Annotation of parser3/src/include/pa_string.h, revision 1.145.4.3
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.145 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:
1.145.4.3! paf 11: static const char* IDENT_STRING_H="$Date: 2003/09/23 10:37:05 $";
1.145 paf 12:
13: // includes
1.1 paf 14:
1.4 paf 15: #include "pa_types.h"
1.145 paf 16: #include "pa_array.h"
17:
18: extern "C" { // cord's author forgot to do that
19: #define CORD_NO_IO
20: #include "cord.h"
21: };
1.4 paf 22:
1.145.4.3! paf 23: /// must use simple version, without optimization for short pieces
! 24: /// it's much more convinient for iterating. we rely on this here: CORD_block_iter
! 25: #define CORD_chars(c, i) CORD_chars_block((c), (i))
! 26: extern "C" CORD CORD_chars_block(char c, size_t i);
1.145.4.1 paf 27:
28: // cord extension
29: /* Returns true if x does contain */
30: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
31: int CORD_nnchr(CORD x, size_t i, size_t n, int not_c);
32:
1.145 paf 33: // forwards
1.9 paf 34:
1.145 paf 35: class Charset;
1.135 paf 36: class Table;
1.71 paf 37: class SQL_Connection;
1.101 parser 38: class Dictionary;
1.145 paf 39: class Request_charsets;
40: class String;
41: typedef Array<const String*> ArrayString;
42:
43: /// this is result of pos functions which mean that substr were not found
44: #define STRING_NOT_FOUND ((size_t)-1)
45:
46: class StringBody {
47:
48: CORD body;
49:
50: public:
51:
52: StringBody(): body(CORD_EMPTY) {}
53: StringBody(CORD abody): body(abody) {
54: assert(!body // no body
55: || *body // ordinary string
56: || body[1]==1 // CONCAT_HDR
57: || body[1]==4 // FN_HDR
58: || body[1]==6 // SUBSTR_HDR
59: );
60: }
61: /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
62: StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
63: append_know_length(str, helper_length?helper_length:strlen(str));
64: }
65: static StringBody Format(int value);
66:
67: void clear() { body=CORD_EMPTY; }
68:
69: bool operator! () const { return is_empty(); }
70:
71: uint hash_code() const;
72:
73: const char* cstr() const { return CORD_to_const_char_star(body); }
74: char* cstrm() const { return CORD_to_char_star(body); }
75:
76: size_t length() const { return CORD_len(body); }
77:
78: bool is_empty() const { return body==CORD_EMPTY; }
79:
80: void append_know_length(const char *str, size_t known_length) {
81: if(known_length)
82: body=CORD_cat_char_star(body, str, known_length);
83: }
84: void append_strdup_know_length(const char* str, size_t known_length) {
85: if(known_length)
86: append_know_length(pa_strdup(str, known_length), known_length);
87: }
88: void append(char c) { body=CORD_cat_char(body, c); }
89: StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
90: StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
91:
92: // could not figure out why this operator is needed [should do this chain: string->simple->==]
93: bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
94: bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
95: bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
96: bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; }
97: bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; }
98: bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; }
99:
100: int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
101: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
102: }
103:
104: char fetch(size_t index) const { return CORD_fetch(body, index); }
105: StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
106: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
107: size_t pos(const StringBody substr, size_t offset=0) const {
108: if(!substr.length())
109: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
110: return CORD_str(body, offset, substr.body);
111: }
112: size_t pos(char c,
113: size_t offset=0) const {
114: return CORD_chr(body, offset, c);
115: }
116:
1.145.4.1 paf 117: /* template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
1.145 paf 118: CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
1.145.4.1 paf 119: }*/
1.145 paf 120:
121: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
122:
1.145.4.2 paf 123: /*StringBody normalize() const {
1.145 paf 124: return StringBody(CORD_balance(body));
1.145.4.2 paf 125: }*/
1.145 paf 126:
127: void dump() const {
128: CORD_dump(body);
129: }
130: };
1.62 paf 131:
1.42 paf 132: /**
1.145.4.1 paf 133: String which knows the lang of all it's langs.
1.41 paf 134:
135: All pieces remember
136: - whether they are tainted or not,
1.145.4.1 paf 137: and the lang which should be used to detaint them
1.41 paf 138: */
1.145 paf 139: class String: public PA_Object {
1.1 paf 140: public:
1.48 paf 141:
1.145.4.1 paf 142: /** piece is tainted or not. the lang to use when detaint
1.106 parser 143: remember to change String_Untaint_lang_name @ untaint.C along
144: */
1.145 paf 145: enum Language {
1.145.4.1 paf 146: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 147: // these two must go before others, there are checks for >L_AS_IS
148: L_CLEAN, ///< clean
149: L_AS_IS, ///< leave all characters intact
1.122 paf 150:
1.145 paf 151: L_PASS_APPENDED,
1.41 paf 152: /**<
1.145.4.1 paf 153: leave lang built into string being appended.
1.41 paf 154: just a flag, that value not stored
155: */
1.145.4.1 paf 156: L_TAINTED, ///< tainted, untaint lang as assigned later
157: // untaint langs. assigned by ^untaint[lang]{...}
1.145 paf 158: L_FILE_SPEC, ///< file specification
159: L_HTTP_HEADER, ///< text in HTTP response header
160: L_MAIL_HEADER, ///< text in mail header
161: L_URI, ///< text in uri
162: L_TABLE, ///< ^table:set body
163: L_SQL, ///< ^table:sql body
164: L_JS, ///< JavaScript code
165: L_XML, ///< ^dom:set xml
166: L_HTML, ///< HTML code (for editing)
1.145.4.1 paf 167: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 168: };
169:
1.145.4.1 paf 170: class Languages {
171:
172: union {
173: struct {
174: Language lang:8;
175: int is_not_just_lang:16-8;
176: };
177: CORD langs;
178: };
179:
180: CORD make_langs(const StringBody body) const {
181: return is_not_just_lang?
182: langs
1.145.4.3! paf 183: :CORD_chars((char)lang, body.length());
1.145 paf 184: }
185:
1.145.4.1 paf 186: CORD make_langs(size_t aoffset, size_t alength) const {
187: return is_not_just_lang?
188: CORD_substr(langs, aoffset, alength)
1.145.4.3! paf 189: :CORD_chars((char)lang, alength);
1.145 paf 190: }
1.145.4.1 paf 191:
1.145.4.3! paf 192: /// appending when 'langs' already contain something [simple cases hanled elsewhere]
! 193: void append(const StringBody current_body,
1.145.4.1 paf 194: const CORD to_nonempty_target_langs) {
1.145.4.3! paf 195: assert(langs);
1.145.4.1 paf 196:
197: if(is_not_just_lang)
198: langs=CORD_cat(langs, to_nonempty_target_langs);
199: else { // we were "just lang"
200: size_t current_size=current_body.length();
201: assert(current_size);
202: langs=CORD_cat(
1.145.4.3! paf 203: CORD_chars((char)lang, current_size), // first piece [making from just 'lang']
1.145.4.1 paf 204: to_nonempty_target_langs); // new piece
205: }
1.145 paf 206: }
1.145.4.1 paf 207:
1.145 paf 208: public:
1.145.4.1 paf 209:
210: Languages(): langs(0) {}
211: Languages(Language alang): lang(alang), is_not_just_lang(0) {}
212:
213: /// MUST be called prior to modification of current_body [uses it's original length]
214: void append(const StringBody current_body, Language alang, size_t asize) {
215: assert(alang);
216: assert(asize);
217:
1.145.4.3! paf 218: if(!is_not_just_lang)
! 219: if(lang) {
! 220: if(lang==alang) // same length? ignoring
! 221: return;
! 222: } else {
! 223: lang=alang; // to uninitialized
! 224: return;
! 225: }
1.145.4.1 paf 226:
1.145.4.3! paf 227: append(current_body, CORD_chars((char)alang, asize));
1.145 paf 228: }
229:
1.145.4.1 paf 230: /// MUST be called prior to modification of current_body [uses it's original length]
231: void append(const StringBody current_body, const StringBody appending_body,
232: const Languages src) {
233: assert(appending_body.length());
234:
1.145.4.3! paf 235: if(!langs)
! 236: langs=src.langs; // to uninitialized
! 237: else if(!src.is_not_just_lang)
! 238: append(current_body, src.lang, appending_body.length()); // simplifying when simple source
! 239: else
! 240: append(current_body, src.make_langs(appending_body));
1.145.4.1 paf 241: }
242:
243: /// MUST be called prior to modification of current_body [uses it's original length]
244: void append(const StringBody current_body,
245: const Languages src, size_t aoffset, size_t alength) {
246: assert(alength);
247:
1.145.4.3! paf 248: if(!langs) // to uninitialized?
1.145.4.1 paf 249: if(src.is_not_just_lang)
1.145.4.3! paf 250: langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex
! 251: else
! 252: lang=src.lang; // to uninitialized simple
! 253: else
! 254: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang) // both simple & of same language?
! 255: return; // ignoring
1.145.4.1 paf 256: else
1.145.4.3! paf 257: append(current_body, src.make_langs(aoffset, alength));
1.145.4.1 paf 258: }
259:
260: /// checks if we have alang all from aoffset to aoffset+alength
261: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
262: if(alang==L_UNSPECIFIED) // ignore lang?
263: return true;
264:
265: if(is_not_just_lang)
266: return CORD_nnchr(langs, aoffset, alength, (unsigned)alang)!=0;
267: else
268: return lang==alang;
269: }
270:
271: template<typename I> void for_each(const StringBody current_body,
1.145.4.2 paf 272: int callback(char, size_t, I), I info) const {
1.145.4.1 paf 273:
274: if(is_not_just_lang)
1.145.4.2 paf 275: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
1.145.4.1 paf 276: else
1.145.4.2 paf 277: callback(lang, current_body.length(), info);
1.145 paf 278: }
279: };
280:
281: struct C {
282: const char *str;
283: size_t length;
284: operator const char *() { return str; }
285: C(const char *astr, size_t asize): str(astr), length(asize) {}
286: };
287:
288: struct Cm {
289: char *str;
290: size_t length;
291: //operator char *() { return str; }
292: Cm(char *astr, size_t asize): str(astr), length(asize) {}
293: };
294:
295: private:
296:
1.145.4.1 paf 297: Languages langs; ///< string characters lang
1.145 paf 298: StringBody body; ///< all characters of string
299:
1.8 paf 300: public:
301:
1.145 paf 302: explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
303: explicit String(const C cstr, bool tainted=false);
1.145.4.1 paf 304: String(StringBody abody, Language alang): body(abody), langs(alang) {}
305: String(const String& src): body(src.body), langs(src.langs) {}
1.145 paf 306:
307: /// for convinient hash lookup
308: operator const StringBody() const { return body; }
309:
310: bool is_empty() const { return body.is_empty(); }
311: size_t length() const { return body.length(); }
312:
313: /// convert to CORD. if 'lang' known, forcing 'lang' to it
314: StringBody cstr_to_string_body(Language lang=L_AS_IS,
315: SQL_Connection* connection=0,
316: const Request_charsets *charsets=0) const;
317:
318: /// convert to constant C string. if 'lang' known, forcing 'lang' to it
319: const char* cstr(Language lang=L_AS_IS,
320: SQL_Connection* connection=0,
321: const Request_charsets *charsets=0) const {
322: return cstr_to_string_body(lang, connection, charsets).cstr();
323: }
324: /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
325: char *cstrm(Language lang=L_AS_IS,
326: SQL_Connection* connection=0,
327: const Request_charsets *charsets=0) const {
328: return cstr_to_string_body(lang, connection, charsets).cstrm();
1.50 paf 329: }
1.108 parser 330: /// puts pieces to buf
1.145 paf 331: Cm serialize(size_t prolog_size) const;
1.108 parser 332: /// appends pieces from buf to self
1.145 paf 333: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
334: /// @see StringBody::append_know_length
335: String& append_know_length(const char* str, size_t known_length, Language lang);
336: /// @see StringBody::append_help_length
337: String& append_help_length(const char* str, size_t helper_length, Language lang);
338: String& append_strdup(const char* str, size_t helper_length, Language lang);
339:
340: bool operator == (const char* y) const { return body==StringBody(y); }
341: bool operator != (const char* y) const { return body!=StringBody(y); }
342:
343: /// this starts with y
344: bool starts_with(const char* y) const {
345: return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
346: }
347: /// x starts with this
348: bool this_starts(const char* x) const {
349: return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 350: }
351:
1.145 paf 352: String& append_to(String& dest, Language lang, bool forced) const;
353: String& append(const String& src, Language lang, bool forced=false) {
354: return src.append_to(*this, lang, forced);
355: }
356: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
357: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
358: String& operator << (const StringBody src) {
1.145.4.1 paf 359: langs.append(body, L_AS_IS, src.length());
1.145 paf 360: body<<src;
361: return *this;
362: }
1.100 parser 363:
1.142 paf 364: /// extracts first char of a string, if any
365: char first_char() const {
1.145 paf 366: return is_empty()?0:body.fetch(0);
1.142 paf 367: }
1.54 paf 368:
1.145 paf 369: bool operator < (const String& src) const { return body<src.body; }
370: bool operator > (const String& src) const { return body>src.body; }
371: bool operator <= (const String& src) const { return body<=src.body; }
372: bool operator >= (const String& src) const { return body>=src.body; }
373: bool operator != (const String& src) const { return body!=src.body; }
374: bool operator == (const String& src) const { return body==src.body; }
375:
1.54 paf 376: /// extracts [start, finish) piece of string
1.145 paf 377: String& mid(size_t substr_begin, size_t substr_end) const;
378:
379: /**
380: ignore lang if it's L_UNSPECIFIED
381: but when specified: look for substring that lies in ONE fragment in THAT lang
382: @return position of substr in string, -1 means "not found" [const char* version]
383: */
384: size_t pos(const StringBody substr,
385: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
386: /// String version of @see pos(const char*, int, Language)
387: size_t pos(const String& substr,
388: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
389: size_t pos(char c,
390: size_t this_offset=0) const {
391: return body.pos(c, this_offset);
392: }
1.55 paf 393:
1.145 paf 394: void split(ArrayString& result,
395: size_t& pos_after,
396: const char* delim,
397: Language lang=L_UNSPECIFIED, int limit=-1) const;
398: void split(ArrayString& result,
399: size_t& pos_after,
1.62 paf 400: const String& delim,
1.145 paf 401: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 402:
1.145 paf 403: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 404: int prestart, int prefinish,
405: int poststart, int postfinish,
1.68 paf 406: void *info);
1.87 parser 407: /**
1.145 paf 408: @return table of found items, if any.
1.87 parser 409: table format is defined and fixed[can be used by others]:
410: @verbatim
411: prematch/match/postmatch/1/2/3/...
412: @endverbatim
413: */
1.145 paf 414: Table* match(Charset& source_charset,
1.64 paf 415: const String& regexp,
1.145 paf 416: const String* options,
1.99 parser 417: Row_action row_action, void *info,
1.145 paf 418: bool& just_matched) const;
1.87 parser 419: enum Change_case_kind {
420: CC_UPPER,
421: CC_LOWER
422: };
1.145 paf 423: String& change_case(Charset& source_charset,
1.87 parser 424: Change_case_kind kind) const;
1.145 paf 425: const String& replace(const Dictionary& dict) const;
1.96 parser 426: double as_double() const;
427: int as_int() const;
1.137 paf 428:
1.7 paf 429: private: //disabled
430:
1.12 paf 431: String& operator = (const String&) { return *this; }
1.7 paf 432:
1.1 paf 433: };
1.119 paf 434:
1.145 paf 435: /// simple hash code of string. used by Hash
436: inline uint hash_code(const StringBody self) {
437: return self.hash_code();
1.119 paf 438: }
1.1 paf 439:
440: #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