sl@0: /* sl@0: ********************************************************************** sl@0: * Copyright (c) 2003-2005, International Business Machines sl@0: * Corporation and others. All Rights Reserved. sl@0: ********************************************************************** sl@0: * Author: Alan Liu sl@0: * Created: September 24 2003 sl@0: * Since: ICU 2.8 sl@0: ********************************************************************** sl@0: */ sl@0: #ifndef _RULEITER_H_ sl@0: #define _RULEITER_H_ sl@0: sl@0: #include "unicode/utypes.h" sl@0: sl@0: U_NAMESPACE_BEGIN sl@0: sl@0: class UnicodeString; sl@0: class ParsePosition; sl@0: class SymbolTable; sl@0: sl@0: /** sl@0: * An iterator that returns 32-bit code points. This class is deliberately sl@0: * not related to any of the ICU character iterator classes sl@0: * in order to minimize complexity. sl@0: * @author Alan Liu sl@0: * @since ICU 2.8 sl@0: */ sl@0: class U_COMMON_API RuleCharacterIterator { sl@0: sl@0: // TODO: Ideas for later. (Do not implement if not needed, lest the sl@0: // code coverage numbers go down due to unused methods.) sl@0: // 1. Add a copy constructor, operator==() method. sl@0: // 2. Rather than return DONE, throw an exception if the end sl@0: // is reached -- this is an alternate usage model, probably not useful. sl@0: sl@0: private: sl@0: /** sl@0: * Text being iterated. sl@0: */ sl@0: const UnicodeString& text; sl@0: sl@0: /** sl@0: * Position of iterator. sl@0: */ sl@0: ParsePosition& pos; sl@0: sl@0: /** sl@0: * Symbol table used to parse and dereference variables. May be 0. sl@0: */ sl@0: const SymbolTable* sym; sl@0: sl@0: /** sl@0: * Current variable expansion, or 0 if none. sl@0: */ sl@0: const UnicodeString* buf; sl@0: sl@0: /** sl@0: * Position within buf. Meaningless if buf == 0. sl@0: */ sl@0: int32_t bufPos; sl@0: sl@0: public: sl@0: /** sl@0: * Value returned when there are no more characters to iterate. sl@0: */ sl@0: enum { DONE = -1 }; sl@0: sl@0: /** sl@0: * Bitmask option to enable parsing of variable names. If (options & sl@0: * PARSE_VARIABLES) != 0, then an embedded variable will be expanded to sl@0: * its value. Variables are parsed using the SymbolTable API. sl@0: */ sl@0: enum { PARSE_VARIABLES = 1 }; sl@0: sl@0: /** sl@0: * Bitmask option to enable parsing of escape sequences. If (options & sl@0: * PARSE_ESCAPES) != 0, then an embedded escape sequence will be expanded sl@0: * to its value. Escapes are parsed using Utility.unescapeAt(). sl@0: */ sl@0: enum { PARSE_ESCAPES = 2 }; sl@0: sl@0: /** sl@0: * Bitmask option to enable skipping of whitespace. If (options & sl@0: * SKIP_WHITESPACE) != 0, then whitespace characters will be silently sl@0: * skipped, as if they were not present in the input. Whitespace sl@0: * characters are defined by UCharacterProperty.isRuleWhiteSpace(). sl@0: */ sl@0: enum { SKIP_WHITESPACE = 4 }; sl@0: sl@0: /** sl@0: * Constructs an iterator over the given text, starting at the given sl@0: * position. sl@0: * @param text the text to be iterated sl@0: * @param sym the symbol table, or null if there is none. If sym is null, sl@0: * then variables will not be deferenced, even if the PARSE_VARIABLES sl@0: * option is set. sl@0: * @param pos upon input, the index of the next character to return. If a sl@0: * variable has been dereferenced, then pos will not increment as sl@0: * characters of the variable value are iterated. sl@0: */ sl@0: RuleCharacterIterator(const UnicodeString& text, const SymbolTable* sym, sl@0: ParsePosition& pos); sl@0: sl@0: /** sl@0: * Returns true if this iterator has no more characters to return. sl@0: */ sl@0: UBool atEnd() const; sl@0: sl@0: /** sl@0: * Returns the next character using the given options, or DONE if there sl@0: * are no more characters, and advance the position to the next sl@0: * character. sl@0: * @param options one or more of the following options, bitwise-OR-ed sl@0: * together: PARSE_VARIABLES, PARSE_ESCAPES, SKIP_WHITESPACE. sl@0: * @param isEscaped output parameter set to TRUE if the character sl@0: * was escaped sl@0: * @param ec input-output error code. An error will only be set by sl@0: * this routing if options includes PARSE_VARIABLES and an unknown sl@0: * variable name is seen, or if options includes PARSE_ESCAPES and sl@0: * an invalid escape sequence is seen. sl@0: * @return the current 32-bit code point, or DONE sl@0: */ sl@0: UChar32 next(int32_t options, UBool& isEscaped, UErrorCode& ec); sl@0: sl@0: /** sl@0: * Returns true if this iterator is currently within a variable expansion. sl@0: */ sl@0: inline UBool inVariable() const; sl@0: sl@0: /** sl@0: * An opaque object representing the position of a RuleCharacterIterator. sl@0: */ sl@0: struct Pos { sl@0: private: sl@0: const UnicodeString* buf; sl@0: int32_t pos; sl@0: int32_t bufPos; sl@0: friend class RuleCharacterIterator; sl@0: }; sl@0: sl@0: /** sl@0: * Sets an object which, when later passed to setPos(), will sl@0: * restore this iterator's position. Usage idiom: sl@0: * sl@0: * RuleCharacterIterator iterator = ...; sl@0: * RuleCharacterIterator::Pos pos; sl@0: * iterator.getPos(pos); sl@0: * for (;;) { sl@0: * iterator.getPos(pos); sl@0: * int c = iterator.next(...); sl@0: * ... sl@0: * } sl@0: * iterator.setPos(pos); sl@0: * sl@0: * @param p a position object to be set to this iterator's sl@0: * current position. sl@0: */ sl@0: void getPos(Pos& p) const; sl@0: sl@0: /** sl@0: * Restores this iterator to the position it had when getPos() sl@0: * set the given object. sl@0: * @param p a position object previously set by getPos() sl@0: */ sl@0: void setPos(const Pos& p); sl@0: sl@0: /** sl@0: * Skips ahead past any ignored characters, as indicated by the given sl@0: * options. This is useful in conjunction with the lookahead() method. sl@0: * sl@0: * Currently, this only has an effect for SKIP_WHITESPACE. sl@0: * @param options one or more of the following options, bitwise-OR-ed sl@0: * together: PARSE_VARIABLES, PARSE_ESCAPES, SKIP_WHITESPACE. sl@0: */ sl@0: void skipIgnored(int32_t options); sl@0: sl@0: /** sl@0: * Returns a string containing the remainder of the characters to be sl@0: * returned by this iterator, without any option processing. If the sl@0: * iterator is currently within a variable expansion, this will only sl@0: * extend to the end of the variable expansion. This method is provided sl@0: * so that iterators may interoperate with string-based APIs. The typical sl@0: * sequence of calls is to call skipIgnored(), then call lookahead(), then sl@0: * parse the string returned by lookahead(), then call jumpahead() to sl@0: * resynchronize the iterator. sl@0: * @param result a string to receive the characters to be returned sl@0: * by future calls to next() sl@0: * @return a reference to result sl@0: */ sl@0: UnicodeString& lookahead(UnicodeString& result) const; sl@0: sl@0: /** sl@0: * Advances the position by the given number of 16-bit code units. sl@0: * This is useful in conjunction with the lookahead() method. sl@0: * @param count the number of 16-bit code units to jump over sl@0: */ sl@0: void jumpahead(int32_t count); sl@0: sl@0: /** sl@0: * Returns a string representation of this object, consisting of the sl@0: * characters being iterated, with a '|' marking the current position. sl@0: * Position within an expanded variable is not indicated. sl@0: * @param result output parameter to receive a string sl@0: * representation of this object sl@0: */ sl@0: // UnicodeString& toString(UnicodeString& result) const; sl@0: sl@0: private: sl@0: /** sl@0: * Returns the current 32-bit code point without parsing escapes, parsing sl@0: * variables, or skipping whitespace. sl@0: * @return the current 32-bit code point sl@0: */ sl@0: UChar32 _current() const; sl@0: sl@0: /** sl@0: * Advances the position by the given amount. sl@0: * @param count the number of 16-bit code units to advance past sl@0: */ sl@0: void _advance(int32_t count); sl@0: }; sl@0: sl@0: inline UBool RuleCharacterIterator::inVariable() const { sl@0: return buf != 0; sl@0: } sl@0: sl@0: U_NAMESPACE_END sl@0: sl@0: #endif // _RULEITER_H_ sl@0: //eof