static int parseEscapedCharacter (void)
{
int d = '\0';
int c = fileGetc ();
switch (c)
{
case 'A': d = '@'; break;
case 'B': d = '\b'; break;
case 'C': d = '^'; break;
case 'D': d = '$'; break;
case 'F': d = '\f'; break;
case 'H': d = '\\'; break;
case 'L': d = '~'; break;
case 'N': d = '\n'; break;
#ifdef QDOS
case 'Q': d = 0x9F; break;
#else
case 'Q': d = '`'; break;
#endif
case 'R': d = '\r'; break;
case 'S': d = '#'; break;
case 'T': d = '\t'; break;
case 'U': d = '\0'; break;
case 'V': d = '|'; break;
case '%': d = '%'; break;
case '\'': d = '\''; break;
case '"': d = '"'; break;
case '(': d = '['; break;
case ')': d = ']'; break;
case '<': d = '{'; break;
case '>': d = '}'; break;
case '\n': skipToCharacter ('%'); break;
case '/':
{
vString *string = parseInteger ('\0');
const char *value = vStringValue (string);
const unsigned long ascii = atol (value);
vStringDelete (string);
c = fileGetc ();
if (c == '/' && ascii < 256)
d = ascii;
break;
}
default: break;
}
return d;
}
static int parseCharacter (void)
{
int c = getcFromInputFile ();
int result = c;
if (c == '%')
result = parseEscapedCharacter ();
c = getcFromInputFile ();
if (c != '\'')
skipToCharacter ('\n');
return result;
}
static void readToken (tokenInfo *const token)
{
int c;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
do
{
c = fileGetc ();
}
while (c == '\t' || c == ' ' || c == '\n');
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
switch (c)
{
case EOF: longjmp (Exception, (int)ExceptionEOF); break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case ',': token->type = TOKEN_COMMA; break;
case '.': token->type = TOKEN_PERIOD; break;
case ':': token->type = TOKEN_COLON; break;
case '{': token->type = TOKEN_OPEN_CURLY; break;
case '}': token->type = TOKEN_CLOSE_CURLY; break;
case '=': token->type = TOKEN_EQUAL_SIGN; break;
case '[': token->type = TOKEN_OPEN_SQUARE; break;
case ']': token->type = TOKEN_CLOSE_SQUARE; break;
case '\'':
case '"':
token->type = TOKEN_STRING;
parseString (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '\\':
c = fileGetc ();
if (c != '\\' && c != '"' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_CHARACTER;
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '/':
{
int d = fileGetc ();
if ( (d != '*') && /* is this the start of a comment? */
(d != '/') ) /* is a one line comment? */
{
fileUngetc (d);
switch (LastTokenType)
{
case TOKEN_CHARACTER:
case TOKEN_KEYWORD:
case TOKEN_IDENTIFIER:
case TOKEN_STRING:
case TOKEN_CLOSE_CURLY:
case TOKEN_CLOSE_PAREN:
case TOKEN_CLOSE_SQUARE:
token->type = TOKEN_FORWARD_SLASH;
break;
default:
token->type = TOKEN_REGEXP;
parseRegExp ();
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
}
}
else
{
if (d == '*')
{
do
{
skipToCharacter ('*');
c = fileGetc ();
if (c == '/')
break;
else
fileUngetc (c);
} while (c != EOF && c != '\0');
goto getNextChar;
}
else if (d == '/') /* is this the start of a comment? */
{
skipToCharacter ('\n');
goto getNextChar;
}
}
break;
}
default:
if (! isIdentChar (c))
token->type = TOKEN_UNDEFINED;
else
{
parseIdentifier (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
token->keyword = analyzeToken (token->string);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
break;
}
LastTokenType = token->type;
}
static void readTokenFull (tokenInfo *const token, boolean include_newlines, vString *const repr)
{
int c;
int i;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
i = 0;
do
{
c = fileGetc ();
i++;
}
while (c == '\t' || c == ' ' ||
((c == '\r' || c == '\n') && ! include_newlines));
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
if (repr)
{
if (i > 1)
vStringPut (repr, ' ');
vStringPut (repr, c);
}
switch (c)
{
case EOF: token->type = TOKEN_EOF; break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case ',': token->type = TOKEN_COMMA; break;
case '.': token->type = TOKEN_PERIOD; break;
case ':': token->type = TOKEN_COLON; break;
case '{': token->type = TOKEN_OPEN_CURLY; break;
case '}': token->type = TOKEN_CLOSE_CURLY; break;
case '=': token->type = TOKEN_EQUAL_SIGN; break;
case '[': token->type = TOKEN_OPEN_SQUARE; break;
case ']': token->type = TOKEN_CLOSE_SQUARE; break;
case '+':
case '-':
{
int d = fileGetc ();
if (d == c) /* ++ or -- */
token->type = TOKEN_POSTFIX_OPERATOR;
else
{
fileUngetc (d);
token->type = TOKEN_BINARY_OPERATOR;
}
break;
}
case '*':
case '%':
case '?':
case '>':
case '<':
case '^':
case '|':
case '&':
token->type = TOKEN_BINARY_OPERATOR;
break;
case '\r':
case '\n':
/* This isn't strictly correct per the standard, but following the
* real rules means understanding all statements, and that's not
* what the parser currently does. What we do here is a guess, by
* avoiding inserting semicolons that would make the statement on
* the left invalid. Hopefully this should not have false negatives
* (e.g. should not miss insertion of a semicolon) but might have
* false positives (e.g. it will wrongfully emit a semicolon for the
* newline in "foo\n+bar").
* This should however be mostly harmless as we only deal with
* newlines in specific situations where we know a false positive
* wouldn't hurt too bad. */
switch (LastTokenType)
{
/* these cannot be the end of a statement, so hold the newline */
case TOKEN_EQUAL_SIGN:
case TOKEN_COLON:
case TOKEN_PERIOD:
case TOKEN_FORWARD_SLASH:
case TOKEN_BINARY_OPERATOR:
/* and these already end one, no need to duplicate it */
case TOKEN_SEMICOLON:
case TOKEN_COMMA:
case TOKEN_CLOSE_CURLY:
case TOKEN_OPEN_CURLY:
include_newlines = FALSE; /* no need to recheck */
goto getNextChar;
break;
default:
token->type = TOKEN_SEMICOLON;
}
break;
case '\'':
case '"':
token->type = TOKEN_STRING;
parseString (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
if (repr)
{
vStringCat (repr, token->string);
vStringPut (repr, c);
}
break;
case '\\':
c = fileGetc ();
if (c != '\\' && c != '"' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_CHARACTER;
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '/':
{
int d = fileGetc ();
if ( (d != '*') && /* is this the start of a comment? */
(d != '/') ) /* is a one line comment? */
{
fileUngetc (d);
switch (LastTokenType)
{
case TOKEN_CHARACTER:
case TOKEN_IDENTIFIER:
case TOKEN_STRING:
case TOKEN_CLOSE_CURLY:
case TOKEN_CLOSE_PAREN:
case TOKEN_CLOSE_SQUARE:
token->type = TOKEN_FORWARD_SLASH;
break;
default:
token->type = TOKEN_REGEXP;
parseRegExp ();
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
}
}
else
{
if (repr) /* remove the / we added */
repr->buffer[--repr->length] = 0;
if (d == '*')
{
do
{
skipToCharacter ('*');
c = fileGetc ();
if (c == '/')
break;
else
fileUngetc (c);
} while (c != EOF && c != '\0');
goto getNextChar;
}
else if (d == '/') /* is this the start of a comment? */
{
skipToCharacter ('\n');
/* if we care about newlines, put it back so it is seen */
if (include_newlines)
fileUngetc ('\n');
goto getNextChar;
}
}
break;
}
case '#':
/* skip shebang in case of e.g. Node.js scripts */
if (token->lineNumber > 1)
token->type = TOKEN_UNDEFINED;
else if ((c = fileGetc ()) != '!')
{
fileUngetc (c);
token->type = TOKEN_UNDEFINED;
}
else
{
skipToCharacter ('\n');
goto getNextChar;
}
break;
default:
if (! isIdentChar (c))
token->type = TOKEN_UNDEFINED;
else
{
parseIdentifier (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
token->keyword = analyzeToken (token->string);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
if (repr && vStringLength (token->string) > 1)
vStringCatS (repr, vStringValue (token->string) + 1);
}
break;
}
LastTokenType = token->type;
}
static void readToken (tokenInfo *const token)
{
int c;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
if (! InPhp)
{
c = findPhpStart ();
if (c != EOF)
InPhp = TRUE;
}
else
c = getcFromInputFile ();
c = skipWhitespaces (c);
token->lineNumber = getInputLineNumber ();
token->filePosition = getInputFilePosition ();
switch (c)
{
case EOF: token->type = TOKEN_EOF; break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case ',': token->type = TOKEN_COMMA; break;
case '.': token->type = TOKEN_PERIOD; break;
case ':': token->type = TOKEN_COLON; break;
case '{': token->type = TOKEN_OPEN_CURLY; break;
case '}': token->type = TOKEN_CLOSE_CURLY; break;
case '[': token->type = TOKEN_OPEN_SQUARE; break;
case ']': token->type = TOKEN_CLOSE_SQUARE; break;
case '&': token->type = TOKEN_AMPERSAND; break;
case '\\': token->type = TOKEN_BACKSLASH; break;
case '=':
{
int d = getcFromInputFile ();
if (d == '=' || d == '>')
token->type = TOKEN_OPERATOR;
else
{
ungetcToInputFile (d);
token->type = TOKEN_EQUAL_SIGN;
}
break;
}
case '\'':
case '"':
token->type = TOKEN_STRING;
parseString (token->string, c);
token->lineNumber = getInputLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '<':
{
int d = getcFromInputFile ();
if (d == '/')
{
/* </script[:white:]*> */
if (tolower ((d = getcFromInputFile ())) == 's' &&
tolower ((d = getcFromInputFile ())) == 'c' &&
tolower ((d = getcFromInputFile ())) == 'r' &&
tolower ((d = getcFromInputFile ())) == 'i' &&
tolower ((d = getcFromInputFile ())) == 'p' &&
tolower ((d = getcFromInputFile ())) == 't' &&
(d = skipWhitespaces (getcFromInputFile ())) == '>')
{
InPhp = FALSE;
goto getNextChar;
}
else
{
ungetcToInputFile (d);
token->type = TOKEN_UNDEFINED;
}
}
else if (d == '<' && (d = getcFromInputFile ()) == '<')
{
token->type = TOKEN_STRING;
parseHeredoc (token->string);
}
else
{
ungetcToInputFile (d);
token->type = TOKEN_UNDEFINED;
}
break;
}
case '#': /* comment */
skipSingleComment ();
goto getNextChar;
break;
case '+':
case '-':
case '*':
case '%':
{
int d = getcFromInputFile ();
if (d != '=' && ! (c == '-' && d == '>'))
ungetcToInputFile (d);
token->type = TOKEN_OPERATOR;
break;
}
case '/': /* division or comment start */
{
int d = getcFromInputFile ();
if (d == '/') /* single-line comment */
{
skipSingleComment ();
goto getNextChar;
}
else if (d == '*')
{
do
{
c = skipToCharacter ('*');
if (c != EOF)
{
c = getcFromInputFile ();
if (c == '/')
break;
else
ungetcToInputFile (c);
}
} while (c != EOF && c != '\0');
goto getNextChar;
}
else
{
if (d != '=')
ungetcToInputFile (d);
token->type = TOKEN_OPERATOR;
}
break;
}
case '$': /* variable start */
{
int d = getcFromInputFile ();
if (! isIdentChar (d))
{
ungetcToInputFile (d);
token->type = TOKEN_UNDEFINED;
}
else
{
parseIdentifier (token->string, d);
token->type = TOKEN_VARIABLE;
}
break;
}
case '?': /* maybe the end of the PHP chunk */
{
int d = getcFromInputFile ();
if (d == '>')
{
InPhp = FALSE;
goto getNextChar;
}
else
{
ungetcToInputFile (d);
token->type = TOKEN_UNDEFINED;
}
break;
}
default:
if (! isIdentChar (c))
token->type = TOKEN_UNDEFINED;
else
{
parseIdentifier (token->string, c);
token->keyword = analyzeToken (token->string, getInputLanguage ());
if (token->keyword == KEYWORD_NONE)
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
break;
}
if (token->type == TOKEN_SEMICOLON ||
token->type == TOKEN_OPEN_CURLY ||
token->type == TOKEN_CLOSE_CURLY)
{
/* reset current statement details on statement end, and when entering
* a deeper scope.
* it is a bit ugly to do this in readToken(), but it makes everything
* a lot simpler. */
CurrentStatement.access = ACCESS_UNDEFINED;
CurrentStatement.impl = IMPL_UNDEFINED;
}
}
static void readToken (tokenInfo *const token)
{
int c;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
do
c = getcFromInputFile ();
while (c == '\t' || c == ' ' || c == '\n');
switch (c)
{
case EOF: token->type = TOKEN_EOF; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case '!': token->type = TOKEN_BANG; break;
case '}': token->type = TOKEN_CLOSE_BRACE; break;
case ']': token->type = TOKEN_CLOSE_BRACKET; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ',': token->type = TOKEN_COMMA; break;
case '$': token->type = TOKEN_DOLLAR; break;
case '.': token->type = TOKEN_DOT; break;
case '{': token->type = TOKEN_OPEN_BRACE; break;
case '[': token->type = TOKEN_OPEN_BRACKET; break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case '~': token->type = TOKEN_TILDE; break;
case '+':
case '*':
case '^':
case '=': token->type = TOKEN_OPERATOR; break;
case '-':
c = getcFromInputFile ();
if (c == '>')
token->type = TOKEN_CONSTRAINT;
else if (c == '-') /* is this the start of a comment? */
{
skipToCharacter ('\n');
goto getNextChar;
}
else
{
if (!isspace (c))
ungetcToInputFile (c);
token->type = TOKEN_OPERATOR;
}
break;
case '?':
case ':':
{
int c2 = getcFromInputFile ();
if (c2 == '=')
token->type = TOKEN_OPERATOR;
else
{
if (!isspace (c2))
ungetcToInputFile (c2);
if (c == ':')
token->type = TOKEN_COLON;
else
token->type = TOKEN_QUESTION;
}
break;
}
case '<':
c = getcFromInputFile ();
if (c != '=' && c != '>' && !isspace (c))
ungetcToInputFile (c);
token->type = TOKEN_OPERATOR;
break;
case '>':
c = getcFromInputFile ();
if (c != '=' && c != '>' && !isspace (c))
ungetcToInputFile (c);
token->type = TOKEN_OPERATOR;
break;
case '/':
c = getcFromInputFile ();
if (c != '/' && c != '=' && !isspace (c))
ungetcToInputFile (c);
token->type = TOKEN_OPERATOR;
break;
case '\\':
c = getcFromInputFile ();
if (c != '\\' && !isspace (c))
ungetcToInputFile (c);
token->type = TOKEN_OPERATOR;
break;
case '"':
token->type = TOKEN_STRING;
parseString (token->string);
break;
case '\'':
token->type = TOKEN_CHARACTER;
parseCharacter ();
break;
default:
if (isalpha (c))
{
parseIdentifier (token->string, c);
token->keyword = analyzeToken (token->string, Lang_eiffel);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
else if (isdigit (c))
{
vString* numeric = parseNumeric (c);
vStringCat (token->string, numeric);
vStringDelete (numeric);
token->type = TOKEN_NUMERIC;
}
else if (isFreeOperatorChar (c))
{
parseFreeOperator (token->string, c);
token->type = TOKEN_OPERATOR;
}
else
token->type = TOKEN_UNDEFINED;
break;
}
}
static void readToken (tokenInfo *const token)
{
int c;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
do
c = fileGetc ();
while (c == '\t' || c == ' ' || c == '\n');
switch (c)
{
case EOF:
longjmp (Exception, (int)ExceptionEOF);
break;
case '!':
token->type = TOKEN_BANG;
break;
case '$':
token->type = TOKEN_DOLLAR;
break;
case '(':
token->type = TOKEN_OPEN_PAREN;
break;
case ')':
token->type = TOKEN_CLOSE_PAREN;
break;
case ',':
token->type = TOKEN_COMMA;
break;
case '.':
token->type = TOKEN_DOT;
break;
case ';':
goto getNextChar;
case '[':
token->type = TOKEN_OPEN_BRACKET;
break;
case ']':
token->type = TOKEN_CLOSE_BRACKET;
break;
case '{':
token->type = TOKEN_OPEN_BRACE;
break;
case '}':
token->type = TOKEN_CLOSE_BRACE;
break;
case '~':
token->type = TOKEN_TILDE;
break;
case '+':
case '*':
case '^':
case '=':
token->type = TOKEN_OPERATOR;
break;
case '-':
c = fileGetc ();
if (c == '>')
token->type = TOKEN_CONSTRAINT;
else if (c == '-') /* is this the start of a comment? */
{
skipToCharacter ('\n');
goto getNextChar;
}
else
{
if (!isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
}
break;
case '?':
case ':':
c = fileGetc ();
if (c == '=')
token->type = TOKEN_OPERATOR;
else
{
token->type = TOKEN_COLON;
if (!isspace (c))
fileUngetc (c);
}
break;
case '<':
c = fileGetc ();
if (c != '=' && c != '>' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
break;
case '>':
c = fileGetc ();
if (c != '=' && c != '>' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
break;
case '/':
c = fileGetc ();
if (c != '/' && c != '=' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
break;
case '\\':
c = fileGetc ();
if (c != '\\' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
break;
case '"':
token->type = TOKEN_STRING;
parseString (token->string);
break;
case '\'':
token->type = TOKEN_CHARACTER;
parseCharacter ();
break;
default:
if (isalpha (c))
{
parseIdentifier (token->string, c);
token->keyword = analyzeToken (token->string);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
else if (isdigit (c))
{
vStringCat (token->string, parseNumeric (c));
token->type = TOKEN_NUMERIC;
}
else if (isFreeOperatorChar (c))
{
parseFreeOperator (token->string, c);
token->type = TOKEN_OPERATOR;
}
else
{
token->type = TOKEN_UNDEFINED;
Assert (! isType (token, TOKEN_UNDEFINED));
}
break;
}
}
static void readToken (tokenInfo *const token)
{
int c;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
do
c = fileGetc ();
while (c == '\t' || c == ' ' || c == '\n');
switch (c)
{
case EOF: longjmp (Exception, (int)ExceptionEOF); break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case ',': token->type = TOKEN_COMMA; break;
case '\'':
case '"':
token->type = TOKEN_STRING;
parseString (token->string, c);
break;
case '-':
c = fileGetc ();
if (c == '-') /* is this the start of a comment? */
{
skipToCharacter ('\n');
goto getNextChar;
}
else
{
if (!isspace (c))
fileUngetc (c);
token->type = TOKEN_OPERATOR;
}
break;
case '/':
{
int d = fileGetc ();
if (d != '*') /* is this the start of a comment? */
fileUngetc (d);
else
{
do
{
skipToCharacter ('*');
c = fileGetc ();
if (c == '/')
break;
else
fileUngetc (c);
} while (c != '\0');
goto getNextChar;
}
break;
}
default:
if (! isIdentChar1 (c))
token->type = TOKEN_UNDEFINED;
else
{
parseIdentifier (token->string, c);
token->keyword = analyzeToken (token->string);
if (isKeyword (token, KEYWORD_rem))
{
vStringClear (token->string);
skipToCharacter ('\n');
goto getNextChar;
}
else if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
break;
}
}
