Annotation of parser3/src/include/pa_string.h, revision 1.145.4.1
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:
1.145.4.1! paf 11: static const char* IDENT_STRING_H="$Date: 2003/07/24 11:31:21 $";
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:
1.145.4.1! paf 23: CORD CORD_chars_gen(char c, size_t i);
! 24:
! 25: // cord extension
! 26: /* Returns true if x does contain */
! 27: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
! 28: int CORD_nnchr(CORD x, size_t i, size_t n, int not_c);
! 29:
1.145 paf 30: // forwards
1.9 paf 31:
1.145 paf 32: class Charset;
1.135 paf 33: class Table;
1.71 paf 34: class SQL_Connection;
1.101 parser 35: class Dictionary;
1.145 paf 36: class Request_charsets;
37: class String;
38: typedef Array<const String*> ArrayString;
39:
40: /// this is result of pos functions which mean that substr were not found
41: #define STRING_NOT_FOUND ((size_t)-1)
42:
43: class StringBody {
44:
45: CORD body;
46:
47: public:
48:
49: StringBody(): body(CORD_EMPTY) {}
50: StringBody(CORD abody): body(abody) {
51: assert(!body // no body
52: || *body // ordinary string
53: || body[1]==1 // CONCAT_HDR
54: || body[1]==4 // FN_HDR
55: || body[1]==6 // SUBSTR_HDR
56: );
57: }
58: /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
59: StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
60: append_know_length(str, helper_length?helper_length:strlen(str));
61: }
62: static StringBody Format(int value);
63:
64: void clear() { body=CORD_EMPTY; }
65:
66: bool operator! () const { return is_empty(); }
67:
68: uint hash_code() const;
69:
70: const char* cstr() const { return CORD_to_const_char_star(body); }
71: char* cstrm() const { return CORD_to_char_star(body); }
72:
73: size_t length() const { return CORD_len(body); }
74:
75: bool is_empty() const { return body==CORD_EMPTY; }
76:
77: void append_know_length(const char *str, size_t known_length) {
78: if(known_length)
79: body=CORD_cat_char_star(body, str, known_length);
80: }
81: void append_strdup_know_length(const char* str, size_t known_length) {
82: if(known_length)
83: append_know_length(pa_strdup(str, known_length), known_length);
84: }
85: void append(char c) { body=CORD_cat_char(body, c); }
86: StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
87: StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
88:
89: // could not figure out why this operator is needed [should do this chain: string->simple->==]
90: bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
91: bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
92: bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
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:
97: int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
98: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
99: }
100:
101: char fetch(size_t index) const { return CORD_fetch(body, index); }
102: StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
103: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
104: size_t pos(const StringBody substr, size_t offset=0) const {
105: if(!substr.length())
106: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
107: return CORD_str(body, offset, substr.body);
108: }
109: size_t pos(char c,
110: size_t offset=0) const {
111: return CORD_chr(body, offset, c);
112: }
113:
1.145.4.1! paf 114: /* template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
1.145 paf 115: CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
1.145.4.1! paf 116: }*/
1.145 paf 117:
118: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
119:
120: StringBody normalize() const {
121: return StringBody(CORD_balance(body));
122: }
123:
124: void dump() const {
125: CORD_dump(body);
126: }
127: };
1.62 paf 128:
1.42 paf 129: /**
1.145.4.1! paf 130: String which knows the lang of all it's langs.
1.41 paf 131:
132: All pieces remember
133: - whether they are tainted or not,
1.145.4.1! paf 134: and the lang which should be used to detaint them
1.41 paf 135: */
1.145 paf 136: class String: public PA_Object {
137:
138: // friend class StringBody;
139:
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
! 183: :CORD_chars_gen((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)
! 189: :CORD_chars_gen((char)lang, alength);
1.145 paf 190: }
1.145.4.1! paf 191:
! 192: /// @returns true if appended, false if not:
! 193: /// when appending can be simplified by just assigning lang/langs
! 194: bool append(const StringBody current_body,
! 195: const CORD to_nonempty_target_langs) {
! 196: if(!lang)
! 197: return false;
! 198:
! 199: if(is_not_just_lang)
! 200: langs=CORD_cat(langs, to_nonempty_target_langs);
! 201: else { // we were "just lang"
! 202: size_t current_size=current_body.length();
! 203: assert(current_size);
! 204: langs=CORD_cat(
! 205: CORD_chars_gen((char)lang, current_size), // first piece [making from just 'lang']
! 206: to_nonempty_target_langs); // new piece
! 207: }
! 208:
! 209: return true;
1.145 paf 210: }
1.145.4.1! paf 211:
1.145 paf 212: public:
1.145.4.1! paf 213:
! 214: Languages(): langs(0) {}
! 215: Languages(Language alang): lang(alang), is_not_just_lang(0) {}
! 216:
! 217: /// MUST be called prior to modification of current_body [uses it's original length]
! 218: void append(const StringBody current_body, Language alang, size_t asize) {
! 219: assert(alang);
! 220: assert(asize);
! 221:
! 222: if(!is_not_just_lang && lang==alang)
! 223: return;
! 224:
! 225: if(!append(current_body, CORD_chars((char)alang, asize)))
! 226: lang=alang;
1.145 paf 227: }
228:
1.145.4.1! paf 229: /// MUST be called prior to modification of current_body [uses it's original length]
! 230: void append(const StringBody current_body, const StringBody appending_body,
! 231: const Languages src) {
! 232: assert(appending_body.length());
! 233:
! 234: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang)
! 235: return;
! 236:
! 237: if(!append(current_body, src.make_langs(appending_body)))
! 238: langs=src.langs;
! 239: }
! 240:
! 241: /// MUST be called prior to modification of current_body [uses it's original length]
! 242: void append(const StringBody current_body,
! 243: const Languages src, size_t aoffset, size_t alength) {
! 244: assert(alength);
! 245: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang)
! 246: return;
! 247:
! 248: if(!append(current_body, src.make_langs(aoffset, alength)))
! 249: if(src.is_not_just_lang)
! 250: langs=CORD_substr(src.langs, aoffset, alength);
! 251: else
! 252: lang=src.lang;
! 253: }
! 254:
! 255: /// checks if we have alang all from aoffset to aoffset+alength
! 256: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
! 257: if(alang==L_UNSPECIFIED) // ignore lang?
! 258: return true;
! 259:
! 260: if(is_not_just_lang)
! 261: return CORD_nnchr(langs, aoffset, alength, (unsigned)alang)!=0;
! 262: else
! 263: return lang==alang;
! 264: }
! 265:
! 266: template<typename I> void for_each(const StringBody current_body,
! 267: int (*callback)(char, size_t, I&), I& info) const {
! 268:
! 269: if(is_not_just_lang)
! 270: CORD_block_iter(langs, 0, (CORD_batched_iter_fn)callback, &info) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/);
! 271: else
! 272: callback(lang, current_body.length());
1.145 paf 273: }
274: };
275:
276: struct C {
277: const char *str;
278: size_t length;
279: operator const char *() { return str; }
280: C(const char *astr, size_t asize): str(astr), length(asize) {}
281: };
282:
283: struct Cm {
284: char *str;
285: size_t length;
286: //operator char *() { return str; }
287: Cm(char *astr, size_t asize): str(astr), length(asize) {}
288: };
289:
290: private:
291:
1.145.4.1! paf 292: Languages langs; ///< string characters lang
1.145 paf 293: StringBody body; ///< all characters of string
294:
1.8 paf 295: public:
296:
1.145 paf 297: explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
298: explicit String(const C cstr, bool tainted=false);
1.145.4.1! paf 299: String(StringBody abody, Language alang): body(abody), langs(alang) {}
! 300: String(const String& src): body(src.body), langs(src.langs) {}
1.145 paf 301:
302: /// for convinient hash lookup
303: operator const StringBody() const { return body; }
304:
305: bool is_empty() const { return body.is_empty(); }
306: size_t length() const { return body.length(); }
307:
308: /// convert to CORD. if 'lang' known, forcing 'lang' to it
309: StringBody cstr_to_string_body(Language lang=L_AS_IS,
310: SQL_Connection* connection=0,
311: const Request_charsets *charsets=0) const;
312:
313: /// convert to constant C string. if 'lang' known, forcing 'lang' to it
314: const char* cstr(Language lang=L_AS_IS,
315: SQL_Connection* connection=0,
316: const Request_charsets *charsets=0) const {
317: return cstr_to_string_body(lang, connection, charsets).cstr();
318: }
319: /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
320: char *cstrm(Language lang=L_AS_IS,
321: SQL_Connection* connection=0,
322: const Request_charsets *charsets=0) const {
323: return cstr_to_string_body(lang, connection, charsets).cstrm();
1.50 paf 324: }
1.108 parser 325: /// puts pieces to buf
1.145 paf 326: Cm serialize(size_t prolog_size) const;
1.108 parser 327: /// appends pieces from buf to self
1.145 paf 328: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
329: /// @see StringBody::append_know_length
330: String& append_know_length(const char* str, size_t known_length, Language lang);
331: /// @see StringBody::append_help_length
332: String& append_help_length(const char* str, size_t helper_length, Language lang);
333: String& append_strdup(const char* str, size_t helper_length, Language lang);
334:
335: bool operator == (const char* y) const { return body==StringBody(y); }
336: bool operator != (const char* y) const { return body!=StringBody(y); }
337:
338: /// this starts with y
339: bool starts_with(const char* y) const {
340: return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
341: }
342: /// x starts with this
343: bool this_starts(const char* x) const {
344: return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 345: }
346:
1.145 paf 347: String& append_to(String& dest, Language lang, bool forced) const;
348: String& append(const String& src, Language lang, bool forced=false) {
349: return src.append_to(*this, lang, forced);
350: }
351: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
352: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
353: String& operator << (const StringBody src) {
1.145.4.1! paf 354: langs.append(body, L_AS_IS, src.length());
1.145 paf 355: body<<src;
356: return *this;
357: }
1.100 parser 358:
1.142 paf 359: /// extracts first char of a string, if any
360: char first_char() const {
1.145 paf 361: return is_empty()?0:body.fetch(0);
1.142 paf 362: }
1.54 paf 363:
1.145 paf 364: bool operator < (const String& src) const { return body<src.body; }
365: bool operator > (const String& src) const { return body>src.body; }
366: bool operator <= (const String& src) const { return body<=src.body; }
367: bool operator >= (const String& src) const { return body>=src.body; }
368: bool operator != (const String& src) const { return body!=src.body; }
369: bool operator == (const String& src) const { return body==src.body; }
370:
1.54 paf 371: /// extracts [start, finish) piece of string
1.145 paf 372: String& mid(size_t substr_begin, size_t substr_end) const;
373:
374: /**
375: ignore lang if it's L_UNSPECIFIED
376: but when specified: look for substring that lies in ONE fragment in THAT lang
377: @return position of substr in string, -1 means "not found" [const char* version]
378: */
379: size_t pos(const StringBody substr,
380: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
381: /// String version of @see pos(const char*, int, Language)
382: size_t pos(const String& substr,
383: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
384: size_t pos(char c,
385: size_t this_offset=0) const {
386: return body.pos(c, this_offset);
387: }
1.55 paf 388:
1.145 paf 389: void split(ArrayString& result,
390: size_t& pos_after,
391: const char* delim,
392: Language lang=L_UNSPECIFIED, int limit=-1) const;
393: void split(ArrayString& result,
394: size_t& pos_after,
1.62 paf 395: const String& delim,
1.145 paf 396: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 397:
1.145 paf 398: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 399: int prestart, int prefinish,
400: int poststart, int postfinish,
1.68 paf 401: void *info);
1.87 parser 402: /**
1.145 paf 403: @return table of found items, if any.
1.87 parser 404: table format is defined and fixed[can be used by others]:
405: @verbatim
406: prematch/match/postmatch/1/2/3/...
407: @endverbatim
408: */
1.145 paf 409: Table* match(Charset& source_charset,
1.64 paf 410: const String& regexp,
1.145 paf 411: const String* options,
1.99 parser 412: Row_action row_action, void *info,
1.145 paf 413: bool& just_matched) const;
1.87 parser 414: enum Change_case_kind {
415: CC_UPPER,
416: CC_LOWER
417: };
1.145 paf 418: String& change_case(Charset& source_charset,
1.87 parser 419: Change_case_kind kind) const;
1.145 paf 420: const String& replace(const Dictionary& dict) const;
1.96 parser 421: double as_double() const;
422: int as_int() const;
1.137 paf 423:
1.7 paf 424: private: //disabled
425:
1.12 paf 426: String& operator = (const String&) { return *this; }
1.7 paf 427:
1.1 paf 428: };
1.119 paf 429:
1.145 paf 430: /// simple hash code of string. used by Hash
431: inline uint hash_code(const StringBody self) {
432: return self.hash_code();
1.119 paf 433: }
1.1 paf 434:
435: #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