From 5f8de423f190bbb79a62f804151bc24824fa32d8 Mon Sep 17 00:00:00 2001 From: "Matt A. Tobin" Date: Fri, 2 Feb 2018 04:16:08 -0500 Subject: Add m-esr52 at 52.6.0 --- intl/icu/source/common/rbbirb.h | 213 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 213 insertions(+) create mode 100644 intl/icu/source/common/rbbirb.h (limited to 'intl/icu/source/common/rbbirb.h') diff --git a/intl/icu/source/common/rbbirb.h b/intl/icu/source/common/rbbirb.h new file mode 100644 index 000000000..e4f736c8f --- /dev/null +++ b/intl/icu/source/common/rbbirb.h @@ -0,0 +1,213 @@ +// Copyright (C) 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html +// +// rbbirb.h +// +// Copyright (C) 2002-2008, International Business Machines Corporation and others. +// All Rights Reserved. +// +// This file contains declarations for several classes from the +// Rule Based Break Iterator rule builder. +// + + +#ifndef RBBIRB_H +#define RBBIRB_H + +#include "unicode/utypes.h" +#include "unicode/uobject.h" +#include "unicode/rbbi.h" +#include "unicode/uniset.h" +#include "unicode/parseerr.h" +#include "uhash.h" +#include "uvector.h" +#include "unicode/symtable.h"// For UnicodeSet parsing, is the interface that + // looks up references to $variables within a set. + + + +U_NAMESPACE_BEGIN + +class RBBIRuleScanner; +struct RBBIRuleTableEl; +class RBBISetBuilder; +class RBBINode; +class RBBITableBuilder; + + + +//-------------------------------------------------------------------------------- +// +// RBBISymbolTable. Implements SymbolTable interface that is used by the +// UnicodeSet parser to resolve references to $variables. +// +//-------------------------------------------------------------------------------- +class RBBISymbolTableEntry : public UMemory { // The symbol table hash table contains one +public: // of these structs for each entry. + RBBISymbolTableEntry(); + UnicodeString key; + RBBINode *val; + ~RBBISymbolTableEntry(); + +private: + RBBISymbolTableEntry(const RBBISymbolTableEntry &other); // forbid copying of this class + RBBISymbolTableEntry &operator=(const RBBISymbolTableEntry &other); // forbid copying of this class +}; + + +class RBBISymbolTable : public UMemory, public SymbolTable { +private: + const UnicodeString &fRules; + UHashtable *fHashTable; + RBBIRuleScanner *fRuleScanner; + + // These next two fields are part of the mechanism for passing references to + // already-constructed UnicodeSets back to the UnicodeSet constructor + // when the pattern includes $variable references. + const UnicodeString ffffString; // = "/uffff" + UnicodeSet *fCachedSetLookup; + +public: + // API inherited from class SymbolTable + virtual const UnicodeString* lookup(const UnicodeString& s) const; + virtual const UnicodeFunctor* lookupMatcher(UChar32 ch) const; + virtual UnicodeString parseReference(const UnicodeString& text, + ParsePosition& pos, int32_t limit) const; + + // Additional Functions + RBBISymbolTable(RBBIRuleScanner *, const UnicodeString &fRules, UErrorCode &status); + virtual ~RBBISymbolTable(); + + virtual RBBINode *lookupNode(const UnicodeString &key) const; + virtual void addEntry (const UnicodeString &key, RBBINode *val, UErrorCode &err); + +#ifdef RBBI_DEBUG + virtual void rbbiSymtablePrint() const; +#else + // A do-nothing inline function for non-debug builds. Member funcs can't be empty + // or the call sites won't compile. + int32_t fFakeField; + #define rbbiSymtablePrint() fFakeField=0; +#endif + +private: + RBBISymbolTable(const RBBISymbolTable &other); // forbid copying of this class + RBBISymbolTable &operator=(const RBBISymbolTable &other); // forbid copying of this class +}; + + +//-------------------------------------------------------------------------------- +// +// class RBBIRuleBuilder The top-level class handling RBBI rule compiling. +// +//-------------------------------------------------------------------------------- +class RBBIRuleBuilder : public UMemory { +public: + + // Create a rule based break iterator from a set of rules. + // This function is the main entry point into the rule builder. The + // public ICU API for creating RBBIs uses this function to do the actual work. + // + static BreakIterator * createRuleBasedBreakIterator( const UnicodeString &rules, + UParseError *parseError, + UErrorCode &status); + +public: + // The "public" functions and data members that appear below are accessed + // (and shared) by the various parts that make up the rule builder. They + // are NOT intended to be accessed by anything outside of the + // rule builder implementation. + RBBIRuleBuilder(const UnicodeString &rules, + UParseError *parseErr, + UErrorCode &status + ); + + virtual ~RBBIRuleBuilder(); + char *fDebugEnv; // controls debug trace output + UErrorCode *fStatus; // Error reporting. Keeping status + UParseError *fParseError; // here avoids passing it everywhere. + const UnicodeString &fRules; // The rule string that we are compiling + + RBBIRuleScanner *fScanner; // The scanner. + RBBINode *fForwardTree; // The parse trees, generated by the scanner, + RBBINode *fReverseTree; // then manipulated by subsequent steps. + RBBINode *fSafeFwdTree; + RBBINode *fSafeRevTree; + + RBBINode **fDefaultTree; // For rules not qualified with a ! + // the tree to which they belong to. + + UBool fChainRules; // True for chained Unicode TR style rules. + // False for traditional regexp rules. + + UBool fLBCMNoChain; // True: suppress chaining of rules on + // chars with LineBreak property == CM. + + UBool fLookAheadHardBreak; // True: Look ahead matches cause an + // immediate break, no continuing for the + // longest match. + + RBBISetBuilder *fSetBuilder; // Set and Character Category builder. + UVector *fUSetNodes; // Vector of all uset nodes. + + RBBITableBuilder *fForwardTables; // State transition tables + RBBITableBuilder *fReverseTables; + RBBITableBuilder *fSafeFwdTables; + RBBITableBuilder *fSafeRevTables; + + UVector *fRuleStatusVals; // The values that can be returned + // from getRuleStatus(). + + RBBIDataHeader *flattenData(); // Create the flattened (runtime format) + // data tables.. +private: + RBBIRuleBuilder(const RBBIRuleBuilder &other); // forbid copying of this class + RBBIRuleBuilder &operator=(const RBBIRuleBuilder &other); // forbid copying of this class +}; + + + + +//---------------------------------------------------------------------------- +// +// RBBISetTableEl is an entry in the hash table of UnicodeSets that have +// been encountered. The val Node will be of nodetype uset +// and contain pointers to the actual UnicodeSets. +// The Key is the source string for initializing the set. +// +// The hash table is used to avoid creating duplicate +// unnamed (not $var references) UnicodeSets. +// +// Memory Management: +// The Hash Table owns these RBBISetTableEl structs and +// the key strings. It does NOT own the val nodes. +// +//---------------------------------------------------------------------------- +struct RBBISetTableEl { + UnicodeString *key; + RBBINode *val; +}; + + +//---------------------------------------------------------------------------- +// +// RBBIDebugPrintf Printf equivalent, for debugging output. +// Conditional compilation of the implementation lets us +// get rid of the stdio dependency in environments where it +// is unavailable. +// +//---------------------------------------------------------------------------- +#ifdef RBBI_DEBUG +#include +#define RBBIDebugPrintf printf +#define RBBIDebugPuts puts +#else +#undef RBBIDebugPrintf +#define RBBIDebugPuts(arg) +#endif + +U_NAMESPACE_END +#endif + + + -- cgit v1.2.3