]> git.saurik.com Git - apple/icu.git/blob - icuSources/common/rbbiscan.h
ICU-8.11.4.tar.gz
[apple/icu.git] / icuSources / common / rbbiscan.h
1 //
2 // rbbiscan.h
3 //
4 // Copyright (C) 2002-2005, International Business Machines Corporation and others.
5 // All Rights Reserved.
6 //
7 // This file contains declarations for class RBBIRuleScanner
8 //
9
10
11 #ifndef RBBISCAN_H
12 #define RBBISCAN_H
13
14 #include "unicode/utypes.h"
15 #include "unicode/uobject.h"
16 #include "unicode/rbbi.h"
17 #include "unicode/uniset.h"
18 #include "unicode/parseerr.h"
19 #include "uhash.h"
20 #include "uvector.h"
21 #include "unicode/symtable.h"// For UnicodeSet parsing, is the interface that
22 // looks up references to $variables within a set.
23 #include "rbbinode.h"
24 //#include "rbbitblb.h"
25
26
27
28 U_NAMESPACE_BEGIN
29
30 class RBBIRuleBuilder;
31 class RBBISymbolTable;
32
33
34 //--------------------------------------------------------------------------------
35 //
36 // class RBBIRuleScanner does the lowest level, character-at-a-time
37 // scanning of break iterator rules.
38 //
39 // The output of the scanner is parse trees for
40 // the rule expressions and a list of all Unicode Sets
41 // encountered.
42 //
43 //--------------------------------------------------------------------------------
44 static const int kStackSize = 100; // The size of the state stack for
45 // rules parsing. Corresponds roughly
46 // to the depth of parentheses nesting
47 // that is allowed in the rules.
48
49 enum EParseAction {dummy01, dummy02}; // Placeholder enum for the specifier for
50 // actions that are specified in the
51 // rule parsing state table.
52
53 class RBBIRuleScanner : public UMemory {
54 public:
55
56 struct RBBIRuleChar {
57 UChar32 fChar;
58 UBool fEscaped;
59 };
60
61 RBBIRuleScanner(RBBIRuleBuilder *rb);
62
63
64 virtual ~RBBIRuleScanner();
65
66 void nextChar(RBBIRuleChar &c); // Get the next char from the input stream.
67 // Return false if at end.
68
69 UBool push(const RBBIRuleChar &c); // Push (unget) one character.
70 // Only a single character may be pushed.
71
72 void parse(); // Parse the rules, generating two parse
73 // trees, one each for the forward and
74 // reverse rules,
75 // and a list of UnicodeSets encountered.
76
77 /**
78 * Return a rules string without unnecessary
79 * characters.
80 */
81 static UnicodeString stripRules(const UnicodeString &rules);
82 private:
83
84 UBool doParseActions(EParseAction a);
85 void error(UErrorCode e); // error reporting convenience function.
86 void fixOpStack(RBBINode::OpPrecedence p);
87 // a character.
88 void findSetFor(const UnicodeString &s, RBBINode *node, UnicodeSet *setToAdopt = NULL);
89
90 UChar32 nextCharLL();
91 #ifdef RBBI_DEBUG
92 void printNodeStack(const char *title);
93 #endif
94 RBBINode *pushNewNode(RBBINode::NodeType t);
95 void scanSet();
96
97
98 RBBIRuleBuilder *fRB; // The rule builder that we are part of.
99
100 int32_t fScanIndex; // Index of current character being processed
101 // in the rule input string.
102 int32_t fNextIndex; // Index of the next character, which
103 // is the first character not yet scanned.
104 UBool fQuoteMode; // Scan is in a 'quoted region'
105 int32_t fLineNum; // Line number in input file.
106 int32_t fCharNum; // Char position within the line.
107 UChar32 fLastChar; // Previous char, needed to count CR-LF
108 // as a single line, not two.
109
110 RBBIRuleChar fC; // Current char for parse state machine
111 // processing.
112 UnicodeString fVarName; // $variableName, valid when we've just
113 // scanned one.
114
115 RBBIRuleTableEl **fStateTable; // State Transition Table for RBBI Rule
116 // parsing. index by p[state][char-class]
117
118 uint16_t fStack[kStackSize]; // State stack, holds state pushes
119 int32_t fStackPtr; // and pops as specified in the state
120 // transition rules.
121
122 RBBINode *fNodeStack[kStackSize]; // Node stack, holds nodes created
123 // during the parse of a rule
124 int32_t fNodeStackPtr;
125
126
127 UBool fReverseRule; // True if the rule currently being scanned
128 // is a reverse direction rule (if it
129 // starts with a '!')
130
131 UBool fLookAheadRule; // True if the rule includes a '/'
132 // somewhere within it.
133
134 RBBISymbolTable *fSymbolTable; // symbol table, holds definitions of
135 // $variable symbols.
136
137 UHashtable *fSetTable; // UnicocodeSet hash table, holds indexes to
138 // the sets created while parsing rules.
139 // The key is the string used for creating
140 // the set.
141
142 UnicodeSet *fRuleSets[10]; // Unicode Sets that are needed during
143 // the scanning of RBBI rules. The
144 // indicies for these are assigned by the
145 // perl script that builds the state tables.
146 // See rbbirpt.h.
147
148 int32_t fRuleNum; // Counts each rule as it is scanned.
149
150 int32_t fOptionStart; // Input index of start of a !!option
151 // keyword, while being scanned.
152
153 UnicodeSet *gRuleSet_rule_char;
154 UnicodeSet *gRuleSet_white_space;
155 UnicodeSet *gRuleSet_name_char;
156 UnicodeSet *gRuleSet_name_start_char;
157
158 RBBIRuleScanner(const RBBIRuleScanner &other); // forbid copying of this class
159 RBBIRuleScanner &operator=(const RBBIRuleScanner &other); // forbid copying of this class
160 };
161
162 U_NAMESPACE_END
163
164 #endif