Enable EFI_BROWSER_ACTION_CHANGED callback type for browser.

[mirror_edk2.git] / BaseTools / Source / Python / Common / Expression.py

## @file\r
# This file is used to parse and evaluate expression in directive or PCD value.\r
#\r
# Copyright (c) 2011, Intel Corporation. All rights reserved.<BR>\r
# This program and the accompanying materials\r
# are licensed and made available under the terms and conditions of the BSD License\r
# which accompanies this distribution.    The full text of the license may be found at\r
# http://opensource.org/licenses/bsd-license.php\r
#\r
# THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,\r
# WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.\r
\r
## Import Modules\r
#\r
from Common.GlobalData import *\r
from CommonDataClass.Exceptions import BadExpression\r
from CommonDataClass.Exceptions import SymbolNotFound\r
from CommonDataClass.Exceptions import WrnExpression\r
from Misc import GuidStringToGuidStructureString\r
\r
ERR_STRING_EXPR         = 'This operator cannot be used in string expression: [%s].'\r
ERR_SNYTAX              = 'Syntax error, the rest of expression cannot be evaluated: [%s].'\r
ERR_MATCH               = 'No matching right parenthesis.'\r
ERR_STRING_TOKEN        = 'Bad string token: [%s].'\r
ERR_MACRO_TOKEN         = 'Bad macro token: [%s].'\r
ERR_EMPTY_TOKEN         = 'Empty token is not allowed.'\r
ERR_PCD_RESOLVE         = 'PCD token cannot be resolved: [%s].'\r
ERR_VALID_TOKEN         = 'No more valid token found from rest of string: [%s].'\r
ERR_EXPR_TYPE           = 'Different types found in expression.'\r
ERR_OPERATOR_UNSUPPORT  = 'Unsupported operator: [%s]'\r
ERR_REL_NOT_IN          = 'Expect "IN" after "not" operator.'\r
WRN_BOOL_EXPR           = 'Operand of boolean type cannot be used in arithmetic expression.'\r
WRN_EQCMP_STR_OTHERS    = '== Comparison between Operand of string type and Boolean/Number Type always return False.'\r
WRN_NECMP_STR_OTHERS    = '!= Comparison between Operand of string type and Boolean/Number Type always return True.'\r
ERR_RELCMP_STR_OTHERS   = 'Operator taking Operand of string type and Boolean/Number Type is not allowed: [%s].'\r
ERR_STRING_CMP          = 'Unicode string and general string cannot be compared: [%s %s %s]'\r
ERR_ARRAY_TOKEN         = 'Bad C array or C format GUID token: [%s].'\r
ERR_ARRAY_ELE           = 'This must be HEX value for NList or Array: [%s].'\r
\r
## SplitString\r
#  Split string to list according double quote\r
#  For example: abc"de\"f"ghi"jkl"mn will be: ['abc', '"de\"f"', 'ghi', '"jkl"', 'mn']\r
#\r
def SplitString(String):\r
    # There might be escaped quote: "abc\"def\\\"ghi"\r
    Str = String.replace('\\\\', '//').replace('\\\"', '\\\'')\r
    RetList = []\r
    InQuote = False\r
    Item = ''\r
    for i, ch in enumerate(Str):\r
        if ch == '"':\r
            InQuote = not InQuote\r
            if not InQuote:\r
                Item += String[i]\r
                RetList.append(Item)\r
                Item = ''\r
                continue\r
            if Item:\r
                RetList.append(Item)\r
                Item = ''\r
        Item += String[i]\r
    if InQuote:\r
        raise BadExpression(ERR_STRING_TOKEN % Item)\r
    if Item:\r
        RetList.append(Item)\r
    return RetList\r
\r
## ReplaceExprMacro\r
#\r
def ReplaceExprMacro(String, Macros, ExceptionList = None):\r
    StrList = SplitString(String)\r
    for i, String in enumerate(StrList):\r
        InQuote = False\r
        if String.startswith('"'):\r
            InQuote = True\r
        MacroStartPos = String.find('$(')\r
        if MacroStartPos < 0:\r
            continue\r
        RetStr = ''\r
        while MacroStartPos >= 0:\r
            RetStr = String[0:MacroStartPos]\r
            MacroEndPos = String.find(')', MacroStartPos)\r
            if MacroEndPos < 0:\r
                raise BadExpression(ERR_MACRO_TOKEN % String[MacroStartPos:])\r
            Macro = String[MacroStartPos+2:MacroEndPos]\r
            if Macro not in Macros:\r
                # From C reference manual:\r
                # If an undefined macro name appears in the constant-expression of\r
                # !if or !elif, it is replaced by the integer constant 0.\r
                RetStr += '0'\r
            elif not InQuote and ExceptionList and Macro in ExceptionList:\r
                # Make sure the macro in exception list is encapsulated by double quote\r
                # For example: DEFINE ARCH = IA32 X64\r
                # $(ARCH) is replaced with "IA32 X64"\r
                RetStr += '"' + Macros[Macro] + '"'\r
            else:\r
                if Macros[Macro].strip() != "":\r
                    RetStr += Macros[Macro]\r
                else:\r
                    RetStr += '""'\r
            RetStr += String[MacroEndPos+1:]\r
            String = RetStr\r
            MacroStartPos = String.find('$(')\r
        StrList[i] = RetStr\r
    return ''.join(StrList)\r
\r
class ValueExpression(object):\r
    # Logical operator mapping\r
    LogicalOperators = {\r
        '&&' : 'and', '||' : 'or',\r
        '!'  : 'not', 'AND': 'and',\r
        'OR' : 'or' , 'NOT': 'not',\r
        'XOR': '^'  , 'xor': '^',\r
        'EQ' : '==' , 'NE' : '!=',\r
        'GT' : '>'  , 'LT' : '<',\r
        'GE' : '>=' , 'LE' : '<=',\r
        'IN' : 'in'\r
    }\r
\r
    NonLetterOpLst = ['+', '-', '&', '|', '^', '!', '=', '>', '<']\r
\r
    PcdPattern = re.compile(r'[_a-zA-Z][0-9A-Za-z_]*\.[_a-zA-Z][0-9A-Za-z_]*$')\r
    HexPattern = re.compile(r'0[xX][0-9a-fA-F]+$')\r
    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}')\r
\r
    SymbolPattern = re.compile("("\r
                                 "\$\([A-Z][A-Z0-9_]*\)|\$\(\w+\.\w+\)|\w+\.\w+|"\r
                                 "&&|\|\||!(?!=)|"\r
                                 "(?<=\W)AND(?=\W)|(?<=\W)OR(?=\W)|(?<=\W)NOT(?=\W)|(?<=\W)XOR(?=\W)|"\r
                                 "(?<=\W)EQ(?=\W)|(?<=\W)NE(?=\W)|(?<=\W)GT(?=\W)|(?<=\W)LT(?=\W)|(?<=\W)GE(?=\W)|(?<=\W)LE(?=\W)"\r
                               ")")\r
\r
    @staticmethod\r
    def Eval(Operator, Oprand1, Oprand2 = None):\r
        WrnExp = None\r
        \r
        if Operator not in ["==", "!=", ">=", "<=", ">", "<", "in", "not in"] and \\r
            (type(Oprand1) == type('') or type(Oprand2) == type('')):\r
            raise BadExpression(ERR_STRING_EXPR % Operator)\r
\r
        TypeDict = {\r
            type(0)  : 0,\r
            type(0L) : 0,\r
            type('') : 1,\r
            type(True) : 2\r
        }\r
\r
        EvalStr = ''\r
        if Operator in ["!", "NOT", "not"]:\r
            if type(Oprand1) == type(''):\r
                raise BadExpression(ERR_STRING_EXPR % Operator)\r
            EvalStr = 'not Oprand1'\r
        else:\r
            if Operator in ["+", "-"] and (type(True) in [type(Oprand1), type(Oprand2)]):\r
                # Boolean in '+'/'-' will be evaluated but raise warning\r
                WrnExp = WrnExpression(WRN_BOOL_EXPR)\r
            elif type('') in [type(Oprand1), type(Oprand2)] and type(Oprand1)!= type(Oprand2):\r
                # == between string and number/boolean will always return False, != return True\r
                if Operator == "==":\r
                    WrnExp = WrnExpression(WRN_EQCMP_STR_OTHERS)\r
                    WrnExp.result = False\r
                    raise WrnExp\r
                elif Operator == "!=":\r
                    WrnExp = WrnExpression(WRN_NECMP_STR_OTHERS)\r
                    WrnExp.result = True\r
                    raise WrnExp\r
                else:\r
                    raise BadExpression(ERR_RELCMP_STR_OTHERS % Operator)\r
            elif TypeDict[type(Oprand1)] != TypeDict[type(Oprand2)]: \r
                if Operator in ["==", "!=", ">=", "<=", ">", "<"] and set((TypeDict[type(Oprand1)], TypeDict[type(Oprand2)])) == set((TypeDict[type(True)], TypeDict[type(0)])):\r
                    # comparison between number and boolean is allowed\r
                    pass\r
                elif Operator in ['&', '|', '^', "&&", "||"] and set((TypeDict[type(Oprand1)], TypeDict[type(Oprand2)])) == set((TypeDict[type(True)], TypeDict[type(0)])):\r
                    # bitwise and logical operation between number and boolean is allowed\r
                    pass                \r
                else:\r
                    raise BadExpression(ERR_EXPR_TYPE)\r
            if type(Oprand1) == type('') and type(Oprand2) == type(''):\r
                if (Oprand1.startswith('L"') and not Oprand2.startswith('L"')) or \\r
                    (not Oprand1.startswith('L"') and Oprand2.startswith('L"')):\r
                    raise BadExpression(ERR_STRING_CMP % (Oprand1, Operator, Oprand2))\r
            if 'in' in Operator and type(Oprand2) == type(''):\r
                Oprand2 = Oprand2.split()\r
            EvalStr = 'Oprand1 ' + Operator + ' Oprand2'\r
\r
        # Local symbols used by built in eval function\r
        Dict = {\r
            'Oprand1' : Oprand1,\r
            'Oprand2' : Oprand2\r
        }\r
        try:\r
            Val = eval(EvalStr, {}, Dict)\r
        except Exception, Excpt:\r
            raise BadExpression(str(Excpt))\r
\r
        if Operator in ['and', 'or']:\r
            if Val:\r
                Val = True\r
            else:\r
                Val = False\r
        \r
        if WrnExp:\r
            WrnExp.result = Val\r
            raise WrnExp\r
        return Val\r
\r
    def __init__(self, Expression, SymbolTable={}):\r
        self._NoProcess = False\r
        if type(Expression) != type(''):\r
            self._Expr = Expression\r
            self._NoProcess = True\r
            return\r
\r
        self._Expr = ReplaceExprMacro(Expression.strip(),\r
                                  SymbolTable,\r
                                  ['TARGET', 'TOOL_CHAIN_TAG', 'ARCH'])\r
\r
        if not self._Expr.strip():\r
            self._NoProcess = True\r
            return\r
\r
        #\r
        # The symbol table including PCD and macro mapping\r
        #\r
        self._Symb = SymbolTable\r
        self._Symb.update(self.LogicalOperators)\r
        self._Idx = 0\r
        self._Len = len(self._Expr)\r
        self._Token = ''\r
\r
        # Literal token without any conversion\r
        self._LiteralToken = ''\r
\r
    # Public entry for this class\r
    def __call__(self):\r
        if self._NoProcess:\r
            return self._Expr\r
\r
        Val = self._OrExpr()\r
        if type(Val) == type('') and Val == 'L""':\r
            Val = ''\r
\r
        # The expression has been parsed, but the end of expression is not reached\r
        # It means the rest does not comply EBNF of <Expression>\r
        if self._Idx != self._Len:\r
            raise BadExpression(ERR_SNYTAX % self._Expr[self._Idx:])\r
\r
        return Val\r
\r
    # Template function to parse binary operators which have same precedence\r
    # Expr [Operator Expr]*\r
    def _ExprFuncTemplate(self, EvalFunc, OpLst):\r
        Val = EvalFunc()\r
        while self._IsOperator(OpLst):\r
            Op = self._Token\r
            Val = self.Eval(Op, Val, EvalFunc())\r
        return Val\r
\r
    # A [|| B]*\r
    def _OrExpr(self):\r
        return self._ExprFuncTemplate(self._AndExpr, ["OR", "or", "||"])\r
\r
    # A [&& B]*\r
    def _AndExpr(self):\r
        return self._ExprFuncTemplate(self._BitOr, ["AND", "and", "&&"])\r
\r
    # A [ | B]*\r
    def _BitOr(self):\r
        return self._ExprFuncTemplate(self._BitXor, ["|"])\r
\r
    # A [ ^ B]*\r
    def _BitXor(self):\r
        return self._ExprFuncTemplate(self._BitAnd, ["XOR", "xor", "^"])\r
\r
    # A [ & B]*\r
    def _BitAnd(self):\r
        return self._ExprFuncTemplate(self._EqExpr, ["&"])\r
\r
    # A [ == B]*\r
    def _EqExpr(self):\r
        Val = self._RelExpr()\r
        while self._IsOperator(["==", "!=", "EQ", "NE", "IN", "in", "!", "NOT", "not"]):\r
            Op = self._Token\r
            if Op in ["!", "NOT", "not"]:\r
                if not self._IsOperator(["IN", "in"]):\r
                    raise BadExpression(ERR_REL_NOT_IN)\r
                Op += ' ' + self._Token\r
            Val = self.Eval(Op, Val, self._RelExpr())\r
        return Val\r
\r
    # A [ > B]*\r
    def _RelExpr(self):\r
        return self._ExprFuncTemplate(self._AddExpr, ["<=", ">=", "<", ">", "LE", "GE", "LT", "GT"])\r
\r
    # A [ + B]*\r
    def _AddExpr(self):\r
        return self._ExprFuncTemplate(self._UnaryExpr, ["+", "-"])\r
\r
    # [!]*A\r
    def _UnaryExpr(self):\r
        if self._IsOperator(["!", "NOT", "not"]):\r
            Val = self._UnaryExpr()\r
            return self.Eval('not', Val)\r
        return self._IdenExpr()\r
\r
    # Parse identifier or encapsulated expression\r
    def _IdenExpr(self):\r
        Tk = self._GetToken()\r
        if Tk == '(':\r
            Val = self._OrExpr()\r
            try:\r
                # _GetToken may also raise BadExpression\r
                if self._GetToken() != ')':\r
                    raise BadExpression(ERR_MATCH)\r
            except BadExpression:\r
                raise BadExpression(ERR_MATCH)\r
            return Val\r
        return Tk\r
\r
    # Skip whitespace or tab\r
    def __SkipWS(self):\r
        for Char in self._Expr[self._Idx:]:\r
            if Char not in ' \t':\r
                break\r
            self._Idx += 1\r
\r
    # Try to convert string to number\r
    def __IsNumberToken(self):\r
        Radix = 10\r
        if self._Token.lower()[0:2] == '0x' and len(self._Token) > 2:\r
            Radix = 16\r
        try:\r
            self._Token = int(self._Token, Radix)\r
            return True\r
        except ValueError:\r
            return False\r
        except TypeError:\r
            return False\r
\r
    # Parse array: {...}\r
    def __GetArray(self):\r
        Token = '{'\r
        self._Idx += 1\r
        self.__GetNList(True)\r
        Token += self._LiteralToken\r
        if self._Idx >= self._Len or self._Expr[self._Idx] != '}':\r
            raise BadExpression(ERR_ARRAY_TOKEN % Token)\r
        Token += '}'\r
\r
        # All whitespace and tabs in array are already stripped.\r
        IsArray = IsGuid = False\r
        if len(Token.split(',')) == 11 and len(Token.split(',{')) == 2 \\r
            and len(Token.split('},')) == 1:\r
            HexLen = [11,6,6,5,4,4,4,4,4,4,6]\r
            HexList= Token.split(',')\r
            if HexList[3].startswith('{') and \\r
                not [Index for Index, Hex in enumerate(HexList) if len(Hex) > HexLen[Index]]:\r
                IsGuid = True\r
        if Token.lstrip('{').rstrip('}').find('{') == -1:\r
            if not [Hex for Hex in Token.lstrip('{').rstrip('}').split(',') if len(Hex) > 4]:\r
                IsArray = True\r
        if not IsArray and not IsGuid:\r
            raise BadExpression(ERR_ARRAY_TOKEN % Token)\r
        self._Idx += 1\r
        self._Token = self._LiteralToken = Token\r
        return self._Token\r
\r
    # Parse string, the format must be: "..."\r
    def __GetString(self):\r
        Idx = self._Idx\r
\r
        # Skip left quote\r
        self._Idx += 1\r
\r
        # Replace escape \\\", \"\r
        Expr = self._Expr[self._Idx:].replace('\\\\', '//').replace('\\\"', '\\\'')\r
        for Ch in Expr:\r
            self._Idx += 1\r
            if Ch == '"':\r
                break\r
        self._Token = self._LiteralToken = self._Expr[Idx:self._Idx]\r
        if not self._Token.endswith('"'):\r
            raise BadExpression(ERR_STRING_TOKEN % self._Token)\r
        self._Token = self._Token[1:-1]\r
        return self._Token\r
\r
    # Get token that is comprised by alphanumeric, underscore or dot(used by PCD)\r
    # @param IsAlphaOp: Indicate if parsing general token or script operator(EQ, NE...)\r
    def __GetIdToken(self, IsAlphaOp = False):\r
        IdToken = ''\r
        for Ch in self._Expr[self._Idx:]:\r
            if not self.__IsIdChar(Ch):\r
                break\r
            self._Idx += 1\r
            IdToken += Ch\r
\r
        self._Token = self._LiteralToken = IdToken\r
        if not IsAlphaOp:\r
            self.__ResolveToken()\r
        return self._Token\r
\r
    # Try to resolve token\r
    def __ResolveToken(self):\r
        if not self._Token:\r
            raise BadExpression(ERR_EMPTY_TOKEN)\r
\r
        # PCD token\r
        if self.PcdPattern.match(self._Token):\r
            if self._Token not in self._Symb:\r
                raise SymbolNotFound(ERR_PCD_RESOLVE % self._Token)\r
            self._Token = ValueExpression(self._Symb[self._Token], self._Symb)()\r
            if type(self._Token) != type(''):\r
                self._LiteralToken = hex(self._Token)\r
                return\r
\r
        if self._Token.startswith('"'):\r
            self._Token = self._Token[1:-1]\r
        elif self._Token in ["FALSE", "false", "False"]:\r
            self._Token = False\r
        elif self._Token in ["TRUE", "true", "True"]:\r
            self._Token = True\r
        else:\r
            self.__IsNumberToken()\r
\r
    def __GetNList(self, InArray=False):\r
        self._GetSingleToken()\r
        if not self.__IsHexLiteral():\r
            if InArray:\r
                raise BadExpression(ERR_ARRAY_ELE % self._Token)\r
            return self._Token\r
\r
        self.__SkipWS()\r
        Expr = self._Expr[self._Idx:]\r
        if not Expr.startswith(','):\r
            return self._Token\r
\r
        NList = self._LiteralToken\r
        while Expr.startswith(','):\r
            NList += ','\r
            self._Idx += 1\r
            self.__SkipWS()\r
            self._GetSingleToken()\r
            if not self.__IsHexLiteral():\r
                raise BadExpression(ERR_ARRAY_ELE % self._Token)\r
            NList += self._LiteralToken\r
            self.__SkipWS()\r
            Expr = self._Expr[self._Idx:]\r
        self._Token = self._LiteralToken = NList\r
        return self._Token\r
\r
    def __IsHexLiteral(self):\r
        if self._LiteralToken.startswith('{') and \\r
            self._LiteralToken.endswith('}'):\r
            return True\r
\r
        if self.HexPattern.match(self._LiteralToken):\r
            Token = self._LiteralToken[2:]\r
            Token = Token.lstrip('0')\r
            if not Token:\r
                self._LiteralToken = '0x0'\r
            else:\r
                self._LiteralToken = '0x' + Token\r
            return True\r
        return False\r
\r
    def _GetToken(self):\r
        return self.__GetNList()\r
\r
    @staticmethod\r
    def __IsIdChar(Ch):\r
        return Ch in '._/:' or Ch.isalnum()\r
\r
    # Parse operand\r
    def _GetSingleToken(self):\r
        self.__SkipWS()\r
        Expr = self._Expr[self._Idx:]\r
        if Expr.startswith('L"'):\r
            # Skip L\r
            self._Idx += 1\r
            UStr = self.__GetString()\r
            self._Token = 'L"' + UStr + '"'\r
            return self._Token\r
\r
        self._Token = ''\r
        if Expr:\r
            Ch = Expr[0]\r
            Match = self.RegGuidPattern.match(Expr)\r
            if Match and not Expr[Match.end():Match.end()+1].isalnum() \\r
                and Expr[Match.end():Match.end()+1] != '_':\r
                self._Idx += Match.end()\r
                self._Token = ValueExpression(GuidStringToGuidStructureString(Expr[0:Match.end()]))()\r
                return self._Token\r
            elif self.__IsIdChar(Ch):\r
                return self.__GetIdToken()\r
            elif Ch == '"':\r
                return self.__GetString()\r
            elif Ch == '{':\r
                return self.__GetArray()\r
            elif Ch == '(' or Ch == ')':\r
                self._Idx += 1\r
                self._Token = Ch\r
                return self._Token\r
\r
        raise BadExpression(ERR_VALID_TOKEN % Expr)\r
\r
    # Parse operator\r
    def _GetOperator(self):\r
        self.__SkipWS()\r
        LegalOpLst = ['&&', '||', '!=', '==', '>=', '<='] + self.NonLetterOpLst\r
\r
        self._Token = ''\r
        Expr = self._Expr[self._Idx:]\r
\r
        # Reach end of expression\r
        if not Expr:\r
            return ''\r
\r
        # Script operator: LT, GT, LE, GE, EQ, NE, and, or, xor, not\r
        if Expr[0].isalpha():\r
            return self.__GetIdToken(True)\r
\r
        # Start to get regular operator: +, -, <, > ...\r
        if Expr[0] not in self.NonLetterOpLst:\r
            return ''\r
\r
        OpToken = ''\r
        for Ch in Expr:\r
            if Ch in self.NonLetterOpLst:\r
                if '!' == Ch and OpToken in ['!=', '!']:\r
                    break\r
                self._Idx += 1\r
                OpToken += Ch\r
            else:\r
                break\r
\r
        if OpToken not in LegalOpLst:\r
            raise BadExpression(ERR_OPERATOR_UNSUPPORT % OpToken)\r
        self._Token = OpToken\r
        return OpToken\r
\r
    # Check if current token matches the operators given from OpList\r
    def _IsOperator(self, OpList):\r
        Idx = self._Idx\r
        self._GetOperator()\r
        if self._Token in OpList:\r
            if self._Token in self.LogicalOperators:\r
                self._Token = self.LogicalOperators[self._Token]\r
            return True\r
        self._Idx = Idx\r
        return False\r
\r
if __name__ == '__main__':\r
    pass\r
\r
\r