1## @file
2# This file is used to parse and evaluate expression in directive or PCD value.
3#
4# Copyright (c) 2011 - 2016, Intel Corporation. All rights reserved.<BR>
5# This program and the accompanying materials
6# are licensed and made available under the terms and conditions of the BSD License
7# which accompanies this distribution.    The full text of the license may be found at
8# http://opensource.org/licenses/bsd-license.php
9#
10# THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,
11# WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.
12
13## Import Modules
14#
15from Common.GlobalData import *
16from CommonDataClass.Exceptions import BadExpression
17from CommonDataClass.Exceptions import WrnExpression
18from Misc import GuidStringToGuidStructureString
19
20ERR_STRING_EXPR         = 'This operator cannot be used in string expression: [%s].'
21ERR_SNYTAX              = 'Syntax error, the rest of expression cannot be evaluated: [%s].'
22ERR_MATCH               = 'No matching right parenthesis.'
23ERR_STRING_TOKEN        = 'Bad string token: [%s].'
24ERR_MACRO_TOKEN         = 'Bad macro token: [%s].'
25ERR_EMPTY_TOKEN         = 'Empty token is not allowed.'
26ERR_PCD_RESOLVE         = 'PCD token cannot be resolved: [%s].'
27ERR_VALID_TOKEN         = 'No more valid token found from rest of string: [%s].'
28ERR_EXPR_TYPE           = 'Different types found in expression.'
29ERR_OPERATOR_UNSUPPORT  = 'Unsupported operator: [%s]'
30ERR_REL_NOT_IN          = 'Expect "IN" after "not" operator.'
31WRN_BOOL_EXPR           = 'Operand of boolean type cannot be used in arithmetic expression.'
32WRN_EQCMP_STR_OTHERS    = '== Comparison between Operand of string type and Boolean/Number Type always return False.'
33WRN_NECMP_STR_OTHERS    = '!= Comparison between Operand of string type and Boolean/Number Type always return True.'
34ERR_RELCMP_STR_OTHERS   = 'Operator taking Operand of string type and Boolean/Number Type is not allowed: [%s].'
35ERR_STRING_CMP          = 'Unicode string and general string cannot be compared: [%s %s %s]'
36ERR_ARRAY_TOKEN         = 'Bad C array or C format GUID token: [%s].'
37ERR_ARRAY_ELE           = 'This must be HEX value for NList or Array: [%s].'
38ERR_EMPTY_EXPR          = 'Empty expression is not allowed.'
39ERR_IN_OPERAND          = 'Macro after IN operator can only be: $(FAMILY), $(ARCH), $(TOOL_CHAIN_TAG) and $(TARGET).'
40
41## SplitString
42#  Split string to list according double quote
43#  For example: abc"de\"f"ghi"jkl"mn will be: ['abc', '"de\"f"', 'ghi', '"jkl"', 'mn']
44#
45def SplitString(String):
46    # There might be escaped quote: "abc\"def\\\"ghi"
47    Str = String.replace('\\\\', '//').replace('\\\"', '\\\'')
48    RetList = []
49    InQuote = False
50    Item = ''
51    for i, ch in enumerate(Str):
52        if ch == '"':
53            InQuote = not InQuote
54            if not InQuote:
55                Item += String[i]
56                RetList.append(Item)
57                Item = ''
58                continue
59            if Item:
60                RetList.append(Item)
61                Item = ''
62        Item += String[i]
63    if InQuote:
64        raise BadExpression(ERR_STRING_TOKEN % Item)
65    if Item:
66        RetList.append(Item)
67    return RetList
68
69## ReplaceExprMacro
70#
71def ReplaceExprMacro(String, Macros, ExceptionList = None):
72    StrList = SplitString(String)
73    for i, String in enumerate(StrList):
74        InQuote = False
75        if String.startswith('"'):
76            InQuote = True
77        MacroStartPos = String.find('$(')
78        if MacroStartPos < 0:
79            for Pcd in gPlatformPcds.keys():
80                if Pcd in String:
81                    if Pcd not in gConditionalPcds:
82                        gConditionalPcds.append(Pcd)
83            continue
84        RetStr = ''
85        while MacroStartPos >= 0:
86            RetStr = String[0:MacroStartPos]
87            MacroEndPos = String.find(')', MacroStartPos)
88            if MacroEndPos < 0:
89                raise BadExpression(ERR_MACRO_TOKEN % String[MacroStartPos:])
90            Macro = String[MacroStartPos+2:MacroEndPos]
91            if Macro not in Macros:
92                # From C reference manual:
93                # If an undefined macro name appears in the constant-expression of
94                # !if or !elif, it is replaced by the integer constant 0.
95                RetStr += '0'
96            elif not InQuote:
97                Tklst = RetStr.split()
98                if Tklst and Tklst[-1] in ['IN', 'in'] and ExceptionList and Macro not in ExceptionList:
99                    raise BadExpression(ERR_IN_OPERAND)
100                # Make sure the macro in exception list is encapsulated by double quote
101                # For example: DEFINE ARCH = IA32 X64
102                # $(ARCH) is replaced with "IA32 X64"
103                if ExceptionList and Macro in ExceptionList:
104                    RetStr += '"' + Macros[Macro] + '"'
105                elif Macros[Macro].strip():
106                    RetStr += Macros[Macro]
107                else:
108                    RetStr += '""'
109            else:
110                RetStr += Macros[Macro]
111            RetStr += String[MacroEndPos+1:]
112            String = RetStr
113            MacroStartPos = String.find('$(')
114        StrList[i] = RetStr
115    return ''.join(StrList)
116
117SupportedInMacroList = ['TARGET', 'TOOL_CHAIN_TAG', 'ARCH', 'FAMILY']
118
119class ValueExpression(object):
120    # Logical operator mapping
121    LogicalOperators = {
122        '&&' : 'and', '||' : 'or',
123        '!'  : 'not', 'AND': 'and',
124        'OR' : 'or' , 'NOT': 'not',
125        'XOR': '^'  , 'xor': '^',
126        'EQ' : '==' , 'NE' : '!=',
127        'GT' : '>'  , 'LT' : '<',
128        'GE' : '>=' , 'LE' : '<=',
129        'IN' : 'in'
130    }
131
132    NonLetterOpLst = ['+', '-', '&', '|', '^', '!', '=', '>', '<']
133
134    PcdPattern = re.compile(r'[_a-zA-Z][0-9A-Za-z_]*\.[_a-zA-Z][0-9A-Za-z_]*$')
135    HexPattern = re.compile(r'0[xX][0-9a-fA-F]+$')
136    RegGuidPattern = re.compile(r'[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}')
137
138    SymbolPattern = re.compile("("
139                                 "\$\([A-Z][A-Z0-9_]*\)|\$\(\w+\.\w+\)|\w+\.\w+|"
140                                 "&&|\|\||!(?!=)|"
141                                 "(?<=\W)AND(?=\W)|(?<=\W)OR(?=\W)|(?<=\W)NOT(?=\W)|(?<=\W)XOR(?=\W)|"
142                                 "(?<=\W)EQ(?=\W)|(?<=\W)NE(?=\W)|(?<=\W)GT(?=\W)|(?<=\W)LT(?=\W)|(?<=\W)GE(?=\W)|(?<=\W)LE(?=\W)"
143                               ")")
144
145    @staticmethod
146    def Eval(Operator, Oprand1, Oprand2 = None):
147        WrnExp = None
148
149        if Operator not in ["==", "!=", ">=", "<=", ">", "<", "in", "not in"] and \
150            (type(Oprand1) == type('') or type(Oprand2) == type('')):
151            raise BadExpression(ERR_STRING_EXPR % Operator)
152
153        TypeDict = {
154            type(0)  : 0,
155            type(0L) : 0,
156            type('') : 1,
157            type(True) : 2
158        }
159
160        EvalStr = ''
161        if Operator in ["!", "NOT", "not"]:
162            if type(Oprand1) == type(''):
163                raise BadExpression(ERR_STRING_EXPR % Operator)
164            EvalStr = 'not Oprand1'
165        else:
166            if Operator in ["+", "-"] and (type(True) in [type(Oprand1), type(Oprand2)]):
167                # Boolean in '+'/'-' will be evaluated but raise warning
168                WrnExp = WrnExpression(WRN_BOOL_EXPR)
169            elif type('') in [type(Oprand1), type(Oprand2)] and type(Oprand1)!= type(Oprand2):
170                # == between string and number/boolean will always return False, != return True
171                if Operator == "==":
172                    WrnExp = WrnExpression(WRN_EQCMP_STR_OTHERS)
173                    WrnExp.result = False
174                    raise WrnExp
175                elif Operator == "!=":
176                    WrnExp = WrnExpression(WRN_NECMP_STR_OTHERS)
177                    WrnExp.result = True
178                    raise WrnExp
179                else:
180                    raise BadExpression(ERR_RELCMP_STR_OTHERS % Operator)
181            elif TypeDict[type(Oprand1)] != TypeDict[type(Oprand2)]:
182                if Operator in ["==", "!=", ">=", "<=", ">", "<"] and set((TypeDict[type(Oprand1)], TypeDict[type(Oprand2)])) == set((TypeDict[type(True)], TypeDict[type(0)])):
183                    # comparison between number and boolean is allowed
184                    pass
185                elif Operator in ['&', '|', '^', "and", "or"] and set((TypeDict[type(Oprand1)], TypeDict[type(Oprand2)])) == set((TypeDict[type(True)], TypeDict[type(0)])):
186                    # bitwise and logical operation between number and boolean is allowed
187                    pass
188                else:
189                    raise BadExpression(ERR_EXPR_TYPE)
190            if type(Oprand1) == type('') and type(Oprand2) == type(''):
191                if (Oprand1.startswith('L"') and not Oprand2.startswith('L"')) or \
192                    (not Oprand1.startswith('L"') and Oprand2.startswith('L"')):
193                    raise BadExpression(ERR_STRING_CMP % (Oprand1, Operator, Oprand2))
194            if 'in' in Operator and type(Oprand2) == type(''):
195                Oprand2 = Oprand2.split()
196            EvalStr = 'Oprand1 ' + Operator + ' Oprand2'
197
198        # Local symbols used by built in eval function
199        Dict = {
200            'Oprand1' : Oprand1,
201            'Oprand2' : Oprand2
202        }
203        try:
204            Val = eval(EvalStr, {}, Dict)
205        except Exception, Excpt:
206            raise BadExpression(str(Excpt))
207
208        if Operator in ['and', 'or']:
209            if Val:
210                Val = True
211            else:
212                Val = False
213
214        if WrnExp:
215            WrnExp.result = Val
216            raise WrnExp
217        return Val
218
219    def __init__(self, Expression, SymbolTable={}):
220        self._NoProcess = False
221        if type(Expression) != type(''):
222            self._Expr = Expression
223            self._NoProcess = True
224            return
225
226        self._Expr = ReplaceExprMacro(Expression.strip(),
227                                  SymbolTable,
228                                  SupportedInMacroList)
229
230        if not self._Expr.strip():
231            raise BadExpression(ERR_EMPTY_EXPR)
232
233        #
234        # The symbol table including PCD and macro mapping
235        #
236        self._Symb = SymbolTable
237        self._Symb.update(self.LogicalOperators)
238        self._Idx = 0
239        self._Len = len(self._Expr)
240        self._Token = ''
241        self._WarnExcept = None
242
243        # Literal token without any conversion
244        self._LiteralToken = ''
245
246    # Public entry for this class
247    #   @param RealValue: False: only evaluate if the expression is true or false, used for conditional expression
248    #                     True : return the evaluated str(value), used for PCD value
249    #
250    #   @return: True or False if RealValue is False
251    #            Evaluated value of string format if RealValue is True
252    #
253    def __call__(self, RealValue=False, Depth=0):
254        if self._NoProcess:
255            return self._Expr
256
257        self._Depth = Depth
258
259        self._Expr = self._Expr.strip()
260        if RealValue and Depth == 0:
261            self._Token = self._Expr
262            if self.__IsNumberToken():
263                return self._Expr
264
265            try:
266                Token = self._GetToken()
267                if type(Token) == type('') and Token.startswith('{') and Token.endswith('}') and self._Idx >= self._Len:
268                    return self._Expr
269            except BadExpression:
270                pass
271
272            self._Idx = 0
273            self._Token = ''
274
275        Val = self._OrExpr()
276        RealVal = Val
277        if type(Val) == type(''):
278            if Val == 'L""':
279                Val = False
280            elif not Val:
281                Val = False
282                RealVal = '""'
283            elif not Val.startswith('L"') and not Val.startswith('{'):
284                Val = True
285                RealVal = '"' + RealVal + '"'
286
287        # The expression has been parsed, but the end of expression is not reached
288        # It means the rest does not comply EBNF of <Expression>
289        if self._Idx != self._Len:
290            raise BadExpression(ERR_SNYTAX % self._Expr[self._Idx:])
291
292        if RealValue:
293            RetVal = str(RealVal)
294        elif Val:
295            RetVal = True
296        else:
297            RetVal = False
298
299        if self._WarnExcept:
300            self._WarnExcept.result = RetVal
301            raise self._WarnExcept
302        else:
303            return RetVal
304
305    # Template function to parse binary operators which have same precedence
306    # Expr [Operator Expr]*
307    def _ExprFuncTemplate(self, EvalFunc, OpLst):
308        Val = EvalFunc()
309        while self._IsOperator(OpLst):
310            Op = self._Token
311            try:
312                Val = self.Eval(Op, Val, EvalFunc())
313            except WrnExpression, Warn:
314                self._WarnExcept = Warn
315                Val = Warn.result
316        return Val
317
318    # A [|| B]*
319    def _OrExpr(self):
320        return self._ExprFuncTemplate(self._AndExpr, ["OR", "or", "||"])
321
322    # A [&& B]*
323    def _AndExpr(self):
324        return self._ExprFuncTemplate(self._BitOr, ["AND", "and", "&&"])
325
326    # A [ | B]*
327    def _BitOr(self):
328        return self._ExprFuncTemplate(self._BitXor, ["|"])
329
330    # A [ ^ B]*
331    def _BitXor(self):
332        return self._ExprFuncTemplate(self._BitAnd, ["XOR", "xor", "^"])
333
334    # A [ & B]*
335    def _BitAnd(self):
336        return self._ExprFuncTemplate(self._EqExpr, ["&"])
337
338    # A [ == B]*
339    def _EqExpr(self):
340        Val = self._RelExpr()
341        while self._IsOperator(["==", "!=", "EQ", "NE", "IN", "in", "!", "NOT", "not"]):
342            Op = self._Token
343            if Op in ["!", "NOT", "not"]:
344                if not self._IsOperator(["IN", "in"]):
345                    raise BadExpression(ERR_REL_NOT_IN)
346                Op += ' ' + self._Token
347            try:
348                Val = self.Eval(Op, Val, self._RelExpr())
349            except WrnExpression, Warn:
350                self._WarnExcept = Warn
351                Val = Warn.result
352        return Val
353
354    # A [ > B]*
355    def _RelExpr(self):
356        return self._ExprFuncTemplate(self._AddExpr, ["<=", ">=", "<", ">", "LE", "GE", "LT", "GT"])
357
358    # A [ + B]*
359    def _AddExpr(self):
360        return self._ExprFuncTemplate(self._UnaryExpr, ["+", "-"])
361
362    # [!]*A
363    def _UnaryExpr(self):
364        if self._IsOperator(["!", "NOT", "not"]):
365            Val = self._UnaryExpr()
366            try:
367                return self.Eval('not', Val)
368            except WrnExpression, Warn:
369                self._WarnExcept = Warn
370                return Warn.result
371        return self._IdenExpr()
372
373    # Parse identifier or encapsulated expression
374    def _IdenExpr(self):
375        Tk = self._GetToken()
376        if Tk == '(':
377            Val = self._OrExpr()
378            try:
379                # _GetToken may also raise BadExpression
380                if self._GetToken() != ')':
381                    raise BadExpression(ERR_MATCH)
382            except BadExpression:
383                raise BadExpression(ERR_MATCH)
384            return Val
385        return Tk
386
387    # Skip whitespace or tab
388    def __SkipWS(self):
389        for Char in self._Expr[self._Idx:]:
390            if Char not in ' \t':
391                break
392            self._Idx += 1
393
394    # Try to convert string to number
395    def __IsNumberToken(self):
396        Radix = 10
397        if self._Token.lower()[0:2] == '0x' and len(self._Token) > 2:
398            Radix = 16
399        try:
400            self._Token = int(self._Token, Radix)
401            return True
402        except ValueError:
403            return False
404        except TypeError:
405            return False
406
407    # Parse array: {...}
408    def __GetArray(self):
409        Token = '{'
410        self._Idx += 1
411        self.__GetNList(True)
412        Token += self._LiteralToken
413        if self._Idx >= self._Len or self._Expr[self._Idx] != '}':
414            raise BadExpression(ERR_ARRAY_TOKEN % Token)
415        Token += '}'
416
417        # All whitespace and tabs in array are already stripped.
418        IsArray = IsGuid = False
419        if len(Token.split(',')) == 11 and len(Token.split(',{')) == 2 \
420            and len(Token.split('},')) == 1:
421            HexLen = [11,6,6,5,4,4,4,4,4,4,6]
422            HexList= Token.split(',')
423            if HexList[3].startswith('{') and \
424                not [Index for Index, Hex in enumerate(HexList) if len(Hex) > HexLen[Index]]:
425                IsGuid = True
426        if Token.lstrip('{').rstrip('}').find('{') == -1:
427            if not [Hex for Hex in Token.lstrip('{').rstrip('}').split(',') if len(Hex) > 4]:
428                IsArray = True
429        if not IsArray and not IsGuid:
430            raise BadExpression(ERR_ARRAY_TOKEN % Token)
431        self._Idx += 1
432        self._Token = self._LiteralToken = Token
433        return self._Token
434
435    # Parse string, the format must be: "..."
436    def __GetString(self):
437        Idx = self._Idx
438
439        # Skip left quote
440        self._Idx += 1
441
442        # Replace escape \\\", \"
443        Expr = self._Expr[self._Idx:].replace('\\\\', '//').replace('\\\"', '\\\'')
444        for Ch in Expr:
445            self._Idx += 1
446            if Ch == '"':
447                break
448        self._Token = self._LiteralToken = self._Expr[Idx:self._Idx]
449        if not self._Token.endswith('"'):
450            raise BadExpression(ERR_STRING_TOKEN % self._Token)
451        self._Token = self._Token[1:-1]
452        return self._Token
453
454    # Get token that is comprised by alphanumeric, underscore or dot(used by PCD)
455    # @param IsAlphaOp: Indicate if parsing general token or script operator(EQ, NE...)
456    def __GetIdToken(self, IsAlphaOp = False):
457        IdToken = ''
458        for Ch in self._Expr[self._Idx:]:
459            if not self.__IsIdChar(Ch):
460                break
461            self._Idx += 1
462            IdToken += Ch
463
464        self._Token = self._LiteralToken = IdToken
465        if not IsAlphaOp:
466            self.__ResolveToken()
467        return self._Token
468
469    # Try to resolve token
470    def __ResolveToken(self):
471        if not self._Token:
472            raise BadExpression(ERR_EMPTY_TOKEN)
473
474        # PCD token
475        if self.PcdPattern.match(self._Token):
476            if self._Token not in self._Symb:
477                Ex = BadExpression(ERR_PCD_RESOLVE % self._Token)
478                Ex.Pcd = self._Token
479                raise Ex
480            self._Token = ValueExpression(self._Symb[self._Token], self._Symb)(True, self._Depth+1)
481            if type(self._Token) != type(''):
482                self._LiteralToken = hex(self._Token)
483                return
484
485        if self._Token.startswith('"'):
486            self._Token = self._Token[1:-1]
487        elif self._Token in ["FALSE", "false", "False"]:
488            self._Token = False
489        elif self._Token in ["TRUE", "true", "True"]:
490            self._Token = True
491        else:
492            self.__IsNumberToken()
493
494    def __GetNList(self, InArray=False):
495        self._GetSingleToken()
496        if not self.__IsHexLiteral():
497            if InArray:
498                raise BadExpression(ERR_ARRAY_ELE % self._Token)
499            return self._Token
500
501        self.__SkipWS()
502        Expr = self._Expr[self._Idx:]
503        if not Expr.startswith(','):
504            return self._Token
505
506        NList = self._LiteralToken
507        while Expr.startswith(','):
508            NList += ','
509            self._Idx += 1
510            self.__SkipWS()
511            self._GetSingleToken()
512            if not self.__IsHexLiteral():
513                raise BadExpression(ERR_ARRAY_ELE % self._Token)
514            NList += self._LiteralToken
515            self.__SkipWS()
516            Expr = self._Expr[self._Idx:]
517        self._Token = self._LiteralToken = NList
518        return self._Token
519
520    def __IsHexLiteral(self):
521        if self._LiteralToken.startswith('{') and \
522            self._LiteralToken.endswith('}'):
523            return True
524
525        if self.HexPattern.match(self._LiteralToken):
526            Token = self._LiteralToken[2:]
527            Token = Token.lstrip('0')
528            if not Token:
529                self._LiteralToken = '0x0'
530            else:
531                self._LiteralToken = '0x' + Token.lower()
532            return True
533        return False
534
535    def _GetToken(self):
536        return self.__GetNList()
537
538    @staticmethod
539    def __IsIdChar(Ch):
540        return Ch in '._/:' or Ch.isalnum()
541
542    # Parse operand
543    def _GetSingleToken(self):
544        self.__SkipWS()
545        Expr = self._Expr[self._Idx:]
546        if Expr.startswith('L"'):
547            # Skip L
548            self._Idx += 1
549            UStr = self.__GetString()
550            self._Token = 'L"' + UStr + '"'
551            return self._Token
552
553        self._Token = ''
554        if Expr:
555            Ch = Expr[0]
556            Match = self.RegGuidPattern.match(Expr)
557            if Match and not Expr[Match.end():Match.end()+1].isalnum() \
558                and Expr[Match.end():Match.end()+1] != '_':
559                self._Idx += Match.end()
560                self._Token = ValueExpression(GuidStringToGuidStructureString(Expr[0:Match.end()]))(True, self._Depth+1)
561                return self._Token
562            elif self.__IsIdChar(Ch):
563                return self.__GetIdToken()
564            elif Ch == '"':
565                return self.__GetString()
566            elif Ch == '{':
567                return self.__GetArray()
568            elif Ch == '(' or Ch == ')':
569                self._Idx += 1
570                self._Token = Ch
571                return self._Token
572
573        raise BadExpression(ERR_VALID_TOKEN % Expr)
574
575    # Parse operator
576    def _GetOperator(self):
577        self.__SkipWS()
578        LegalOpLst = ['&&', '||', '!=', '==', '>=', '<='] + self.NonLetterOpLst
579
580        self._Token = ''
581        Expr = self._Expr[self._Idx:]
582
583        # Reach end of expression
584        if not Expr:
585            return ''
586
587        # Script operator: LT, GT, LE, GE, EQ, NE, and, or, xor, not
588        if Expr[0].isalpha():
589            return self.__GetIdToken(True)
590
591        # Start to get regular operator: +, -, <, > ...
592        if Expr[0] not in self.NonLetterOpLst:
593            return ''
594
595        OpToken = ''
596        for Ch in Expr:
597            if Ch in self.NonLetterOpLst:
598                if '!' == Ch and OpToken:
599                    break
600                self._Idx += 1
601                OpToken += Ch
602            else:
603                break
604
605        if OpToken not in LegalOpLst:
606            raise BadExpression(ERR_OPERATOR_UNSUPPORT % OpToken)
607        self._Token = OpToken
608        return OpToken
609
610    # Check if current token matches the operators given from OpList
611    def _IsOperator(self, OpList):
612        Idx = self._Idx
613        self._GetOperator()
614        if self._Token in OpList:
615            if self._Token in self.LogicalOperators:
616                self._Token = self.LogicalOperators[self._Token]
617            return True
618        self._Idx = Idx
619        return False
620
621if __name__ == '__main__':
622    pass
623    while True:
624        input = raw_input('Input expr: ')
625        if input in 'qQ':
626            break
627        try:
628            print ValueExpression(input)(True)
629            print ValueExpression(input)(False)
630        except WrnExpression, Ex:
631            print Ex.result
632            print str(Ex)
633        except Exception, Ex:
634            print str(Ex)
635