;
-		}
-		else if(token.str=="{")
-			sub=true;
-		else if(token.str==";")
-			break;
-		else if(token.type==Token::INTEGER)
-			result.args.push_back(Value(Value::INTEGER, token.str));
-		else if(token.type==Token::FLOAT)
-			result.args.push_back(Value(Value::FLOAT, token.str));
-		else if(token.type==Token::STRING)
-			result.args.push_back(Value(Value::STRING, token.str));
-		else if(token.type==Token::IDENTIFIER)
-		{
-			if(token.str=="true")
-				result.args.push_back(Value(Value::BOOLEAN, "1"));
-			else if(token.str=="false")
-				result.args.push_back(Value(Value::BOOLEAN, "0"));
 			else
-				result.args.push_back(Value(Value::ENUM, token.str));
-			//result.args.push_back(resolve_identifiertoken.str);
+				return st;
 		}
-		else if(token.str=="")
-			throw ParseError(src+": Unexcepted EOF", src, in.get_line_number());
-		else
-			throw ParseError(get_location()+": Syntax error", src, in.get_line_number());
 	}
-
-	return result;
-}
-
-Token Parser::parse_token()
-{
-	int c=0;
-	unsigned comment=0;
-
-	// Skip over comments and whitespace
-	while(in)
-	{
-		c=in.get();
-		int next=in.peek();
-
-		if(c=='/' && next=='/')
-			comment=1;
-		else if(c=='/' && next=='*')
-			comment=2;
-		else if(c=='\n' && comment==1)
-			comment=0;
-		else if(c=='*' && next=='/' && comment==2)
-			comment=3;
-		else if(comment==3)   // Skip the second character of block comment end
-			comment=0;
-		else if(!isspace(c) && !comment)
-			break;
-	}
-
-	if(comment)  // Didn't hit any non-whitespace
-		throw ParseError(src+": Unfinished comment", src, in.get_line_number());
-
-	enum ParseState
-	{
-		INIT,
-		SIGN,
-		FLOATEXPINIT,
-		FLOATEXPSIGN,
-		STRING,
-		ACCEPT,
-		ZERO,
-		DECIMAL,
-		HEXADECIMAL,
-		OCTAL,
-		FLOAT,
-		FLOATEXP,
-		IDENTIFIER
-	};
-
-	static Token::Type token_type[]=
-	{
-		Token::SPECIAL,
-		Token::SPECIAL,
-		Token::SPECIAL,
-		Token::SPECIAL,
-		Token::STRING,
-		Token::SPECIAL,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::FLOAT,
-		Token::FLOAT,
-		Token::IDENTIFIER
-	};
-
-	ParseState state=INIT;
-	string     buf;
-	bool       escape=false;
-
-	while(in)
-	{
-		if(state!=INIT)
-			c=in.get();
-		int next=in.peek();
-
-		buf+=c;
-
-		switch(state)
-		{
-		case INIT:
-			if(c=='0')
-				state=ZERO;
-			else if(c=='-' || c=='+')
-				state=SIGN;
-			else if(c=='.')
-				state=FLOAT;
-			else if(c=='"')
-				state=STRING;
-			else if(c=='{' || c=='}' || c==';')
-				return Token(Token::SPECIAL, string(1, c));
-			else if(isdigit(c))
-				state=DECIMAL;
-			else if(isalpha(c))
-				state=IDENTIFIER;
-			else
-				parse_error(c, state);
-			break;
-
-		case SIGN:
-			if(c=='0')
-				state=ZERO;
-			else if(isdigit(c))
-				state=DECIMAL;
-			else if(c=='.')
-				state=FLOAT;
-			else
-				parse_error(c, state);
-			break;
-
-		case ZERO:
-			if(c=='x')
-				state=HEXADECIMAL;
-			else if(isdigit(c))
-				state=OCTAL;
-			else if(c=='.')
-				state=FLOAT;
-			else
-				parse_error(c, state);
-			break;
-
-		case DECIMAL:
-			if(c=='.')
-				state=FLOAT;
-			else if(!isdigit(c))
-				parse_error(c, state);
-			break;
-
-		case HEXADECIMAL:
-			if(!isxdigit(c))
-				parse_error(c, state);
-			break;
-
-		case OCTAL:
-			if(!isodigit(c))
-				parse_error(c, state);
-			break;
-
-		case FLOAT:
-			if(c=='e' || c=='E')
-				state=FLOATEXPINIT;
-			else if(!isdigit(c))
-				parse_error(c, state);
-			break;
-
-		case FLOATEXPINIT:
-			if(c=='+' || c=='-')
-				state=FLOATEXPSIGN;
-			else if(isdigit(c))
-				state=FLOATEXP;
-			else
-				parse_error(c, state);
-			break;
-
-		case FLOATEXPSIGN:
-			if(isdigit(c))
-				state=FLOATEXP;
-			else
-				parse_error(c, state);
-			break;
-
-		case FLOATEXP:
-			if(!isdigit(c))
-				parse_error(c, state);
-			break;
-
-		case STRING:
-			if(c=='\\')
-				escape=!escape;
-			else if(c=='"' && !escape)
-				return Token(Token::STRING, unescape_string(buf));
-			else
-				escape=false;
-			break;
-
-		case IDENTIFIER:
-			if(!isalpha(c) && !isdigit(c) && c!='_')
-				parse_error(c, state);
-			break;
-
-		default:
-			throw Exception(get_location()+": Internal error (bad state)");
-		}
-
-		if(is_delimiter(next) && state>=ACCEPT)
-			return Token(token_type[state], buf);
-	}
-
-	return Token(Token::SPECIAL, "");
-}
-
-bool Parser::is_delimiter(int c)
-{
-	return (isspace(c) || c=='{' || c=='}' || c==';' || c=='/');
-}
-
-bool Parser::isodigit(int c)
-{
-	return (c>='0' && c<='7');
-}
-
-string Parser::unescape_string(const string &str)
-{
-	string   result;
-	bool     escape=false;
-	unsigned hexcape=0;
-	for(string::const_iterator i=str.begin()+1; i!=str.end()-1; ++i)
+	catch(const Exception &e)
 	{
-		if(escape)
-		{
-			if(*i=='n')
-				result+='\n';
-			else if(*i=='t')
-				result+='\t';
-			else if(*i=='\\')
-				result+='\\';
-			else if(*i=='"')
-				result+='"';
-			else if(*i=='x')
-				hexcape=0x100;
-			else
-				throw ParseError("Invalid escape", src, in.get_line_number());
-			escape=false;
-		}
-		else if(hexcape)
-		{
-			unsigned digit=0;
-			if(*i>='0' && *i<='9')
-				digit=*i-'0';
-			else if(*i>='a' && *i<='f')
-				digit=*i-'a'+10;
-			else if(*i>='A' && *i<='F')
-				digit=*i-'A'+10;
-			else
-				throw ParseError("Invalid hex digit", src, in.get_line_number());
-
-			hexcape=(hexcape<<4)|digit;
-			if(hexcape&0x10000)
-			{
-				result+=hexcape&0xFF;
-				hexcape=0;
-			}
-		}
-		else if(*i=='\\')
-			escape=true;
-		else
-			result+=*i;
+		good=false;
+		throw;
 	}
-
-	return result;
-}
-
-string Parser::get_location()
-{
-	ostringstream ss;
-	ss<<src<<':'<<in.get_line_number();
-	return ss.str();
-}
-
-void Parser::parse_error(int c, int state)
-{
-	ostringstream ss;
-	ss<<get_location()<<": Parse error at '"<<(char)c<<"' (state "<<state<<')';
-	throw ParseError(ss.str(), src, in.get_line_number());
 }
 
 } // namespace DataFile