--- parser3/src/main/pa_string.C 2003/03/27 10:53:01 1.172.2.21.2.33 +++ parser3/src/main/pa_string.C 2003/09/25 09:15:03 1.176 @@ -5,7 +5,7 @@ Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char* IDENT_STRING_C="$Date: 2003/03/27 10:53:01 $"; +static const char* IDENT_STRING_C="$Date: 2003/09/25 09:15:03 $"; #include "pcre.h" @@ -15,6 +15,32 @@ static const char* IDENT_STRING_C="$Date #include "pa_dictionary.h" #include "pa_charset.h" +// 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'*/); +} + // helpers /// String::match uses this as replace & global search table columns @@ -28,61 +54,19 @@ public: *this+=new String("match"); *this+=new String("postmatch"); for(int i=0; i i(src); ; fragment_begin=fragment_end) { - const element_type fragment=i.next(); - fragment_end=fragment_begin+fragment.length; - - if(substr_beginfragment_end) // 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(); - fragment_end=fragment_begin+fragment.length; - - if(substr_end>fragment_end) // are there still more? - append(Fragment(fragment.lang, fragment.length)); // appending whole fragment - else { // no, it was last - append(Fragment(fragment.lang, substr_end-fragment_begin)); - return; - } - } - - break; - } -} +// String::Body methods -// StringBody methods - -StringBody StringBody::Format(int value) { +String::Body String::Body::Format(int value) { char local[MAX_NUMBER]; size_t length=snprintf(local, MAX_NUMBER, "%d", value); - return StringBody(pa_strdup(local, length), length); + return String::Body(pa_strdup(local, length), length); } static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) { @@ -95,7 +79,7 @@ static int CORD_batched_iter_fn_generic_ generic_hash_code(result, s); return 0; }; -uint StringBody::hash_code() const { +uint String::Body::hash_code() const { uint result=0; CORD_iter5(body, 0, CORD_batched_iter_fn_generic_hash_code, @@ -105,24 +89,45 @@ uint StringBody::hash_code() const { // String methods -String::String(const char* cstr, size_t length, bool tainted): body(CORD_EMPTY) { - append(cstr, length, tainted?L_TAINTED:L_CLEAN); +String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) { + append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN); } - -String::String(const String& src): body(src.body) { - fragments.append(src.fragments); +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::append(const char* str, size_t length, Language lang) { +String& String::append_know_length(const char* str, size_t known_length, Language lang) { + if(!known_length) + return *this; + + // 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) return *this; - size_t use_length=length?length:strlen(str); - if(!use_length) + size_t known_length=helper_length?helper_length:strlen(str); + if(!known_length) + return *this; + + return append_know_length(str, known_length, lang); +} +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; - body.append(str, use_length); - fragments+=Fragment(lang, use_length); + // 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; } @@ -131,67 +136,29 @@ String& String::mid(size_t substr_begin, String& result=*new String; 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); - if(substr_begin==substr_end) + size_t substr_length=substr_end-substr_begin; + if(!substr_length) return result; - // first: letters themselves - result.body=body.mid(substr_begin, substr_end-substr_begin); - - // next: their langs - result.fragments.append_positions(fragments, substr_begin, substr_end); + // first: their langs + result.langs.append(result.body, langs, substr_begin, substr_length); + // next: letters themselves + result.body=body.mid(substr_begin, substr_length); // SAPI::log("piece of '%s' from %d to %d is '%s'", //cstr(), substr_begin, substr_end, result.cstr()); + ASSERT_STRING_INVARIANT(result); return result; } -size_t String::pos(const StringBody substr, - size_t this_offset, Language lang) const { - // first: letters themselves +size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { size_t substr_begin=body.pos(substr, this_offset); - if(substr_begin==CORD_NOT_FOUND) + if(substr_begin==CORD_NOT_FOUND || !langs.check_lang(lang, substr_begin, substr.length())) 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 i(fragments); i.has_next(); fragment_begin=fragment_end) { - const Fragment fragment=i.next(); - fragment_end=fragment_begin+fragment.length; - - 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? - throw Exception(0, // (*) see below - 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) - 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; + return substr_begin; } size_t String::pos(const String& substr, @@ -200,14 +167,14 @@ size_t String::pos(const String& substr, } void String::split(ArrayString& result, - size_t& pos_after, - const char* delim, - Language lang, int limit) const { + size_t& pos_after, + const char* delim, + Language lang, int limit) const { size_t self_length=length(); if(size_t delim_length=strlen(delim)) { int pos_before; // 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); pos_after=pos_before+delim_length; } @@ -223,13 +190,13 @@ void String::split(ArrayString& result, } void String::split(ArrayString& result, - size_t& pos_after, - const String& delim, Language lang, - int limit) const { + size_t& pos_after, + const String& delim, Language lang, + int limit) const { if(!delim.is_empty()) { int pos_before; // 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); pos_after=pos_before+delim.length(); } @@ -265,8 +232,8 @@ static void regex_options(const String* if(options && !options->is_empty()) for(Regex_option *o=regex_option; o->keyL; o++) - if(options->pos(o->keyL)>=0 - || (o->keyU && options->pos(o->keyU)>=0)) { + 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 @@ -277,10 +244,10 @@ static void regex_options(const String* } Table* String::match(Charset& source_charset, - const String& regexp, - const String* options, - Row_action row_action, void *info, - bool& just_matched) const { + const String& regexp, + const String* options, + Row_action row_action, void *info, + bool& just_matched) const { if(regexp.is_empty()) throw Exception(0, 0, @@ -316,7 +283,8 @@ Table* String::match(Charset& source_cha int ovector[oveclength]; // 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 prestart=0; @@ -416,8 +384,8 @@ String& String::change_case(Charset& sou *dest++=(char)c; } + result.langs=langs; result.body=new_cstr; - result.fragments.append(fragments); return result; } @@ -432,8 +400,8 @@ const String& String::replace(const Dict if(Table::element_type row=dict.first_that_begins(current)) { // prematch if(size_t prematch_length=current-prematch_begin) { - result.body.append(prematch_begin, prematch_length); - result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr); + result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr); + result.body.append_strdup_know_length(prematch_begin, prematch_length); } // match @@ -444,10 +412,7 @@ const String& String::replace(const Dict if(row->count()>1) { // are there any b? const String* b=row->get(1); - if(size_t b_length=b->length()) { - result.body<cstr_to_string_body(); - result.fragments.append(b->fragments); - } + result<<*b; } } else // simply advance current++; @@ -455,10 +420,11 @@ const String& String::replace(const Dict // postmatch if(size_t postmatch_length=current-prematch_begin) { - result.body.append(prematch_begin, postmatch_length); - result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr); + result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr); + result.body.append_strdup_know_length(prematch_begin, postmatch_length); } + ASSERT_STRING_INVARIANT(result); return result; } @@ -535,23 +501,26 @@ static int serialize_body_piece(const ch memcpy(*cur, s, length); *cur+=length; return 0; }; -String::C String::serialize(size_t prolog_length) const { +String::Cm String::serialize(size_t prolog_length) const { +#if TODO + //_asm int 3; size_t buf_length= prolog_length - +fragments.count()*(sizeof(Language)+sizeof(size_t)) + +sizeof(size_t) + +langs.count()*(sizeof(Language)+sizeof(size_t)) +length(); - C result(new(PointerFreeGC) char[buf_length], buf_length); + String::Cm result(new(PointerFreeGC) char[buf_length], buf_length); // 1: prolog char *cur=result.str+prolog_length; - // 2: fragments.count - size_t fragments_count=fragments.count(); + // 2: langs.count + size_t fragments_count=langs.count(); memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count); // 3: lang info - for(Array_iterator i(fragments); i.has_next(); ) { + for(Array_iterator i(langs); i.has_next(); ) { const Fragment fragment=i.next(); // lang memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang); @@ -563,8 +532,11 @@ String::C String::serialize(size_t prolo body.for_each(serialize_body_piece, &cur); return result; +#endif + return String::Cm(0, 0); } bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) { +#if TODO if(buf_length<=prolog_length) return false; buf_length-=prolog_length; @@ -572,32 +544,62 @@ bool String::deserialize(size_t prolog_l // 1: prolog const char* cur=(const char* )buf+prolog_length; - // 2: fragments.count - if(buf_length(cur); cur+=sizeof(size_t); buf_length-=sizeof(size_t); + + if(fragments_count) { + // 3: lang info + size_t total_length=0; + for(size_t f=0; f(cur); cur+=sizeof(Language); + size_t fragment_length=*reinterpret_cast(cur); cur+=sizeof(size_t); + langs+=Fragment(lang, fragment_length); + total_length+=fragment_length; - // 3: lang info - size_t total_length=0; - for(size_t f=0; f(cur); cur+=sizeof(Language); - size_t fragment_length=*reinterpret_cast(cur); cur+=sizeof(size_t); - fragments+=Fragment(lang, fragment_length); - total_length+=fragment_length; + // 4: letters + if(buf_length!=total_length) + return false; - buf_length-=piece_length; + body=String::Body(cur, buf_length); } - // 4: letters - if(buf_length!=total_length) - return false; + ASSERT_STRING_INVARIANT(*this); + return true; +#endif + 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); +} +const char* String::Languages::v() const { + if(is_not_just_lang) + return CORD_to_const_char_star(langs); + else + return (const char*)&langs; +} +const char* String::v() const { +#define 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, + "%.*s%s} " + "{%d:%s", + LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"", + strlen(body_view), body_view + ); - return true; + return buf; +#undef LIMIT_VIEW }