X-Git-Url: http://git.tdb.fi/?a=blobdiff_plain;f=source%2Ftextparser.cpp;h=56c9bb2b07d75fbbdac9394b4edf42c314d6fdb2;hb=5c98c1f499413c41fd46ddf71a46b481338d7d6b;hp=78d2f1df6334b32dbf67a7871975deda4dafb0a4;hpb=2f79370bffe0bac865dc97c5114dc87c1936fbb4;p=libs%2Fdatafile.git diff --git a/source/textparser.cpp b/source/textparser.cpp index 78d2f1d..56c9bb2 100644 --- a/source/textparser.cpp +++ b/source/textparser.cpp @@ -1,11 +1,4 @@ -/* $Id$ - -This file is part of libmspdatafile -Copyright © 2007 Mikko Rasa, Mikkosoft Productions -Distributed under the LGPL -*/ - -#include +#include #include #include "input.h" #include "textparser.h" @@ -28,74 +21,73 @@ Statement TextParser::parse() 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(Value(INTEGER, token.str)); + result.append(lexical_cast(token.str)); else if(token.type==Token::FLOAT) - result.args.push_back(Value(FLOAT, token.str)); + result.append(lexical_cast(token.str)); else if(token.type==Token::STRING) - result.args.push_back(Value(STRING, token.str)); + result.append(token.str); else if(token.type==Token::IDENTIFIER) { if(token.str=="true") - result.args.push_back(Value(BOOLEAN, "1")); + result.append(true); else if(token.str=="false") - result.args.push_back(Value(BOOLEAN, "0")); + result.append(false); else - result.args.push_back(Value(ENUM, token.str)); - //result.args.push_back(resolve_identifiertoken.str); + result.append(Symbol(token.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; @@ -103,31 +95,33 @@ Statement TextParser::parse_statement(const Token *t) 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; - else if(!isspace(c) && !comment) - break; + comment = 0; + else if(c!=-1 && !isspace(c) && !comment) + 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 { @@ -143,6 +137,7 @@ Token TextParser::parse_token() OCTAL, FLOAT, FLOATEXP, + STRING_END, IDENTIFIER }; @@ -152,7 +147,7 @@ Token TextParser::parse_token() Token::SPECIAL, Token::SPECIAL, Token::SPECIAL, - Token::STRING, + Token::SPECIAL, Token::SPECIAL, Token::INTEGER, Token::INTEGER, @@ -160,138 +155,150 @@ Token TextParser::parse_token() Token::INTEGER, Token::FLOAT, Token::FLOAT, + Token::STRING, Token::IDENTIFIER }; - ParseState state=INIT; - string buf; - bool escape=false; + ParseState state = INIT; + string buf; + 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; - else if(isalpha(c) || c=='_') - state=IDENTIFIER; + state = DECIMAL; + else if(isalpha(c) || c=='_' || c=='\\') + 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) - { - try - { - return Token(Token::STRING, c_unescape(buf.substr(1, buf.size()-2))); - } - catch(const Exception &e) - { - throw ParseError(format("%s: %s", get_location(), e.what()), src, in.get_line_number()); - } - } + state = STRING_END; else - escape=false; + escape = false; break; case IDENTIFIER: - if(!isalpha(c) && !isdigit(c) && c!='_') - parse_error(c, state); + if(!isalpha(c) && !isdigit(c) && c!='_' && c!='-' && c!='/') + parse_error(c, "0-9A-Za-z_/-"); break; + case STRING_END: + throw_at(ParseError("Garbage after string"), get_location()); + 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 Token(token_type[state], buf); + { + if(state==IDENTIFIER && buf[0]=='\\') + return Token(Token::IDENTIFIER, buf.substr(1)); + else if(state==STRING_END) + { + try + { + return Token(Token::STRING, c_unescape(buf.substr(1, buf.size()-2))); + } + catch(Exception &e) + { + e.at(get_location()); + throw; + } + } + else + return Token(token_type[state], buf); + } } return Token(Token::SPECIAL, ""); @@ -314,9 +321,9 @@ string TextParser::get_location() 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(c), state), src, in.get_line_number()); + throw_at(ParseError(format("Parse error at '%c', expected one of \"%s\"", static_cast(c), e)), get_location()); } } // namespace DataFile