X-Git-Url: http://git.tdb.fi/?p=libs%2Fdatafile.git;a=blobdiff_plain;f=source%2Fparser.cpp;h=5972e2d98f832d312afa4373bfe66d4204d10f8f;hp=7b9af11d60ca292a860ae346cebcb813220e5469;hb=ed78b585cfc4ecb44972e346857e887b183fd7a7;hpb=36e888dd1208822ae6df13b63f32c97d36b0a616

diff --git a/source/parser.cpp b/source/parser.cpp
index 7b9af11..5972e2d 100644
--- a/source/parser.cpp
+++ b/source/parser.cpp
@@ -1,349 +1,111 @@
-/*
-This file is part of libmspparser
-Copyright © 2006  Mikko Rasa, Mikkosoft Productions
-Distributed under the LGPL
-*/
-#include <cctype>
-#include <sstream>
-#include <msp/error.h>
+#include <msp/strings/format.h>
+#include "binaryparser.h"
+#include "except.h"
 #include "parser.h"
 #include "statement.h"
-#include "token.h"
+#include "textparser.h"
 
 using namespace std;
 
 namespace Msp {
-namespace Parser {
+namespace DataFile {
 
-Parser::Parser(istream &i, const string &s):
+Parser::Parser(IO::Base &i, const string &s):
 	in(i),
+	main_src(s),
 	src(s),
-	good(true)
+	good(true),
+	mode(new TextParser(in, src))
 { }
 
-Statement Parser::parse()
+Parser::~Parser()
 {
-	if(!good)
-		throw Exception("Parser is not good");
-	
-	try
-	{
-		return parse_(0);
-	}
-	catch(const Exception &e)
-	{
-		good=false;
-		throw;
-	}
+	delete mode;
 }
 
-Statement Parser::parse_(const Token *t)
+Statement Parser::parse(bool raw)
 {
-	Statement result;
-	bool      sub=false;
-	bool      finish=false;
-	
-	while(in)
+	if(!good)
+		throw logic_error("Parser::parse() !good");
+
+	try
 	{
-		Token token;
-		if(t)
-		{
-			token=*t;
-			t=0;
-		}
-		else
-			token=parse_token();
-		
-		if(result.keyword.empty())
+		while(1)
 		{
-			if(token.str.empty())
-				break;
-			else if(token.type!=Token::IDENTIFIER)
-				throw DataError(get_location()+": Syntax error at token '"+token.str+"' (expected an identifier)");
-			result.keyword=token.str;
-			result.valid=true;
-			result.source=src;
-			result.line=in.get_line_number();
-		}
-		else if(sub)
-		{
-			if(token.str=="}")
-			{
-				sub=false;
-				finish=true;
-			}
-			else
+			Statement st = mode->parse();
+			if(!st.keyword.compare(0, 2, "__"))
 			{
-				Statement ss=parse_(&token);
-				result.sub.push_back(ss);
+				st.control = true;
+				process_control_statement(st);
 			}
+
+			if(raw || !st.control)
+				return st;
+			else if(!good)  // This will occur with an __end statement
+				return Statement();
 		}
-		else if(finish)
-		{
-			if(token.str!=";")
-				throw DataError(get_location()+": Syntax error at token '"+token.str+"' (Expected a ';')");
-			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);
-		}
-		else if(token.str=="")
-			throw DataError(src+": Unexcepted EOF");
+	}
+	catch(const exception &e)
+	{
+		good = false;
+		if(dynamic_cast<const data_error *>(&e))
+			throw;
 		else
-			throw DataError(get_location()+": Syntax error");
+			throw data_error(src, in.get_line_number(), e);
 	}
-
-	return result;
 }
 
-Token Parser::parse_token()
+void Parser::process_control_statement(const Statement &st)
 {
-	int c;
-	unsigned comment=0;
-	while(in)
+	if(st.keyword=="__bin")
 	{
-		c=in.get();
-		int next=in.peek();
-
-		//cout<<c<<' '<<next<<'\n';
+		delete mode;
+		mode = new BinaryParser(in, src);
 
-		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;
+		while(in.peek()=='\n')
+			in.get();
 	}
-	if(comment)
-		throw DataError(src+": Unfinished comment");
-	
-	enum ParseState
-	{
-		INIT,
-		NEGATIVE,
-		STRING,
-		ACCEPT,
-		ZERO,
-		DECIMAL,
-		HEXADECIMAL,
-		OCTAL,
-		FLOAT,
-		IDENTIFIER
-	};
-
-	static Token::Type token_type[]=
+	else if(st.keyword=="__text")
 	{
-		Token::SPECIAL,
-		Token::SPECIAL,
-		Token::STRING,
-		Token::SPECIAL,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::INTEGER,
-		Token::FLOAT,
-		Token::IDENTIFIER
-	};
-
-	ParseState state=INIT;
-	string     buf;
-	bool       escape=false;
-
-	while(in)
+		delete mode;
+		mode = new TextParser(in, src);
+	}
+	else if(st.keyword=="__z")
+		in.set_decompress();
+	else if(st.keyword=="__src")
 	{
-		if(state!=INIT)
-			c=in.get();
-		int next=in.peek();
-		
-		buf+=c;
-
-		switch(state)
-		{
-		case INIT:
-			if(c=='0')
-				state=ZERO;
-			else if(c=='-')
-				state=NEGATIVE;
-			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 NEGATIVE:
-			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(!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);
+		string s = st.args[0].get<string>();
+		if(s.empty())
+			src = main_src;
+		else
+			src = format("%s[%s]", main_src, s);
 	}
-
-	return Token(Token::SPECIAL, "");
+	else if(st.keyword=="__end")
+		good = false;
+	else
+		mode->process_control_statement(st);
 }
 
-bool Parser::is_delimiter(int c)
+const StatementKey *Parser::peek(unsigned level)
 {
-	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)
+	while(good)
 	{
-		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 DataError("Invalid escape");
-			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 DataError("Invalid hex digit");
-
-			hexcape=(hexcape<<4)|digit;
-			if(hexcape&0x10000)
-			{
-				result+=hexcape&0xFF;
-				hexcape=0;
-			}
-		}
-		else if(*i=='\\')
-			escape=true;
+		const StatementKey *key = mode->peek(level);
+		if(key && !key->keyword.compare(0, 2, "__"))
+			process_control_statement(mode->parse());
 		else
-			result+=*i;
+			return key;
 	}
 
-	return result;
-}
-
-string Parser::get_location()
-{
-	ostringstream ss;
-	ss<<src<<':'<<in.get_line_number();
-	return ss.str();
+	return 0;
 }
 
-void Parser::parse_error(int c, int state)
+bool Parser::parse_and_load(unsigned level, Loader &ldr, const LoaderAction &act)
 {
-	ostringstream ss;
-	ss<<get_location()<<": Parse error at '"<<(char)c<<"' (state "<<state<<')';
-	throw DataError(ss.str());
+	// Peek first to get any control statements processed
+	peek(level);
+	return mode->parse_and_load(level, ldr, act);
 }
 
-} // namespace Parser
+} // namespace DataFile
 } // namespace Msp