%PDF- %PDF-
Direktori : /usr/lib/python3.6/site-packages/pycparser/ |
Current File : //usr/lib/python3.6/site-packages/pycparser/c_lexer.py |
#------------------------------------------------------------------------------ # pycparser: c_lexer.py # # CLexer class: lexer for the C language # # Copyright (C) 2008-2015, Eli Bendersky # License: BSD #------------------------------------------------------------------------------ import re import sys from ply import lex from ply.lex import TOKEN class CLexer(object): """ A lexer for the C language. After building it, set the input text with input(), and call token() to get new tokens. The public attribute filename can be set to an initial filaneme, but the lexer will update it upon #line directives. """ def __init__(self, error_func, on_lbrace_func, on_rbrace_func, type_lookup_func): """ Create a new Lexer. error_func: An error function. Will be called with an error message, line and column as arguments, in case of an error during lexing. on_lbrace_func, on_rbrace_func: Called when an LBRACE or RBRACE is encountered (likely to push/pop type_lookup_func's scope) type_lookup_func: A type lookup function. Given a string, it must return True IFF this string is a name of a type that was defined with a typedef earlier. """ self.error_func = error_func self.on_lbrace_func = on_lbrace_func self.on_rbrace_func = on_rbrace_func self.type_lookup_func = type_lookup_func self.filename = '' # Keeps track of the last token returned from self.token() self.last_token = None # Allow either "# line" or "# <num>" to support GCC's # cpp output # self.line_pattern = re.compile('([ \t]*line\W)|([ \t]*\d+)') self.pragma_pattern = re.compile('[ \t]*pragma\W') def build(self, **kwargs): """ Builds the lexer from the specification. Must be called after the lexer object is created. This method exists separately, because the PLY manual warns against calling lex.lex inside __init__ """ self.lexer = lex.lex(object=self, **kwargs) def reset_lineno(self): """ Resets the internal line number counter of the lexer. """ self.lexer.lineno = 1 def input(self, text): self.lexer.input(text) def token(self): self.last_token = self.lexer.token() return self.last_token def find_tok_column(self, token): """ Find the column of the token in its line. """ last_cr = self.lexer.lexdata.rfind('\n', 0, token.lexpos) return token.lexpos - last_cr ######################-- PRIVATE --###################### ## ## Internal auxiliary methods ## def _error(self, msg, token): location = self._make_tok_location(token) self.error_func(msg, location[0], location[1]) self.lexer.skip(1) def _make_tok_location(self, token): return (token.lineno, self.find_tok_column(token)) ## ## Reserved keywords ## keywords = ( '_BOOL', '_COMPLEX', 'AUTO', 'BREAK', 'CASE', 'CHAR', 'CONST', 'CONTINUE', 'DEFAULT', 'DO', 'DOUBLE', 'ELSE', 'ENUM', 'EXTERN', 'FLOAT', 'FOR', 'GOTO', 'IF', 'INLINE', 'INT', 'LONG', 'REGISTER', 'OFFSETOF', 'RESTRICT', 'RETURN', 'SHORT', 'SIGNED', 'SIZEOF', 'STATIC', 'STRUCT', 'SWITCH', 'TYPEDEF', 'UNION', 'UNSIGNED', 'VOID', 'VOLATILE', 'WHILE', ) keyword_map = {} for keyword in keywords: if keyword == '_BOOL': keyword_map['_Bool'] = keyword elif keyword == '_COMPLEX': keyword_map['_Complex'] = keyword else: keyword_map[keyword.lower()] = keyword ## ## All the tokens recognized by the lexer ## tokens = keywords + ( # Identifiers 'ID', # Type identifiers (identifiers previously defined as # types with typedef) 'TYPEID', # constants 'INT_CONST_DEC', 'INT_CONST_OCT', 'INT_CONST_HEX', 'INT_CONST_BIN', 'FLOAT_CONST', 'HEX_FLOAT_CONST', 'CHAR_CONST', 'WCHAR_CONST', # String literals 'STRING_LITERAL', 'WSTRING_LITERAL', # Operators 'PLUS', 'MINUS', 'TIMES', 'DIVIDE', 'MOD', 'OR', 'AND', 'NOT', 'XOR', 'LSHIFT', 'RSHIFT', 'LOR', 'LAND', 'LNOT', 'LT', 'LE', 'GT', 'GE', 'EQ', 'NE', # Assignment 'EQUALS', 'TIMESEQUAL', 'DIVEQUAL', 'MODEQUAL', 'PLUSEQUAL', 'MINUSEQUAL', 'LSHIFTEQUAL','RSHIFTEQUAL', 'ANDEQUAL', 'XOREQUAL', 'OREQUAL', # Increment/decrement 'PLUSPLUS', 'MINUSMINUS', # Structure dereference (->) 'ARROW', # Conditional operator (?) 'CONDOP', # Delimeters 'LPAREN', 'RPAREN', # ( ) 'LBRACKET', 'RBRACKET', # [ ] 'LBRACE', 'RBRACE', # { } 'COMMA', 'PERIOD', # . , 'SEMI', 'COLON', # ; : # Ellipsis (...) 'ELLIPSIS', # pre-processor 'PPHASH', # '#' ) ## ## Regexes for use in tokens ## ## # valid C identifiers (K&R2: A.2.3), plus '$' (supported by some compilers) identifier = r'[a-zA-Z_$][0-9a-zA-Z_$]*' hex_prefix = '0[xX]' hex_digits = '[0-9a-fA-F]+' bin_prefix = '0[bB]' bin_digits = '[01]+' # integer constants (K&R2: A.2.5.1) integer_suffix_opt = r'(([uU]ll)|([uU]LL)|(ll[uU]?)|(LL[uU]?)|([uU][lL])|([lL][uU]?)|[uU])?' decimal_constant = '(0'+integer_suffix_opt+')|([1-9][0-9]*'+integer_suffix_opt+')' octal_constant = '0[0-7]*'+integer_suffix_opt hex_constant = hex_prefix+hex_digits+integer_suffix_opt bin_constant = bin_prefix+bin_digits+integer_suffix_opt bad_octal_constant = '0[0-7]*[89]' # character constants (K&R2: A.2.5.2) # Note: a-zA-Z and '.-~^_!=&;,' are allowed as escape chars to support #line # directives with Windows paths as filenames (..\..\dir\file) # For the same reason, decimal_escape allows all digit sequences. We want to # parse all correct code, even if it means to sometimes parse incorrect # code. # simple_escape = r"""([a-zA-Z._~!=&\^\-\\?'"])""" decimal_escape = r"""(\d+)""" hex_escape = r"""(x[0-9a-fA-F]+)""" bad_escape = r"""([\\][^a-zA-Z._~^!=&\^\-\\?'"x0-7])""" escape_sequence = r"""(\\("""+simple_escape+'|'+decimal_escape+'|'+hex_escape+'))' cconst_char = r"""([^'\\\n]|"""+escape_sequence+')' char_const = "'"+cconst_char+"'" wchar_const = 'L'+char_const unmatched_quote = "('"+cconst_char+"*\\n)|('"+cconst_char+"*$)" bad_char_const = r"""('"""+cconst_char+"""[^'\n]+')|('')|('"""+bad_escape+r"""[^'\n]*')""" # string literals (K&R2: A.2.6) string_char = r"""([^"\\\n]|"""+escape_sequence+')' string_literal = '"'+string_char+'*"' wstring_literal = 'L'+string_literal bad_string_literal = '"'+string_char+'*'+bad_escape+string_char+'*"' # floating constants (K&R2: A.2.5.3) exponent_part = r"""([eE][-+]?[0-9]+)""" fractional_constant = r"""([0-9]*\.[0-9]+)|([0-9]+\.)""" floating_constant = '(((('+fractional_constant+')'+exponent_part+'?)|([0-9]+'+exponent_part+'))[FfLl]?)' binary_exponent_part = r'''([pP][+-]?[0-9]+)''' hex_fractional_constant = '((('+hex_digits+r""")?\."""+hex_digits+')|('+hex_digits+r"""\.))""" hex_floating_constant = '('+hex_prefix+'('+hex_digits+'|'+hex_fractional_constant+')'+binary_exponent_part+'[FfLl]?)' ## ## Lexer states: used for preprocessor \n-terminated directives ## states = ( # ppline: preprocessor line directives # ('ppline', 'exclusive'), # pppragma: pragma # ('pppragma', 'exclusive'), ) def t_PPHASH(self, t): r'[ \t]*\#' if self.line_pattern.match(t.lexer.lexdata, pos=t.lexer.lexpos): t.lexer.begin('ppline') self.pp_line = self.pp_filename = None elif self.pragma_pattern.match(t.lexer.lexdata, pos=t.lexer.lexpos): t.lexer.begin('pppragma') else: t.type = 'PPHASH' return t ## ## Rules for the ppline state ## @TOKEN(string_literal) def t_ppline_FILENAME(self, t): if self.pp_line is None: self._error('filename before line number in #line', t) else: self.pp_filename = t.value.lstrip('"').rstrip('"') @TOKEN(decimal_constant) def t_ppline_LINE_NUMBER(self, t): if self.pp_line is None: self.pp_line = t.value else: # Ignore: GCC's cpp sometimes inserts a numeric flag # after the file name pass def t_ppline_NEWLINE(self, t): r'\n' if self.pp_line is None: self._error('line number missing in #line', t) else: self.lexer.lineno = int(self.pp_line) if self.pp_filename is not None: self.filename = self.pp_filename t.lexer.begin('INITIAL') def t_ppline_PPLINE(self, t): r'line' pass t_ppline_ignore = ' \t' def t_ppline_error(self, t): self._error('invalid #line directive', t) ## ## Rules for the pppragma state ## def t_pppragma_NEWLINE(self, t): r'\n' t.lexer.lineno += 1 t.lexer.begin('INITIAL') def t_pppragma_PPPRAGMA(self, t): r'pragma' pass t_pppragma_ignore = ' \t<>.-{}();=+-*/$%@&^~!?:,0123456789' @TOKEN(string_literal) def t_pppragma_STR(self, t): pass @TOKEN(identifier) def t_pppragma_ID(self, t): pass def t_pppragma_error(self, t): self._error('invalid #pragma directive', t) ## ## Rules for the normal state ## t_ignore = ' \t' # Newlines def t_NEWLINE(self, t): r'\n+' t.lexer.lineno += t.value.count("\n") # Operators t_PLUS = r'\+' t_MINUS = r'-' t_TIMES = r'\*' t_DIVIDE = r'/' t_MOD = r'%' t_OR = r'\|' t_AND = r'&' t_NOT = r'~' t_XOR = r'\^' t_LSHIFT = r'<<' t_RSHIFT = r'>>' t_LOR = r'\|\|' t_LAND = r'&&' t_LNOT = r'!' t_LT = r'<' t_GT = r'>' t_LE = r'<=' t_GE = r'>=' t_EQ = r'==' t_NE = r'!=' # Assignment operators t_EQUALS = r'=' t_TIMESEQUAL = r'\*=' t_DIVEQUAL = r'/=' t_MODEQUAL = r'%=' t_PLUSEQUAL = r'\+=' t_MINUSEQUAL = r'-=' t_LSHIFTEQUAL = r'<<=' t_RSHIFTEQUAL = r'>>=' t_ANDEQUAL = r'&=' t_OREQUAL = r'\|=' t_XOREQUAL = r'\^=' # Increment/decrement t_PLUSPLUS = r'\+\+' t_MINUSMINUS = r'--' # -> t_ARROW = r'->' # ? t_CONDOP = r'\?' # Delimeters t_LPAREN = r'\(' t_RPAREN = r'\)' t_LBRACKET = r'\[' t_RBRACKET = r'\]' t_COMMA = r',' t_PERIOD = r'\.' t_SEMI = r';' t_COLON = r':' t_ELLIPSIS = r'\.\.\.' # Scope delimiters # To see why on_lbrace_func is needed, consider: # typedef char TT; # void foo(int TT) { TT = 10; } # TT x = 5; # Outside the function, TT is a typedef, but inside (starting and ending # with the braces) it's a parameter. The trouble begins with yacc's # lookahead token. If we open a new scope in brace_open, then TT has # already been read and incorrectly interpreted as TYPEID. So, we need # to open and close scopes from within the lexer. # Similar for the TT immediately outside the end of the function. # @TOKEN(r'\{') def t_LBRACE(self, t): self.on_lbrace_func() return t @TOKEN(r'\}') def t_RBRACE(self, t): self.on_rbrace_func() return t t_STRING_LITERAL = string_literal # The following floating and integer constants are defined as # functions to impose a strict order (otherwise, decimal # is placed before the others because its regex is longer, # and this is bad) # @TOKEN(floating_constant) def t_FLOAT_CONST(self, t): return t @TOKEN(hex_floating_constant) def t_HEX_FLOAT_CONST(self, t): return t @TOKEN(hex_constant) def t_INT_CONST_HEX(self, t): return t @TOKEN(bin_constant) def t_INT_CONST_BIN(self, t): return t @TOKEN(bad_octal_constant) def t_BAD_CONST_OCT(self, t): msg = "Invalid octal constant" self._error(msg, t) @TOKEN(octal_constant) def t_INT_CONST_OCT(self, t): return t @TOKEN(decimal_constant) def t_INT_CONST_DEC(self, t): return t # Must come before bad_char_const, to prevent it from # catching valid char constants as invalid # @TOKEN(char_const) def t_CHAR_CONST(self, t): return t @TOKEN(wchar_const) def t_WCHAR_CONST(self, t): return t @TOKEN(unmatched_quote) def t_UNMATCHED_QUOTE(self, t): msg = "Unmatched '" self._error(msg, t) @TOKEN(bad_char_const) def t_BAD_CHAR_CONST(self, t): msg = "Invalid char constant %s" % t.value self._error(msg, t) @TOKEN(wstring_literal) def t_WSTRING_LITERAL(self, t): return t # unmatched string literals are caught by the preprocessor @TOKEN(bad_string_literal) def t_BAD_STRING_LITERAL(self, t): msg = "String contains invalid escape code" self._error(msg, t) @TOKEN(identifier) def t_ID(self, t): t.type = self.keyword_map.get(t.value, "ID") if t.type == 'ID' and self.type_lookup_func(t.value): t.type = "TYPEID" return t def t_error(self, t): msg = 'Illegal character %s' % repr(t.value[0]) self._error(msg, t)