Annotation of parser3/src/include/pa_string.h, revision 1.145.4.6
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:
1.145.4.6! paf 11: static const char* IDENT_STRING_H="$Date: 2003/09/24 07:13:52 $";
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))
1.145.4.1 paf 26:
27: // cord extension
28: /* Returns true if x does contain */
29: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
1.145.4.5 paf 30: int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c);
1.145.4.1 paf 31:
1.145 paf 32: // forwards
1.9 paf 33:
1.145 paf 34: class Charset;
1.135 paf 35: class Table;
1.71 paf 36: class SQL_Connection;
1.101 parser 37: class Dictionary;
1.145 paf 38: class Request_charsets;
39: class String;
40: typedef Array<const String*> ArrayString;
41:
42: /// this is result of pos functions which mean that substr were not found
43: #define STRING_NOT_FOUND ((size_t)-1)
44:
45: class StringBody {
46:
47: CORD body;
48:
49: public:
50:
51: StringBody(): body(CORD_EMPTY) {}
52: StringBody(CORD abody): body(abody) {
53: assert(!body // no body
54: || *body // ordinary string
55: || body[1]==1 // CONCAT_HDR
56: || body[1]==4 // FN_HDR
57: || body[1]==6 // SUBSTR_HDR
58: );
59: }
60: /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
61: StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
62: append_know_length(str, helper_length?helper_length:strlen(str));
63: }
64: static StringBody Format(int value);
65:
66: void clear() { body=CORD_EMPTY; }
67:
68: bool operator! () const { return is_empty(); }
69:
70: uint hash_code() const;
71:
72: const char* cstr() const { return CORD_to_const_char_star(body); }
73: char* cstrm() const { return CORD_to_char_star(body); }
74:
75: size_t length() const { return CORD_len(body); }
76:
77: bool is_empty() const { return body==CORD_EMPTY; }
78:
79: void append_know_length(const char *str, size_t known_length) {
80: if(known_length)
81: body=CORD_cat_char_star(body, str, known_length);
82: }
83: void append_strdup_know_length(const char* str, size_t known_length) {
84: if(known_length)
85: append_know_length(pa_strdup(str, known_length), known_length);
86: }
87: void append(char c) { body=CORD_cat_char(body, c); }
88: StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
89: StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
90:
91: // could not figure out why this operator is needed [should do this chain: string->simple->==]
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: 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:
99: int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
100: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
101: }
102:
103: char fetch(size_t index) const { return CORD_fetch(body, index); }
104: StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
105: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
106: size_t pos(const StringBody substr, size_t offset=0) const {
107: if(!substr.length())
108: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
109: return CORD_str(body, offset, substr.body);
110: }
111: size_t pos(char c,
112: size_t offset=0) const {
113: return CORD_chr(body, offset, c);
114: }
115:
1.145.4.1 paf 116: /* template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
1.145 paf 117: CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
1.145.4.1 paf 118: }*/
1.145 paf 119:
120: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
121:
1.145.4.2 paf 122: /*StringBody normalize() const {
1.145 paf 123: return StringBody(CORD_balance(body));
1.145.4.2 paf 124: }*/
1.145 paf 125:
126: void dump() const {
127: CORD_dump(body);
128: }
129: };
1.62 paf 130:
1.145.4.4 paf 131: template<typename T>
132: inline size_t get_length(T current) {
133: return current;
134: }
1.42 paf 135: /**
1.145.4.1 paf 136: String which knows the lang of all it's langs.
1.41 paf 137:
138: All pieces remember
139: - whether they are tainted or not,
1.145.4.1 paf 140: and the lang which should be used to detaint them
1.41 paf 141: */
1.145 paf 142: class String: public PA_Object {
1.1 paf 143: public:
1.48 paf 144:
1.145.4.1 paf 145: /** piece is tainted or not. the lang to use when detaint
1.106 parser 146: remember to change String_Untaint_lang_name @ untaint.C along
147: */
1.145 paf 148: enum Language {
1.145.4.1 paf 149: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 150: // these two must go before others, there are checks for >L_AS_IS
151: L_CLEAN, ///< clean
152: L_AS_IS, ///< leave all characters intact
1.122 paf 153:
1.145 paf 154: L_PASS_APPENDED,
1.41 paf 155: /**<
1.145.4.1 paf 156: leave lang built into string being appended.
1.41 paf 157: just a flag, that value not stored
158: */
1.145.4.1 paf 159: L_TAINTED, ///< tainted, untaint lang as assigned later
160: // untaint langs. assigned by ^untaint[lang]{...}
1.145 paf 161: L_FILE_SPEC, ///< file specification
162: L_HTTP_HEADER, ///< text in HTTP response header
163: L_MAIL_HEADER, ///< text in mail header
164: L_URI, ///< text in uri
165: L_TABLE, ///< ^table:set body
166: L_SQL, ///< ^table:sql body
167: L_JS, ///< JavaScript code
168: L_XML, ///< ^dom:set xml
169: L_HTML, ///< HTML code (for editing)
1.145.4.1 paf 170: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 171: };
172:
1.145.4.1 paf 173: class Languages {
174:
175: union {
176: struct {
177: Language lang:8;
1.145.4.5 paf 178: int is_not_just_lang:sizeof(CORD)*8-8;
1.145.4.1 paf 179: };
180: CORD langs;
181: };
182:
1.145.4.4 paf 183: template<typename C>
184: CORD make_langs(C current) const {
1.145.4.1 paf 185: return is_not_just_lang?
186: langs
1.145.4.4 paf 187: :CORD_chars((char)lang, get_length(current));
1.145 paf 188: }
189:
1.145.4.1 paf 190: CORD make_langs(size_t aoffset, size_t alength) const {
191: return is_not_just_lang?
192: CORD_substr(langs, aoffset, alength)
1.145.4.3 paf 193: :CORD_chars((char)lang, alength);
1.145 paf 194: }
1.145.4.1 paf 195:
1.145.4.3 paf 196: /// appending when 'langs' already contain something [simple cases hanled elsewhere]
1.145.4.4 paf 197: template<typename C>
198: void append(C current,
1.145.4.1 paf 199: const CORD to_nonempty_target_langs) {
1.145.4.3 paf 200: assert(langs);
1.145.4.1 paf 201:
202: if(is_not_just_lang)
203: langs=CORD_cat(langs, to_nonempty_target_langs);
204: else { // we were "just lang"
1.145.4.4 paf 205: size_t current_size=get_length(current);
1.145.4.1 paf 206: assert(current_size);
207: langs=CORD_cat(
1.145.4.3 paf 208: CORD_chars((char)lang, current_size), // first piece [making from just 'lang']
1.145.4.1 paf 209: to_nonempty_target_langs); // new piece
210: }
1.145 paf 211: }
1.145.4.1 paf 212:
1.145 paf 213: public:
1.145.4.1 paf 214:
215: Languages(): langs(0) {}
216: Languages(Language alang): lang(alang), is_not_just_lang(0) {}
217:
1.145.4.4 paf 218: /// MUST be called exactly prior to modification of current [uses it's length]
219: template<typename C>
220: void append(C current, Language alang, size_t asize) {
1.145.4.1 paf 221: assert(alang);
222: assert(asize);
223:
1.145.4.3 paf 224: if(!is_not_just_lang)
225: if(lang) {
226: if(lang==alang) // same length? ignoring
227: return;
228: } else {
229: lang=alang; // to uninitialized
230: return;
231: }
1.145.4.1 paf 232:
1.145.4.4 paf 233: append(current, CORD_chars((char)alang, asize));
1.145 paf 234: }
235:
1.145.4.4 paf 236: /// MUST be called exactly prior to modification of current [uses it's length]
237: template<typename C>
238: void append(C current, size_t appending_length,
1.145.4.1 paf 239: const Languages src) {
1.145.4.4 paf 240: assert(appending_length);
1.145.4.1 paf 241:
1.145.4.3 paf 242: if(!langs)
243: langs=src.langs; // to uninitialized
244: else if(!src.is_not_just_lang)
1.145.4.4 paf 245: append(current, src.lang, appending_length); // simplifying when simple source
1.145.4.3 paf 246: else
1.145.4.4 paf 247: append(current, src.make_langs(appending_length));
1.145.4.1 paf 248: }
249:
1.145.4.4 paf 250: /// MUST be called exactly prior to modification of current [uses it's length]
251: template<typename C>
252: void append(C current,
1.145.4.1 paf 253: const Languages src, size_t aoffset, size_t alength) {
254: assert(alength);
255:
1.145.4.3 paf 256: if(!langs) // to uninitialized?
1.145.4.1 paf 257: if(src.is_not_just_lang)
1.145.4.3 paf 258: langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex
259: else
260: lang=src.lang; // to uninitialized simple
261: else
262: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang) // both simple & of same language?
263: return; // ignoring
1.145.4.1 paf 264: else
1.145.4.4 paf 265: append(current, src.make_langs(aoffset, alength));
1.145.4.1 paf 266: }
267:
1.145.4.5 paf 268: /// checks if we have lang<=alang all from aoffset to aoffset+alength
1.145.4.1 paf 269: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
270: if(alang==L_UNSPECIFIED) // ignore lang?
271: return true;
272:
273: if(is_not_just_lang)
1.145.4.5 paf 274: return CORD_range_contains_chr_greater_then(langs, aoffset, alength, (unsigned)alang)==0;
1.145.4.1 paf 275: else
1.145.4.5 paf 276: return lang<=alang;
1.145.4.1 paf 277: }
278:
1.145.4.4 paf 279: template<typename C, typename I>
280: void for_each(C current,
1.145.4.2 paf 281: int callback(char, size_t, I), I info) const {
1.145.4.1 paf 282:
283: if(is_not_just_lang)
1.145.4.2 paf 284: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
1.145.4.1 paf 285: else
1.145.4.4 paf 286: callback(lang, get_length(current), info);
287: }
288:
289: bool invariant(size_t current_length) {
290: if(!langs)
291: return current_length==0;
292: if(is_not_just_lang)
1.145.4.6! paf 293: return !*langs && CORD_len(langs)==current_length;
1.145.4.4 paf 294: return true; // uncheckable, actually
1.145 paf 295: }
296: };
297:
298: struct C {
299: const char *str;
300: size_t length;
301: operator const char *() { return str; }
302: C(const char *astr, size_t asize): str(astr), length(asize) {}
303: };
304:
305: struct Cm {
306: char *str;
307: size_t length;
308: //operator char *() { return str; }
309: Cm(char *astr, size_t asize): str(astr), length(asize) {}
310: };
311:
312: private:
313:
1.145.4.1 paf 314: Languages langs; ///< string characters lang
1.145 paf 315: StringBody body; ///< all characters of string
316:
1.145.4.4 paf 317: #define ASSERT_STRING_INVARIANT(string) \
318: assert((string).langs.invariant((string).body.length()))
319:
1.8 paf 320: public:
321:
1.145 paf 322: explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
323: explicit String(const C cstr, bool tainted=false);
1.145.4.4 paf 324: String(StringBody abody, Language alang): body(abody), langs(alang) {
325: assert(!body.is_empty());
326: ASSERT_STRING_INVARIANT(*this);
327: }
328: String(const String& src): body(src.body), langs(src.langs) {
329: ASSERT_STRING_INVARIANT(*this);
330: }
1.145 paf 331:
332: /// for convinient hash lookup
333: operator const StringBody() const { return body; }
334:
335: bool is_empty() const { return body.is_empty(); }
336: size_t length() const { return body.length(); }
337:
338: /// convert to CORD. if 'lang' known, forcing 'lang' to it
339: StringBody cstr_to_string_body(Language lang=L_AS_IS,
340: SQL_Connection* connection=0,
341: const Request_charsets *charsets=0) const;
342:
343: /// convert to constant C string. if 'lang' known, forcing 'lang' to it
344: const char* cstr(Language lang=L_AS_IS,
345: SQL_Connection* connection=0,
346: const Request_charsets *charsets=0) const {
347: return cstr_to_string_body(lang, connection, charsets).cstr();
348: }
349: /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
350: char *cstrm(Language lang=L_AS_IS,
351: SQL_Connection* connection=0,
352: const Request_charsets *charsets=0) const {
353: return cstr_to_string_body(lang, connection, charsets).cstrm();
1.50 paf 354: }
1.108 parser 355: /// puts pieces to buf
1.145 paf 356: Cm serialize(size_t prolog_size) const;
1.108 parser 357: /// appends pieces from buf to self
1.145 paf 358: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
359: /// @see StringBody::append_know_length
360: String& append_know_length(const char* str, size_t known_length, Language lang);
361: /// @see StringBody::append_help_length
362: String& append_help_length(const char* str, size_t helper_length, Language lang);
363: String& append_strdup(const char* str, size_t helper_length, Language lang);
364:
365: bool operator == (const char* y) const { return body==StringBody(y); }
366: bool operator != (const char* y) const { return body!=StringBody(y); }
367:
368: /// this starts with y
369: bool starts_with(const char* y) const {
370: return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
371: }
372: /// x starts with this
373: bool this_starts(const char* x) const {
374: return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 375: }
376:
1.145 paf 377: String& append_to(String& dest, Language lang, bool forced) const;
378: String& append(const String& src, Language lang, bool forced=false) {
379: return src.append_to(*this, lang, forced);
380: }
381: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
382: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
383: String& operator << (const StringBody src) {
1.145.4.1 paf 384: langs.append(body, L_AS_IS, src.length());
1.145 paf 385: body<<src;
386: return *this;
387: }
1.100 parser 388:
1.142 paf 389: /// extracts first char of a string, if any
390: char first_char() const {
1.145 paf 391: return is_empty()?0:body.fetch(0);
1.142 paf 392: }
1.54 paf 393:
1.145 paf 394: bool operator < (const String& src) const { return body<src.body; }
395: bool operator > (const String& src) const { return body>src.body; }
396: bool operator <= (const String& src) const { return body<=src.body; }
397: bool operator >= (const String& src) const { return body>=src.body; }
398: bool operator != (const String& src) const { return body!=src.body; }
399: bool operator == (const String& src) const { return body==src.body; }
400:
1.54 paf 401: /// extracts [start, finish) piece of string
1.145 paf 402: String& mid(size_t substr_begin, size_t substr_end) const;
403:
404: /**
405: ignore lang if it's L_UNSPECIFIED
406: but when specified: look for substring that lies in ONE fragment in THAT lang
407: @return position of substr in string, -1 means "not found" [const char* version]
408: */
409: size_t pos(const StringBody substr,
410: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
411: /// String version of @see pos(const char*, int, Language)
412: size_t pos(const String& substr,
413: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
414: size_t pos(char c,
415: size_t this_offset=0) const {
416: return body.pos(c, this_offset);
417: }
1.55 paf 418:
1.145 paf 419: void split(ArrayString& result,
420: size_t& pos_after,
421: const char* delim,
422: Language lang=L_UNSPECIFIED, int limit=-1) const;
423: void split(ArrayString& result,
424: size_t& pos_after,
1.62 paf 425: const String& delim,
1.145 paf 426: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 427:
1.145 paf 428: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 429: int prestart, int prefinish,
430: int poststart, int postfinish,
1.68 paf 431: void *info);
1.87 parser 432: /**
1.145 paf 433: @return table of found items, if any.
1.87 parser 434: table format is defined and fixed[can be used by others]:
435: @verbatim
436: prematch/match/postmatch/1/2/3/...
437: @endverbatim
438: */
1.145 paf 439: Table* match(Charset& source_charset,
1.64 paf 440: const String& regexp,
1.145 paf 441: const String* options,
1.99 parser 442: Row_action row_action, void *info,
1.145 paf 443: bool& just_matched) const;
1.87 parser 444: enum Change_case_kind {
445: CC_UPPER,
446: CC_LOWER
447: };
1.145 paf 448: String& change_case(Charset& source_charset,
1.87 parser 449: Change_case_kind kind) const;
1.145 paf 450: const String& replace(const Dictionary& dict) const;
1.96 parser 451: double as_double() const;
452: int as_int() const;
1.137 paf 453:
1.7 paf 454: private: //disabled
455:
1.12 paf 456: String& operator = (const String&) { return *this; }
1.7 paf 457:
1.1 paf 458: };
1.145.4.4 paf 459:
460: template<>
461: inline size_t get_length<StringBody>(StringBody body) {
462: return body.length();
463: }
1.119 paf 464:
1.145 paf 465: /// simple hash code of string. used by Hash
466: inline uint hash_code(const StringBody self) {
467: return self.hash_code();
1.119 paf 468: }
1.1 paf 469:
470: #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