Annotation of parser3/src/include/pa_string.h, revision 1.144.2.28.2.40
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.144.2.18 paf 4: Copyright (c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com)
1.124 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.1 paf 6: */
7:
8: #ifndef PA_STRING_H
9: #define PA_STRING_H
1.140 paf 10:
1.144.2.28.2. 0(paf 11:3): static const char* IDENT_STRING_H="$Date: 2003/03/27 10:04:06 $";
1.144.2.10 paf 12:
13: // includes
1.1 paf 14:
1.144.2.12 paf 15: #include "pa_types.h"
16: #include "pa_array.h"
1.4 paf 17:
1.144.2.28.2. (paf 18:): extern "C" { // cord's author forgot to do that
7(paf 19:3): #define CORD_NO_IO
(paf 20:): #include "cord.h"
21:): };
22:):
1.144.2.12 paf 23: // forwards
1.144.2.10 paf 24:
1.144.2.28 paf 25: class Charset;
1.144.2.28.2. (paf 26:): class Table;
1.144.2.5 paf 27: class SQL_Connection;
28: class Dictionary;
1.144.2.15 paf 29: class Request_charsets;
1.144.2.28.2. (paf 30:): class String;
7(paf 31:3): typedef Array<const String*> ArrayString;
1.144.2.5 paf 32:
33: // helpers
1.144.2.28.2. 0(paf 34:3): /*
(paf 35:): /// appends clean piece to String @see String::append
7(paf 36:3): #define APPEND_CLEAN(src, length) append(src, length, String::L_CLEAN)
(paf 37:): /// appends piece to String as-is @see String::append
7(paf 38:3): #define APPEND_AS_IS(src, length) append(src, length, String::L_AS_IS)
(paf 39:): /// appends tainted piece to String @see String::append
7(paf 40:3): #define APPEND_TAINTED(src, length) append(src, length, String::L_TAINTED)
(paf 41:): /// handy: appends const char* piece to String @see String::append
42:): #define APPEND_CONST(src) APPEND_AS_IS(src, 0)
0(paf 43:3): */
(paf 44:): /// this is result of pos functions which mean that substr were not found
45:): #define STRING_NOT_FOUND ((size_t)-1)
46:):
7(paf 47:3): class StringBody {
48:3):
49:3): CORD body;
50:3):
51:3): public:
52:3):
53:3): StringBody(): body(CORD_EMPTY) {}
5(paf 54:3): StringBody(CORD abody): body(abody) {
55:3): #ifdef _DEBUG
56:3): if(body && !*body)
57:3): switch(*(char*)(body+1)) {
58:3): case 1: case 4: case 6: break;
59:3): default: _asm int 3; break;
60:3): }
61:3): #endif
62:3): }
63:3): StringBody(const char* str, size_t length): body(CORD_EMPTY) {
2(paf 64:3): append(str, length);
65:3): }
4(paf 66:3): static StringBody Format(int value);
7(paf 67:3):
68:3): void clear() { body=CORD_EMPTY; }
69:3):
70:3): bool operator! () const { return is_empty(); }
71:3):
72:3): uint hash_code() const;
73:3):
0(paf 74:3): const char* cstr() const { return CORD_to_const_char_star(body); }
1(paf 75:3): char* cstrm() const { return CORD_to_char_star(body); }
0(paf 76:3):
7(paf 77:3): size_t length() const { return CORD_len(body); }
78:3):
79:3): bool is_empty() const { return body==CORD_EMPTY; }
80:3):
81:3): void append(const char *str, size_t length) {
5(paf 82:3): #ifdef _DEBUG
83:3): if(!*str)
84:3): _asm int 3;
85:3): bool were_empty=cstr==CORD_EMPTY;
86:3): #endif
2(paf 87:3): body=CORD_cat_char_star(body, str, length);
5(paf 88:3): #ifdef _DEBUG
89:3): if(were_empty && !*str)
90:3): _asm int 3;
91:3): #endif
7(paf 92:3): }
2(paf 93:3): void append(char c) { body=CORD_cat_char(body, c); }
6(paf 94:3): StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
95:3): StringBody& operator << (const char* str) { body=CORD_cat_char_star(body, str, strlen(str)); return *this; }
7(paf 96:3):
97:3): // could not figure out why this operator is needed [should do this chain: string->simple->==]
8(paf 98:3): bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
99:3): bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
100:3): bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
101:3): bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; }
102:3): bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; }
103:3): bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; }
7(paf 104:3):
105:3): int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
106:3): return CORD_ncmp(body, x_begin, y.body, y_begin, size);
107:3): }
108:3):
109:3): char fetch(size_t index) const { return CORD_fetch(body, index); }
1(paf 110:3): StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
3(paf 111:3): size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
1(paf 112:3): size_t pos(const StringBody substr, size_t offset=0) const {
113:3): if(!substr.length())
114:3): return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
115:3): return CORD_str(body, offset, substr.body);
116:3): }
2(paf 117:3): size_t pos(char c,
118:3): size_t offset=0) const {
119:3): return CORD_chr(body, offset, c);
120:3): }
1(paf 121:3):
122:3): template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
123:3): CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
124:3): }
125:3):
126:3): void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
7(paf 127:3):
128:3): };
(paf 129:):
1.144.2.2 paf 130: /**
131: String which knows the language of all it's fragments.
1.41 paf 132:
133: All pieces remember
134: - whether they are tainted or not,
135: and the language which should be used to detaint them
136: */
1.144.2.28.2. (paf 137:): class String: public PA_Object {
138:):
2(paf 139:3): // friend class StringBody;
7(paf 140:3):
1.144.2.2 paf 141: public:
1.1 paf 142:
1.106 parser 143: /** piece is tainted or not. the language to use when detaint
144: remember to change String_Untaint_lang_name @ untaint.C along
145: */
1.144.2.28.2. 7(paf 146:3): enum Language {
147:3): L_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum
148:3): // these two must go before others, there are checks for >L_AS_IS
149:3): L_CLEAN, ///< clean
150:3): L_AS_IS, ///< leave all characters intact
1.122 paf 151:
1.144.2.28.2. 7(paf 152:3): L_PASS_APPENDED,
1.41 paf 153: /**<
154: leave language built into string being appended.
155: just a flag, that value not stored
156: */
1.144.2.28.2. 7(paf 157:3): L_TAINTED, ///< tainted, untaint language as assigned later
1.122 paf 158: // untaint languages. assigned by ^untaint[lang]{...}
1.144.2.28.2. 7(paf 159:3): L_FILE_SPEC, ///< file specification
160:3): L_HTTP_HEADER, ///< text in HTTP response header
161:3): L_MAIL_HEADER, ///< text in mail header
162:3): L_URI, ///< text in uri
163:3): L_TABLE, ///< ^table:set body
164:3): L_SQL, ///< ^table:sql body
165:3): L_JS, ///< JavaScript code
166:3): L_XML, ///< ^dom:set xml
167:3): L_HTML, ///< HTML code (for editing)
168:3): L_OPTIMIZE_BIT = 0x8000 ///< flag, requiring cstr whitespace optimization
1.27 paf 169: };
170:
1.144.2.28.2. (paf 171:): struct Fragment {
7(paf 172:3): Language lang; ///< untaint flag, later untaint language
0(paf 173:3): size_t length; ///< length
9(paf 174:3): Fragment(Language alang, size_t asize): lang(alang), length(asize) {
0(paf 175:3): #ifdef _DEBUG
9(paf 176:3): if(alang==L_UNSPECIFIED)
0(paf 177:3): _asm int 3; // MUST NOT
178:3): if(asize==(size_t)-1)
179:3): _asm int 3; // MUST NOT
180:3): #endif
9(paf 181:3): }
(paf 182:): };
183:):
184:): class ArrayFragment: public Array<Fragment> {
6(paf 185:3): void append(element_type src) {
(paf 186:): *static_cast<Array<Fragment> *>(this)+=src;
187:): }
6(paf 188:3): /// hiding from accidental USE, use append_positions
189:3): void append(const ArrayFragment& src, int offset, int limit) {
190:3): static_cast<Array<Fragment> *>(this)->append(src, offset, limit);
191:3): }
(paf 192:): public:
193:): ArrayFragment& operator += (element_type src) {
8(paf 194:3): if(size_t lcount=count()) { // not empty?
(paf 195:): // try to join with last
0(paf 196:3): Fragment& last=unchecked_get_ref(lcount-1);
(paf 197:): if(last.lang==src.lang) {
0(paf 198:3): last.length+=src.length;
(paf 199:): return *this;
200:): }
201:): }
6(paf 202:3): append(src);
(paf 203:): return *this;
204:): }
6(paf 205:3): void append(const ArrayFragment& src) { append(src, 0, 0); }
206:3): void append_positions(const ArrayFragment& src, size_t substr_begin, size_t substr_end);
(paf 207:): };
208:):
209:): struct C {
210:): char *str;
0(paf 211:3): size_t length;
(paf 212:): operator char *() { return str; }
0(paf 213:3): C(char *astr, size_t asize): str(astr), length(asize) {}
(paf 214:): };
215:):
216:): private:
217:):
7(paf 218:3): StringBody body; ///< all characters of string
0(paf 219:3): ArrayFragment fragments; ///< fragment language+length info
(paf 220:):
1.8 paf 221: public:
222:
1.144.2.28.2. 9(paf 223:3): explicit String(const char* cstr=0, size_t length=0, bool tainted=false);
1.14 paf 224: String(const String& src);
1.144.2.28.2. 7(paf 225:3): String(StringBody abody, Language alang): body(abody) {
226:3): fragments+=Fragment(alang, abody.length());
227:3): }
228:3):
229:3): /// for convinient hash lookup
2(paf 230:3): operator const StringBody() const { return body; }
7(paf 231:3):
232:3): bool is_empty() const { return body.is_empty(); }
233:3): size_t length() const { return body.length(); }
(paf 234:):
2(paf 235:3): /// convert to CORD. if 'lang' known, forcing 'lang' to it
1(paf 236:3): StringBody cstr_to_string_body(Language lang=L_AS_IS,
7(paf 237:3): SQL_Connection* connection=0,
(paf 238:): const Request_charsets *charsets=0) const;
239:):
2(paf 240:3): /// convert to constant C string. if 'lang' known, forcing 'lang' to it
7(paf 241:3): const char* cstr(Language lang=L_AS_IS,
242:3): SQL_Connection* connection=0,
1.144.2.14 paf 243: const Request_charsets *charsets=0) const {
1.144.2.28.2. 1(paf 244:3): return cstr_to_string_body(lang, connection, charsets).cstr();
(paf 245:): }
2(paf 246:3): /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
7(paf 247:3): char *cstrm(Language lang=L_AS_IS,
248:3): SQL_Connection* connection=0,
(paf 249:): const Request_charsets *charsets=0) const {
1(paf 250:3): return cstr_to_string_body(lang, connection, charsets).cstrm();
1.144.2.9 paf 251: }
1.108 parser 252: /// puts pieces to buf
1.144.2.28.2. (paf 253:): C serialize(size_t prolog_size) const;
1.108 parser 254: /// appends pieces from buf to self
1.144.2.28.2. 4(paf 255:3): bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
6(paf 256:3): /** append fragment,
9(paf 257:3): @c length can be zero, it will be autocalced
6(paf 258:3): when length<strlen(str) makes heap copy of @c str piece
1.46 paf 259: */
1.144.2.28.2. 9(paf 260:3): String& append(const char* str, size_t length, Language lang);
(paf 261:):
7(paf 262:3): bool operator == (const char* y) const { return body==StringBody(y); }
0(paf 263:3): bool operator != (const char* y) const { return body!=StringBody(y); }
(paf 264:):
265:): /// this starts with y
7(paf 266:3): bool starts_with(const char* y) const {
267:3): return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
1.80 paf 268: }
1.144.2.28.2. 9(paf 269:3): /// x starts with this
270:3): bool this_starts(const char* x) const {
271:3): return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
(paf 272:): }
1.80 paf 273:
1.144.2.28.2. 7(paf 274:3): String& append_to(String& dest, Language lang, bool forced) const;
275:3): String& append(const String& src, Language lang, bool forced=false) {
1.144.2.11 paf 276: return src.append_to(*this, lang, forced);
277: }
1.144.2.28.2. 7(paf 278:3): String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
279:3): String& operator << (const char* src) { return append(src, 0, L_AS_IS); }
8(paf 280:3): String& operator << (const StringBody src) {
6(paf 281:3): body<<src;
9(paf 282:3): fragments+=Fragment(L_AS_IS, src.length());
283:3): return *this;
284:3): }
1.100 parser 285:
1.142 paf 286: /// extracts first char of a string, if any
287: char first_char() const {
1.144.2.28.2. 7(paf 288:3): return is_empty()?0:body.fetch(0);
1.142 paf 289: }
1.144.2.28.2. 8(paf 290:3):
291:3): bool operator < (const String& src) const { return body<src.body; }
292:3): bool operator > (const String& src) const { return body>src.body; }
293:3): bool operator <= (const String& src) const { return body<=src.body; }
294:3): bool operator >= (const String& src) const { return body>=src.body; }
295:3): bool operator != (const String& src) const { return body!=src.body; }
296:3): bool operator == (const String& src) const { return body==src.body; }
1.54 paf 297:
298: /// extracts [start, finish) piece of string
1.144.2.28.2. 2(paf 299:3): String& mid(size_t substr_begin, size_t substr_end) const;
(paf 300:):
301:): /**
7(paf 302:3): ignore lang if it's L_UNSPECIFIED
(paf 303:): but when specified: look for substring that lies in ONE fragment in THAT lang
304:): @return position of substr in string, -1 means "not found" [const char* version]
305:): */
1(paf 306:3): size_t pos(const StringBody substr,
7(paf 307:3): size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
308:3): /// String version of @see pos(const char*, int, Language)
(paf 309:): size_t pos(const String& substr,
7(paf 310:3): size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
2(paf 311:3): size_t pos(char c,
312:3): size_t this_offset=0) const {
313:3): return body.pos(c, this_offset);
314:3): }
1.62 paf 315:
1.144.2.22 paf 316: void split(ArrayString& result,
1.144.2.28.2. 2(paf 317:3): size_t& pos_after,
(paf 318:): const char* delim,
2(paf 319:3): Language lang=L_UNSPECIFIED, int limit=-1) const;
1.144.2.22 paf 320: void split(ArrayString& result,
1.144.2.28.2. 5(paf 321:3): size_t& pos_after,
1.62 paf 322: const String& delim,
1.144.2.28.2. 7(paf 323:3): Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 324:
1.144.2.28.2. 6(paf 325:3): typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 326: int prestart, int prefinish,
327: int poststart, int postfinish,
1.144.2.9 paf 328: void *info);
1.87 parser 329: /**
1.144.2.5 paf 330: @return table of found items, if any.
1.87 parser 331: table format is defined and fixed[can be used by others]:
332: @verbatim
333: prematch/match/postmatch/1/2/3/...
334: @endverbatim
335: */
1.144.2.28.2. (paf 336:): Table* match(Charset& source_charset,
1.64 paf 337: const String& regexp,
1.144.2.28.2. 5(paf 338:3): const String* options,
1.99 parser 339: Row_action row_action, void *info,
1.144.2.27 paf 340: bool& just_matched) const;
1.87 parser 341: enum Change_case_kind {
342: CC_UPPER,
343: CC_LOWER
344: };
1.144.2.28.2. (paf 345:): String& change_case(Charset& source_charset,
5(paf 346:3): Change_case_kind kind) const;
(paf 347:): const String& replace(const Dictionary& dict) const;
348:): double as_double() const;
349:): int as_int() const;
1.137 paf 350:
1.7 paf 351: private: //disabled
352:
1.12 paf 353: String& operator = (const String&) { return *this; }
1.7 paf 354:
1.1 paf 355: };
1.144.2.7 paf 356:
357: /// simple hash code of string. used by Hash
1.144.2.28.2. 8(paf 358:3): inline uint hash_code(const StringBody self) {
1.144.2.7 paf 359: return self.hash_code();
360: }
1.1 paf 361:
362: #endif
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_string.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / include