|
|
| version 1.201, 2004/10/07 09:22:01 | version 1.267, 2021/12/28 15:58:31 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class. @see untalength_t.C. | Parser: string class. @see untalength_t.C. |
| Copyright (c) 2001-2004 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001-2020 Art. Lebedev Studio (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| */ | */ |
| static const char * const IDENT_STRING_C="$Date$"; | |
| #include "pcre.h" | |
| #include "pa_string.h" | #include "pa_string.h" |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_table.h" | #include "pa_table.h" |
| #include "pa_dictionary.h" | #include "pa_dictionary.h" |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #include "pa_vregex.h" | |
| volatile const char * IDENT_PA_STRING_C="$Id$" IDENT_PA_STRING_H; | |
| const String String::Empty; | const String String::Empty; |
| int pa_atoi(const char* str, const String* problem_source) { | #define COMPILE_ASSERT(x) extern int assert_checker[(x) ? 1 : -1] |
| COMPILE_ASSERT(sizeof(String::Languages) == sizeof(CORD)); | |
| // pa_atoui is based on Manuel Novoa III _strto_l for uClibc | |
| template<typename T> inline T pa_ato_any(const char *str, int base, const String* problem_source,const T max){ | |
| T result = 0; | |
| const char *pos = str; | |
| while (isspace(*pos)) /* skip leading whitespace */ | |
| ++pos; | |
| if (base == 16 && *pos == '0') { /* handle option prefix */ | |
| ++pos; | |
| if (*pos == 'x' || *pos == 'X') { | |
| ++pos; | |
| } | |
| } | |
| if (base == 0) { /* dynamic base */ | |
| base = 10; /* default is 10 */ | |
| if (*pos == '0') { | |
| ++pos; | |
| if (*pos == 'x' || *pos == 'X'){ | |
| ++pos; | |
| base=16; | |
| } | |
| } | |
| } | |
| if (base < 2 || base > 16) { /* illegal base */ | |
| throw Exception(PARSER_RUNTIME, 0, "base to must be an integer from 2 to 16"); | |
| } | |
| T cutoff = max / base; | |
| int cutoff_digit = (int)(max - cutoff * base); | |
| while(true) { | |
| int digit; | |
| if ((*pos >= '0') && (*pos <= '9')) { | |
| digit = (*pos - '0'); | |
| } else if (*pos >= 'a') { | |
| digit = (*pos - 'a' + 10); | |
| } else if (*pos >= 'A') { | |
| digit = (*pos - 'A' + 10); | |
| } else break; | |
| if (digit >= base) { | |
| break; | |
| } | |
| ++pos; | |
| /* adjust number, with overflow check */ | |
| if ((result > cutoff) || ((result == cutoff) && (digit > cutoff_digit))) { | |
| throw Exception("number.format", problem_source, problem_source ? "out of range (int)" : "'%s' is out of range (int)", str); | |
| } else { | |
| result = result * base + digit; | |
| } | |
| } | |
| while(char c=*pos++) | |
| if(!isspace(c)) | |
| throw Exception("number.format", problem_source, problem_source ? "invalid number (int)" : "'%s' is invalid number (int)", str); | |
| return result; | |
| } | |
| unsigned int pa_atoui(const char *str, int base, const String* problem_source){ | |
| if(!str) | |
| return 0; | |
| return pa_ato_any<unsigned int>(str, base, problem_source, UINT_MAX); | |
| } | |
| uint64_t pa_atoul(const char *str, int base, const String* problem_source){ | |
| if(!str) | if(!str) |
| return 0; | return 0; |
| while(*str && isspace((unsigned char)*str)) | return pa_ato_any<uint64_t>(str, base, problem_source, ULLONG_MAX); |
| } | |
| int pa_atoi(const char* str, int base, const String* problem_source) { | |
| if(!str) | |
| return 0; | |
| while(isspace(*str)) | |
| str++; | str++; |
| if(!*str) | if(!*str) |
| return 0; | return 0; |
| int result; | |
| char *error_pos; | |
| bool negative=false; | bool negative=false; |
| if(str[0]=='-') { | if(str[0]=='-') { |
| negative=true; | negative=true; |
| Line 35 int pa_atoi(const char* str, const Strin | Line 116 int pa_atoi(const char* str, const Strin |
| } else if(str[0]=='+') { | } else if(str[0]=='+') { |
| str++; | str++; |
| } | } |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(int)(unsigned long)strtol(str, &error_pos, 0); | |
| else { | |
| // skip leading 0000, to disable octal interpretation | |
| do str++; while(*str=='0'); | |
| result=(int)strtol(str, &error_pos, 0); | |
| } | |
| else | |
| result=(int)strtol(str, &error_pos, 0); | |
| if(negative) | |
| result=-result; | |
| while(char c=*error_pos++) | unsigned int result=pa_atoui(str, base, problem_source); |
| if(!isspace((unsigned char)c)) | |
| throw Exception("number.format", | |
| problem_source, | |
| problem_source?"invalid number (int)": "'%s' is invalid number (int)", str); | |
| return result; | if(negative && result <= ((unsigned int)(-(1+INT_MIN)))+1) |
| return -(int)result; | |
| if(result<=INT_MAX) | |
| return (int)result; | |
| throw Exception("number.format", problem_source, problem_source ? "out of range (int)" : "'%s' is out of range (int)", str); | |
| } | } |
| double pa_atod(const char* str, const String* problem_source) { | double pa_atod(const char* str, const String* problem_source) { |
| if(!str) | if(!str) |
| return 0; | return 0; |
| while(*str && isspace((unsigned char)*str)) | while(isspace(*str)) |
| str++; | str++; |
| if(!*str) | if(!*str) |
| return 0; | return 0; |
| double result; | |
| char *error_pos; | |
| bool negative=false; | bool negative=false; |
| if(str[0]=='-') { | if(str[0]=='-') { |
| negative=true; | negative=true; |
| Line 76 double pa_atod(const char* str, const St | Line 145 double pa_atod(const char* str, const St |
| } else if(str[0]=='+') { | } else if(str[0]=='+') { |
| str++; | str++; |
| } | } |
| // 0xABC | |
| if(str[0]=='0') | double result; |
| if(str[1]=='x' || str[1]=='X') | if(str[0]=='0') { |
| result=(double)(unsigned long)strtol(str, &error_pos, 0); | if(str[1]=='x' || str[1]=='X') { |
| else { | // 0xABC |
| result=(double)pa_atoul(str, 0, problem_source); | |
| return negative ? -result : result; | |
| } else { | |
| // skip leading 0000, to disable octal interpretation | // skip leading 0000, to disable octal interpretation |
| do str++; while(*str=='0'); | do str++; while(*str=='0'); |
| result=(double)strtod(str, &error_pos); | |
| } | } |
| else | } |
| result=(double)strtod(str, &error_pos); | |
| if(negative) | char *error_pos; |
| result=-result; | result=strtod(str, &error_pos); |
| while(char c=*error_pos++) | while(char c=*error_pos++) |
| if(!isspace((unsigned char)c)) | if(!isspace((unsigned char)c)) |
| throw Exception("number.format", | throw Exception("number.format", problem_source, problem_source ? "invalid number (double)" : "'%s' is invalid number (double)", str); |
| problem_source, | |
| problem_source?"invalid number (double)": "'%s' is invalid number (double)", str); | |
| return result; | return negative ? -result : result; |
| } | } |
| // cord lib extension | // cord lib extension |
| #ifndef DOXYGEN | #ifndef DOXYGEN |
| typedef struct { | typedef struct { |
| ssize_t countdown; | ssize_t countdown; |
| char target; /* Character we're looking for */ | int target; /* Character we're looking for */ |
| } chr_data; | } chr_data; |
| #endif | #endif |
| static int CORD_range_contains_chr_greater_then_proc(char c, size_t size, void* client_data) | static int CORD_range_contains_chr_greater_then_proc(char c, size_t size, void* client_data) |
| { | { |
| register chr_data * d = (chr_data *)client_data; | register chr_data * d = (chr_data *)client_data; |
| if (d -> countdown<=0) return(2); | if (d -> countdown<=0) return(2); |
| d -> countdown -= size; | d -> countdown -= size; |
| if (c > d -> target) return(1); | if (c > d -> target) return(1); |
| return(0); | return(0); |
| } | } |
| int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c) | int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c) |
| { | { |
| chr_data d; | chr_data d; |
| d.countdown = n; | d.countdown = n; |
| d.target = c; | d.target = c; |
| return(CORD_block_iter(x, i, CORD_range_contains_chr_greater_then_proc, &d) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/); | return(CORD_block_iter(x, i, CORD_range_contains_chr_greater_then_proc, &d) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/); |
| } | } |
| static int CORD_block_count_proc(char /*c*/, size_t /*size*/, void* client_data) | static int CORD_block_count_proc(char /*c*/, size_t /*size*/, void* client_data) |
| { | { |
| int* result=(int*)client_data; | int* result=(int*)client_data; |
| (*result)++; | (*result)++; |
| return(0); // 0=continue | return(0); // 0=continue |
| } | } |
| size_t CORD_block_count(CORD x) | size_t CORD_block_count(CORD x) |
| { | { |
| size_t result=0; | size_t result=0; |
| CORD_block_iter(x, 0, CORD_block_count_proc, &result); | CORD_block_iter(x, 0, CORD_block_count_proc, &result); |
| return result; | return result; |
| } | } |
| // helpers | // helpers |
| Line 163 Table string_match_table_template(new St | Line 235 Table string_match_table_template(new St |
| String::Body String::Body::Format(int value) { | String::Body String::Body::Format(int value) { |
| char local[MAX_NUMBER]; | char local[MAX_NUMBER]; |
| size_t length=snprintf(local, MAX_NUMBER, "%d", value); | size_t length=snprintf(local, MAX_NUMBER, "%d", value); |
| return String::Body(pa_strdup(local, length), length); | return String::Body(pa_strdup(local, length)); |
| } | } |
| String::Body String::Body::trim(String::Trim_kind kind, const char* chars, | String::Body String::Body::trim(String::Trim_kind kind, const char* chars, size_t* out_start, size_t* out_length, Charset* source_charset) const { |
| size_t* out_start, size_t* out_length) const { | |
| size_t our_length=length(); | size_t our_length=length(); |
| if(!our_length) | if(!our_length) |
| return *this; | return *this; |
| if(!chars) | |
| chars=" \t\n"; // white space | // check if any UTF-8 in chars |
| bool fast=true; | |
| if(chars && source_charset && source_charset->isUTF8()){ | |
| const char* pos=chars; | |
| while(unsigned char c=*pos++) | |
| if(c>127){ | |
| fast=false; | |
| break; | |
| } | |
| } | |
| size_t start=0; | size_t start=0; |
| size_t end=our_length; | size_t end=our_length; |
| // from left... | if(!chars) |
| if(kind!=TRIM_END) { | chars=" \t\n"; // white space |
| CORD_pos pos; set_pos(pos, 0); | |
| while(true) { | if(fast){ |
| char c=CORD_pos_fetch(pos); | // from left... |
| if(strchr(chars, c)) { | if(kind!=TRIM_END) { |
| if(++start==our_length) | CORD_pos pos; set_pos(pos, 0); |
| return 0; // all chars are empty, just return empty string | while(true) { |
| } else | char c=CORD_pos_fetch(pos); |
| break; | if(strchr(chars, c)) { |
| if(++start==our_length) | |
| CORD_next(pos); | return 0; // all chars are empty, just return empty string |
| } | } else |
| } | break; |
| // from right.. | |
| if(kind!=TRIM_START) { | CORD_next(pos); |
| CORD_pos pos; set_pos(pos, end-1); | } |
| while(true) { | } |
| char c=CORD_pos_fetch(pos); | |
| if(strchr(chars, c)) { | // from right.. |
| if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible | if(kind!=TRIM_START) { |
| return 0; // all chars are empty, just return empty string | CORD_pos pos; set_pos(pos, end-1); |
| } else | while(true) { |
| break; | char c=CORD_pos_fetch(pos); |
| if(strchr(chars, c)) { | |
| if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible | |
| return 0; // all chars are empty, just return empty string | |
| } else | |
| break; | |
| CORD_prev(pos); | |
| } | |
| } | |
| } else { | |
| const XMLByte* src_begin=(const XMLByte*)cstr(); | |
| const XMLByte* src_end=src_begin+our_length; | |
| // from left... | |
| if(kind!=TRIM_END) { | |
| while(src_begin<src_end){ | |
| uint char_length=1; | |
| const XMLByte* ptr=src_begin; | |
| // searching first UTF-8 byte: http://tools.ietf.org/html/rfc3629#section-3 | |
| while(++src_begin<=src_end && (*src_begin>127 && *src_begin<0xC0)) | |
| char_length++; | |
| bool found=false; | |
| for(const char* chars_byte=chars; chars_byte=strchr(chars_byte, *ptr); chars_byte++) | |
| if(strncmp(chars_byte, (const char*)ptr, char_length)==0){ | |
| found=true; | |
| break; | |
| } | |
| if(found){ | |
| start+=char_length; | |
| if(start==our_length) | |
| return 0; // all chars are empty, just return empty string | |
| } else | |
| break; | |
| } | |
| } | |
| CORD_prev(pos); | // from right.. |
| if(kind!=TRIM_START) { | |
| while(src_begin<src_end){ | |
| uint char_length=1; | |
| // searching first UTF-8 byte: http://tools.ietf.org/html/rfc3629#section-3 | |
| while(src_begin<=--src_end && (*src_end>127 && *src_end<0xC0)) | |
| char_length++; | |
| bool found=false; | |
| for(const char* chars_byte=chars; chars_byte=strchr(chars_byte, *src_end); chars_byte++) | |
| if(strncmp(chars_byte, (const char*)src_end, char_length)==0){ | |
| found=true; | |
| break; | |
| } | |
| if(found){ | |
| end-=char_length; | |
| if(end==0) | |
| return 0; // all chars are empty, just return empty string | |
| } else | |
| break; | |
| } | |
| } | } |
| } | } |
| Line 222 static int CORD_batched_iter_fn_generic_ | Line 360 static int CORD_batched_iter_fn_generic_ |
| generic_hash_code(result, c); | generic_hash_code(result, c); |
| return 0; | return 0; |
| } | } |
| static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) { | static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) { |
| uint& result=*static_cast<uint*>(client_data); | uint& result=*static_cast<uint*>(client_data); |
| generic_hash_code(result, s); | generic_hash_code(result, s); |
| return 0; | return 0; |
| } | |
| uint String::Body::get_hash_code() const { | |
| #ifdef HASH_CODE_CACHING | |
| if(hash_code) | |
| return hash_code; | |
| #else | |
| uint hash_code=0; | |
| #endif | |
| if (body && CORD_IS_STRING(body)){ | |
| generic_hash_code(hash_code, (const char *)body); | |
| } else { | |
| CORD_iter5(body, 0, | |
| CORD_batched_iter_fn_generic_hash_code, | |
| CORD_batched_iter_fn_generic_hash_code, &hash_code); | |
| } | |
| return hash_code; | |
| } | |
| struct CORD_pos_info { | |
| const char* chars; | |
| size_t left; | |
| size_t pos; | |
| }; | }; |
| uint String::Body::hash_code() const { | |
| uint result=0; | // can be called only for IS_FUNCTION(CORD) which is used in String::Body::strrpbrk |
| CORD_iter5(body, 0, | static int CORD_iter_fn_rpos(char c, CORD_pos_info* info) { |
| CORD_batched_iter_fn_generic_hash_code, | if(info->pos < info->left){ |
| CORD_batched_iter_fn_generic_hash_code, &result); | info->pos=STRING_NOT_FOUND; |
| return result; | return 1; |
| } | |
| if(strchr(info->chars, c)) | |
| return 1; | |
| --(info->pos); | |
| return 0; | |
| } | |
| size_t String::Body::strrpbrk(const char* chars, size_t left, size_t right) const { | |
| if(is_empty() || !chars || !strlen(chars)) | |
| return STRING_NOT_FOUND; | |
| CORD_pos_info info={chars, left, right}; | |
| if(CORD_riter4(body, right, (CORD_iter_fn)CORD_iter_fn_rpos, &info)) | |
| return info.pos; | |
| else | |
| return STRING_NOT_FOUND; | |
| } | } |
| // String methods | |
| String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) { | // can be called only for IS_FUNCTION(CORD) which is used in String::Body::rskipchars |
| append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN); | static int CORD_iter_fn_rskip(char c, CORD_pos_info* info) { |
| if(info->pos < info->left) { | |
| info->pos=STRING_NOT_FOUND; | |
| return 1; | |
| } | |
| if(!strchr(info->chars, c)) | |
| return 1; | |
| --(info->pos); | |
| return 0; | |
| } | } |
| String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) { | |
| append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN); | size_t String::Body::rskipchars(const char* chars, size_t left, size_t right) const { |
| if(is_empty() || !chars || !strlen(chars)) | |
| return STRING_NOT_FOUND; | |
| CORD_pos_info info={chars, left, right}; | |
| if(CORD_riter4(body, right, (CORD_iter_fn)CORD_iter_fn_rskip, &info)) | |
| return info.pos; | |
| else | |
| return STRING_NOT_FOUND; | |
| } | } |
| // String methods | |
| String& String::append_know_length(const char* str, size_t known_length, Language lang) { | String& String::append_know_length(const char* str, size_t known_length, Language lang) { |
| if(!known_length) | if(!known_length) |
| return *this; | return *this; |
| Line 256 String& String::append_know_length(const | Line 449 String& String::append_know_length(const |
| ASSERT_STRING_INVARIANT(*this); | ASSERT_STRING_INVARIANT(*this); |
| return *this; | return *this; |
| } | } |
| String& String::append_help_length(const char* str, size_t helper_length, Language lang) { | String& String::append_help_length(const char* str, size_t helper_length, Language lang) { |
| if(!str) | if(!str) |
| return *this; | return *this; |
| Line 265 String& String::append_help_length(const | Line 459 String& String::append_help_length(const |
| return append_know_length(str, known_length, lang); | return append_know_length(str, known_length, lang); |
| } | } |
| String::String(int value, const char *format) : langs(L_CLEAN){ | |
| char buf[MAX_NUMBER]; | |
| body.append_strdup_know_length(buf, snprintf(buf, MAX_NUMBER, format, value)); | |
| } | |
| String& String::append_strdup(const char* str, size_t helper_length, Language lang) { | String& String::append_strdup(const char* str, size_t helper_length, Language lang) { |
| size_t known_length=helper_length?helper_length:strlen(str); | size_t known_length=helper_length?helper_length:strlen(str); |
| if(!known_length) | if(!known_length) |
| Line 279 String& String::append_strdup(const char | Line 479 String& String::append_strdup(const char |
| return *this; | return *this; |
| } | } |
| struct CORD_length_info { | |
| size_t len; | |
| size_t skip; | |
| }; | |
| int CORD_batched_len(const char* s, CORD_length_info* info){ | |
| info->len += lengthUTF8( (const XMLByte *)s, (const XMLByte *)s+strlen(s)); | |
| return 0; | |
| } | |
| // can be called only for IS_FUNCTION(CORD) which are used in large String::Body::mid | |
| int CORD_batched_len(const char c, CORD_length_info* info){ | |
| if (info->skip==0){ | |
| info->len++; | |
| info->skip = lengthUTF8Char(c)-1; | |
| } else { | |
| info->skip--; | |
| } | |
| return 0; | |
| } | |
| size_t String::length(Charset& charset) const { | |
| if(charset.isUTF8()){ | |
| CORD_length_info info = {0, 0}; | |
| body.for_each<CORD_length_info *>(CORD_batched_len, CORD_batched_len, &info); | |
| return info.len; | |
| } else | |
| return body.length(); | |
| } | |
| /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab" | /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab" |
| String& String::mid(size_t substr_begin, size_t substr_end) const { | String& String::mid(size_t substr_begin, size_t substr_end) const { |
| String& result=*new String; | String& result=*new String; |
| Line 299 String& String::mid(size_t substr_begin, | Line 529 String& String::mid(size_t substr_begin, |
| return result; | return result; |
| } | } |
| // from, to and helper_length in characters, not in bytes (it's important for utf-8) | |
| String& String::mid(Charset& charset, size_t from, size_t to, size_t helper_length) const { | |
| String& result=*new String; | |
| size_t self_length=helper_length ? helper_length : length(charset); | |
| if(!self_length) | |
| return result; | |
| from=min(min(to, from), self_length); | |
| to=min(max(to, from), self_length); | |
| size_t substr_length=to-from; | |
| if(!substr_length) | |
| return result; | |
| if(charset.isUTF8()){ | |
| const XMLByte* src_begin=(const XMLByte*)cstr(); | |
| const XMLByte* src_end=src_begin+body.length(); | |
| // convert 'from' and 'substr_length' from 'characters' to 'bytes' | |
| from=getUTF8BytePos(src_begin, src_end, from); | |
| substr_length=getUTF8BytePos(src_begin+from, src_end, substr_length); | |
| if(!substr_length) | |
| return result; | |
| } | |
| // first: their langs | |
| result.langs.append(result.body, langs, from, substr_length); | |
| // next: letters themselves | |
| result.body=body.mid(from, substr_length); | |
| ASSERT_STRING_INVARIANT(result); | |
| return result; | |
| } | |
| size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { | size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { |
| size_t substr_length=substr.length(); | size_t substr_length=substr.length(); |
| while(true) { | while(true) { |
| Line 314 size_t String::pos(const String::Body su | Line 581 size_t String::pos(const String::Body su |
| } | } |
| } | } |
| size_t String::pos(const String& substr, | size_t String::pos(const String& substr, size_t this_offset, Language lang) const { |
| size_t this_offset, Language lang) const { | |
| return pos(substr.body, this_offset, lang); | return pos(substr.body, this_offset, lang); |
| } | } |
| void String::split(ArrayString& result, | size_t String::pos(Charset& charset, const String& substr, size_t this_offset, Language lang) const { |
| size_t& pos_after, | |
| const char* delim, | if(charset.isUTF8()){ |
| Language lang, int limit) const { | const XMLByte* srcPtr=(const XMLByte*)cstr(); |
| const XMLByte* srcEnd=srcPtr+body.length(); | |
| // convert 'this_offset' from 'characters' to 'bytes' | |
| this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset); | |
| size_t result=pos(substr.body, this_offset, lang); | |
| return (result==CORD_NOT_FOUND) | |
| ? STRING_NOT_FOUND | |
| : getUTF8CharPos(srcPtr, srcEnd, result); // convert 'result' from 'bytes' to 'characters' | |
| } else { | |
| size_t result=pos(substr.body, this_offset, lang); | |
| return (result==CORD_NOT_FOUND) | |
| ? STRING_NOT_FOUND | |
| : result; | |
| } | |
| } | |
| void String::split(ArrayString& result, size_t pos_after, const char* delim, Language lang) const { | |
| if(is_empty()) | |
| return; | |
| size_t self_length=length(); | size_t self_length=length(); |
| if(size_t delim_length=strlen(delim)) { | if(size_t delim_length=strlen(delim)) { |
| size_t pos_before; | size_t pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) { | while((pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND) { |
| result+=&mid(pos_after, pos_before); | result+=&mid(pos_after, pos_before); |
| pos_after=pos_before+delim_length; | pos_after=pos_before+delim_length; |
| } | } |
| // last piece | // last piece |
| if(pos_after<self_length && limit) { | if(pos_after<self_length) |
| result+=&mid(pos_after, self_length); | result+=&mid(pos_after, self_length); |
| pos_after=self_length; | |
| } | |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=this; |
| pos_after+=self_length; | |
| } | } |
| } | } |
| void String::split(ArrayString& result, | void String::split(ArrayString& result, size_t pos_after, const String& delim, Language lang) const { |
| size_t& pos_after, | if(is_empty()) |
| const String& delim, Language lang, | return; |
| int limit) const { | |
| if(!delim.is_empty()) { | if(!delim.is_empty()) { |
| size_t pos_before; | size_t pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) { | while((pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND) { |
| result+=&mid(pos_after, pos_before); | result+=&mid(pos_after, pos_before); |
| pos_after=pos_before+delim.length(); | pos_after=pos_before+delim.length(); |
| } | } |
| // last piece | // last piece |
| if(pos_after<length() && limit) { | if(pos_after<length()) |
| result+=&mid(pos_after, length()); | result+=&mid(pos_after, length()); |
| pos_after=length(); | |
| } | |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=this; |
| pos_after+=length(); | |
| } | } |
| } | } |
| static void regex_options(const String* options, int *result, bool& need_pre_post_match){ | Table* String::match(VRegex* vregex, Row_action row_action, void *info, int& matches_count) const { |
| struct Regex_option { | |
| const char* keyL; | |
| const char* keyU; | |
| int clear, set; | |
| int *result; | |
| bool *flag; | |
| } regex_option[]={ | |
| {"i", "I", 0, PCRE_CASELESS, result, 0}, // a=A | |
| {"s", "S", 0, PCRE_DOTALL, result, 0}, // \n\n$ [default] | |
| {"x", "U", 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored | |
| {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$ | |
| {"g", "G", 0, 1, result+1, 0}, // many rows | |
| {"'", 0, 0, 0, 0, &need_pre_post_match}, | |
| {0, 0, 0, 0, 0, 0} | |
| }; | |
| result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY; | |
| result[1]=0; | |
| if(options && !options->is_empty()) | |
| for(Regex_option *o=regex_option; o->keyL; o++) | |
| if(options->pos(o->keyL)!=STRING_NOT_FOUND | |
| || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) { | |
| if(o->flag) | |
| *o->flag=true; | |
| else { // result | |
| *o->result &= ~o->clear; | |
| *o->result |= o->set; | |
| } | |
| } | |
| } | |
| Table* String::match(Charset& source_charset, | // vregex->info(); // I have no idea what does it for? |
| const String& regexp, | |
| const String* options, | bool need_pre_post_match=vregex->is_pre_post_match_needed(); |
| Row_action row_action, void *info, | bool global=vregex->is_global_search(); |
| bool& just_matched) const { | |
| if(regexp.is_empty()) | |
| throw Exception(0, | |
| 0, | |
| "regexp is empty"); | |
| const char* pattern=regexp.cstr(); | |
| const char* errptr; | |
| int erroffset; | |
| bool need_pre_post_match=false; | |
| int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match); | |
| bool global=option_bits[1]!=0; | |
| pcre *code=pcre_compile(pattern, option_bits[0], | |
| &errptr, &erroffset, | |
| source_charset.pcre_tables); | |
| if(!code) | |
| throw Exception(0, | |
| ®exp.mid(erroffset, regexp.length()), | |
| "regular expression syntax error - %s", errptr); | |
| int subpatterns=pcre_info(code, 0, 0); | |
| if(subpatterns<0) { | |
| pcre_free(code); | |
| throw Exception(0, | |
| ®exp, | |
| "pcre_info error (%d)", | |
| subpatterns); | |
| } | |
| const char* subject=cstr(); | const char* subject=cstr(); |
| size_t subject_length=strlen(subject); | size_t subject_length=length(); |
| const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3; | const int ovector_size=(1/*match*/+MAX_MATCH_GROUPS)*3; |
| int ovector[oveclength]; | int ovector[ovector_size]; |
| // create table | |
| Table::Action_options table_options; | Table::Action_options table_options; |
| Table& table=*new Table(string_match_table_template, table_options); | Table& table=*new Table(string_match_table_template, table_options); |
| int exec_option_bits=0; | |
| int prestart=0; | int prestart=0; |
| int poststart=0; | int poststart=0; |
| int postfinish=length(); | int postfinish=length(); |
| int action_was_executed=-1; | |
| while(true) { | while(true) { |
| int exec_substrings=pcre_exec(code, 0, | int exec_result=vregex->exec(subject, subject_length, ovector, ovector_size, prestart); |
| subject, subject_length, prestart, | |
| exec_option_bits, ovector, oveclength); | |
| if(exec_substrings==PCRE_ERROR_NOMATCH) { | |
| pcre_free(code); | |
| row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); | |
| if(global || subpatterns) | |
| return &table; // global or with subpatterns=true+result | |
| else { | |
| just_matched=false; return 0; // not global=no result | |
| } | |
| } | |
| if(exec_substrings<0) { | if(exec_result<0) // only PCRE_ERROR_NOMATCH might be here, other negative results cause an exception |
| pcre_free(code); | break; |
| throw Exception(0, | |
| ®exp, | |
| "regular expression execute error (%d)", | |
| exec_substrings); | |
| } | |
| int prefinish=ovector[0]; | int prefinish=ovector[0]; |
| poststart=ovector[1]; | poststart=ovector[1]; |
| ArrayString* row=new ArrayString; | |
| if (prestart==poststart && action_was_executed==1){ | |
| prestart++; | |
| action_was_executed=0; | |
| continue; | |
| } | |
| ArrayString* row=new ArrayString(3); | |
| if(need_pre_post_match) { | if(need_pre_post_match) { |
| *row+=&mid(0, prefinish); // .prematch column value | *row+=&mid(0, prefinish); // .prematch column value |
| *row+=&mid(prefinish, poststart); // .match | *row+=&mid(prefinish, poststart); // .match |
| Line 479 Table* String::match(Charset& source_cha | Line 688 Table* String::match(Charset& source_cha |
| *row+=&Empty; // .postmatch | *row+=&Empty; // .postmatch |
| } | } |
| for(int i=1; i<exec_substrings; i++) { | for(int i=1; i<exec_result; i++) { |
| // -1:-1 case handled peacefully by mid() itself | // -1:-1 case handled peacefully by mid() itself |
| *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value | *row+=(ovector[i*2+0]>=0 && ovector[i*2+1]>0)?&mid(ovector[i*2+0], ovector[i*2+1]):new String; // .i column value |
| } | } |
| row_action(table, row, prestart, prefinish, poststart, postfinish, info); | matches_count++; |
| row_action(table, row, prestart - !action_was_executed, prefinish, poststart, postfinish, info); | |
| if(!global || prestart==poststart) { // not global | going to hang | if(!global) // last step |
| pcre_free(code); | break; |
| row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); | |
| return &table; | |
| } | |
| prestart=poststart; | |
| /* | prestart=poststart; |
| if(option_bits[0] & PCRE_MULTILINE) | action_was_executed=1; |
| exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL | |
| */ | |
| } | } |
| row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); | |
| return vregex->is_just_count() ? 0 : &table; | |
| } | } |
| String& String::change_case(Charset& source_charset, Change_case_kind kind) const { | String& String::change_case(Charset& source_charset, Change_case_kind kind) const { |
| Line 506 String& String::change_case(Charset& sou | Line 713 String& String::change_case(Charset& sou |
| return result; | return result; |
| char* new_cstr=cstrm(); | char* new_cstr=cstrm(); |
| size_t new_cstr_len=length(); | |
| if(source_charset.isUTF8()) { | if(source_charset.isUTF8()) { |
| size_t new_cstr_len=length(); | |
| switch(kind) { | switch(kind) { |
| case CC_UPPER: | case CC_UPPER: |
| change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper); | change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper); |
| Line 556 String& String::change_case(Charset& sou | Line 764 String& String::change_case(Charset& sou |
| return result; | return result; |
| } | } |
| const String& String::escape(Charset& source_charset) const { | |
| if(is_empty()) | |
| return *this; | |
| return Charset::escape(*this, source_charset); | |
| } | |
| #define STRING_APPEND(result, from_cstr, langs, langs_offset, length) \ | |
| result.langs.append(result.body, langs, langs_offset, length); \ | |
| result.body.append_strdup_know_length(from_cstr, length); | |
| const String& String::replace(const Dictionary& dict) const { | const String& String::replace(const Dictionary& dict) const { |
| if(!dict.count() || is_empty()) | |
| return *this; | |
| String& result=*new String(); | String& result=*new String(); |
| const char* old_cstr=cstr(); | const char* old_cstr=cstr(); |
| const char* prematch_begin=old_cstr; | const char* prematch_begin=old_cstr; |
| const char* current=old_cstr; | if(dict.count()==1) { |
| while(*current) { | // optimized simple case |
| if(Dictionary::Subst subst=dict.first_that_begins(current)) { | |
| Dictionary::Subst subst=dict.get(0); | |
| while(const char* p=strstr(prematch_begin, subst.from)) { | |
| // prematch | // prematch |
| if(size_t prematch_length=current-prematch_begin) { | if(size_t prematch_length=p-prematch_begin) { |
| result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length); | STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, prematch_length) |
| result.body.append_strdup_know_length(prematch_begin, prematch_length); | |
| } | } |
| // match | // match |
| // skip 'a' in 'current'; move prematch_begin | prematch_begin=p+subst.from_length; |
| current+=subst.from_length; prematch_begin=current; | |
| if(const String* b=subst.to) // are there any b? | if(const String* b=subst.to) // are there any b? |
| result<<*b; | result<<*b; |
| } else // simply advance | } |
| current++; | |
| } else { | |
| const char* current=old_cstr; | |
| while(*current) { | |
| if(Dictionary::Subst subst=dict.first_that_begins(current)) { | |
| // prematch | |
| if(size_t prematch_length=current-prematch_begin) { | |
| STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, prematch_length) | |
| } | |
| // match | |
| // skip 'a' in 'current'; move prematch_begin | |
| current+=subst.from_length; prematch_begin=current; | |
| if(const String* b=subst.to) // are there any b? | |
| result<<*b; | |
| } else // simply advance | |
| current++; | |
| } | |
| } | } |
| if(prematch_begin==old_cstr) // not modified | |
| return *this; | |
| // postmatch | // postmatch |
| if(size_t postmatch_length=current-prematch_begin) { | if(size_t postmatch_length=old_cstr+length()-prematch_begin) { |
| result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length); | STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, postmatch_length) |
| result.body.append_strdup_know_length(prematch_begin, postmatch_length); | |
| } | } |
| ASSERT_STRING_INVARIANT(result); | ASSERT_STRING_INVARIANT(result); |
| Line 593 const String& String::replace(const Dict | Line 837 const String& String::replace(const Dict |
| static int serialize_body_char(char c, char** cur) { | static int serialize_body_char(char c, char** cur) { |
| *((*cur)++)=c; | *((*cur)++)=c; |
| return 0; // 0=continue | return 0; // 0=continue |
| }; | } |
| static int serialize_body_piece(const char* s, char** cur) { | static int serialize_body_piece(const char* s, char** cur) { |
| size_t length=strlen(s); | size_t length=strlen(s); |
| memcpy(*cur, s, length); *cur+=length; | memcpy(*cur, s, length); *cur+=length; |
| return 0; // 0=continue | return 0; // 0=continue |
| }; | } |
| static int serialize_lang_piece(char alang, size_t asize, char** cur) { | static int serialize_lang_piece(char alang, size_t asize, char** cur) { |
| // lang | // lang |
| **cur=alang; (*cur)++; | **cur=alang; (*cur)++; |
| Line 607 static int serialize_lang_piece(char ala | Line 853 static int serialize_lang_piece(char ala |
| return 0; // 0=continue | return 0; // 0=continue |
| } | } |
| String::Cm String::serialize(size_t prolog_length) const { | String::Cm String::serialize(size_t prolog_length) const { |
| size_t fragments_count=langs.count(); | size_t fragments_count=langs.count(); |
| size_t body_length=body.length(); | |
| size_t buf_length= | size_t buf_length= |
| prolog_length //1 | prolog_length //1 |
| +sizeof(size_t) //2 | +sizeof(size_t) //2 |
| +fragments_count*(sizeof(char)+sizeof(size_t)) //3 | +body_length //3 |
| +body.length() //4 | +1 // 4 for zero terminator used in deserialize |
| +1; // for zero terminator used in deserialize | +sizeof(size_t) //5 |
| +fragments_count*(sizeof(char)+sizeof(size_t)); //6 | |
| String::Cm result(new(PointerFreeGC) char[buf_length], buf_length); | String::Cm result(new(PointerFreeGC) char[buf_length], buf_length); |
| // 1: prolog | // 1: prolog |
| char *cur=result.str+prolog_length; | char *cur=result.str+prolog_length; |
| // 2: langs.count [WARNING: not cast, addresses must be %4=0 on sparc] | // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc] |
| memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length); | |
| // 3: letters | |
| body.for_each(serialize_body_char, serialize_body_piece, &cur); | |
| // 4: zero terminator | |
| *cur++=0; | |
| // 5: langs.count [WARNING: not cast, addresses must be %4=0 on sparc] | |
| memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count); | memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count); |
| // 3: lang info | // 6: lang info |
| langs.for_each(body, serialize_lang_piece, &cur); | langs.for_each(body, serialize_lang_piece, &cur); |
| // 4: letters | |
| body.for_each(serialize_body_char, serialize_body_piece, &cur); | |
| // 5: zero terminator | |
| *cur=0; | |
| return result; | return result; |
| } | } |
| bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) { | |
| if(buf_length<=prolog_length) | bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) { |
| size_t in_buf=buf_size; | |
| if(in_buf<=prolog_size) | |
| return false; | return false; |
| buf_length-=prolog_length; | in_buf-=prolog_size; |
| buf_length-=1; // 5: zero terminator | |
| // 1: prolog | // 1: prolog |
| const char* cur=(const char* )buf+prolog_length; | const char* cur=(const char* )buf+prolog_size; |
| // 2: chars.count | |
| size_t body_length; | |
| if(in_buf<sizeof(body_length)) // body.length don't fit? | |
| return false; | |
| // [WARNING: not cast, addresses must be %4=0 on sparc] | |
| memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length); | |
| in_buf-=sizeof(body_length); | |
| if(in_buf<body_length+1) // letters+terminator don't fit? | |
| return false; | |
| // 4: zero terminator | |
| if(cur[body_length] != 0) // in place? | |
| return false; | |
| // 3: letters | |
| body=String::Body(String::C(cur, body_length)); | |
| cur+=body_length+1; | |
| in_buf-=body_length+1; | |
| // 2: langs.count | // 5: langs.count |
| size_t fragments_count; | size_t fragments_count; |
| if(buf_length<sizeof(fragments_count)) // langs.count don't fit? | if(in_buf<sizeof(fragments_count)) // langs.count don't fit? |
| return false; | return false; |
| // [WARNING: not cast, addresses must be %4=0 on sparc] | // [WARNING: not cast, addresses must be %4=0 on sparc] |
| memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count); | memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count); |
| buf_length-=sizeof(fragments_count); | in_buf-=sizeof(fragments_count); |
| if(fragments_count) { | if(fragments_count) { |
| // 3: lang info | // 6: lang info |
| size_t total_length=0; | size_t total_length=0; |
| for(size_t f=0; f<fragments_count; f++) { | for(size_t f=0; f<fragments_count; f++) { |
| char lang; | char lang; |
| size_t fragment_length; | size_t fragment_length; |
| size_t piece_length=sizeof(lang)+sizeof(fragment_length); | size_t piece_length=sizeof(lang)+sizeof(fragment_length); |
| if(buf_length<piece_length) // lang+length | if(in_buf<piece_length) // lang+length |
| return false; | return false; |
| // lang | // lang |
| Line 662 bool String::deserialize(size_t prolog_l | Line 933 bool String::deserialize(size_t prolog_l |
| // length [WARNING: not cast, addresses must be %4=0 on sparc] | // length [WARNING: not cast, addresses must be %4=0 on sparc] |
| memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length); | memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length); |
| size_t combined_length=total_length+fragment_length; | |
| if(combined_length>body_length) | |
| return false; // file curruption | |
| // uchar needed to prevent propagating 0x80 bit to upper bytes | // uchar needed to prevent propagating 0x80 bit to upper bytes |
| langs.append(total_length, (String::Language)(uchar)lang, fragment_length); | langs.append(total_length, (String::Language)(uchar)lang, fragment_length); |
| total_length+=fragment_length; | total_length=combined_length; |
| in_buf-=piece_length; | |
| buf_length-=piece_length; | |
| } | } |
| // 4: letters | if(total_length!=body_length) // length(all language fragments) vs length(letters) |
| if(buf_length!=total_length) | |
| return false; | return false; |
| // serialize wrote extra zero byte there, we can rely on that | |
| body=String::Body(cur, buf_length); | |
| } | } |
| if(in_buf!=0) // some strange extra bytes | |
| return false; | |
| ASSERT_STRING_INVARIANT(*this); | ASSERT_STRING_INVARIANT(*this); |
| return true; | return true; |
| } | } |
| const char* String::Body::v() const { | |
| return CORD_to_const_char_star(body); | |
| } | |
| void String::Body::dump() const { | void String::Body::dump() const { |
| CORD_dump(body); | CORD_dump(body); |
| } | } |
| const char* String::Languages::v() const { | const char* String::Languages::visualize() const { |
| if(opt.is_not_just_lang) | if(opt.is_not_just_lang) |
| return CORD_to_const_char_star(langs); | return CORD_to_const_char_star(langs, 0); |
| else | else |
| return (const char*)&langs; | return 0; |
| } | } |
| void String::Languages::dump() const { | void String::Languages::dump() const { |
| if(opt.is_not_just_lang) | if(opt.is_not_just_lang) |
| CORD_dump(langs); | CORD_dump(langs); |
| else | else |
| puts((const char*)&langs); | puts((const char*)&langs); |
| } | } |
| const char* String::v() const { | |
| const uint LIMIT_VIEW=20; | |
| char* buf=(char*)malloc(MAX_STRING); | |
| const char*body_view=body.v(); | |
| const char*langs_view=langs.v(); | |
| snprintf(buf, MAX_STRING, | |
| "%d:%.*s%s} " | |
| "{%d:%s", | |
| langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"", | |
| strlen(body_view), body_view | |
| ); | |
| return buf; | |
| } | |
| void String::dump() const { | void String::dump() const { |
| body.dump(); | body.dump(); |
| langs.dump(); | langs.dump(); |
| } | } |
| const String& String::trim(String::Trim_kind kind, const char* chars) const { | |
| if(!length()) | static char *n_chars(char c, size_t length){ |
| char *result=(char *)pa_malloc_atomic(length+1); | |
| memset(result, c, length); | |
| result[length] = '\0'; | |
| return result; | |
| } | |
| char* String::visualize_langs() const { | |
| return is_not_just_lang() ? pa_strdup(langs.visualize()) : n_chars((char)just_lang(), length()); | |
| } | |
| const String& String::trim(String::Trim_kind kind, const char* chars, Charset* source_charset) const { | |
| if(is_empty()) | |
| return *this; | return *this; |
| size_t substr_begin, substr_length; | size_t substr_begin, substr_length; |
| Body new_body=body.trim(kind, chars, &substr_begin, &substr_length); | Body new_body=body.trim(kind, chars, &substr_begin, &substr_length, source_charset); |
| if(new_body==body) // we received unchanged pointer, do likewise | if(new_body==body) // we received unchanged pointer, do likewise |
| return *this; | return *this; |
| // new_body differs from body, adjust langs along | // new_body differs from body, adjust langs along |