Some refactoring of TextParser logic

[libs/datafile.git] / source / textparser.cpp

diff --git a/source/textparser.cpp b/source/textparser.cpp
index c4eb03c7d91ef21e2b1af8d707e2f1363cc7fc4e..a7e8ee9d6abaf4d5198824fde47e811651996777 100644 (file)
--- a/source/textparser.cpp
+++ b/source/textparser.cpp
@@ -1,5 +1,6 @@
 #include <msp/strings/format.h>
 #include <msp/strings/utils.h>
+#include "except.h"
 #include "input.h"
 #include "textparser.h"
 #include "token.h"
@@ -9,28 +10,6 @@ using namespace std;
 namespace Msp {
 namespace DataFile {
 
-class parse_error: public runtime_error
-{
-public:
-       parse_error(const std::string &t):
-               runtime_error(t.empty() ? "at end of input" : format("after '%s'", t))
-       { }
-
-       virtual ~parse_error() throw() { }
-};
-
-
-class syntax_error: public runtime_error
-{
-public:
-       syntax_error(const std::string &t):
-               runtime_error(t.empty() ? "at end of input" : format("at '%s'", t))
-       { }
-
-       virtual ~syntax_error() throw() { }
-};
-
-
 TextParser::TextParser(Input &i, const string &s):
        ParserMode(i, s)
 { }
@@ -43,8 +22,7 @@ Statement TextParser::parse()
 Statement TextParser::parse_statement(const Token *t)
 {
        Statement result;
-       bool sub = false;
-       bool finish = false;
+       unsigned sub = 0;
 
        while(in)
        {
@@ -68,27 +46,24 @@ Statement TextParser::parse_statement(const Token *t)
                        result.source = src;
                        result.line = in.get_line_number();
                }
-               else if(sub)
+               else if(sub==1)
                {
                        if(token.str=="}")
-                       {
-                               sub = false;
-                               finish = true;
-                       }
+                               sub = 2;
                        else
                        {
                                Statement ss = parse_statement(&token);
                                result.sub.push_back(ss);
                        }
                }
-               else if(finish)
+               else if(sub==2)
                {
                        if(token.str!=";")
                                throw syntax_error(token.str);
                        break;
                }
                else if(token.str=="{")
-                       sub = true;
+                       sub = 1;
                else if(token.str==";")
                        break;
                else if(token.type==Token::INTEGER)
@@ -124,9 +99,9 @@ Token TextParser::parse_token()
                c = in.get();
                int next = in.peek();
 
-               if(c=='/' && next=='/')
+               if(c=='/' && next=='/' && !comment)
                        comment = 1;
-               else if(c=='/' && next=='*')
+               else if(c=='/' && next=='*' && !comment)
                        comment = 2;
                else if(c=='\n' && comment==1)
                        comment = 0;
@@ -150,6 +125,7 @@ Token TextParser::parse_token()
                FLOATEXPINIT,
                FLOATEXPSIGN,
                STRING,
+               STRING_ESCAPE,
                ACCEPT,
                ZERO,
                DECIMAL,
@@ -169,6 +145,7 @@ Token TextParser::parse_token()
                Token::SPECIAL,
                Token::SPECIAL,
                Token::SPECIAL,
+               Token::SPECIAL,
                Token::INTEGER,
                Token::INTEGER,
                Token::INTEGER,
@@ -181,7 +158,6 @@ Token TextParser::parse_token()
 
        ParseState state = INIT;
        string buf;
-       bool escape = false;
 
        while(in || state==INIT)
        {
@@ -237,6 +213,8 @@ Token TextParser::parse_token()
                case DECIMAL:
                        if(c=='.')
                                state = FLOAT;
+                       else if(c=='e' || c=='E')
+                               state = FLOATEXPINIT;
                        else if(!isdigit(c))
                                throw parse_error(buf);
                        break;
@@ -281,11 +259,13 @@ Token TextParser::parse_token()
 
                case STRING:
                        if(c=='\\')
-                               escape = !escape;
-                       else if(c=='"' && !escape)
+                               state = STRING_ESCAPE;
+                       else if(c=='"')
                                state = STRING_END;
-                       else
-                               escape = false;
+                       break;
+
+               case STRING_ESCAPE:
+                       state = STRING;
                        break;
 
                case IDENTIFIER: