|
|
| version 1.95, 2001/05/21 16:38:46 | version 1.185, 2009/05/15 06:57:43 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class decl. | Parser: string class decl. |
| Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | |
| Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | |
| $Id$ | |
| */ | */ |
| #ifndef PA_STRING_H | #ifndef PA_STRING_H |
| #define PA_STRING_H | #define PA_STRING_H |
| #include "pa_config_includes.h" | static const char * const IDENT_STRING_H="$Date$"; |
| #include "pa_pool.h" | |
| // includes | |
| #include "pa_types.h" | #include "pa_types.h" |
| #include "pa_array.h" | |
| class Table; | extern "C" { // cord's author forgot to do that |
| #define CORD_NO_IO | |
| #include "cord.h" | |
| }; | |
| /** | // defines |
| $MAIN:html-typo table elements must enlarge string not more that that. | |
| that's a tradeoff - otherwise we'd have to scan string twice: | // cache hash code in String::Body for faster hash access |
| - first for buffer length | #define HASH_CODE_CACHING |
| - second for replacements themselves | |
| */ | |
| #define UNTAINT_TIMES_BIGGER 10 | |
| #ifndef NO_STRING_ORIGIN | // cord extension |
| # define STRING_APPEND_PARAMS \ | /* Returns true if x does contain */ |
| const char *src, size_t size, \ | /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */ |
| String::Untaint_lang lang, \ | int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c); |
| const char *file, uint line | size_t CORD_block_count(CORD x); |
| /// appends piece to String @see String::real_append | |
| # define APPEND(src, size, lang, file, line) \ | // forwards |
| real_append(src, size, lang, file, line) | |
| #else | |
| # define STRING_APPEND_PARAMS \ | |
| const char *src, \ | |
| size_t size, \ | |
| String::Untaint_lang lang | |
| /// appends piece to String @see String::real_append | |
| # define APPEND(src, size, lang, file, line) \ | |
| real_append(src, size, lang) | |
| #endif | |
| /// appends clean piece to String @see String::real_append | |
| #define APPEND_CLEAN(src, size, file, line) \ | |
| APPEND(src, size, String::UL_CLEAN, file, line) | |
| /// appends piece to String as-is @see String::real_append | |
| #define APPEND_AS_IS(src, size, file, line) \ | |
| APPEND(src, size, String::UL_AS_IS, file, line) | |
| /// appends tainted piece to String @see String::real_append | |
| #define APPEND_TAINTED(src, size, file, line) \ | |
| APPEND(src, size, String::UL_TAINTED, file, line) | |
| /// handy: appends const char* piece to String @see String::real_append | |
| #define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) | |
| class Array; | class Charset; |
| class Table; | |
| class SQL_Connection; | class SQL_Connection; |
| class Dictionary; | |
| class Request_charsets; | |
| class String; | |
| typedef Array<const String*> ArrayString; | |
| class VRegex; | |
| // generally useful | |
| int pa_atoi(const char* str, const String* problem_source=0); | |
| double pa_atod(const char* str, const String* problem_source=0); | |
| /// this is result of pos functions which mean that substr were not found | |
| #define STRING_NOT_FOUND ((size_t)-1) | |
| template<typename T> | |
| inline size_t get_length(T current) { | |
| return current; | |
| } | |
| /** | /** |
| Pooled string. | String which knows the lang of all it's langs. |
| Internal structure: | |
| @verbatim | |
| String Chunk0 | |
| ====== ======== | |
| head--------------->[ptr, size, ...] | |
| append_here-------->[ptr, size, ...] | |
| . | |
| . | |
| [ptr, size, ...] | |
| link_row----------->[link to the next chunk] | |
| @endverbatim | |
| All pieces remember | All pieces remember |
| - the file and its line they are from [can be turned off by NO_STRING_ORIGIN] | |
| - whether they are tainted or not, | - whether they are tainted or not, |
| and the language which should be used to detaint them | and the lang which should be used to detaint them |
| */ | */ |
| class String : public Pooled { | |
| class String: public PA_Object { | |
| public: | public: |
| enum { | /** piece is tainted or not. the lang to use when detaint |
| CR_PREALLOCATED_COUNT=2, ///< default preallocated item count | remember to change String_Untaint_lang_name @ untaint.C along |
| CR_GROW_COUNT=2 ///< each time the String chunk_is_full() string expanded() | |
| }; | WARNING WARNING WARNING WARNING WARNING WARNING |
| pos function compares(<=) languages, that is used in searching | |
| for table column separator being L_CLEAN or L_AS_IS. | |
| they search for AS_IS, meaning AS_IS|CLEAN [doing <=L_AS_IS check]. | |
| letters assigned for debugging, but it's important for no language-letter | |
| come before L_AS_IS other then L_CLEAN | |
| WARNING WARNING WARNING WARNING WARNING WARNING | |
| */ | |
| enum Language { | |
| L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty | |
| // these two must go before others, there are checks for >L_AS_IS | |
| L_CLEAN='0', ///< clean WARNING: read above warning before changing | |
| L_AS_IS='A', ///< leave all characters intact WARNING: read above warning before changing | |
| /// piece is tainted or not. the language to use when detaint | L_PASS_APPENDED='P', |
| enum Untaint_lang { | |
| UL_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum | |
| UL_CLEAN, ///< clean | |
| UL_TAINTED, ///< tainted, untaint language as assigned later | |
| // untaint languages. assigned by ^untaint[lang]{...} | |
| UL_PASS_APPENDED, | |
| /**< | /**< |
| leave language built into string being appended. | leave lang built into string being appended. |
| just a flag, that value not stored | just a flag, that value not stored |
| */ | */ |
| UL_AS_IS, ///< leave all characters intact | L_TAINTED='T', ///< tainted, untaint lang as assigned later |
| UL_FILE_NAME, ///< filename | // untaint langs. assigned by ^untaint[lang]{...} |
| UL_HTTP_HEADER, ///< text in HTTP response header | L_FILE_SPEC='F', ///< file specification |
| UL_MAIL_HEADER, ///< text in mail header | L_HTTP_HEADER='h', ///< text in HTTP response header |
| UL_URI, ///< text in uri | L_MAIL_HEADER='m', ///< text in mail header |
| UL_TABLE, ///< ^table:set body | L_URI='U', ///< text in uri |
| UL_SQL, ///< ^table:sql body | L_SQL='Q', ///< ^table:sql body |
| UL_JS, ///< JavaScript code | L_JS='J', ///< JavaScript code |
| UL_HTML, ///< HTML code (for editing) | L_XML='X', ///< ^dom:set xml |
| UL_USER_HTML ///< HTML code with USER chars | L_HTML='H', ///< HTML code |
| L_REGEX='R', ///< RegEx expression | |
| L_HTTP_COOKIE='C', ///< cookies encoded as %uXXXX for compartibility with js functions encode/decode | |
| L_FILE_POST='f', ///temporary escaping zero-char | |
| // READ WARNING ABOVE BEFORE ADDING ANYTHING | |
| L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization | |
| }; | |
| enum Trim_kind { | |
| TRIM_BOTH, | |
| TRIM_START, | |
| TRIM_END | |
| }; | }; |
| union Languages { | |
| struct { | |
| #ifdef PA_LITTLE_ENDIAN | |
| Language lang:8; | |
| size_t is_not_just_lang:sizeof(CORD)*8-8; | |
| #elif defined(PA_BIG_ENDIAN) | |
| size_t is_not_just_lang:sizeof(CORD)*8-8; | |
| Language lang:8; | |
| #else | |
| # error word endianness not determined for some obscure reason | |
| #endif | |
| } opt; | |
| CORD langs; | |
| template<typename C> | |
| CORD make_langs(C current) const { | |
| return opt.is_not_just_lang? | |
| langs | |
| :CORD_chars((char)opt.lang, get_length(current)); | |
| } | |
| CORD make_langs(size_t aoffset, size_t alength) const { | |
| return opt.is_not_just_lang? | |
| CORD_substr(langs, aoffset, alength) | |
| :CORD_chars((char)opt.lang, alength); | |
| } | |
| /// appending when 'langs' already contain something [simple cases handled elsewhere] | |
| template<typename C> | |
| void append(C current, | |
| const CORD to_nonempty_target_langs) { | |
| assert(langs); | |
| if(opt.is_not_just_lang) | |
| langs=CORD_cat(langs, to_nonempty_target_langs); | |
| else { // we were "just lang" | |
| size_t current_size=get_length(current); | |
| assert(current_size); | |
| langs=CORD_cat( | |
| CORD_chars((char)opt.lang, current_size), // first piece [making from just 'lang'] | |
| to_nonempty_target_langs); // new piece | |
| } | |
| } | |
| public: | |
| const char* v() const; | |
| void dump() const; | |
| Languages(): langs(0) {} | |
| Languages(Language alang) { | |
| opt.lang=alang; | |
| opt.is_not_just_lang=0; | |
| } | |
| /// MUST be called exactly prior to modification of current [uses it's length] | |
| template<typename C> | |
| void append(C current, Language alang, size_t asize) { | |
| assert(alang); | |
| assert(asize); | |
| if(!opt.is_not_just_lang) | |
| if(opt.lang) { | |
| if(opt.lang==alang) // same language? ignoring | |
| return; | |
| } else { | |
| opt.lang=alang; // to uninitialized | |
| return; | |
| } | |
| append(current, CORD_chars((char)alang, asize)); | |
| } | |
| template<typename C> | |
| void appendHelper(C current, Language alang, C asize_helper) { | |
| assert(alang); | |
| if(!opt.is_not_just_lang) | |
| if(opt.lang) { | |
| if(opt.lang==alang) // same language? ignoring | |
| return; | |
| } else { | |
| opt.lang=alang; // to uninitialized | |
| return; | |
| } | |
| append(current, CORD_chars((char)alang, asize_helper.length())); | |
| } | |
| template<typename C> | |
| void appendHelper(C current, C length_helper, const Languages src) { | |
| if(!langs) | |
| langs=src.langs; // to uninitialized | |
| else if(!src.opt.is_not_just_lang) | |
| appendHelper(current, src.opt.lang, length_helper); // simplifying when simple source | |
| else | |
| append(current, src.make_langs(length_helper)); | |
| } | |
| /// MUST be called exactly prior to modification of current [uses it's length] | |
| template<typename C> | |
| void append(C current, | |
| const Languages src, size_t aoffset, size_t alength) { | |
| assert(alength); | |
| if(!langs) // to uninitialized? | |
| if(src.opt.is_not_just_lang) | |
| langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex | |
| else | |
| opt.lang=src.opt.lang; // to uninitialized simple | |
| else | |
| if(!opt.is_not_just_lang && !src.opt.is_not_just_lang && opt.lang==src.opt.lang) // both simple & of same language? | |
| return; // ignoring | |
| else | |
| append(current, src.make_langs(aoffset, alength)); | |
| } | |
| /// checks if we have lang<=alang all from aoffset to aoffset+alength | |
| bool check_lang(Language alang, size_t aoffset, size_t alength) const { | |
| if(alang==L_UNSPECIFIED) // ignore lang? | |
| return true; | |
| if(opt.is_not_just_lang) | |
| return CORD_range_contains_chr_greater_then(langs, aoffset, alength, (unsigned)alang)==0; | |
| else | |
| return (unsigned)opt.lang<=(unsigned)alang; | |
| } | |
| /// @returns count of blocks | |
| /// @todo currently there can be adjucent blocks of same language. someday merge them | |
| size_t count() const { | |
| return opt.is_not_just_lang? | |
| CORD_block_count(langs) | |
| : opt.lang? | |
| 1 | |
| : 0; | |
| }; | |
| template<typename C, typename I> | |
| void for_each(C current, | |
| int callback(char, size_t, I), I info) const { | |
| if(opt.is_not_just_lang) | |
| CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info); | |
| else | |
| callback(opt.lang, get_length(current), info); | |
| } | |
| bool invariant(size_t current_length) const { | |
| if(!langs) | |
| return current_length==0; | |
| if(opt.is_not_just_lang) | |
| return CORD_len(langs)==current_length; | |
| return true; // uncheckable, actually | |
| } | |
| }; | |
| class Body { | |
| CORD body; | |
| #ifdef HASH_CODE_CACHING | |
| // cached hash code is not reseted on write operations as test shows | |
| // that string body does not change after it is stored as a hash key | |
| mutable uint hash_code; | |
| #endif | |
| public: | |
| const char* v() const; | |
| void dump() const; | |
| #ifdef HASH_CODE_CACHING | |
| Body(): body(CORD_EMPTY), hash_code(0) {} | |
| Body(CORD abody, uint ahash_code): body(abody), hash_code(ahash_code) {} | |
| Body(CORD abody): body(abody), hash_code(0) { | |
| #else | |
| Body(): body(CORD_EMPTY) {} | |
| Body(CORD abody): body(abody) { | |
| #endif | |
| assert(!body // no body | |
| || *body // ordinary string | |
| || body[1]==1 // CONCAT_HDR | |
| || body[1]==4 // FN_HDR | |
| || body[1]==6 // SUBSTR_HDR | |
| ); | |
| } | |
| static Body Format(int value); | |
| void clear() { body=CORD_EMPTY; } | |
| bool operator! () const { return is_empty(); } | |
| CORD get_cord() const { return body; } | |
| uint get_hash_code() const; | |
| const char* cstr() const { return CORD_to_const_char_star(body); } | |
| char* cstrm() const { return CORD_to_char_star(body); } | |
| size_t length() const { return CORD_len(body); } | |
| bool is_empty() const { return body==CORD_EMPTY; } | |
| void append_know_length(const char *str, size_t known_length) { | |
| if(known_length) | |
| body=CORD_cat_char_star(body, str, known_length); | |
| } | |
| void append_strdup_know_length(const char* str, size_t known_length) { | |
| if(known_length) | |
| append_know_length(pa_strdup(str, known_length), known_length); | |
| } | |
| void append(char c) { body=CORD_cat_char(body, c); } | |
| Body& operator << (const Body src) { body=CORD_cat(body, src.body); return *this; } | |
| Body& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; } | |
| // could not figure out why this operator is needed [should do this chain: string->simple->==] | |
| bool operator < (const Body src) const { return CORD_cmp(body, src.body)<0; } | |
| bool operator > (const Body src) const { return CORD_cmp(body, src.body)>0; } | |
| bool operator <= (const Body src) const { return CORD_cmp(body, src.body)<=0; } | |
| bool operator >= (const Body src) const { return CORD_cmp(body, src.body)>=0; } | |
| bool operator != (const Body src) const { return CORD_cmp(body, src.body)!=0; } | |
| bool operator == (const Body src) const { return CORD_cmp(body, src.body)==0; } | |
| int ncmp(size_t x_begin, const Body y, size_t y_begin, size_t size) const { | |
| return CORD_ncmp(body, x_begin, y.body, y_begin, size); | |
| } | |
| char fetch(size_t index) const { return CORD_fetch(body, index); } | |
| Body mid(size_t index, size_t length) const { return CORD_substr(body, index, length); } | |
| size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); } | |
| size_t pos(const Body substr, size_t offset=0) const { | |
| if(substr.is_empty()) | |
| return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1] | |
| // CORD_str checks for bad offset [CORD_chr does not] | |
| return CORD_str(body, offset, substr.body); | |
| } | |
| size_t pos(char c, | |
| size_t offset=0) const { | |
| if(offset>=length()) // CORD_chr does not check that [and ABORT's in that case] | |
| return STRING_NOT_FOUND; | |
| return CORD_chr(body, offset, c); | |
| } | |
| template<typename I> int for_each( | |
| int (*f)(char c, I), | |
| I info) const { | |
| return CORD_iter(body, (CORD_iter_fn)f, (void*)info); | |
| } | |
| template<typename I> int for_each( | |
| int (*f1)(char c, I), | |
| int (*f2)(const char* s, I), | |
| I info) const { | |
| return CORD_iter5(body, 0, (CORD_iter_fn)f1, (CORD_batched_iter_fn)f2, info); | |
| } | |
| void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); } | |
| /*Body normalize() const { | |
| return Body(CORD_balance(body)); | |
| }*/ | |
| /// @returns this or 0 or mid. if returns this or 0 out_* are not filled | |
| Body trim(Trim_kind kind=TRIM_BOTH, const char* chars=0, | |
| size_t* out_start=0, size_t* out_length=0) const; | |
| }; | |
| struct C { | |
| const char *str; | |
| size_t length; | |
| operator const char *() { return str; } | |
| C(): str(0), length(0) {} | |
| C(const char *astr, size_t asize): str(astr), length(asize) {} | |
| }; | |
| struct Cm { | |
| char *str; | |
| size_t length; | |
| //operator char *() { return str; } | |
| Cm(): str(0), length(0) {} | |
| Cm(char *astr, size_t asize): str(astr), length(asize) {} | |
| }; | |
| private: | |
| Body body; ///< all characters of string | |
| Languages langs; ///< string characters lang | |
| const char* v() const; | |
| void dump() const; | |
| #define ASSERT_STRING_INVARIANT(string) \ | |
| assert((string).langs.invariant((string).body.length())) | |
| public: | public: |
| String(Pool& apool, const char *src=0, size_t src_size=0, bool tainted=false); | static const String Empty; |
| String(const String& src); | |
| size_t size() const { return fsize; } | explicit String(){}; |
| /// convert to C string. if 'lang' known, forcing 'lang' to it | explicit String(const char* cstr, Language alang=L_CLEAN){ |
| char *cstr(Untaint_lang lang=UL_UNSPECIFIED, | if(cstr && *cstr){ |
| SQL_Connection *connection=0, | body=cstr; |
| const char *charset=0) const { | langs=alang; |
| } | |
| char *result=(char *)malloc(cstr_bufsize(lang)); | |
| char *eol=store_to(result, lang, connection, charset); | |
| *eol=0; | |
| return result; | |
| } | } |
| /** append fragment | explicit String(const String::C cstr, Language alang=L_CLEAN){ |
| @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST | if(cstr.length){ |
| */ | body=cstr.str; |
| String& real_append(STRING_APPEND_PARAMS); | langs=alang; |
| /// @return <0 ==0 or >0 depending on comparison result | } |
| int cmp (int& partial, const String& src, | |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | |
| bool operator < (const String& src) const { int p; return cmp(p, src)<0; } | |
| bool operator > (const String& src) const { int p; return cmp(p, src)>0; } | |
| bool operator <= (const String& src) const { int p; return cmp(p, src)<=0; } | |
| bool operator >= (const String& src) const { int p; return cmp(p, src)>=0; } | |
| bool operator == (const String& src) const { | |
| if(size()!=src.size()) // can speed up in trivial case | |
| return false; | |
| int p; return cmp(p, src)==0; | |
| } | } |
| bool operator != (const String& src) const { int p; return cmp(p, src)!=0; } | String(int value, char *format); |
| String(Body abody, Language alang): body(abody), langs(alang) { | |
| /** | ASSERT_STRING_INVARIANT(*this); |
| @param partial | } |
| returns partial match status. | String(const String& src): body(src.body), langs(src.langs) { |
| - -1: strings too different | ASSERT_STRING_INVARIANT(*this); |
| - 0: full match | |
| - 1: means @c this starts @c src | |
| - 2: means @src starts @this | |
| */ | |
| int cmp(int& partial, const char* src_ptr, size_t src_size=0, | |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | |
| bool operator == (const char* src_ptr) const { | |
| size_t src_size=src_ptr?strlen(src_ptr):0; | |
| if(size() != src_size) | |
| return false; | |
| int partial; // unused | |
| return cmp(partial, src_ptr, src_size)==0; | |
| } | |
| bool operator != (const char* src_ptr) const { | |
| int partial; // unused | |
| return cmp(partial, src_ptr, 0)!=0; | |
| } | } |
| /// for convinient hash lookup | |
| #ifdef HASH_CODE_CACHING | |
| operator const Body&() const { return body; } | |
| #else | |
| operator const Body() const { return body; } | |
| #endif | |
| /** | bool is_empty() const { return body.is_empty(); } |
| appends other String. | size_t length() const { return body.length(); } |
| size_t length(Charset& charset) const; | |
| /// convert to CORD. if 'lang' known, forcing 'lang' to it | |
| Body cstr_to_string_body(Language lang=L_AS_IS, | |
| SQL_Connection* connection=0, | |
| const Request_charsets *charsets=0) const; | |
| /// convert to constant C string. if 'lang' known, forcing 'lang' to it | |
| const char* cstr(Language lang=L_AS_IS, | |
| SQL_Connection* connection=0, | |
| const Request_charsets *charsets=0) const { | |
| return cstr_to_string_body(lang, connection, charsets).cstr(); | |
| } | |
| /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it | |
| char *cstrm(Language lang=L_AS_IS, | |
| SQL_Connection* connection=0, | |
| const Request_charsets *charsets=0) const { | |
| return cstr_to_string_body(lang, connection, charsets).cstrm(); | |
| } | |
| /// puts pieces to buf | |
| Cm serialize(size_t prolog_size) const; | |
| /// appends pieces from buf to self | |
| bool deserialize(size_t prolog_size, void *buf, size_t buf_size); | |
| /// @see Body::append_know_length | |
| String& append_know_length(const char* str, size_t known_length, Language lang); | |
| /// @see Body::append_help_length | |
| String& append_help_length(const char* str, size_t helper_length, Language lang); | |
| String& append_strdup(const char* str, size_t helper_length, Language lang); | |
| bool operator == (const char* y) const { return body==Body(y); } | |
| bool operator != (const char* y) const { return body!=Body(y); } | |
| /// this starts with y | |
| bool starts_with(const char* y) const { | |
| return body.ncmp(0/*x_begin*/, Body(y), 0/*y_begin*/, strlen(y))==0; | |
| } | |
| /// x starts with this | |
| bool this_starts(const char* x) const { | |
| return Body(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0; | |
| } | |
| marking all tainted pieces of it with @a lang. | String& append_to(String& dest, Language lang, bool forced) const; |
| or marking ALL pieces of it with a @a lang when @a forced to. | String& append(const String& src, Language lang, bool forced=false) { |
| */ | return src.append_to(*this, lang, forced); |
| String& append(const String& src, Untaint_lang lang, bool forced=false); | } |
| String& operator << (const String& src) { return append(src, UL_PASS_APPENDED); } | String& operator << (const String& src) { return append(src, L_PASS_APPENDED); } |
| String& operator << (const char *src) { return APPEND_CONST(src); } | String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); } |
| String& operator << (const Body src); | |
| /// extracts first char of a string, if any | |
| char first_char() const { | |
| return is_empty()?0:body.fetch(0); | |
| } | |
| /// simple hash code of string. used by Hash | bool operator < (const String& src) const { return body<src.body; } |
| uint hash_code() const; | bool operator > (const String& src) const { return body>src.body; } |
| bool operator <= (const String& src) const { return body<=src.body; } | |
| bool operator >= (const String& src) const { return body>=src.body; } | |
| bool operator != (const String& src) const { return body!=src.body; } | |
| bool operator == (const String& src) const { return body==src.body; } | |
| /// extracts [start, finish) piece of string | /// extracts [start, finish) piece of string |
| String& mid(size_t start, size_t finish) const; | String& mid(size_t substr_begin, size_t substr_end) const; |
| String& mid(Charset& charset, size_t from, size_t to, size_t helper_length=0) const; | |
| /// @return position of substr in string, -1 means "not found" [String version] | /** |
| int pos(const String& substr, | ignore lang if it's L_UNSPECIFIED |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | but when specified: look for substring that lies in ONE fragment in THAT lang |
| /// @return position of substr in string, -1 means "not found" [const char* version] | @return position of substr in string, -1 means "not found" [const char* version] |
| int pos(const char *substr, size_t substr_size, | */ |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | size_t pos(const Body substr, |
| size_t this_offset=0, Language lang=L_UNSPECIFIED) const; | |
| void split(Array& result, | /// String version of @see pos(const char*, int, Language) |
| size_t *pos_after_ref, | size_t pos(const String& substr, |
| const char *delim, size_t delim_size, | size_t this_offset=0, Language lang=L_UNSPECIFIED) const; |
| Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; | size_t pos(char c, |
| void split(Array& result, | size_t this_offset=0) const { |
| size_t *pos_after_ref, | return body.pos(c, this_offset); |
| } | |
| size_t pos(Charset& charset, | |
| const String& substr, | |
| size_t this_offset=0, Language lang=L_UNSPECIFIED) const; | |
| void split(ArrayString& result, | |
| size_t& pos_after, | |
| const char* delim, | |
| Language lang=L_UNSPECIFIED, int limit=-1) const; | |
| void split(ArrayString& result, | |
| size_t& pos_after, | |
| const String& delim, | const String& delim, |
| Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; | Language lang=L_UNSPECIFIED, int limit=-1) const; |
| typedef void (*Row_action)(Table& table, Array *row, int start, int finish, | typedef void (*Row_action)(Table& table, ArrayString* row, |
| int prestart, int prefinish, | |
| int poststart, int postfinish, | |
| void *info); | void *info); |
| /** | /** |
| @return true if fills table. | @return table of found items, if any. |
| table format is defined and fixed[can be used by others]: | table format is defined and fixed[can be used by others]: |
| @verbatim | @verbatim |
| prematch/match/postmatch/1/2/3/... | prematch/match/postmatch/1/2/3/... |
| @endverbatim | @endverbatim |
| */ | */ |
| bool match(const unsigned char *pcre_tables, | Table* match(VRegex* vregex, |
| const String *aorigin, | Row_action row_action, void *info, |
| const String& regexp, | int& matches_count) const; |
| const String *options, | |
| Table **table, | |
| Row_action row_action, void *info) const; | |
| enum Change_case_kind { | enum Change_case_kind { |
| CC_UPPER, | CC_UPPER, |
| CC_LOWER | CC_LOWER |
| }; | }; |
| String& change_case(Pool& pool, const unsigned char *pcre_tables, | String& change_case(Charset& source_charset, |
| Change_case_kind kind) const; | Change_case_kind kind) const; |
| double as_double(); | const String& replace(const Dictionary& dict) const; |
| int as_int(); | const String& trim(Trim_kind kind=TRIM_BOTH, const char* chars=0) const; |
| double as_double() const { return pa_atod(cstr(), this); } | |
| #ifndef NO_STRING_ORIGIN | int as_int() const { return pa_atoi(cstr(), this); } |
| /// origin of string. calculated by first row | bool as_bool() const { return as_int()!=0; } |
| const Origin& origin() const; | const String& escape(Charset& source_charset) const; |
| #endif | |
| private: | |
| /// several String fragments | |
| struct Chunk { | |
| size_t count; ///< the number of rows in chunk | |
| /// string fragment or a link to next chunk union | |
| union Row { | |
| /// fragment | |
| struct { | |
| const char *ptr; ///< pointer to the start | |
| size_t size; ///< length | |
| Untaint_lang lang; ///< untaint flag, later untaint language | |
| #ifndef NO_STRING_ORIGIN | |
| Origin origin; ///< origin | |
| #endif | |
| } item; | |
| Chunk *link; ///< link to the next chunk in chain | |
| } rows[CR_PREALLOCATED_COUNT]; | |
| Chunk *preallocated_link; ///< next rows are here | |
| } | |
| head; ///< the head chunk of the chunk chain | |
| /// next append would write to this record | |
| Chunk::Row *append_here; | |
| /** the address of place where lies address | |
| of the link to the next chunk to allocate | |
| */ | |
| Chunk::Row *link_row; | |
| private: | |
| // last chunk | |
| Chunk *last_chunk; | |
| // string size | private: //disabled |
| size_t fsize; | |
| // used rows in all chunks | |
| int fused_rows; | |
| private: | String& operator = (const String&) { return *this; } |
| bool chunk_is_full() { | }; |
| return append_here == link_row; | |
| } | |
| void expand(); | |
| size_t cstr_bufsize(Untaint_lang lang) const; | template<> |
| /// convert to C string, store to 'dest' which must be big enough for proper untaint | inline size_t get_length<String::Body>(String::Body body) { |
| char *store_to(char *dest, Untaint_lang lang=UL_UNSPECIFIED, | return body.length(); |
| SQL_Connection *connection=0, | } |
| const char *charset=0) const; | |
| #ifndef HASH_CODE_CACHING | |
| /// simple hash code of string. used by Hash | |
| inline uint hash_code(const String::Body self) { | |
| return self.get_hash_code(); | |
| } | |
| #endif | |
| private: //disabled | |
| String& operator = (const String&) { return *this; } | /// now that we've declared specialization we can use it |
| inline String& String::operator << (const String::Body src) { | |
| langs.append(body, L_AS_IS, src.length()); | |
| body<<src; | |
| return *this; | |
| } | |
| }; | |
| #endif | #endif |