1 | /* |
---|
2 | * Licensed to the Apache Software Foundation (ASF) under one or more |
---|
3 | * contributor license agreements. See the NOTICE file distributed with |
---|
4 | * this work for additional information regarding copyright ownership. |
---|
5 | * The ASF licenses this file to You under the Apache License, Version 2.0 |
---|
6 | * (the "License"); you may not use this file except in compliance with |
---|
7 | * the License. You may obtain a copy of the License at |
---|
8 | * |
---|
9 | * http://www.apache.org/licenses/LICENSE-2.0 |
---|
10 | * |
---|
11 | * Unless required by applicable law or agreed to in writing, software |
---|
12 | * distributed under the License is distributed on an "AS IS" BASIS, |
---|
13 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
---|
14 | * See the License for the specific language governing permissions and |
---|
15 | * limitations under the License. |
---|
16 | */ |
---|
17 | |
---|
18 | /* |
---|
19 | * $Id: RegxParser.hpp 568078 2007-08-21 11:43:25Z amassari $ |
---|
20 | */ |
---|
21 | |
---|
22 | /* |
---|
23 | * A regular expression parser |
---|
24 | */ |
---|
25 | #if !defined(REGXPARSER_HPP) |
---|
26 | #define REGXPARSER_HPP |
---|
27 | |
---|
28 | // --------------------------------------------------------------------------- |
---|
29 | // Includes |
---|
30 | // --------------------------------------------------------------------------- |
---|
31 | #include <xercesc/util/RefVectorOf.hpp> |
---|
32 | #include <xercesc/util/XMLUniDefs.hpp> |
---|
33 | |
---|
34 | XERCES_CPP_NAMESPACE_BEGIN |
---|
35 | |
---|
36 | // --------------------------------------------------------------------------- |
---|
37 | // Forward Declaration |
---|
38 | // --------------------------------------------------------------------------- |
---|
39 | class Token; |
---|
40 | class RangeToken; |
---|
41 | class TokenFactory; |
---|
42 | |
---|
43 | class XMLUTIL_EXPORT RegxParser : public XMemory |
---|
44 | { |
---|
45 | public: |
---|
46 | |
---|
47 | // ----------------------------------------------------------------------- |
---|
48 | // Public constant data |
---|
49 | // ----------------------------------------------------------------------- |
---|
50 | // Parse tokens |
---|
51 | enum { |
---|
52 | REGX_T_CHAR = 0, |
---|
53 | REGX_T_EOF = 1, |
---|
54 | REGX_T_OR = 2, |
---|
55 | REGX_T_STAR = 3, |
---|
56 | REGX_T_PLUS = 4, |
---|
57 | REGX_T_QUESTION = 5, |
---|
58 | REGX_T_LPAREN = 6, |
---|
59 | REGX_T_RPAREN = 7, |
---|
60 | REGX_T_DOT = 8, |
---|
61 | REGX_T_LBRACKET = 9, |
---|
62 | REGX_T_BACKSOLIDUS = 10, |
---|
63 | REGX_T_CARET = 11, |
---|
64 | REGX_T_DOLLAR = 12, |
---|
65 | REGX_T_LPAREN2 = 13, |
---|
66 | REGX_T_LOOKAHEAD = 14, |
---|
67 | REGX_T_NEGATIVELOOKAHEAD = 15, |
---|
68 | REGX_T_LOOKBEHIND = 16, |
---|
69 | REGX_T_NEGATIVELOOKBEHIND = 17, |
---|
70 | REGX_T_INDEPENDENT = 18, |
---|
71 | REGX_T_SET_OPERATIONS = 19, |
---|
72 | REGX_T_POSIX_CHARCLASS_START = 20, |
---|
73 | REGX_T_COMMENT = 21, |
---|
74 | REGX_T_MODIFIERS = 22, |
---|
75 | REGX_T_CONDITION = 23, |
---|
76 | REGX_T_XMLSCHEMA_CC_SUBTRACTION = 24 |
---|
77 | }; |
---|
78 | |
---|
79 | static const unsigned short S_NORMAL; |
---|
80 | static const unsigned short S_INBRACKETS; |
---|
81 | static const unsigned short S_INXBRACKETS; |
---|
82 | |
---|
83 | // ----------------------------------------------------------------------- |
---|
84 | // Public Constructors and Destructor |
---|
85 | // ----------------------------------------------------------------------- |
---|
86 | RegxParser(MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager); |
---|
87 | virtual ~RegxParser(); |
---|
88 | |
---|
89 | // ----------------------------------------------------------------------- |
---|
90 | // Getter methods |
---|
91 | // ----------------------------------------------------------------------- |
---|
92 | unsigned short getParseContext() const; |
---|
93 | unsigned short getState() const; |
---|
94 | XMLInt32 getCharData() const; |
---|
95 | int getNoParen() const; |
---|
96 | int getOffset() const; |
---|
97 | bool hasBackReferences() const; |
---|
98 | TokenFactory* getTokenFactory() const; |
---|
99 | |
---|
100 | // ----------------------------------------------------------------------- |
---|
101 | // Setter methods |
---|
102 | // ----------------------------------------------------------------------- |
---|
103 | void setParseContext(const unsigned short value); |
---|
104 | void setTokenFactory(TokenFactory* const tokFactory); |
---|
105 | |
---|
106 | // ----------------------------------------------------------------------- |
---|
107 | // Public Parsing methods |
---|
108 | // ----------------------------------------------------------------------- |
---|
109 | Token* parse(const XMLCh* const regxStr, const int options); |
---|
110 | |
---|
111 | protected: |
---|
112 | // ----------------------------------------------------------------------- |
---|
113 | // Protected Helper methods |
---|
114 | // ----------------------------------------------------------------------- |
---|
115 | virtual bool checkQuestion(const int off); |
---|
116 | virtual XMLInt32 decodeEscaped(); |
---|
117 | MemoryManager* getMemoryManager() const; |
---|
118 | // ----------------------------------------------------------------------- |
---|
119 | // Protected Parsing/Processing methods |
---|
120 | // ----------------------------------------------------------------------- |
---|
121 | void processNext(); |
---|
122 | Token* parseRegx(const bool matchingRParen = false); |
---|
123 | virtual Token* processCaret(); |
---|
124 | virtual Token* processDollar(); |
---|
125 | virtual Token* processLook(const unsigned short tokType); |
---|
126 | virtual Token* processBacksolidus_A(); |
---|
127 | virtual Token* processBacksolidus_z(); |
---|
128 | virtual Token* processBacksolidus_Z(); |
---|
129 | virtual Token* processBacksolidus_b(); |
---|
130 | virtual Token* processBacksolidus_B(); |
---|
131 | virtual Token* processBacksolidus_lt(); |
---|
132 | virtual Token* processBacksolidus_gt(); |
---|
133 | virtual Token* processBacksolidus_c(); |
---|
134 | virtual Token* processBacksolidus_C(); |
---|
135 | virtual Token* processBacksolidus_i(); |
---|
136 | virtual Token* processBacksolidus_I(); |
---|
137 | virtual Token* processBacksolidus_g(); |
---|
138 | virtual Token* processBacksolidus_X(); |
---|
139 | virtual Token* processBackReference(); |
---|
140 | virtual Token* processStar(Token* const tok); |
---|
141 | virtual Token* processPlus(Token* const tok); |
---|
142 | virtual Token* processQuestion(Token* const tok); |
---|
143 | virtual Token* processParen(); |
---|
144 | virtual Token* processParen2(); |
---|
145 | virtual Token* processCondition(); |
---|
146 | virtual Token* processModifiers(); |
---|
147 | virtual Token* processIndependent(); |
---|
148 | virtual RangeToken* parseCharacterClass(const bool useNRange); |
---|
149 | virtual RangeToken* parseSetOperations(); |
---|
150 | virtual XMLInt32 processCInCharacterClass(RangeToken* const tok, |
---|
151 | const XMLInt32 ch); |
---|
152 | RangeToken* processBacksolidus_pP(const XMLInt32 ch); |
---|
153 | |
---|
154 | // ----------------------------------------------------------------------- |
---|
155 | // Protected PreCreated RangeToken access methods |
---|
156 | // ----------------------------------------------------------------------- |
---|
157 | virtual Token* getTokenForShorthand(const XMLInt32 ch); |
---|
158 | |
---|
159 | private: |
---|
160 | // ----------------------------------------------------------------------- |
---|
161 | // Private parsing/processing methods |
---|
162 | // ----------------------------------------------------------------------- |
---|
163 | Token* parseTerm(const bool matchingRParen = false); |
---|
164 | Token* parseFactor(); |
---|
165 | Token* parseAtom(); |
---|
166 | |
---|
167 | // ----------------------------------------------------------------------- |
---|
168 | // Unimplemented constructors and operators |
---|
169 | // ----------------------------------------------------------------------- |
---|
170 | RegxParser(const RegxParser&); |
---|
171 | RegxParser& operator=(const RegxParser&); |
---|
172 | |
---|
173 | // ----------------------------------------------------------------------- |
---|
174 | // Private data types |
---|
175 | // ----------------------------------------------------------------------- |
---|
176 | class ReferencePosition : public XMemory |
---|
177 | { |
---|
178 | public : |
---|
179 | ReferencePosition(const int refNo, const int position); |
---|
180 | |
---|
181 | int fReferenceNo; |
---|
182 | int fPosition; |
---|
183 | }; |
---|
184 | |
---|
185 | // ----------------------------------------------------------------------- |
---|
186 | // Private Helper methods |
---|
187 | // ----------------------------------------------------------------------- |
---|
188 | bool isSet(const int flag); |
---|
189 | int hexChar(const XMLInt32 ch); |
---|
190 | |
---|
191 | // ----------------------------------------------------------------------- |
---|
192 | // Private data members |
---|
193 | // ----------------------------------------------------------------------- |
---|
194 | MemoryManager* fMemoryManager; |
---|
195 | bool fHasBackReferences; |
---|
196 | int fOptions; |
---|
197 | int fOffset; |
---|
198 | int fNoGroups; |
---|
199 | unsigned short fParseContext; |
---|
200 | int fStringLen; |
---|
201 | unsigned short fState; |
---|
202 | XMLInt32 fCharData; |
---|
203 | XMLCh* fString; |
---|
204 | RefVectorOf<ReferencePosition>* fReferences; |
---|
205 | TokenFactory* fTokenFactory; |
---|
206 | }; |
---|
207 | |
---|
208 | |
---|
209 | // --------------------------------------------------------------------------- |
---|
210 | // RegxParser: Getter Methods |
---|
211 | // --------------------------------------------------------------------------- |
---|
212 | inline unsigned short RegxParser::getParseContext() const { |
---|
213 | |
---|
214 | return fParseContext; |
---|
215 | } |
---|
216 | |
---|
217 | inline unsigned short RegxParser::getState() const { |
---|
218 | |
---|
219 | return fState; |
---|
220 | } |
---|
221 | |
---|
222 | inline XMLInt32 RegxParser::getCharData() const { |
---|
223 | |
---|
224 | return fCharData; |
---|
225 | } |
---|
226 | |
---|
227 | inline int RegxParser::getNoParen() const { |
---|
228 | |
---|
229 | return fNoGroups; |
---|
230 | } |
---|
231 | |
---|
232 | inline int RegxParser::getOffset() const { |
---|
233 | |
---|
234 | return fOffset; |
---|
235 | } |
---|
236 | |
---|
237 | inline bool RegxParser::hasBackReferences() const { |
---|
238 | |
---|
239 | return fHasBackReferences; |
---|
240 | } |
---|
241 | |
---|
242 | inline TokenFactory* RegxParser::getTokenFactory() const { |
---|
243 | |
---|
244 | return fTokenFactory; |
---|
245 | } |
---|
246 | |
---|
247 | inline MemoryManager* RegxParser::getMemoryManager() const { |
---|
248 | return fMemoryManager; |
---|
249 | } |
---|
250 | // --------------------------------------------------------------------------- |
---|
251 | // RegxParser: Setter Methods |
---|
252 | // --------------------------------------------------------------------------- |
---|
253 | inline void RegxParser::setParseContext(const unsigned short value) { |
---|
254 | |
---|
255 | fParseContext = value; |
---|
256 | } |
---|
257 | |
---|
258 | inline void RegxParser::setTokenFactory(TokenFactory* const tokFactory) { |
---|
259 | |
---|
260 | fTokenFactory = tokFactory; |
---|
261 | } |
---|
262 | |
---|
263 | // --------------------------------------------------------------------------- |
---|
264 | // RegxParser: Helper Methods |
---|
265 | // --------------------------------------------------------------------------- |
---|
266 | inline bool RegxParser::isSet(const int flag) { |
---|
267 | |
---|
268 | return (fOptions & flag) == flag; |
---|
269 | } |
---|
270 | |
---|
271 | |
---|
272 | inline int RegxParser::hexChar(const XMLInt32 ch) { |
---|
273 | |
---|
274 | if (ch < chDigit_0 || ch > chLatin_f) |
---|
275 | return -1; |
---|
276 | |
---|
277 | if (ch <= chDigit_9) |
---|
278 | return ch - chDigit_0; |
---|
279 | |
---|
280 | if (ch < chLatin_A) |
---|
281 | return -1; |
---|
282 | |
---|
283 | if (ch <= chLatin_F) |
---|
284 | return ch - chLatin_A + 10; |
---|
285 | |
---|
286 | if (ch < chLatin_a) |
---|
287 | return -1; |
---|
288 | |
---|
289 | return ch - chLatin_a + 10; |
---|
290 | } |
---|
291 | |
---|
292 | XERCES_CPP_NAMESPACE_END |
---|
293 | |
---|
294 | #endif |
---|
295 | |
---|
296 | /** |
---|
297 | * End file RegxParser.hpp |
---|
298 | */ |
---|
299 | |
---|