|
|
| version 1.172.2.21.2.10, 2003/03/20 14:34:01 | version 1.201, 2004/10/07 09:22:01 |
|---|---|
| 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) 2001-2004 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 "pcre.h" |
| Line 15 static const char* IDENT_STRING_C="$Date | Line 15 static const char* IDENT_STRING_C="$Date |
| #include "pa_dictionary.h" | #include "pa_dictionary.h" |
| #include "pa_charset.h" | #include "pa_charset.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; | |
| char 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 |
| String::ArrayFragment& String::ArrayFragment::append(const ArrayFragment& src, | String::Body String::Body::Format(int value) { |
| size_t substr_start, size_t substr_end) { | char local[MAX_NUMBER]; |
| if(substr_start==substr_end) | size_t length=snprintf(local, MAX_NUMBER, "%d", value); |
| return String::Body(pa_strdup(local, length), length); | |
| } | |
| String::Body String::Body::trim(String::Trim_kind kind, const char* chars, | |
| size_t* out_start, size_t* out_length) const { | |
| size_t our_length=length(); | |
| if(!our_length) | |
| return *this; | return *this; |
| if(!chars) | |
| chars=" \t\n"; // white space | |
| size_t fragment_start=0; | size_t start=0; |
| size_t fragment_end; | size_t end=our_length; |
| for(Array_iterator<element_type> i(src); ; fragment_start=fragment_end) { | // from left... |
| const element_type& fragment=i.next(); | if(kind!=TRIM_END) { |
| fragment_end=fragment_start+fragment.length; | CORD_pos pos; set_pos(pos, 0); |
| while(true) { | |
| if(substr_start<fragment_start) // not reached fragments which may include 'substr'? | char c=CORD_pos_fetch(pos); |
| continue; | if(strchr(chars, c)) { |
| if(++start==our_length) | |
| // found first fragment including piece of 'substr' | return 0; // all chars are empty, just return empty string |
| size_t piece_end=min(substr_end, fragment_end); | } else |
| *this+=Fragment(fragment.lang, piece_end-substr_start); | break; |
| while(substr_end>fragment_end) { // are there more fragments including pieces of 'substr'? | CORD_next(pos); |
| const element_type& fragment=i.next(); | |
| fragment_end=fragment_start+fragment.length; | |
| if(substr_end>fragment_end) // are there still more? | |
| simple_append(Fragment(fragment.lang, fragment.length)); // appending whole fragment | |
| else { // no, it was last | |
| simple_append(Fragment(fragment.lang, substr_end-fragment_start)); | |
| goto break2; | |
| } | |
| } | } |
| } | |
| // 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); |
| } | } |
| break2: | } |
| return *this; | if(start==0 && end==our_length) // nobody moved a thing |
| return *this; | |
| if(out_start) | |
| *out_start=start; | |
| size_t new_length=end-start; | |
| if(out_length) | |
| *out_length=new_length; | |
| return mid(start, new_length); | |
| } | } |
| /* | |
| void String::ArrayFragment::mid(ArrayFragment& result, size_t substr_start, size_t substr_end) { | static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) { |
| uint& result=*static_cast<uint*>(client_data); | |
| generic_hash_code(result, c); | |
| return 0; | |
| } | |
| static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) { | |
| uint& result=*static_cast<uint*>(client_data); | |
| generic_hash_code(result, s); | |
| return 0; | |
| }; | |
| uint String::Body::hash_code() const { | |
| uint result=0; | |
| CORD_iter5(body, 0, | |
| CORD_batched_iter_fn_generic_hash_code, | |
| CORD_batched_iter_fn_generic_hash_code, &result); | |
| return result; | |
| } | } |
| */ | |
| // 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(cstr, helper_length, tainted?UL_TAINTED:UL_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) | |
| String& String::append_cstr(const char* heap_variable, size_t helper_length, Untaint_lang lang) { | |
| if(!heap_variable) | |
| return *this; | |
| size_t use_length; | |
| if(helper_length) { | |
| #ifdef _DEBUG | |
| size_t real_length=strlen(heap_variable); | |
| if(helper_length!=real_length) | |
| throw Exception(0, | |
| this, | |
| "bad String::String(real=%d, helper%d", | |
| real_length, helper_length); | |
| #endif | |
| use_length=helper_length; | |
| } else | |
| use_length=strlen(heap_variable); | |
| size_t length=helper_length?helper_length:strlen(heap_variable); | |
| if(!length) | |
| return *this; | return *this; |
| body=CORD_cat_char_star(body, heap_variable, length); | // first: langs |
| fragments+=Fragment(lang, length); | langs.append(body, lang, known_length); |
| // next: letters themselves | |
| body.append_know_length(str, known_length); | |
| 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_copy(const char* auto_variable, size_t helper_length, Untaint_lang lang) { | if(!str) |
| if(!auto_variable) | |
| return *this; | return *this; |
| size_t use_length=helper_length?helper_length:strlen(auto_variable); | size_t known_length=helper_length?helper_length:strlen(str); |
| if(!use_length) | if(!known_length) |
| return *this; | return *this; |
| body=pa_strdup(auto_variable, use_length); | return append_know_length(str, known_length, lang); |
| fragments+=Fragment(lang, use_length); | |
| return *this; | |
| } | } |
| 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; | |
| static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) { | // first: langs |
| uint& result=*static_cast<uint*>(client_data); | langs.append(body, lang, known_length); |
| generic_hash_code(result, s); | // next: letters themselves |
| return 0; | body.append_strdup_know_length(str, known_length); |
| }; | |
| uint String::hash_code() const { | ASSERT_STRING_INVARIANT(*this); |
| uint result=0; | return *this; |
| CORD_iter5(body, 0, 0, CORD_batched_iter_fn_generic_hash_code, &result); | |
| return result; | |
| } | } |
| const String& String::mid(size_t substr_start, size_t substr_end) const { | /// @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& result=*new String; | String& result=*new String; |
| substr_start=min(substr_start, length()); | size_t self_length=length(); |
| substr_end=max(substr_start, substr_end); | substr_begin=min(substr_begin, self_length); |
| if(substr_start==substr_end) | substr_end=min(max(substr_end, substr_begin), self_length); |
| size_t substr_length=substr_end-substr_begin; | |
| if(!substr_length) | |
| return result; | return result; |
| // first: letters themselves | // first: their langs |
| result.body=CORD_substr(body, substr_start, substr_end-substr_start); | result.langs.append(result.body, langs, substr_begin, substr_length); |
| // next: letters themselves | |
| // next: their langs | result.body=body.mid(substr_begin, substr_length); |
| result.fragments.append(fragments, substr_start, substr_end); | |
| // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", | ASSERT_STRING_INVARIANT(result); |
| //cstr(), substr_start, substr_end, result.cstr()); | |
| return result; | return result; |
| } | } |
| size_t String::pos(CORD substr, | size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { |
| size_t this_offset, Untaint_lang lang) const { | size_t substr_length=substr.length(); |
| // first: letters themselves | while(true) { |
| size_t substr_start=CORD_str(body, this_offset, substr); | size_t substr_begin=body.pos(substr, this_offset); |
| if(substr_start==CORD_NOT_FOUND) | |
| return STRING_NOT_FOUND; | |
| // next: check the lang when specified | |
| if(lang==UL_UNSPECIFIED) // ignore lang? | |
| return substr_start; | |
| // substr must be in one fragment, and fragments' lang must = lang | |
| size_t substr_end=substr_start+CORD_len(substr); | |
| size_t fragment_start=0; | |
| size_t fragment_end; | |
| for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_start=fragment_end) { | |
| const Fragment& fragment=i.next(); | |
| fragment_end=fragment_start+fragment.length; | |
| if(substr_start<fragment_start) // not reached fragments which may include 'result'? | |
| continue; | |
| if(substr_end>fragment_end) // end of substr OUT of current fragment? | if(substr_begin==CORD_NOT_FOUND) |
| break; | return STRING_NOT_FOUND; |
| return fragment.lang==lang?substr_end:STRING_NOT_FOUND; | if(langs.check_lang(lang, substr_begin, substr_length)) |
| } | return substr_begin; |
| return STRING_NOT_FOUND; | this_offset=substr_begin+substr_length; |
| } | |
| } | } |
| size_t String::pos(const String& substr, | size_t String::pos(const String& substr, |
| size_t this_offset, Untaint_lang 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, | void String::split(ArrayString& result, |
| size_t& pos_after, | size_t& pos_after, |
| const char* delim, | const char* delim, |
| Untaint_lang lang, int limit) { | 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 343 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, Untaint_lang lang, | const String& delim, Language lang, |
| int limit) const { | int limit) const { |
| if(delim) { | 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 364 void String::split(ArrayString& result, |
| } | } |
| } | } |
| static void regex_options(const String& options, int *result, bool& need_pre_post_match){ | static void regex_options(const String* options, int *result, bool& need_pre_post_match){ |
| struct Regex_option { | struct Regex_option { |
| const char* keyL; | const char* keyL; |
| const char* keyU; | const char* keyU; |
| Line 246 static void regex_options(const String& | Line 372 static void regex_options(const String& |
| int *result; | int *result; |
| bool *flag; | bool *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 | {"x", "U", 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored |
| {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$ | {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$ |
| {"g", "G", 0, true, result+1}, // many rows | {"g", "G", 0, 1, result+1, 0}, // many rows |
| {"'", 0, 0, 0, 0, &need_pre_post_match}, | {"'", 0, 0, 0, 0, &need_pre_post_match}, |
| {0} | {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) | if(options && !options->is_empty()) |
| for(Regex_option *o=regex_option; o->keyL; o++) | for(Regex_option *o=regex_option; o->keyL; o++) |
| if(options.pos(o->keyL)>=0 | if(options->pos(o->keyL)!=STRING_NOT_FOUND |
| || (o->keyU && options.pos(o->keyU)>=0)) { | || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) { |
| if(o->flag) | if(o->flag) |
| *o->flag=true; | *o->flag=true; |
| else { // result | else { // result |
| Line 271 static void regex_options(const String& | Line 397 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 { | bool& just_matched) const { |
| if(!regexp) | if(regexp.is_empty()) |
| throw Exception(0, | throw Exception(0, |
| 0, | 0, |
| "regexp is empty"); | "regexp is empty"); |
| Line 284 Table* String::match(Charset& source_cha | Line 410 Table* String::match(Charset& source_cha |
| const char* errptr; | const char* errptr; |
| int erroffset; | int erroffset; |
| bool need_pre_post_match=false; | bool need_pre_post_match=false; |
| int option_bits[2]; regex_options(options, option_bits, need_pre_post_match); | int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match); |
| bool global=option_bits[1]!=0; | bool global=option_bits[1]!=0; |
| pcre *code=pcre_compile(pattern, option_bits[0], | pcre *code=pcre_compile(pattern, option_bits[0], |
| &errptr, &erroffset, | &errptr, &erroffset, |
| Line 310 Table* String::match(Charset& source_cha | Line 436 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 325 Table* String::match(Charset& source_cha | Line 452 Table* String::match(Charset& source_cha |
| 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) | if(global || subpatterns) |
| return table; // global or with subpatterns=true+result | return &table; // global or with subpatterns=true+result |
| else { | else { |
| just_matched=false; return 0; // not global=no result | just_matched=false; return 0; // not global=no result |
| } | } |
| Line 347 Table* String::match(Charset& source_cha | Line 474 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 362 Table* String::match(Charset& source_cha | Line 489 Table* String::match(Charset& source_cha |
| if(!global || prestart==poststart) { // not global | going to hang | if(!global || prestart==poststart) { // not global | going to hang |
| 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 &table; |
| } | } |
| prestart=poststart; | prestart=poststart; |
| Line 373 Table* String::match(Charset& source_cha | Line 500 Table* String::match(Charset& source_cha |
| } | } |
| } | } |
| String& String::change_case(Charset& source_charset, Change_case_kind kind) { | String& String::change_case(Charset& source_charset, Change_case_kind kind) const { |
| String& result=*new String(); | String& result=*new String(); |
| if(is_empty()) | |
| const unsigned char *tables=source_charset.pcre_tables; | return result; |
| const unsigned char *a; | |
| const unsigned char *b; | |
| switch(kind) { | |
| case CC_UPPER: | |
| a=tables+lcc_offset; | |
| b=tables+fcc_offset; | |
| break; | |
| case CC_LOWER: | |
| a=tables+lcc_offset; | |
| b=0; | |
| break; | |
| default: | |
| throw Exception(0, | |
| this, | |
| "unknown change case kind #%d", | |
| static_cast<int>(kind)); // never | |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| char* new_cstr=cstrm(); | char* new_cstr=cstrm(); |
| char *dest=new_cstr; | size_t new_cstr_len=length(); |
| unsigned char index; | if(source_charset.isUTF8()) { |
| for(const char* current=new_cstr; index=(unsigned char)*current; current++) { | switch(kind) { |
| unsigned char c=a[index]; | case CC_UPPER: |
| if(b) | change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper); |
| c=b[c]; | break; |
| case CC_LOWER: | |
| change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower); | |
| break; | |
| default: | |
| assert(!"unknown change case kind"); | |
| break; // never | |
| } | |
| } else { | |
| const unsigned char *tables=source_charset.pcre_tables; | |
| const unsigned char *a; | |
| const unsigned char *b; | |
| switch(kind) { | |
| case CC_UPPER: | |
| a=tables+lcc_offset; | |
| b=tables+fcc_offset; | |
| 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, 0, fragments.count()); | |
| return result; | return result; |
| } | } |
| Line 417 String& String::change_case(Charset& sou | Line 559 String& String::change_case(Charset& sou |
| 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_start=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_starts(current)) { | while(*current) { |
| if(Dictionary::Subst subst=dict.first_that_begins(current)) { | |
| // prematch | // prematch |
| if(size_t prematch_length=current-prematch_start) { | if(size_t prematch_length=current-prematch_begin) { |
| result.body=CORD_cat_char_star(result.body, prematch_start, prematch_length); | result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length); |
| result.fragments.append(fragments, prematch_start-old_cstr, prematch_length); | result.body.append_strdup_know_length(prematch_begin, prematch_length); |
| } | } |
| // match | // match |
| // skip 'a' in 'current'; move prematch_begin | |
| current+=subst.from_length; prematch_begin=current; | |
| const String* a=row->get(0); | if(const String* b=subst.to) // are there any b? |
| // skip 'a' in 'current'; move prematch_start | result<<*b; |
| current+=a->length(); prematch_start=current; | |
| if(row->count()>1) { // are there any b? | |
| const String* b=row->get(1); | |
| if(size_t b_length=b->length()) { | |
| result.body=CORD_cat(result.body, b->cstr_to_cord()); | |
| result.fragments.append(b->fragments, 0, b->fragments.count()); | |
| } | |
| } | |
| } else // simply advance | } else // simply advance |
| current++; | current++; |
| } | } |
| // postmatch | // postmatch |
| if(size_t prematch_length=current-prematch_start) { | if(size_t postmatch_length=current-prematch_begin) { |
| result.body=CORD_cat_char_star(result.body, prematch_start, prematch_length); | result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length); |
| result.fragments.append(fragments, prematch_start-old_cstr, prematch_length); | 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)) | static int serialize_body_piece(const char* s, char** cur) { |
| 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 CORD_batched_iter_fn_append(const char* s, void* client_data) { | |
| char*& cur=*static_cast<char**>(client_data); | |
| 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 buf_length= | size_t buf_length= |
| prolog_length | prolog_length //1 |
| +fragments.count()*(sizeof(Untaint_lang)+sizeof(size_t)) | +sizeof(size_t) //2 |
| +length(); | +fragments_count*(sizeof(char)+sizeof(size_t)) //3 |
| C result(new(PointerFreeGC) char[buf_length], buf_length); | +body.length() //4 |
| +1; // for zero terminator used in deserialize | |
| 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: fragments.count | |
| size_t fragments_count=fragments.count(); | |
| 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 | // 3: lang info |
| for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) { | langs.for_each(body, serialize_lang_piece, &cur); |
| 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 | // 4: letters |
| CORD_iter5(body, 0, 0, CORD_batched_iter_fn_append, &cur); | 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, const char* file) { | bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) { |
| if(buf_length<=prolog_length) | if(buf_length<=prolog_length) |
| return false; | return false; |
| buf_length-=prolog_length; | buf_length-=prolog_length; |
| 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_length; |
| // 2: fragments.count | // 2: langs.count |
| if(buf_length<sizeof(size_t)) // fragments.count don't fit? | size_t fragments_count; |
| if(buf_length<sizeof(fragments_count)) // langs.count 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(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count); |
| buf_length-=sizeof(fragments_count); | |
| if(fragments_count) { | |
| // 3: 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(buf_length<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); | |
| // uchar needed to prevent propagating 0x80 bit to upper bytes | |
| langs.append(total_length, (String::Language)(uchar)lang, fragment_length); | |
| total_length+=fragment_length; | |
| // 3: lang info | buf_length-=piece_length; |
| size_t total_length=0; | } |
| for(size_t f=0; f<fragments_count; f++) { | |
| size_t piece_length=sizeof(Untaint_lang)+sizeof(size_t); | |
| if(buf_length<piece_length) // lang+length | |
| return false; | |
| Untaint_lang lang=*reinterpret_cast<const Untaint_lang *>(cur); cur+=sizeof(Untaint_lang); | // 4: letters |
| size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t); | if(buf_length!=total_length) |
| fragments+=Fragment(lang, fragment_length); | return false; |
| total_length+=fragment_length; | |
| buf_length-=piece_length; | // serialize wrote extra zero byte there, we can rely on that |
| body=String::Body(cur, buf_length); | |
| } | } |
| // 4: letters | ASSERT_STRING_INVARIANT(*this); |
| if(buf_length!=total_length) | return true; |
| return false; | } |
| body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length); | const char* String::Body::v() const { |
| return CORD_to_const_char_star(body); | |
| } | |
| void String::Body::dump() const { | |
| CORD_dump(body); | |
| } | |
| return true; | const char* String::Languages::v() const { |
| if(opt.is_not_just_lang) | |
| 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 | |
| // first: their langs | |
| result.langs.append(result.body, langs, substr_begin, substr_length); | |
| // next: letters themselves | |
| result.body=new_body; | |
| ASSERT_STRING_INVARIANT(result); | |
| return result; | |
| } | } |