Annotation of parser3/src/include/pa_string.h, revision 1.144.2.28.2.48
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. 8(paf 11:3): static const char* IDENT_STRING_H="$Date: 2003/04/08 12:47:30 $";
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:
1.144.2.28.2. (paf 33:): /// this is result of pos functions which mean that substr were not found
34:): #define STRING_NOT_FOUND ((size_t)-1)
35:):
7(paf 36:3): class StringBody {
37:3):
38:3): CORD body;
39:3):
40:3): public:
41:3):
42:3): StringBody(): body(CORD_EMPTY) {}
5(paf 43:3): StringBody(CORD abody): body(abody) {
3(paf 44:3): assert(!body // no body
45:3): || *body // ordinary string
46:3): || body[1]==1 // CONCAT_HDR
47:3): || body[1]==4 // FN_HDR
48:3): || body[1]==6 // SUBSTR_HDR
49:3): );
5(paf 50:3): }
1(paf 51:3): /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
52:3): StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
6(paf 53:3): append_know_length(str, helper_length?helper_length:strlen(str));
2(paf 54:3): }
4(paf 55:3): static StringBody Format(int value);
7(paf 56:3):
57:3): void clear() { body=CORD_EMPTY; }
58:3):
59:3): bool operator! () const { return is_empty(); }
60:3):
61:3): uint hash_code() const;
62:3):
0(paf 63:3): const char* cstr() const { return CORD_to_const_char_star(body); }
1(paf 64:3): char* cstrm() const { return CORD_to_char_star(body); }
0(paf 65:3):
7(paf 66:3): size_t length() const { return CORD_len(body); }
67:3):
68:3): bool is_empty() const { return body==CORD_EMPTY; }
69:3):
6(paf 70:3): void append_know_length(const char *str, size_t known_length) {
71:3): body=CORD_cat_char_star(body, str, known_length);
7(paf 72:3): }
6(paf 73:3): void append_strdup_know_length(const char* str, size_t known_length) {
74:3): append_know_length(pa_strdup(str, known_length), known_length);
1(paf 75:3): }
2(paf 76:3): void append(char c) { body=CORD_cat_char(body, c); }
6(paf 77:3): StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
78:3): StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
7(paf 79:3):
80:3): // could not figure out why this operator is needed [should do this chain: string->simple->==]
8(paf 81:3): bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
82:3): bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
83:3): bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
84:3): bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; }
85:3): bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; }
86:3): bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; }
7(paf 87:3):
88:3): int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
89:3): return CORD_ncmp(body, x_begin, y.body, y_begin, size);
90:3): }
91:3):
92:3): char fetch(size_t index) const { return CORD_fetch(body, index); }
1(paf 93:3): StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
3(paf 94:3): size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
1(paf 95:3): size_t pos(const StringBody substr, size_t offset=0) const {
96:3): if(!substr.length())
97:3): return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
98:3): return CORD_str(body, offset, substr.body);
99:3): }
2(paf 100:3): size_t pos(char c,
101:3): size_t offset=0) const {
102:3): return CORD_chr(body, offset, c);
103:3): }
1(paf 104:3):
105:3): template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
106:3): CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
107:3): }
108:3):
109:3): void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
7(paf 110:3):
5(paf 111:3): StringBody normalize() const {
112:3): return StringBody(CORD_balance(body));
113:3): }
114:3):
115:3): void dump() const {
116:3): CORD_dump(body);
117:3): }
7(paf 118:3): };
(paf 119:):
1.144.2.2 paf 120: /**
121: String which knows the language of all it's fragments.
1.41 paf 122:
123: All pieces remember
124: - whether they are tainted or not,
125: and the language which should be used to detaint them
126: */
1.144.2.28.2. (paf 127:): class String: public PA_Object {
128:):
2(paf 129:3): // friend class StringBody;
7(paf 130:3):
1.144.2.2 paf 131: public:
1.1 paf 132:
1.106 parser 133: /** piece is tainted or not. the language to use when detaint
134: remember to change String_Untaint_lang_name @ untaint.C along
135: */
1.144.2.28.2. 7(paf 136:3): enum Language {
137:3): L_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum
138:3): // these two must go before others, there are checks for >L_AS_IS
139:3): L_CLEAN, ///< clean
140:3): L_AS_IS, ///< leave all characters intact
1.122 paf 141:
1.144.2.28.2. 7(paf 142:3): L_PASS_APPENDED,
1.41 paf 143: /**<
144: leave language built into string being appended.
145: just a flag, that value not stored
146: */
1.144.2.28.2. 7(paf 147:3): L_TAINTED, ///< tainted, untaint language as assigned later
1.122 paf 148: // untaint languages. assigned by ^untaint[lang]{...}
1.144.2.28.2. 7(paf 149:3): L_FILE_SPEC, ///< file specification
150:3): L_HTTP_HEADER, ///< text in HTTP response header
151:3): L_MAIL_HEADER, ///< text in mail header
152:3): L_URI, ///< text in uri
153:3): L_TABLE, ///< ^table:set body
154:3): L_SQL, ///< ^table:sql body
155:3): L_JS, ///< JavaScript code
156:3): L_XML, ///< ^dom:set xml
157:3): L_HTML, ///< HTML code (for editing)
158:3): L_OPTIMIZE_BIT = 0x8000 ///< flag, requiring cstr whitespace optimization
1.27 paf 159: };
160:
1.144.2.28.2. (paf 161:): struct Fragment {
7(paf 162:3): Language lang; ///< untaint flag, later untaint language
0(paf 163:3): size_t length; ///< length
9(paf 164:3): Fragment(Language alang, size_t asize): lang(alang), length(asize) {
3(paf 165:3): assert(alang!=L_UNSPECIFIED);
166:3): assert(asize!=(size_t)-1);
9(paf 167:3): }
(paf 168:): };
169:):
170:): class ArrayFragment: public Array<Fragment> {
6(paf 171:3): void append(element_type src) {
(paf 172:): *static_cast<Array<Fragment> *>(this)+=src;
173:): }
6(paf 174:3): /// hiding from accidental USE, use append_positions
175:3): void append(const ArrayFragment& src, int offset, int limit) {
176:3): static_cast<Array<Fragment> *>(this)->append(src, offset, limit);
177:3): }
(paf 178:): public:
179:): ArrayFragment& operator += (element_type src) {
8(paf 180:3): if(size_t lcount=count()) { // not empty?
(paf 181:): // try to join with last
4(paf 182:3): Fragment& last=get_ref(lcount-1);
(paf 183:): if(last.lang==src.lang) {
0(paf 184:3): last.length+=src.length;
(paf 185:): return *this;
186:): }
187:): }
6(paf 188:3): append(src);
(paf 189:): return *this;
190:): }
6(paf 191:3): void append(const ArrayFragment& src) { append(src, 0, 0); }
192:3): void append_positions(const ArrayFragment& src, size_t substr_begin, size_t substr_end);
7(paf 193:3):
194:3): size_t length() {
195:3): size_t result=0;
196:3): for(Array_iterator<element_type> i(*this); i.has_next(); ) {
197:3): const Fragment fragment=i.next();
198:3): result+=fragment.length;
199:3): }
200:3): return result;
201:3): }
(paf 202:): };
203:):
204:): struct C {
7(paf 205:3): const char *str;
206:3): size_t length;
207:3): operator const char *() { return str; }
208:3): C(const char *astr, size_t asize): str(astr), length(asize) {}
209:3): };
210:3):
211:3): struct Cm {
(paf 212:): char *str;
0(paf 213:3): size_t length;
7(paf 214:3): //operator char *() { return str; }
215:3): Cm(char *astr, size_t asize): str(astr), length(asize) {}
(paf 216:): };
217:):
218:): private:
219:):
7(paf 220:3): StringBody body; ///< all characters of string
0(paf 221:3): ArrayFragment fragments; ///< fragment language+length info
(paf 222:):
1.8 paf 223: public:
224:
1.144.2.28.2. 6(paf 225:3): explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
8(paf 226:3): explicit String(const C cstr, bool tainted=false);
1.14 paf 227: String(const String& src);
1.144.2.28.2. 7(paf 228:3): String(StringBody abody, Language alang): body(abody) {
229:3): fragments+=Fragment(alang, abody.length());
230:3): }
8(paf 231:3):
232:3): #define ASSERT_STRING_INVARIANT(string) \
233:3): assert((string).body.length()==(string).fragments.length())
7(paf 234:3):
235:3): /// for convinient hash lookup
2(paf 236:3): operator const StringBody() const { return body; }
7(paf 237:3):
238:3): bool is_empty() const { return body.is_empty(); }
239:3): size_t length() const { return body.length(); }
(paf 240:):
2(paf 241:3): /// convert to CORD. if 'lang' known, forcing 'lang' to it
1(paf 242:3): StringBody cstr_to_string_body(Language lang=L_AS_IS,
7(paf 243:3): SQL_Connection* connection=0,
(paf 244:): const Request_charsets *charsets=0) const;
245:):
2(paf 246:3): /// convert to constant C string. if 'lang' known, forcing 'lang' to it
7(paf 247:3): const char* cstr(Language lang=L_AS_IS,
248:3): SQL_Connection* connection=0,
1.144.2.14 paf 249: const Request_charsets *charsets=0) const {
1.144.2.28.2. 1(paf 250:3): return cstr_to_string_body(lang, connection, charsets).cstr();
(paf 251:): }
2(paf 252:3): /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
7(paf 253:3): char *cstrm(Language lang=L_AS_IS,
254:3): SQL_Connection* connection=0,
(paf 255:): const Request_charsets *charsets=0) const {
1(paf 256:3): return cstr_to_string_body(lang, connection, charsets).cstrm();
1.144.2.9 paf 257: }
1.108 parser 258: /// puts pieces to buf
1.144.2.28.2. 7(paf 259:3): Cm serialize(size_t prolog_size) const;
1.108 parser 260: /// appends pieces from buf to self
1.144.2.28.2. 4(paf 261:3): bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
7(paf 262:3): /// @see StringBody::append_know_length
263:3): String& append_know_length(const char* str, size_t known_length, Language lang);
1(paf 264:3): /// @see StringBody::append_help_length
265:3): String& append_help_length(const char* str, size_t helper_length, Language lang);
6(paf 266:3): String& append_strdup(const char* str, size_t helper_length, Language lang);
(paf 267:):
7(paf 268:3): bool operator == (const char* y) const { return body==StringBody(y); }
0(paf 269:3): bool operator != (const char* y) const { return body!=StringBody(y); }
(paf 270:):
271:): /// this starts with y
7(paf 272:3): bool starts_with(const char* y) const {
273:3): return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
1.80 paf 274: }
1.144.2.28.2. 9(paf 275:3): /// x starts with this
276:3): bool this_starts(const char* x) const {
277:3): return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
(paf 278:): }
1.80 paf 279:
1.144.2.28.2. 7(paf 280:3): String& append_to(String& dest, Language lang, bool forced) const;
281:3): String& append(const String& src, Language lang, bool forced=false) {
1.144.2.11 paf 282: return src.append_to(*this, lang, forced);
283: }
1.144.2.28.2. 7(paf 284:3): String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
1(paf 285:3): String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
8(paf 286:3): String& operator << (const StringBody src) {
6(paf 287:3): body<<src;
9(paf 288:3): fragments+=Fragment(L_AS_IS, src.length());
289:3): return *this;
290:3): }
1.100 parser 291:
1.142 paf 292: /// extracts first char of a string, if any
293: char first_char() const {
1.144.2.28.2. 7(paf 294:3): return is_empty()?0:body.fetch(0);
1.142 paf 295: }
1.144.2.28.2. 8(paf 296:3):
297:3): bool operator < (const String& src) const { return body<src.body; }
298:3): bool operator > (const String& src) const { return body>src.body; }
299:3): bool operator <= (const String& src) const { return body<=src.body; }
300:3): bool operator >= (const String& src) const { return body>=src.body; }
301:3): bool operator != (const String& src) const { return body!=src.body; }
302:3): bool operator == (const String& src) const { return body==src.body; }
1.54 paf 303:
304: /// extracts [start, finish) piece of string
1.144.2.28.2. 2(paf 305:3): String& mid(size_t substr_begin, size_t substr_end) const;
(paf 306:):
307:): /**
7(paf 308:3): ignore lang if it's L_UNSPECIFIED
(paf 309:): but when specified: look for substring that lies in ONE fragment in THAT lang
310:): @return position of substr in string, -1 means "not found" [const char* version]
311:): */
1(paf 312:3): size_t pos(const StringBody substr,
7(paf 313:3): size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
314:3): /// String version of @see pos(const char*, int, Language)
(paf 315:): size_t pos(const String& substr,
7(paf 316:3): size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
2(paf 317:3): size_t pos(char c,
318:3): size_t this_offset=0) const {
319:3): return body.pos(c, this_offset);
320:3): }
1.62 paf 321:
1.144.2.22 paf 322: void split(ArrayString& result,
1.144.2.28.2. 2(paf 323:3): size_t& pos_after,
(paf 324:): const char* delim,
2(paf 325:3): Language lang=L_UNSPECIFIED, int limit=-1) const;
1.144.2.22 paf 326: void split(ArrayString& result,
1.144.2.28.2. 5(paf 327:3): size_t& pos_after,
1.62 paf 328: const String& delim,
1.144.2.28.2. 7(paf 329:3): Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 330:
1.144.2.28.2. 6(paf 331:3): typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 332: int prestart, int prefinish,
333: int poststart, int postfinish,
1.144.2.9 paf 334: void *info);
1.87 parser 335: /**
1.144.2.5 paf 336: @return table of found items, if any.
1.87 parser 337: table format is defined and fixed[can be used by others]:
338: @verbatim
339: prematch/match/postmatch/1/2/3/...
340: @endverbatim
341: */
1.144.2.28.2. (paf 342:): Table* match(Charset& source_charset,
1.64 paf 343: const String& regexp,
1.144.2.28.2. 5(paf 344:3): const String* options,
1.99 parser 345: Row_action row_action, void *info,
1.144.2.27 paf 346: bool& just_matched) const;
1.87 parser 347: enum Change_case_kind {
348: CC_UPPER,
349: CC_LOWER
350: };
1.144.2.28.2. (paf 351:): String& change_case(Charset& source_charset,
5(paf 352:3): Change_case_kind kind) const;
(paf 353:): const String& replace(const Dictionary& dict) const;
354:): double as_double() const;
355:): int as_int() const;
1.137 paf 356:
1.7 paf 357: private: //disabled
358:
1.12 paf 359: String& operator = (const String&) { return *this; }
1.7 paf 360:
1.1 paf 361: };
1.144.2.7 paf 362:
363: /// simple hash code of string. used by Hash
1.144.2.28.2. 8(paf 364:3): inline uint hash_code(const StringBody self) {
1.144.2.7 paf 365: return self.hash_code();
366: }
1.1 paf 367:
368: #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