/* $Id$
This file is part of libmspdatafile
-Copyright © 2007 Mikko Rasa, Mikkosoft Productions
+Copyright © 2007-2008 Mikko Rasa, Mikkosoft Productions
Distributed under the LGPL
*/
#include <msp/strings/formatter.h>
+#include <msp/strings/utils.h>
#include "input.h"
#include "textparser.h"
#include "token.h"
Statement TextParser::parse_statement(const Token *t)
{
Statement result;
- bool sub=false;
- bool finish=false;
+ bool sub = false;
+ bool finish = false;
while(in)
{
Token token;
if(t)
{
- token=*t;
- t=0;
+ token = *t;
+ t = 0;
}
else
- token=parse_token();
+ token = parse_token();
if(result.keyword.empty())
{
if(token.str.empty())
break;
else if(token.type!=Token::IDENTIFIER)
- throw ParseError(format("%s: Syntax error at token '%s' (expected an identifier)", get_location(), token.str), src, in.get_line_number());
- result.keyword=token.str;
- result.valid=true;
- result.source=src;
- result.line=in.get_line_number();
+ throw_at(ParseError(format("Syntax error at token '%s' (expected an identifier)", token.str)), get_location());
+ 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;
+ sub = false;
+ finish = true;
}
else
{
- Statement ss=parse_statement(&token);
+ Statement ss = parse_statement(&token);
result.sub.push_back(ss);
}
}
else if(finish)
{
if(token.str!=";")
- throw ParseError(format("%s: Syntax error at token '%s' (Expected a ';')", get_location(), token.str), src, in.get_line_number());
+ throw_at(ParseError(format("Syntax error at token '%s' (Expected a ';')", token.str)), get_location());
break;
}
else if(token.str=="{")
- sub=true;
+ sub = true;
else if(token.str==";")
break;
else if(token.type==Token::INTEGER)
//result.args.push_back(resolve_identifiertoken.str);
}
else if(token.str=="")
- throw ParseError(src+": Unexcepted end of input", src, in.get_line_number());
+ throw_at(ParseError("Unexcepted end of input"), get_location());
else
- throw ParseError(get_location()+": Syntax error", src, in.get_line_number());
+ throw_at(ParseError("Syntax error"), get_location());
}
return result;
Token TextParser::parse_token()
{
- int c=0;
- unsigned comment=0;
+ int c = 0;
+ int comment = 0;
// Skip over comments and whitespace
- while(in)
+ while(in && comment>=0)
{
- c=in.get();
- int next=in.peek();
+ c = in.get();
+ int next = in.peek();
if(c=='/' && next=='/')
- comment=1;
+ comment = 1;
else if(c=='/' && next=='*')
- comment=2;
+ comment = 2;
else if(c=='\n' && comment==1)
- comment=0;
+ comment = 0;
else if(c=='*' && next=='/' && comment==2)
- comment=3;
+ comment = 3;
else if(comment==3) // Skip the second character of block comment end
- comment=0;
+ comment = 0;
else if(!isspace(c) && !comment)
- break;
+ comment = -1;
}
- if(comment) // Didn't hit any non-whitespace
- throw ParseError(src+": Unfinished comment at end of input", src, in.get_line_number());
+ if(comment>0) // EOF while in comment
+ throw_at(ParseError("Unfinished comment at end of input"), get_location());
+ else if(comment==0) // Didn't hit any non-whitespace
+ return Token(Token::SPECIAL, "");
enum ParseState
{
Token::IDENTIFIER
};
- ParseState state=INIT;
+ ParseState state = INIT;
string buf;
- bool escape=false;
+ bool escape = false;
- while(in)
+ while(in || state==INIT)
{
if(state!=INIT)
- c=in.get();
- int next=in.peek();
+ c = in.get();
+ int next = in.peek();
- buf+=c;
+ buf += c;
switch(state)
{
case INIT:
if(c=='0')
- state=ZERO;
+ state = ZERO;
else if(c=='-' || c=='+')
- state=SIGN;
+ state = SIGN;
else if(c=='.')
- state=FLOAT;
+ state = FLOAT;
else if(c=='"')
- state=STRING;
+ state = STRING;
else if(c=='{' || c=='}' || c==';')
return Token(Token::SPECIAL, string(1, c));
else if(isdigit(c))
- state=DECIMAL;
+ state = DECIMAL;
else if(isalpha(c) || c=='_')
- state=IDENTIFIER;
+ state = IDENTIFIER;
else
- parse_error(c, state);
+ parse_error(c, "0-9A-Za-z_.\"{};+-");
break;
case SIGN:
if(c=='0')
- state=ZERO;
+ state = ZERO;
else if(isdigit(c))
- state=DECIMAL;
+ state = DECIMAL;
else if(c=='.')
- state=FLOAT;
+ state = FLOAT;
else
- parse_error(c, state);
+ parse_error(c, "0-9.");
break;
case ZERO:
if(c=='x')
- state=HEXADECIMAL;
+ state = HEXADECIMAL;
else if(isdigit(c))
- state=OCTAL;
+ state = OCTAL;
else if(c=='.')
- state=FLOAT;
+ state = FLOAT;
else
- parse_error(c, state);
+ parse_error(c, "0-9A-Fa-f.");
break;
case DECIMAL:
if(c=='.')
- state=FLOAT;
+ state = FLOAT;
else if(!isdigit(c))
- parse_error(c, state);
+ parse_error(c, "0-9.");
break;
case HEXADECIMAL:
if(!isxdigit(c))
- parse_error(c, state);
+ parse_error(c, "0-9A-Fa-f");
break;
case OCTAL:
if(!isodigit(c))
- parse_error(c, state);
+ parse_error(c, "0-7");
break;
case FLOAT:
if(c=='e' || c=='E')
- state=FLOATEXPINIT;
+ state = FLOATEXPINIT;
else if(!isdigit(c))
- parse_error(c, state);
+ parse_error(c, "0-9Ee");
break;
case FLOATEXPINIT:
if(c=='+' || c=='-')
- state=FLOATEXPSIGN;
+ state = FLOATEXPSIGN;
else if(isdigit(c))
- state=FLOATEXP;
+ state = FLOATEXP;
else
- parse_error(c, state);
+ parse_error(c, "0-9+-");
break;
case FLOATEXPSIGN:
if(isdigit(c))
- state=FLOATEXP;
+ state = FLOATEXP;
else
- parse_error(c, state);
+ parse_error(c, "0-9");
break;
case FLOATEXP:
if(!isdigit(c))
- parse_error(c, state);
+ parse_error(c, "0-9");
break;
case STRING:
if(c=='\\')
- escape=!escape;
+ escape = !escape;
else if(c=='"' && !escape)
- return Token(Token::STRING, unescape_string(buf));
+ {
+ try
+ {
+ return Token(Token::STRING, c_unescape(buf.substr(1, buf.size()-2)));
+ }
+ catch(Exception &e)
+ {
+ e.at(get_location());
+ throw;
+ }
+ }
else
- escape=false;
+ escape = false;
break;
case IDENTIFIER:
if(!isalpha(c) && !isdigit(c) && c!='_')
- parse_error(c, state);
+ parse_error(c, "0-9A-Za-z_");
break;
default:
- throw Exception(get_location()+": Internal error (bad state)");
+ throw_at(InvalidState("Internal error (bad state)"), get_location());
}
if(is_delimiter(next) && state>=ACCEPT)
return (c>='0' && c<='7');
}
-string TextParser::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)
- {
- 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(format("%s: Invalid escape sequence '\\%c'", get_location(), *i), 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(get_location()+": 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;
- }
-
- return result;
-}
-
string TextParser::get_location()
{
ostringstream ss;
return ss.str();
}
-void TextParser::parse_error(int c, int state)
+void TextParser::parse_error(int c, const char *e)
{
- throw ParseError(format("%s: Parse error at '%c' (state %d)", get_location(), static_cast<char>(c), state), src, in.get_line_number());
+ throw_at(ParseError(format("Parse error at '%c', expected one of \"%s\"", static_cast<char>(c), e)), get_location());
}
} // namespace DataFile
projects / libs / datafile.git / blobdiff