recognizers.py

来自「antlr最新版本V3源代码」· Python 代码 · 共 1,189 行 · 第 1/3 页

        and, hence, the follow context stack is:

        depth  local follow set     after call to rule
          0         \<EOF>                    a (from main())
          1          ']'                     b
          3          '^'                     c

        Notice that ')' is not included, because b would have to have
        been called from a different context in rule a for ')' to be
        included.

        For error recovery, we cannot consider FOLLOW(c)
        (context-sensitive or otherwise).  We need the combined set of
        all context-sensitive FOLLOW sets--the set of all tokens that
        could follow any reference in the call chain.  We need to
        resync to one of those tokens.  Note that FOLLOW(c)='^' and if
        we resync'd to that token, we'd consume until EOF.  We need to
        sync to context-sensitive FOLLOWs for a, b, and c: {']','^'}.
        In this case, for input "[]", LA(1) is in this set so we would
        not consume anything and after printing an error rule c would
        return normally.  It would not find the required '^' though.
        At this point, it gets a mismatched token error and throws an
        exception (since LA(1) is not in the viable following token
        set).  The rule exception handler tries to recover, but finds
        the same recovery set and doesn't consume anything.  Rule b
        exits normally returning to rule a.  Now it finds the ']' (and
        with the successful match exits errorRecovery mode).

        So, you cna see that the parser walks up call chain looking
        for the token that was a member of the recovery set.

        Errors are not generated in errorRecovery mode.

        ANTLR's error recovery mechanism is based upon original ideas:

        "Algorithms + Data Structures = Programs" by Niklaus Wirth

        and

        "A note on error recovery in recursive descent parsers":
        http://portal.acm.org/citation.cfm?id=947902.947905

        Later, Josef Grosch had some good ideas:

        "Efficient and Comfortable Error Recovery in Recursive Descent
        Parsers":
        ftp://www.cocolab.com/products/cocktail/doca4.ps/ell.ps.zip

        Like Grosch I implemented local FOLLOW sets that are combined
        at run-time upon error to avoid overhead during parsing.
        """
        
        return self.combineFollows(False)

        
    def computeContextSensitiveRuleFOLLOW(self):
        """
        Compute the context-sensitive FOLLOW set for current rule.
        This is set of token types that can follow a specific rule
        reference given a specific call chain.  You get the set of
        viable tokens that can possibly come next (lookahead depth 1)
        given the current call chain.  Contrast this with the
        definition of plain FOLLOW for rule r:

         FOLLOW(r)={x | S=>*alpha r beta in G and x in FIRST(beta)}

        where x in T* and alpha, beta in V*; T is set of terminals and
        V is the set of terminals and nonterminals.  In other words,
        FOLLOW(r) is the set of all tokens that can possibly follow
        references to r in *any* sentential form (context).  At
        runtime, however, we know precisely which context applies as
        we have the call chain.  We may compute the exact (rather
        than covering superset) set of following tokens.

        For example, consider grammar:

        stat : ID '=' expr ';'      // FOLLOW(stat)=={EOF}
             | "return" expr '.'
             ;
        expr : atom ('+' atom)* ;   // FOLLOW(expr)=={';','.',')'}
        atom : INT                  // FOLLOW(atom)=={'+',')',';','.'}
             | '(' expr ')'
             ;

        The FOLLOW sets are all inclusive whereas context-sensitive
        FOLLOW sets are precisely what could follow a rule reference.
        For input input "i=(3);", here is the derivation:

        stat => ID '=' expr ';'
             => ID '=' atom ('+' atom)* ';'
             => ID '=' '(' expr ')' ('+' atom)* ';'
             => ID '=' '(' atom ')' ('+' atom)* ';'
             => ID '=' '(' INT ')' ('+' atom)* ';'
             => ID '=' '(' INT ')' ';'

        At the "3" token, you'd have a call chain of

          stat -> expr -> atom -> expr -> atom

        What can follow that specific nested ref to atom?  Exactly ')'
        as you can see by looking at the derivation of this specific
        input.  Contrast this with the FOLLOW(atom)={'+',')',';','.'}.

        You want the exact viable token set when recovering from a
        token mismatch.  Upon token mismatch, if LA(1) is member of
        the viable next token set, then you know there is most likely
        a missing token in the input stream.  "Insert" one by just not
        throwing an exception.
        """

        return self.combineFollows(True)


    def combineFollows(self, exact):
        followSet = set()
        for localFollowSet in reversed(self.following):
            followSet |= localFollowSet
            if exact and EOR_TOKEN_TYPE not in localFollowSet:
                break

        followSet -= set([EOR_TOKEN_TYPE])
        return followSet


    def recoverFromMismatchedToken(self, input, e, ttype, follow):
        """Attempt to recover from a single missing or extra token.

        EXTRA TOKEN

        LA(1) is not what we are looking for.  If LA(2) has the right token,
        however, then assume LA(1) is some extra spurious token.  Delete it
        and LA(2) as if we were doing a normal match(), which advances the
        input.

        MISSING TOKEN

        If current token is consistent with what could come after
        ttype then it is ok to "insert" the missing token, else throw
        exception For example, Input "i=(3;" is clearly missing the
        ')'.  When the parser returns from the nested call to expr, it
        will have call chain:

          stat -> expr -> atom

        and it will be trying to match the ')' at this point in the
        derivation:

             => ID '=' '(' INT ')' ('+' atom)* ';'
                                ^
        match() will see that ';' doesn't match ')' and report a
        mismatched token error.  To recover, it sees that LA(1)==';'
        is in the set of tokens that can follow the ')' token
        reference in rule atom.  It can assume that you forgot the ')'.
        """
                                         
        # if next token is what we are looking for then "delete" this token
        if input.LA(2) == ttype:
            self.reportError(e)

            self.beginResync()
            input.consume() # simply delete extra token
            self.endResync()
            input.consume()  # move past ttype token as if all were ok
            return

        if not self.recoverFromMismatchedElement(input, e, follow):
            raise e



    def recoverFromMismatchedSet(self, input, e, follow):
        # TODO do single token deletion like above for Token mismatch
        if not self.recoverFromMismatchedElement(input, e, follow):
            raise e


    def recoverFromMismatchedElement(self, input, e, follow):
        """
        This code is factored out from mismatched token and mismatched set
        recovery.  It handles "single token insertion" error recovery for
        both.  No tokens are consumed to recover from insertions.  Return
        true if recovery was possible else return false.
        """
        
        if follow is None:
            # we have no information about the follow; we can only consume
            # a single token and hope for the best
            return False

        # compute what can follow this grammar element reference
        if EOR_TOKEN_TYPE in follow:
            viableTokensFollowingThisRule = \
                self.computeContextSensitiveRuleFOLLOW()
            
            follow = (follow | viableTokensFollowingThisRule) \
                     - set([EOR_TOKEN_TYPE])

        # if current token is consistent with what could come after set
        # then it is ok to "insert" the missing token, else throw exception
        if input.LA(1) in follow:
            self.reportError(e)
            return True

        # nothing to do; throw exception
        return False


    def consumeUntil(self, input, tokenTypes):
        """
        Consume tokens until one matches the given token or token set

        tokenTypes can be a single token type or a set of token types
        
        """
        
        if not isinstance(tokenTypes, (set, frozenset)):
            tokenTypes = frozenset([tokenTypes])

        ttype = input.LA(1)
        while ttype != EOF and ttype not in tokenTypes:
            input.consume()
            ttype = input.LA(1)


    def getRuleInvocationStack(self):
        """
        Return List<String> of the rules in your parser instance
        leading up to a call to this method.  You could override if
        you want more details such as the file/line info of where
        in the parser java code a rule is invoked.

        This is very useful for error messages and for context-sensitive
        error recovery.

        You must be careful, if you subclass a generated recognizers.
        The default implementation will only search the module of self
        for rules, but the subclass will not contain any rules.
        You probably want to override this method to look like

        def getRuleInvocationStack(self):
            return self._getRuleInvocationStack(<class>.__module__)

        where <class> is the class of the generated recognizer, e.g.
        the superclass of self.
	"""

        return self._getRuleInvocationStack(self.__module__)


    def _getRuleInvocationStack(cls, module):
        """
        A more general version of getRuleInvocationStack where you can
        pass in, for example, a RecognitionException to get it's rule
        stack trace.  This routine is shared with all recognizers, hence,
        static.

        TODO: move to a utility class or something; weird having lexer call
        this
        """

        # mmmhhh,... perhaps look at the first argument
        # (f_locals[co_varnames[0]]?) and test if it's a (sub)class of
        # requested recognizer...
        
        rules = []
        for frame in reversed(inspect.stack()):
            code = frame[0].f_code
            codeMod = inspect.getmodule(code)
            if codeMod is None:
                continue

            # skip frames not in requested module
            if codeMod.__name__ != module:
                continue

            # skip some unwanted names
            if code.co_name in ('nextToken', '<module>'):
                continue

            rules.append(code.co_name)

        return rules
        
    _getRuleInvocationStack = classmethod(_getRuleInvocationStack)
    

    def getBacktrackingLevel(self):
        return self.backtracking


    def getGrammarFileName(self):
        """For debugging and other purposes, might want the grammar name.
        
        Have ANTLR generate an implementation for this method.
        """

        return None


    def toStrings(self, tokens):
        """A convenience method for use most often with template rewrites.

        Convert a List<Token> to List<String>
        """

        if tokens is None:
            return None

        return [token.text for token in tokens]


    def getRuleMemoization(self, ruleIndex, ruleStartIndex):
        """
        Given a rule number and a start token index number, return
        MEMO_RULE_UNKNOWN if the rule has not parsed input starting from
        start index.  If this rule has parsed input starting from the
        start index before, then return where the rule stopped parsing.
        It returns the index of the last token matched by the rule.

        For now we use a hashtable and just the slow Object-based one.
        Later, we can make a special one for ints and also one that
        tosses out data after we commit past input position i.
	"""
        
        if ruleIndex not in self.ruleMemo:
            self.ruleMemo[ruleIndex] = {}
		
        stopIndex = self.ruleMemo[ruleIndex].get(ruleStartIndex, None)
        if stopIndex is None:
            return self.MEMO_RULE_UNKNOWN

        return stopIndex


    def alreadyParsedRule(self, input, ruleIndex):
        """
        Has this rule already parsed input at the current index in the
        input stream?  Return the stop token index or MEMO_RULE_UNKNOWN.
        If we attempted but failed to parse properly before, return
        MEMO_RULE_FAILED.

        This method has a side-effect: if we have seen this input for
        this rule and successfully parsed before, then seek ahead to
        1 past the stop token matched for this rule last time.
        """
        
        stopIndex = self.getRuleMemoization(ruleIndex, input.index())
        if stopIndex == self.MEMO_RULE_UNKNOWN:
            return False

        if stopIndex == self.MEMO_RULE_FAILED:
            self.failed = True

        else:
            input.seek(stopIndex + 1)

        return True


    def memoize(self, input, ruleIndex, ruleStartIndex):
        """
        Record whether or not this rule parsed the input at this position
	successfully.
	"""

        if self.failed:
            stopTokenIndex = self.MEMO_RULE_FAILED
        else:
            stopTokenIndex = input.index() - 1
        
        if ruleIndex in self.ruleMemo:
            self.ruleMemo[ruleIndex][ruleStartIndex] = stopTokenIndex


    def traceIn(self, ruleName, ruleIndex, inputSymbol):
        sys.stdout.write("enter %s %s" % (ruleName, inputSymbol))
        
        if self.failed:
            sys.stdout.write(" failed=%s" % self.failed)

        if self.backtracking > 0:
            sys.stdout.write(" backtracking=%s" % self.backtracking)

        sys.stdout.write('\n')


    def traceOut(self, ruleName, ruleIndex, inputSymbol):
        sys.stdout.write("exit %s %s" % (ruleName, inputSymbol))
        
        if self.failed:
            sys.stdout.write(" failed=%s" % self.failed)

        if self.backtracking > 0:
            sys.stdout.write(" backtracking=%s" % self.backtracking)

        sys.stdout.write('\n')