|
|
| version 1.172.2.21.2.23, 2003/03/24 15:26:03 | version 1.219, 2009/04/10 11:33:16 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class. @see untalength_t.C. | Parser: string class. @see untalength_t.C. |
| Copyright (c) 2001-2003 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: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| */ | */ |
| static const char* IDENT_STRING_C="$Date$"; | 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" |
| Line 15 static const char* IDENT_STRING_C="$Date | Line 13 static const char* IDENT_STRING_C="$Date |
| #include "pa_dictionary.h" | #include "pa_dictionary.h" |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #include "pcre.h" | |
| const String String::Empty; | |
| int pa_atoi(const char* str, const String* problem_source) { | |
| if(!str) | |
| return 0; | |
| while(*str && isspace((unsigned char)*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| int result; | |
| char *error_pos; | |
| bool negative=false; | |
| if(str[0]=='-') { | |
| negative=true; | |
| str++; | |
| } else if(str[0]=='+') { | |
| 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++) | |
| if(!isspace((unsigned char)c)) | |
| throw Exception("number.format", | |
| problem_source, | |
| problem_source?"invalid number (int)": "'%s' is invalid number (int)", str); | |
| return result; | |
| } | |
| double pa_atod(const char* str, const String* problem_source) { | |
| if(!str) | |
| return 0; | |
| while(*str && isspace((unsigned char)*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| double result; | |
| char *error_pos; | |
| bool negative=false; | |
| if(str[0]=='-') { | |
| negative=true; | |
| str++; | |
| } else if(str[0]=='+') { | |
| str++; | |
| } | |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(double)(unsigned long)strtol(str, &error_pos, 0); | |
| else { | |
| // skip leading 0000, to disable octal interpretation | |
| do str++; while(*str=='0'); | |
| result=(double)strtod(str, &error_pos); | |
| } | |
| else | |
| result=(double)strtod(str, &error_pos); | |
| if(negative) | |
| result=-result; | |
| while(char c=*error_pos++) | |
| if(!isspace((unsigned char)c)) | |
| throw Exception("number.format", | |
| problem_source, | |
| problem_source?"invalid number (double)": "'%s' is invalid number (double)", str); | |
| return result; | |
| } | |
| // cord lib extension | |
| #ifndef DOXYGEN | |
| typedef struct { | |
| ssize_t countdown; | |
| int target; /* Character we're looking for */ | |
| } chr_data; | |
| #endif | |
| 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; | |
| if (d -> countdown<=0) return(2); | |
| d -> countdown -= size; | |
| if (c > d -> target) return(1); | |
| return(0); | |
| } | |
| int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c) | |
| { | |
| chr_data d; | |
| d.countdown = n; | |
| 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'*/); | |
| } | |
| static int CORD_block_count_proc(char /*c*/, size_t /*size*/, void* client_data) | |
| { | |
| int* result=(int*)client_data; | |
| (*result)++; | |
| return(0); // 0=continue | |
| } | |
| size_t CORD_block_count(CORD x) | |
| { | |
| size_t result=0; | |
| CORD_block_iter(x, 0, CORD_block_count_proc, &result); | |
| return result; | |
| } | |
| // helpers | // helpers |
| /// String::match uses this as replace & global search table columns | /// String::match uses this as replace & global search table columns |
| Line 28 public: | Line 151 public: |
| *this+=new String("match"); | *this+=new String("match"); |
| *this+=new String("postmatch"); | *this+=new String("postmatch"); |
| for(int i=0; i<MAX_MATCH_GROUPS; i++) { | for(int i=0; i<MAX_MATCH_GROUPS; i++) { |
| char *cname=new(PointerFreeGC) char[3/*strlen("100")*/+1/*terminating 0*/]; | *this+=new String(String::Body::Format(1+i), String::L_CLEAN); |
| *this+=new String(cname, sprintf(cname, "%d", 1+i)); | |
| } | } |
| } | } |
| }; | }; |
| Table string_match_table_template(new String_match_table_template_columns); | Table string_match_table_template(new String_match_table_template_columns); |
| // String::ArrayFragment methods | // String::Body methods |
| void String::ArrayFragment::append_positions(const ArrayFragment& src, | String::Body String::Body::Format(int value) { |
| size_t substr_begin, size_t substr_end) { | char local[MAX_NUMBER]; |
| if(substr_begin==substr_end) | size_t length=snprintf(local, MAX_NUMBER, "%d", value); |
| return; | return String::Body(pa_strdup(local, length), length); |
| } | |
| size_t fragment_begin=0; | |
| size_t fragment_end; | |
| for(Array_iterator<element_type> i(src); ; fragment_begin=fragment_end) { | |
| const element_type& fragment=i.next(); | |
| fragment_end=fragment_begin+fragment.length; | |
| if(fragment_begin<substr_begin) // not reached fragments which may include 'substr'? | |
| continue; | |
| // found first fragment including piece of 'substr' | |
| size_t piece_end=min(substr_end, fragment_end); | |
| *this+=Fragment(fragment.lang, piece_end-substr_begin); | |
| while(substr_end>fragment_end) { // are there more fragments including pieces of 'substr'? | |
| fragment_begin=fragment_end; | |
| const element_type& fragment=i.next(); | String::Body String::Body::trim(String::Trim_kind kind, const char* chars, |
| fragment_end=fragment_begin+fragment.length; | size_t* out_start, size_t* out_length) const { |
| size_t our_length=length(); | |
| if(!our_length) | |
| return *this; | |
| if(!chars) | |
| chars=" \t\n"; // white space | |
| if(substr_end>fragment_end) // are there still more? | size_t start=0; |
| append(Fragment(fragment.lang, fragment.length)); // appending whole fragment | size_t end=our_length; |
| else { // no, it was last | // from left... |
| append(Fragment(fragment.lang, substr_end-fragment_begin)); | if(kind!=TRIM_END) { |
| return; | CORD_pos pos; set_pos(pos, 0); |
| } | while(true) { |
| char c=CORD_pos_fetch(pos); | |
| if(strchr(chars, c)) { | |
| if(++start==our_length) | |
| return 0; // all chars are empty, just return empty string | |
| } else | |
| break; | |
| CORD_next(pos); | |
| } | } |
| } | |
| // from right.. | |
| if(kind!=TRIM_START) { | |
| CORD_pos pos; set_pos(pos, end-1); | |
| while(true) { | |
| 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; | |
| break; | CORD_prev(pos); |
| } | } |
| } | } |
| /* | |
| void String::ArrayFragment::mid(ArrayFragment& result, size_t substr_begin, size_t substr_end) { | if(start==0 && end==our_length) // nobody moved a thing |
| } | return *this; |
| */ | |
| // StringSimple methods | if(out_start) |
| *out_start=start; | |
| size_t new_length=end-start; | |
| if(out_length) | |
| *out_length=new_length; | |
| return mid(start, new_length); | |
| } | |
| static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) { | static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) { |
| uint& result=*static_cast<uint*>(client_data); | uint& result=*static_cast<uint*>(client_data); |
| Line 91 static int CORD_batched_iter_fn_generic_ | Line 227 static int CORD_batched_iter_fn_generic_ |
| generic_hash_code(result, s); | generic_hash_code(result, s); |
| return 0; | return 0; |
| }; | }; |
| uint StringBody::hash_code() const { | uint String::Body::hash_code() const { |
| uint result=0; | uint result=0; |
| CORD_iter5(body, 0, | CORD_iter5(body, 0, |
| CORD_batched_iter_fn_generic_hash_code, | CORD_batched_iter_fn_generic_hash_code, |
| Line 102 uint StringBody::hash_code() const { | Line 238 uint StringBody::hash_code() const { |
| // String methods | // String methods |
| String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) { | String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) { |
| append(cstr, helper_length, tainted?L_TAINTED:L_CLEAN); | append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN); |
| } | |
| String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) { | |
| append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN); | |
| } | } |
| String::String(const String& src): body(src.body), fragments(src.fragments) {} | String& String::append_know_length(const char* str, size_t known_length, Language lang) { |
| if(!known_length) | |
| return *this; | |
| String& String::append(const char* str, size_t length, Language lang) { | // first: langs |
| langs.append(body, lang, known_length); | |
| // next: letters themselves | |
| body.append_know_length(str, known_length); | |
| ASSERT_STRING_INVARIANT(*this); | |
| return *this; | |
| } | |
| String& String::append_help_length(const char* str, size_t helper_length, Language lang) { | |
| if(!str) | if(!str) |
| return *this; | return *this; |
| size_t use_length=length?length:strlen(str); | size_t known_length=helper_length?helper_length:strlen(str); |
| if(!use_length) | if(!known_length) |
| return *this; | return *this; |
| body.append(str, use_length); | return append_know_length(str, known_length, lang); |
| fragments+=Fragment(lang, use_length); | } |
| String& String::append_strdup(const char* str, size_t helper_length, Language lang) { | |
| size_t known_length=helper_length?helper_length:strlen(str); | |
| if(!known_length) | |
| return *this; | |
| // first: langs | |
| langs.append(body, lang, known_length); | |
| // next: letters themselves | |
| body.append_strdup_know_length(str, known_length); | |
| ASSERT_STRING_INVARIANT(*this); | |
| return *this; | return *this; |
| } | } |
| size_t String::length(Charset& charset) const { | |
| if(charset.isUTF8()){ | |
| const XMLByte* srcPtr=(const XMLByte*)cstrm(); | |
| return lengthUTF8(srcPtr, srcPtr+body.length()); | |
| } 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; |
| size_t self_length=length(); | size_t self_length=length(); |
| substr_begin=max(min(substr_begin, self_length), (size_t)0); | substr_begin=min(substr_begin, self_length); |
| substr_end=min(max(substr_end, substr_begin), self_length); | substr_end=min(max(substr_end, substr_begin), self_length); |
| if(substr_begin==substr_end) | size_t substr_length=substr_end-substr_begin; |
| if(!substr_length) | |
| return result; | return result; |
| // first: letters themselves | // first: their langs |
| result.body=body.mid(substr_begin, substr_end-substr_begin); | result.langs.append(result.body, langs, substr_begin, substr_length); |
| // next: letters themselves | |
| result.body=body.mid(substr_begin, substr_length); | |
| ASSERT_STRING_INVARIANT(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; | |
| // next: their langs | size_t self_length=(helper_length)?helper_length:length(charset); |
| result.fragments.append_positions(fragments, substr_begin, substr_end); | |
| // SAPI::log("piece of '%s' from %d to %d is '%s'", | if(!self_length) |
| //cstr(), substr_begin, substr_end, result.cstr()); | 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* srcPtr=(const XMLByte*)cstrm(); | |
| const XMLByte* srcEnd=srcPtr+body.length(); | |
| // convert 'from' and 'substr_length' from 'characters' to 'bytes' | |
| from=getUTF8BytePos(srcPtr, srcEnd, from); | |
| substr_length=getUTF8BytePos(srcPtr+from, srcEnd, 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; | return result; |
| } | } |
| size_t String::pos(const StringBody substr, | size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { |
| size_t this_offset, Language lang) const { | size_t substr_length=substr.length(); |
| // first: letters themselves | while(true) { |
| size_t substr_begin=body.pos(substr, this_offset); | size_t substr_begin=body.pos(substr, this_offset); |
| if(substr_begin==CORD_NOT_FOUND) | |
| return STRING_NOT_FOUND; | |
| // next: check the lang when specified | |
| if(lang==L_UNSPECIFIED) // ignore lang? | |
| return substr_begin; | |
| // substr must be in one fragment, and fragments' lang must = lang | |
| size_t substr_end=substr_begin+substr.length(); | |
| size_t fragment_begin=0; | |
| size_t fragment_end; | |
| for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) { | |
| const Fragment& fragment=i.next(); | |
| fragment_end=fragment_begin+fragment.length; | |
| if(substr_begin<fragment_begin) // not reached fragments which may include 'result'? | |
| continue; | |
| if(substr_begin>=fragment_end) // begin of substr OUT of current fragment? | |
| continue; | |
| if(substr_end>fragment_end) // end of substr OUT of current fragment? | if(substr_begin==CORD_NOT_FOUND) |
| throw Exception(0, // (*) see below | return STRING_NOT_FOUND; |
| this, | |
| "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another", | |
| substr.cstr(), this_offset); | |
| if(fragment.lang==lang) | if(langs.check_lang(lang, substr_begin, substr_length)) |
| return substr_begin; | return substr_begin; |
| else { // bad lang... | |
| /// WARNING: this possibly skips assert (*), but it's fast | |
| substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/); | |
| if(substr_begin==CORD_NOT_FOUND) | |
| return STRING_NOT_FOUND; | |
| size_t substr_end=substr_begin+substr.length(); | |
| // and continuing with next fragment | |
| } | |
| } | |
| return STRING_NOT_FOUND; | this_offset=substr_begin+substr_length; |
| } | |
| } | } |
| size_t String::pos(const String& substr, | size_t String::pos(const String& substr, |
| Line 193 size_t String::pos(const String& substr, | Line 364 size_t String::pos(const String& substr, |
| return pos(substr.body, this_offset, lang); | return pos(substr.body, this_offset, lang); |
| } | } |
| size_t String::pos(Charset& charset, const String& substr, | |
| size_t this_offset, Language lang) const { | |
| size_t result; | |
| if(charset.isUTF8()){ | |
| const XMLByte* srcPtr=(const XMLByte*)cstrm(); | |
| const XMLByte* srcEnd=srcPtr+body.length(); | |
| // convert 'this_offset' from 'characters' to 'bytes' | |
| this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset); | |
| result=pos(substr.body, this_offset, lang); | |
| if(result==CORD_NOT_FOUND) | |
| return STRING_NOT_FOUND; | |
| // convert 'result' from 'bytes' to 'characters' | |
| result=getUTF8CharPos(srcPtr, srcEnd, result); | |
| } else { | |
| result=pos(substr.body, this_offset, lang); | |
| if(result==CORD_NOT_FOUND) | |
| return STRING_NOT_FOUND; | |
| } | |
| return result; | |
| } | |
| void String::split(ArrayString& result, | void String::split(ArrayString& result, |
| size_t& pos_after, | size_t& pos_after, |
| const char* delim, | const char* delim, |
| Language lang, int limit) const { | Language lang, int limit) const { |
| size_t self_length=length(); | size_t self_length=length(); |
| if(size_t delim_length=strlen(delim)) { | if(size_t delim_length=strlen(delim)) { |
| int pos_before; | size_t pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) { | for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) { |
| result+=&mid(pos_after, pos_before); | result+=&mid(pos_after, pos_before); |
| pos_after=pos_before+delim_length; | pos_after=pos_before+delim_length; |
| } | } |
| Line 217 void String::split(ArrayString& result, | Line 414 void String::split(ArrayString& result, |
| } | } |
| void String::split(ArrayString& result, | void String::split(ArrayString& result, |
| size_t& pos_after, | size_t& pos_after, |
| const String& delim, Language lang, | const String& delim, Language lang, |
| int limit) const { | int limit) const { |
| if(!delim.is_empty()) { | if(!delim.is_empty()) { |
| int pos_before; | size_t pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) { | for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) { |
| result+=&mid(pos_after, pos_before); | result+=&mid(pos_after, pos_before); |
| pos_after=pos_before+delim.length(); | pos_after=pos_before+delim.length(); |
| } | } |
| Line 238 void String::split(ArrayString& result, | Line 435 void String::split(ArrayString& result, |
| } | } |
| } | } |
| static void regex_options(const String* options, int *result, bool& need_pre_post_match){ | enum Match_feature { |
| MF_NEED_PRE_POST_MATCH = 0x01, | |
| MF_JUST_COUNT_MATCHES = 0x02 | |
| }; | |
| static void regex_options(const String* options, int* result, int* match_features){ | |
| struct Regex_option { | struct Regex_option { |
| const char* keyL; | const char* key; |
| const char* keyU; | const char* keyAlt; |
| int clear, set; | int clear; |
| int set; | |
| int *result; | int *result; |
| bool *flag; | int flag; |
| } regex_option[]={ | } regex_option[]={ |
| {"i", "I", 0, PCRE_CASELESS, result}, // a=A | {"i", "I", 0, PCRE_CASELESS, result, 0}, // a=A |
| {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default] | {"s", "S", 0, PCRE_DOTALL, result, 0}, // ^\n\n$ [default] |
| {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored | {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$ |
| {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$ | {"x", 0, 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored |
| {"g", "G", 0, true, result+1}, // many rows | {"U", 0, 0, PCRE_UNGREEDY, result, 0}, // ungreedy patterns (greedy by default) |
| {"'", 0, 0, 0, 0, &need_pre_post_match}, | {"g", "G", 0, 1, result+1, 0}, // many rows |
| {0} | {"'", 0, 0, 0, 0, MF_NEED_PRE_POST_MATCH}, |
| {"n", 0, 0, 0, 0, MF_JUST_COUNT_MATCHES}, | |
| {0, 0, 0, 0, 0, 0} | |
| }; | }; |
| result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY; | result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY; |
| result[1]=0; | result[1]=0; |
| if(options && !options->is_empty()) | if(options && !options->is_empty()) |
| for(Regex_option *o=regex_option; o->keyL; o++) | for(Regex_option *o=regex_option; o->key; o++) |
| if(options->pos(o->keyL)>=0 | if( |
| || (o->keyU && options->pos(o->keyU)>=0)) { | options->pos(o->key)!=STRING_NOT_FOUND |
| if(o->flag) | || (o->keyAlt && options->pos(o->keyAlt)!=STRING_NOT_FOUND) |
| *o->flag=true; | ){ |
| else { // result | if(o->flag){ |
| (*match_features) |= o->flag; | |
| } else { | |
| *o->result &= ~o->clear; | *o->result &= ~o->clear; |
| *o->result |= o->set; | *o->result |= o->set; |
| } | } |
| Line 271 static void regex_options(const String* | Line 478 static void regex_options(const String* |
| } | } |
| Table* String::match(Charset& source_charset, | Table* String::match(Charset& source_charset, |
| const String& regexp, | const String& regexp, |
| const String* options, | const String* options, |
| Row_action row_action, void *info, | Row_action row_action, void *info, |
| bool& just_matched) const { | int& matches_count) const { |
| if(regexp.is_empty()) | if(regexp.is_empty()) |
| throw Exception(0, | throw Exception(PARSER_RUNTIME, |
| 0, | 0, |
| "regexp is empty"); | "regexp is empty"); |
| const char* pattern=regexp.cstr(); | const char* pattern=regexp.cstr(String::L_UNSPECIFIED); // fix any tainted with L_REGEX |
| const char* errptr; | const char* errptr; |
| int erroffset; | int erroffset; |
| bool need_pre_post_match=false; | int option_bits[2]={0}; |
| int option_bits[2]; regex_options(options, option_bits, need_pre_post_match); | int match_features=0; |
| regex_options(options, option_bits, &match_features); | |
| bool need_pre_post_match=(match_features & MF_NEED_PRE_POST_MATCH) != 0; | |
| bool just_count_matches=(match_features & MF_JUST_COUNT_MATCHES) != 0; | |
| bool global=option_bits[1]!=0; | bool global=option_bits[1]!=0; |
| if(source_charset.isUTF8()){ | |
| // @todo (for UTF-8): check string & pattern and use PCRE_NO_UTF8_CHECK option | |
| option_bits[0]|=PCRE_UTF8; | |
| } | |
| pcre *code=pcre_compile(pattern, option_bits[0], | pcre *code=pcre_compile(pattern, option_bits[0], |
| &errptr, &erroffset, | &errptr, &erroffset, |
| source_charset.pcre_tables); | source_charset.pcre_tables); |
| if(!code) | if(!code) |
| throw Exception(0, | throw Exception(PCRE_EXCEPTION_TYPE, |
| ®exp.mid(erroffset, regexp.length()), | ®exp.mid(erroffset, regexp.length()), |
| "regular expression syntax error - %s", errptr); | "regular expression syntax error - %s", errptr); |
| int subpatterns=pcre_info(code, 0, 0); | int subpatterns=pcre_info(code, 0, 0); |
| if(subpatterns<0) { | if(subpatterns<0) { |
| pcre_free(code); | pcre_free(code); |
| throw Exception(0, | throw Exception(PCRE_EXCEPTION_TYPE, |
| ®exp, | ®exp, |
| "pcre_info error (%d)", | "pcre_info error (%d)", |
| subpatterns); | subpatterns); |
| Line 310 Table* String::match(Charset& source_cha | Line 526 Table* String::match(Charset& source_cha |
| int ovector[oveclength]; | int ovector[oveclength]; |
| // create table | // create table |
| Table& table=*new Table(string_match_table_template); | Table::Action_options table_options; |
| Table& table=*new Table(string_match_table_template, table_options); | |
| int exec_option_bits=0; | int exec_option_bits=0; |
| int prestart=0; | int prestart=0; |
| Line 324 Table* String::match(Charset& source_cha | Line 541 Table* String::match(Charset& source_cha |
| if(exec_substrings==PCRE_ERROR_NOMATCH) { | if(exec_substrings==PCRE_ERROR_NOMATCH) { |
| pcre_free(code); | pcre_free(code); |
| row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); | row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); |
| if(global || subpatterns) | return just_count_matches ? 0 : &table; |
| return &table; // global or with subpatterns=true+result | |
| else { | |
| just_matched=false; return 0; // not global=no result | |
| } | |
| } | } |
| if(exec_substrings<0) { | if(exec_substrings<0) { |
| pcre_free(code); | pcre_free(code); |
| throw Exception(0, | throw Exception(PCRE_EXCEPTION_TYPE, |
| ®exp, | ®exp, |
| "regular expression execute error (%d)", | print_pcre_exec_error_text(exec_substrings), |
| exec_substrings); | exec_substrings); |
| } | } |
| Line 347 Table* String::match(Charset& source_cha | Line 560 Table* String::match(Charset& source_cha |
| *row+=&mid(prefinish, poststart); // .match | *row+=&mid(prefinish, poststart); // .match |
| *row+=&mid(poststart, postfinish); // .postmatch | *row+=&mid(poststart, postfinish); // .postmatch |
| } else { | } else { |
| *row+=0; // .prematch column value | *row+=&Empty; // .prematch column value |
| *row+=0; // .match | *row+=&Empty; // .match |
| *row+=0; // .postmatch | *row+=&Empty; // .postmatch |
| } | } |
| for(int i=1; i<exec_substrings; i++) { | for(int i=1; i<exec_substrings; i++) { |
| Line 357 Table* String::match(Charset& source_cha | Line 570 Table* String::match(Charset& source_cha |
| *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value | *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value |
| } | } |
| matches_count++; | |
| row_action(table, row, prestart, prefinish, poststart, postfinish, info); | row_action(table, row, prestart, prefinish, poststart, postfinish, info); |
| if(!global || prestart==poststart) { // not global | going to hang | if(!global || prestart==poststart) { // last step |
| pcre_free(code); | pcre_free(code); |
| row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); | row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); |
| return &table; | return just_count_matches ? 0 : &table; |
| } | } |
| prestart=poststart; | prestart=poststart; |
| Line 375 Table* String::match(Charset& source_cha | Line 589 Table* String::match(Charset& source_cha |
| String& String::change_case(Charset& source_charset, Change_case_kind kind) const { | String& String::change_case(Charset& source_charset, Change_case_kind kind) const { |
| String& result=*new String(); | String& result=*new String(); |
| if(is_empty()) | |
| return result; | |
| const unsigned char *tables=source_charset.pcre_tables; | char* new_cstr=cstrm(); |
| const unsigned char *a; | if(source_charset.isUTF8()) { |
| const unsigned char *b; | size_t new_cstr_len=length(); |
| switch(kind) { | switch(kind) { |
| case CC_UPPER: | case CC_UPPER: |
| a=tables+lcc_offset; | change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper); |
| b=tables+fcc_offset; | break; |
| break; | case CC_LOWER: |
| case CC_LOWER: | change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower); |
| a=tables+lcc_offset; | break; |
| b=0; | default: |
| break; | assert(!"unknown change case kind"); |
| default: | break; // never |
| throw Exception(0, | } |
| this, | |
| "unknown change case kind #%d", | } else { |
| static_cast<int>(kind)); // never | const unsigned char *tables=source_charset.pcre_tables; |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| char* new_cstr=cstrm(); | const unsigned char *a; |
| char *dest=new_cstr; | const unsigned char *b; |
| unsigned char index; | switch(kind) { |
| for(const char* current=new_cstr; index=(unsigned char)*current; current++) { | case CC_UPPER: |
| unsigned char c=a[index]; | a=tables+lcc_offset; |
| if(b) | b=tables+fcc_offset; |
| c=b[c]; | break; |
| case CC_LOWER: | |
| a=tables+lcc_offset; | |
| b=0; | |
| break; | |
| default: | |
| assert(!"unknown change case kind"); | |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| char *dest=new_cstr; | |
| unsigned char index; | |
| for(const char* current=new_cstr; (index=(unsigned char)*current); current++) { | |
| unsigned char c=a[index]; | |
| if(b) | |
| c=b[c]; | |
| *dest++=(char)c; | *dest++=(char)c; |
| } | |
| } | } |
| result.langs=langs; | |
| result.body=new_cstr; | result.body=new_cstr; |
| result.fragments.append(fragments); | |
| return result; | return result; |
| } | } |
| const String& String::escape(Charset& source_charset) const { | |
| if(is_empty()) | |
| return *this; | |
| return Charset::escape(*this, source_charset); | |
| } | |
| const String& String::replace(const Dictionary& dict) const { | const String& String::replace(const Dictionary& dict) const { |
| 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; |
| for(const char* current=old_cstr; *current; ) { | const char* current=old_cstr; |
| if(Table::element_type row=dict.first_that_begins(current)) { | while(*current) { |
| if(Dictionary::Subst subst=dict.first_that_begins(current)) { | |
| // prematch | // prematch |
| if(size_t prematch_length=current-prematch_begin) { | if(size_t prematch_length=current-prematch_begin) { |
| result.body.append(prematch_begin, prematch_length); | result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length); |
| result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr); | result.body.append_strdup_know_length(prematch_begin, prematch_length); |
| } | } |
| // match | // match |
| const String* a=row->get(0); | |
| // skip 'a' in 'current'; move prematch_begin | // skip 'a' in 'current'; move prematch_begin |
| current+=a->length(); prematch_begin=current; | current+=subst.from_length; prematch_begin=current; |
| if(row->count()>1) { // are there any b? | if(const String* b=subst.to) // are there any b? |
| const String* b=row->get(1); | result<<*b; |
| if(size_t b_length=b->length()) { | |
| result.body.append(b->cstr_to_string_body()); | |
| result.fragments.append(b->fragments); | |
| } | |
| } | |
| } else // simply advance | } else // simply advance |
| current++; | current++; |
| } | } |
| // postmatch | // postmatch |
| if(size_t postmatch_length=current-prematch_begin) { | if(size_t postmatch_length=current-prematch_begin) { |
| result.body.append(prematch_begin, postmatch_length); | result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length); |
| result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr); | result.body.append_strdup_know_length(prematch_begin, postmatch_length); |
| } | } |
| ASSERT_STRING_INVARIANT(result); | |
| return result; | return result; |
| } | } |
| double String::as_double() const { | static int serialize_body_char(char c, char** cur) { |
| double result; | *((*cur)++)=c; |
| const char *str=cstr(); | return 0; // 0=continue |
| }; | |
| while(*str && isspace(*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| char *error_pos; | |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(double)(unsigned long)strtol(str, &error_pos, 0); | |
| else | |
| result=(double)strtod(str+1/*skip leading 0*/, &error_pos); | |
| else | |
| result=(double)strtod(str, &error_pos); | |
| while(char c=*error_pos++) | |
| if(!isspace(c)) | |
| throw Exception("number.format", | |
| this, | |
| "invalid number (double)"); | |
| return result; | |
| } | |
| int String::as_int() const { | |
| int result; | |
| const char *str=cstr(); | |
| while(*str && isspace(*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| char *error_pos; | |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(int)(unsigned long)strtol(str, &error_pos, 0); | |
| else | |
| result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0); | |
| else | |
| result=(int)strtol(str, &error_pos, 0); | |
| while(char c=*error_pos++) | |
| if(!isspace(c)) | |
| throw Exception("number.format", | |
| this, | |
| "invalid number (int)"); | |
| return result; | |
| } | |
| inline void uint2uchars(uint word, uchar *bytes) { | |
| bytes[0]=word&0xFF; | |
| bytes[1]=(word>>8)&0xFF; | |
| bytes[2]=(word>>16)&0xFF; | |
| bytes[3]=(word>>24)&0xFF; | |
| } | |
| inline uint uchars2uint(uchar *bytes) { | |
| return bytes[3]<<24 | |
| | bytes[2]<<16 | |
| | bytes[1]<<8 | |
| | bytes[0]; | |
| } | |
| 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; | return 0; // 0=continue |
| }; | }; |
| String::C String::serialize(size_t prolog_length) const { | static int serialize_lang_piece(char alang, size_t asize, char** cur) { |
| // lang | |
| **cur=alang; (*cur)++; | |
| // length [WARNING: not cast, addresses must be %4=0 on sparc] | |
| memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize); | |
| return 0; // 0=continue | |
| } | |
| String::Cm String::serialize(size_t prolog_length) const { | |
| size_t fragments_count=langs.count(); | |
| size_t body_length=body.length(); | |
| size_t buf_length= | size_t buf_length= |
| prolog_length | prolog_length //1 |
| +fragments.count()*(sizeof(Language)+sizeof(size_t)) | +sizeof(size_t) //2 |
| +length(); | +body_length //3 |
| C result(new(PointerFreeGC) char[buf_length], buf_length); | +1 // 4 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); | |
| // 1: prolog | // 1: prolog |
| char *cur=result.str+prolog_length; | char *cur=result.str+prolog_length; |
| // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc] | |
| memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length); | |
| // 2: fragments.count | // 3: letters |
| size_t fragments_count=fragments.count(); | 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); |
| // 6: lang info | |
| // 3: lang info | langs.for_each(body, serialize_lang_piece, &cur); |
| for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) { | |
| const Fragment& fragment=i.next(); | |
| // lang | |
| memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang); | |
| // length | |
| memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length); | |
| } | |
| // 4: letters | |
| body.for_each(serialize_body_piece, &cur); | |
| return result; | return result; |
| } | } |
| bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) { | bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) { |
| if(buf_length<=prolog_length) | size_t in_buf=buf_size; |
| if(in_buf<=prolog_size) | |
| return false; | return false; |
| buf_length-=prolog_length; | in_buf-=prolog_size; |
| // 1: prolog | // 1: prolog |
| const char* cur=(const char* )buf+prolog_length; | const char* cur=(const char* )buf+prolog_size; |
| // 2: fragments.count | // 2: chars.count |
| if(buf_length<sizeof(size_t)) // fragments.count don't fit? | size_t body_length; |
| if(in_buf<sizeof(body_length)) // body.length don't fit? | |
| return false; | return false; |
| size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t); | // [WARNING: not cast, addresses must be %4=0 on sparc] |
| buf_length-=sizeof(size_t); | memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length); |
| in_buf-=sizeof(body_length); | |
| // 3: lang info | if(in_buf<body_length+1) // letters+terminator don't fit? |
| size_t total_length=0; | return false; |
| for(size_t f=0; f<fragments_count; f++) { | // 4: zero terminator |
| size_t piece_length=sizeof(Language)+sizeof(size_t); | if(cur[body_length] != 0) // in place? |
| if(buf_length<piece_length) // lang+length | return false; |
| // 3: letters | |
| body=String::Body(cur, body_length); | |
| cur+=body_length+1; | |
| in_buf-=body_length+1; | |
| // 5: langs.count | |
| size_t fragments_count; | |
| if(in_buf<sizeof(fragments_count)) // langs.count don't fit? | |
| return false; | |
| // [WARNING: not cast, addresses must be %4=0 on sparc] | |
| memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count); | |
| in_buf-=sizeof(fragments_count); | |
| if(fragments_count) { | |
| // 6: lang info | |
| size_t total_length=0; | |
| for(size_t f=0; f<fragments_count; f++) { | |
| char lang; | |
| size_t fragment_length; | |
| size_t piece_length=sizeof(lang)+sizeof(fragment_length); | |
| if(in_buf<piece_length) // lang+length | |
| return false; | |
| // lang | |
| lang=*cur++; | |
| // length [WARNING: not cast, addresses must be %4=0 on sparc] | |
| 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 | |
| langs.append(total_length, (String::Language)(uchar)lang, fragment_length); | |
| total_length=combined_length; | |
| in_buf-=piece_length; | |
| } | |
| if(total_length!=body_length) // length(all language fragments) vs length(letters) | |
| return false; | return false; |
| } | |
| if(in_buf!=0) // some strange extra bytes | |
| return false; | |
| Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language); | ASSERT_STRING_INVARIANT(*this); |
| size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t); | return true; |
| fragments+=Fragment(lang, fragment_length); | } |
| total_length+=fragment_length; | |
| buf_length-=piece_length; | const char* String::Body::v() const { |
| } | return CORD_to_const_char_star(body); |
| } | |
| void String::Body::dump() const { | |
| CORD_dump(body); | |
| } | |
| // 4: letters | const char* String::Languages::v() const { |
| if(buf_length!=total_length) | if(opt.is_not_just_lang) |
| return false; | return CORD_to_const_char_star(langs); |
| else | |
| return (const char*)&langs; | |
| } | |
| void String::Languages::dump() const { | |
| if(opt.is_not_just_lang) | |
| CORD_dump(langs); | |
| else | |
| 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 { | |
| body.dump(); | |
| langs.dump(); | |
| } | |
| const String& String::trim(String::Trim_kind kind, const char* chars) const { | |
| if(!length()) | |
| return *this; | |
| size_t substr_begin, substr_length; | |
| Body new_body=body.trim(kind, chars, &substr_begin, &substr_length); | |
| if(new_body==body) // we received unchanged pointer, do likewise | |
| return *this; | |
| // new_body differs from body, adjust langs along | |
| String& result=*new String; | |
| if(!new_body) // body.trim produced empty result | |
| return result; | |
| // body.trim produced nonempty result | |
| body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length); | // first: their langs |
| result.langs.append(result.body, langs, substr_begin, substr_length); | |
| // next: letters themselves | |
| result.body=new_body; | |
| return true; | ASSERT_STRING_INVARIANT(result); |
| return result; | |
| } | } |