|
|
| version 1.92, 2001/06/28 07:41:59 | version 1.131, 2001/12/13 10:47:34 |
|---|---|
| Line 2 | Line 2 |
| Parser: string class. @see untasize_t.C. | Parser: string class. @see untasize_t.C. |
| Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexander Petrosyan <paf@design.ru> (http://paf.design.ru) | |
| Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | |
| $Id$ | $Id$ |
| */ | */ |
| static char *RCSId="$Id$"; | |
| #include "pa_config_includes.h" | #include "pa_config_includes.h" |
| Line 22 static char *RCSId="$Id$"; | Line 20 static char *RCSId="$Id$"; |
| #include "pa_array.h" | #include "pa_array.h" |
| #include "pa_globals.h" | #include "pa_globals.h" |
| #include "pa_table.h" | #include "pa_table.h" |
| #include "pa_dictionary.h" | |
| String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : | String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : |
| Pooled(apool) { | Pooled(apool) { |
| last_chunk=&head; | last_chunk=&head; |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| append_here=head.rows; | append_here=head.rows; |
| head.preallocated_link=0; | head_link=0; |
| link_row=&head.rows[head.count]; | link_row=&head.rows[head.count]; |
| fused_rows=fsize=0; | |
| if(src) | if(src) |
| if(tainted) | if(tainted) |
| Line 39 String::String(Pool& apool, const char * | Line 37 String::String(Pool& apool, const char * |
| APPEND_CLEAN(src, src_size, 0, 0); | APPEND_CLEAN(src, src_size, 0, 0); |
| } | } |
| void String::expand() { | String::String(const String& src) : |
| size_t new_chunk_count=last_chunk->count+CR_GROW_COUNT; | Pooled(src.pool()) { |
| last_chunk=static_cast<Chunk *>( | |
| malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *))); | |
| last_chunk->count=new_chunk_count; | |
| link_row->link=last_chunk; | |
| append_here=last_chunk->rows; | |
| link_row=&last_chunk->rows[last_chunk->count]; | |
| link_row->link=0; | |
| } | |
| String::String(const String& src) : Pooled(src.pool()) { | |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| size_t src_used_rows=src.fused_rows; | uint src_used_rows=src.used_rows(); |
| if(src_used_rows<=head.count) { | if(src_used_rows<=head.count) { |
| // all new rows fit size_to preallocated area | // all new rows fit size_to preallocated area |
| size_t curr_chunk_rows=head.count; | last_chunk=&head; |
| uint curr_chunk_rows=head.count; | |
| memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows); | memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows); |
| append_here=&head.rows[src_used_rows]; | append_here=&head.rows[src_used_rows]; |
| link_row=&head.rows[curr_chunk_rows]; | link_row=&head.rows[curr_chunk_rows]; |
| Line 73 String::String(const String& src) : Pool | Line 62 String::String(const String& src) : Pool |
| // preallocated chunk src to constructing head | // preallocated chunk src to constructing head |
| memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count); | memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count); |
| // remaining rows size_to new_chunk | // remaining rows size_to new_chunk |
| size_t curr_chunk_rows=src_used_rows-head.count; | uint curr_chunk_rows=src_used_rows-head.count; |
| Chunk *new_chunk=static_cast<Chunk *>( | last_chunk=static_cast<Chunk *>( |
| malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); | malloc(sizeof(uint)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *), 9)); |
| new_chunk->count=curr_chunk_rows; | last_chunk->count=curr_chunk_rows; |
| head.preallocated_link=new_chunk; | head_link=last_chunk; |
| append_here=link_row=&new_chunk->rows[new_chunk->count]; | append_here=link_row=&last_chunk->rows[last_chunk->count]; |
| Chunk *old_chunk=src.head.preallocated_link; | Chunk *old_chunk=src.head_link; |
| Chunk::Row *new_rows=new_chunk->rows; | Chunk::Row *new_rows=last_chunk->rows; |
| size_t rows_left_to_copy=new_chunk->count; | uint rows_left_to_copy=last_chunk->count; |
| while(true) { | while(true) { |
| size_t old_count=old_chunk->count; | uint old_count=old_chunk->count; |
| Chunk *next_chunk=old_chunk->rows[old_count].link; | Chunk *next_chunk=old_chunk->rows[old_count].link; |
| if(next_chunk) { | if(next_chunk) { |
| // not last source chunk | // not last source chunk |
| Line 103 String::String(const String& src) : Pool | Line 92 String::String(const String& src) : Pool |
| } | } |
| } | } |
| link_row->link=0; | link_row->link=0; |
| fused_rows=src_used_rows; | src_used_rows; |
| fsize=src.fsize; | |
| } | } |
| String& String::append(const String& src, Untaint_lang lang, bool forced) { | size_t String::size() const { |
| const Chunk *chunk=&src.head; | size_t result=0; |
| do { | STRING_FOREACH_ROW( |
| const Chunk::Row *row=chunk->rows; | result+=row->item.size; |
| for(size_t i=0; i<chunk->count; i++, row++) { | ); |
| if(row==src.append_here) | |
| goto break2; | |
| APPEND(row->item.ptr, row->item.size, | |
| (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | break2: |
| return *this; | return result; |
| } | |
| /// @todo not very optimal | |
| uint String::used_rows() const { | |
| uint result=0; | |
| STRING_FOREACH_ROW( | |
| result++; | |
| ); | |
| break2: | |
| return result; | |
| } | |
| void String::expand() { | |
| uint new_chunk_count=last_chunk->count+CR_GROW_COUNT; | |
| if(new_chunk_count>MAX_USHORT) | |
| new_chunk_count=MAX_USHORT; | |
| last_chunk=static_cast<Chunk *>( | |
| malloc(sizeof(uint)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *), 10)); | |
| last_chunk->count=new_chunk_count; | |
| link_row->link=last_chunk; | |
| append_here=last_chunk->rows; | |
| link_row=&last_chunk->rows[last_chunk->count]; | |
| link_row->link=0; | |
| } | } |
| String& String::real_append(STRING_APPEND_PARAMS) { | String& String::real_append(STRING_APPEND_PARAMS) { |
| Line 133 String& String::real_append(STRING_APPEN | Line 135 String& String::real_append(STRING_APPEN |
| if(!size) | if(!size) |
| return *this; | return *this; |
| while(size>MAX_USHORT) { | |
| if(chunk_is_full()) | |
| expand(); | |
| append_here->item.ptr=src; | |
| append_here->item.size=MAX_USHORT; | |
| append_here->item.lang=lang; | |
| #ifndef NO_STRING_ORIGIN | |
| append_here->item.origin.file=file; | |
| append_here->item.origin.line=line; | |
| #endif | |
| append_here++; | |
| src+=MAX_USHORT; | |
| size-=MAX_USHORT; | |
| } | |
| if(chunk_is_full()) | if(chunk_is_full()) |
| expand(); | expand(); |
| append_here->item.ptr=src; | append_here->item.ptr=src; |
| fsize+=append_here->item.size=size; | append_here->item.size=size; |
| append_here->item.lang=lang; | append_here->item.lang=lang; |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| append_here->item.origin.file=file; | append_here->item.origin.file=file; |
| append_here->item.origin.line=line; | append_here->item.origin.line=line; |
| #endif | #endif |
| append_here++; fused_rows++; | append_here++; |
| return *this; | return *this; |
| } | } |
| uint String::hash_code() const { | char String::first_char() const { |
| uint result=0; | if(!used_rows()) |
| throw Exception(0, 0, | |
| this, | |
| "getting first char of empty string"); | |
| const Chunk *chunk=&head; | return *head.rows[0].item.ptr; |
| do { | } |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; i++) { | |
| if(row==append_here) | |
| goto break2; | |
| uint String::hash_code() const { | |
| uint result=0; | |
| STRING_FOREACH_ROW( | |
| result=Hash::generic_code(result, row->item.ptr, row->item.size); | result=Hash::generic_code(result, row->item.ptr, row->item.size); |
| row++; | ); |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | break2: |
| return result; | return result; |
| } | } |
| Line 171 break2: | Line 189 break2: |
| int String::cmp(int& partial, const String& src, | int String::cmp(int& partial, const String& src, |
| size_t this_offset, Untaint_lang lang) const { | size_t this_offset, Untaint_lang lang) const { |
| partial=-1; | partial=-1; |
| this_offset=min(this_offset, size()-1); | size_t a_size=size(); |
| this_offset=min(this_offset, a_size-1); | |
| const Chunk *a_chunk=&head; | const Chunk *a_chunk=&head; |
| const Chunk *b_chunk=&src.head; | const Chunk *b_chunk=&src.head; |
| Line 181 int String::cmp(int& partial, const Stri | Line 200 int String::cmp(int& partial, const Stri |
| size_t b_offset=0; | size_t b_offset=0; |
| Chunk::Row *a_end=append_here; | Chunk::Row *a_end=append_here; |
| Chunk::Row *b_end=src.append_here; | Chunk::Row *b_end=src.append_here; |
| size_t a_countdown=a_chunk->count; | uint a_countdown=a_chunk->count; |
| size_t b_countdown=b_chunk->count; | uint b_countdown=b_chunk->count; |
| size_t result; | int result; |
| size_t pos=0; | size_t pos=0; |
| bool a_break=size()==0; | bool a_break=a_size==0; |
| bool b_break=src.size()==0; | bool b_break=src.size()==0; |
| if(!(a_break || b_break)) while(true) { | if(!(a_break || b_break)) while(true) { |
| if(pos+a_row->item.size > this_offset) { | if(pos+a_row->item.size > this_offset) { |
| Line 259 int String::cmp(int& partial, const Stri | Line 278 int String::cmp(int& partial, const Stri |
| int String::cmp(int& partial, const char* b_ptr, size_t src_size, | int String::cmp(int& partial, const char* b_ptr, size_t src_size, |
| size_t this_offset, Untaint_lang lang) const { | size_t this_offset, Untaint_lang lang) const { |
| partial=-1; | partial=-1; |
| size_t a_size=size(); | |
| size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0; | size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0; |
| this_offset=min(this_offset, size()-1); | this_offset=min(this_offset, a_size-1); |
| const Chunk *a_chunk=&head; | const Chunk *a_chunk=&head; |
| const Chunk::Row *a_row=a_chunk->rows; | const Chunk::Row *a_row=a_chunk->rows; |
| size_t a_offset=this_offset; | size_t a_offset=this_offset; |
| size_t b_offset=0; | size_t b_offset=0; |
| Chunk::Row *a_end=append_here; | Chunk::Row *a_end=append_here; |
| size_t a_countdown=a_chunk->count; | uint a_countdown=a_chunk->count; |
| size_t pos=0; | size_t pos=0; |
| bool a_break=size()==0; | bool a_break=a_size==0; |
| bool b_break=b_size==0; | bool b_break=b_size==0; |
| if(!(a_break || b_break)) while(true) { | if(!(a_break || b_break)) while(true) { |
| if(pos+a_row->item.size > this_offset) { | if(pos+a_row->item.size > this_offset) { |
| Line 282 int String::cmp(int& partial, const char | Line 302 int String::cmp(int& partial, const char |
| (b_size-b_offset); | (b_size-b_offset); |
| if(size_diff==0) { // a has same size as b | if(size_diff==0) { // a has same size as b |
| if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, |
| a_row->item.size-a_offset)!=0) | a_row->item.size-a_offset)!=0) |
| return result; | return result; |
| pos+=a_row->item.size; | pos+=a_row->item.size; |
| a_row++; a_countdown--; a_offset=0; | a_row++; a_countdown--; a_offset=0; |
| b_break=true; | b_break=true; |
| } else if (size_diff>0) { // a longer | } else if (size_diff>0) { // a longer |
| if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, |
| b_size-b_offset)!=0) | b_size-b_offset)!=0) |
| return result; | return result; |
| a_offset+=b_size-b_offset; | a_offset+=b_size-b_offset; |
| b_break=true; | b_break=true; |
| } else { // b longer | } else { // b longer |
| if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, |
| a_row->item.size-a_offset)!=0) | a_row->item.size-a_offset)!=0) |
| return result; | return result; |
| b_offset+=a_row->item.size-a_offset; | b_offset+=a_row->item.size-a_offset; |
| Line 328 int String::cmp(int& partial, const char | Line 348 int String::cmp(int& partial, const char |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| const Origin& String::origin() const { | const Origin& String::origin() const { |
| if(!fused_rows) | if(!used_rows()) { |
| THROW(0, 0, | static const Origin empty_origin={"empty string"}; |
| 0, | return empty_origin; |
| "String::origin() of empty string called"); | } |
| // determining origin by last appended piece | // determining origin by last appended piece |
| // because first one frequently constant. | // because first one frequently constant. |
| Line 345 const Origin& String::origin() const { | Line 365 const Origin& String::origin() const { |
| #endif | #endif |
| String& String::mid(size_t start, size_t finish) const { | String& String::mid(size_t start, size_t finish) const { |
| String& result=*NEW String(pool()); | |
| start=max(0, start); | start=max(0, start); |
| finish=min(size(), finish); | finish=min(size(), finish); |
| if(start==finish) | if(start==finish) |
| return *empty_string; | return result; |
| String& result=*NEW String(pool()); | |
| size_t pos=0; | size_t pos=0; |
| const Chunk *chunk=&head; | STRING_FOREACH_ROW( |
| do { | size_t item_finish=pos+row->item.size; |
| const Chunk::Row *row=chunk->rows; | if(item_finish > start) { // started now or already? |
| for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) { | bool started=result.size()==0; // started now? |
| if(row==append_here) | bool finished=finish <= item_finish; // finished now? |
| size_t offset=started?start-pos:0; | |
| size_t size=finished?finish-pos:row->item.size; | |
| result.APPEND( | |
| row->item.ptr+offset, size-offset, | |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| if(finished) | |
| goto break2; | goto break2; |
| size_t item_finish=pos+row->item.size; | |
| if(item_finish > start) { // started now or already? | |
| bool started=result.size()==0; // started now? | |
| bool finished=finish <= item_finish; // finished now? | |
| size_t offset=started?start-pos:0; | |
| size_t size=finished?finish-pos:row->item.size; | |
| result.APPEND( | |
| row->item.ptr+offset, size-offset, | |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| if(finished) | |
| goto break2; | |
| } | |
| } | } |
| chunk=row->link; | pos+=row->item.size; |
| } while(chunk); | ); |
| break2: | break2: |
| // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", | // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", |
| //cstr(), start, finish, result.cstr()); | //cstr(), start, finish, result.cstr()); |
| Line 383 break2: | Line 396 break2: |
| } | } |
| int String::pos(const String& substr, | int String::pos(const String& substr, |
| size_t result, Untaint_lang lang) const { | int result, Untaint_lang lang) const { |
| for(; result<size(); result++) { | size_t self_size=size(); |
| for(; size_t(result)<self_size; result++) { | |
| int partial; cmp(partial, substr, result, lang); | int partial; cmp(partial, substr, result, lang); |
| if( | if( |
| partial==0 || // full match | partial==0 || // full match |
| Line 396 int String::pos(const String& substr, | Line 410 int String::pos(const String& substr, |
| } | } |
| int String::pos(const char *substr, size_t substr_size, | int String::pos(const char *substr, size_t substr_size, |
| size_t result, Untaint_lang lang) const { | int result, Untaint_lang lang) const { |
| for(; result<size(); result++) { | size_t self_size=size(); |
| for(; size_t(result)<self_size; result++) { | |
| int partial; cmp(partial, substr, substr_size, result, lang); | int partial; cmp(partial, substr, substr_size, result, lang); |
| if( | if( |
| partial==0 || // full match | partial==0 || // full match |
| Line 412 void String::split(Array& result, | Line 427 void String::split(Array& result, |
| size_t* pos_after_ref, | size_t* pos_after_ref, |
| const char *delim, size_t delim_size, | const char *delim, size_t delim_size, |
| Untaint_lang lang, int limit) const { | Untaint_lang lang, int limit) const { |
| size_t self_size=size(); | |
| if(delim_size) { | if(delim_size) { |
| size_t pos_after=pos_after_ref?*pos_after_ref:0; | size_t pos_after=pos_after_ref?*pos_after_ref:0; |
| int pos_before; | int pos_before; |
| Line 421 void String::split(Array& result, | Line 437 void String::split(Array& result, |
| pos_after=pos_before+delim_size; | pos_after=pos_before+delim_size; |
| } | } |
| // last piece | // last piece |
| if(pos_after<size() && limit) { | if(pos_after<self_size && limit) { |
| result+=&mid(pos_after, size()); | result+=&mid(pos_after, self_size); |
| pos_after=size(); | pos_after=self_size; |
| } | } |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref=pos_after; | *pos_after_ref=pos_after; |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=this; |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref+=size(); | *pos_after_ref+=self_size; |
| } | } |
| } | } |
| Line 492 bool String::match(const unsigned char * | Line 508 bool String::match(const unsigned char * |
| const String& regexp, | const String& regexp, |
| const String *options, | const String *options, |
| Table **table, | Table **table, |
| Row_action row_action, void *info) const { | Row_action row_action, void *info, |
| bool *was_global) const { | |
| if(!regexp.size()) | if(!regexp.size()) |
| THROW(0, 0, | throw Exception(0, 0, |
| aorigin, | aorigin, |
| "regexp is empty"); | "regexp is empty"); |
| const char *pattern=regexp.cstr(UL_AS_IS); | const char *pattern=regexp.cstr(); |
| const char *errptr; | const char *errptr; |
| int erroffset; | int erroffset; |
| int option_bits[2]; regex_options(options?options->cstr():0, option_bits); | int option_bits[2]; regex_options(options?options->cstr():0, option_bits); |
| if(was_global) | |
| *was_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, |
| pcre_tables); | pcre_tables); |
| if(!code) | if(!code) |
| THROW(0, 0, | throw Exception(0, 0, |
| ®exp.mid(erroffset, regexp.size()), | ®exp.mid(erroffset, regexp.size()), |
| "regular expression syntax error - %s", errptr); | "regular expression syntax error - %s", errptr); |
| int info_substrings=pcre_info(code, 0, 0); | int info_substrings=pcre_info(code, 0, 0); |
| if(info_substrings<0) { | if(info_substrings<0) { |
| (*pcre_free)(code); | pcre_free(code); |
| THROW(0, 0, | throw Exception(0, 0, |
| aorigin, | aorigin, |
| "pcre_info error (%d)", | "pcre_info error (%d)", |
| info_substrings); | info_substrings); |
| } | } |
| int startoffset=0; | int startoffset=0; |
| const char *subject=cstr(UL_AS_IS); | const char *subject=cstr(); |
| int length=strlen(subject); | int length=strlen(subject); |
| int ovecsize; | int ovecsize; |
| int *ovector=(int *)malloc(sizeof(int)* | int *ovector=(int *)malloc(sizeof(int)* |
| (ovecsize=(1/*match*/+info_substrings)*3)); | (ovecsize=(1/*match*/+info_substrings)*3), 11); |
| { // create table | { // create table |
| Array& columns=*NEW Array(pool()); | Array& columns=*NEW Array(pool()); |
| Line 547 bool String::match(const unsigned char * | Line 566 bool String::match(const unsigned char * |
| exec_option_bits, ovector, ovecsize); | exec_option_bits, ovector, ovecsize); |
| 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 row*/, 0, 0, info); | row_action(**table, 0/*last time, no row*/, 0, 0, info); |
| return option_bits[1]!=0; // global=true+table, not global=false | return option_bits[1]!=0; // global=true+table, not global=false |
| } | } |
| if(exec_substrings<0) { | if(exec_substrings<0) { |
| (*pcre_free)(code); | pcre_free(code); |
| THROW(0, 0, | throw Exception(0, 0, |
| aorigin, | aorigin, |
| "regular expression execute error (%d)", | "regular expression execute error (%d)", |
| exec_substrings); | exec_substrings); |
| Line 570 bool String::match(const unsigned char * | Line 589 bool String::match(const unsigned char * |
| 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 |
| } | } |
| (*row_action)(**table, &row, startoffset, ovector[0], info); | row_action(**table, &row, startoffset, ovector[0], info); |
| if(!option_bits[1] || !(startoffset=ovector[1])) { // not global | going to hang | if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang |
| (*pcre_free)(code); | pcre_free(code); |
| (*row_action)(**table, 0/*last time, no row*/, 0, 0, info); | row_action(**table, 0/*last time, no row*/, 0, 0, info); |
| return true; | return true; |
| } | } |
| startoffset=ovector[1]; | |
| /* | /* |
| if(option_bits[0] & PCRE_MULTILINE) | if(option_bits[0] & PCRE_MULTILINE) |
| Line 601 String& String::change_case(Pool& pool, | Line 621 String& String::change_case(Pool& pool, |
| b=0; | b=0; |
| break; | break; |
| default: | default: |
| PTHROW(0, 0, | throw Exception(0, 0, |
| this, | this, |
| "unknown change case kind #%d", | "unknown change case kind #%d", |
| static_cast<int>(kind)); // never | static_cast<int>(kind)); // never |
| Line 612 String& String::change_case(Pool& pool, | Line 632 String& String::change_case(Pool& pool, |
| const Chunk *chunk=&head; | const Chunk *chunk=&head; |
| do { | do { |
| const Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(size_t i=0; i<chunk->count; i++, row++) { | for(uint i=0; i<chunk->count; i++, row++) { |
| if(row==append_here) | if(row==append_here) |
| goto break2; | goto break2; |
| char *new_cstr=(char *)pool.malloc(row->item.size); | char *new_cstr=(char *)pool.malloc(row->item.size, 12); |
| char *dest=new_cstr; | char *dest=new_cstr; |
| const char *src=row->item.ptr; | const char *src=row->item.ptr; |
| for(int size=row->item.size; size--; src++) { | for(int size=row->item.size; size--; src++) { |
| Line 638 break2: | Line 658 break2: |
| return result; | return result; |
| } | } |
| void String::join_chain(Pool& pool, | |
| uint& ai, const Chunk*& achunk, const Chunk::Row*& arow, | |
| uchar& joined_lang, const char *& joined_ptr, size_t& joined_size) const { | |
| joined_lang=arow->item.lang; | |
| // calc size | |
| joined_size=0; | |
| { | |
| uint start_i=ai; | |
| const Chunk::Row *start_row=arow; | |
| const Chunk *chunk=achunk; | |
| do { | |
| const Chunk::Row *row=start_row; | |
| for(uint i=start_i; i<chunk->count; i++, row++) { | |
| if(row==append_here) | |
| goto break21; | |
| if(row->item.lang==joined_lang) | |
| joined_size+=row->item.size; | |
| else | |
| goto break21; | |
| } | |
| if(chunk=row->link) { | |
| start_i=0; | |
| start_row=chunk->rows; | |
| } else | |
| break; | |
| } while(true); | |
| break21:; | |
| } | |
| // if one row, return simply itself | |
| if(joined_size==arow->item.size) { | |
| joined_ptr=arow->item.ptr; | |
| ai++; arow++; | |
| if(ai==achunk->count) | |
| achunk=arow->link; | |
| } else { | |
| // join adjacent rows | |
| char *ptr=(char *)pool.malloc(joined_size,13); | |
| joined_ptr=ptr; | |
| uint start_i=ai; | |
| const Chunk::Row *start_row=arow; | |
| const Chunk *chunk=achunk; | |
| uint i; | |
| const Chunk::Row *row; | |
| do { | |
| row=start_row; | |
| for(i=start_i; i<chunk->count; i++, row++) { | |
| if(row==append_here) | |
| goto break22; | |
| if(row->item.lang==joined_lang) { | |
| memcpy(ptr, row->item.ptr, row->item.size); | |
| ptr+=row->item.size; | |
| } else | |
| goto break22; | |
| } | |
| if(chunk=row->link) { | |
| start_i=0; | |
| start_row=chunk->rows; | |
| } else | |
| break; | |
| } while(true); | |
| break22:; | |
| // return joined rows | |
| ai=i; | |
| arow=row; | |
| achunk=chunk; | |
| } | |
| } | |
| String& String::reconstruct(Pool& pool) const { | |
| //_asm int 3; | |
| String& result=*new(pool) String(pool); | |
| const Chunk *chunk=&head; | |
| while(true) { | |
| const Chunk::Row *row=chunk->rows; | |
| for(uint i=0; i<chunk->count; ) { | |
| if(row==append_here) | |
| goto break2; | |
| uchar joined_lang; | |
| const char *joined_ptr; | |
| size_t joined_size; | |
| #ifndef NO_STRING_ORIGIN | |
| const char *joined_origin_file=row->item.origin.file; | |
| const size_t joined_origin_line=row->item.origin.line; | |
| #endif | |
| join_chain(pool, i, chunk, row, | |
| joined_lang, joined_ptr, joined_size); | |
| result.APPEND(joined_ptr, joined_size, joined_lang, | |
| joined_origin_file, joined_origin_line); | |
| if(!chunk) | |
| goto break2; | |
| } | |
| } | |
| break2: | |
| return result; | |
| }; | |
| String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const { | |
| //_asm int 3; | |
| String& result=*new(pool) String(pool); | |
| STRING_FOREACH_ROW( | |
| const char *src=row->item.ptr; | |
| size_t src_size=row->item.size; | |
| char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14); | |
| char *dest=new_cstr; | |
| while(src_size) { | |
| // there is a row where first column starts 'src' | |
| if(Table::Item *item=dict.first_that_starts(src, src_size)) { | |
| // get a=>b values | |
| const String& a=*static_cast<Array *>(item)->get_string(0); | |
| const String& b=*static_cast<Array *>(item)->get_string(1); | |
| // skip 'a' in 'src' && reduce work size | |
| src+=a.size(); src_size-=a.size(); | |
| // write 'b' to 'dest' && skip 'b' in 'dest' | |
| b.store_to(dest); dest+=b.size(); | |
| } else { | |
| // write a char to b && reduce work size | |
| *dest++=*src++; src_size--; | |
| } | |
| } | |
| result.APPEND(new_cstr, dest-new_cstr, | |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| ); | |
| break2: | |
| return result; | |
| } | |
| String& String::replace(Pool& pool, Dictionary& dict) const { | |
| return reconstruct(pool).replace_in_reconstructed(pool, dict); | |
| } | |
| double String::as_double() const { | double String::as_double() const { |
| double result; | double result; |
| const char *cstr=this->cstr(); | const char *cstr; |
| char *error_pos=0; | char buf[MAX_NUMBER]; |
| if(head.rows+1==append_here) { | |
| int size=min(head.rows[0].item.size, MAX_NUMBER-1); | |
| memcpy(buf, head.rows[0].item.ptr, size); | |
| buf[size]=0; | |
| cstr=buf; | |
| } else | |
| cstr=this->cstr(); | |
| char *error_pos; | |
| // 0xABC | // 0xABC |
| if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X')) | if(cstr[0]=='0') |
| result=(double)(unsigned long)strtol(cstr, &error_pos, 0); | if(cstr[1]=='x' || cstr[1]=='X') |
| result=(double)(unsigned long)strtol(cstr, &error_pos, 0); | |
| else | |
| result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos); | |
| else | else |
| result=strtod(cstr, &error_pos); | result=(double)strtod(cstr, &error_pos); |
| if(error_pos && *error_pos) | if(*error_pos/*not EOS*/) |
| THROW(0, 0, | throw Exception(0, 0, |
| this, | this, |
| "invalid number (double)"); | "invalid number (double)"); |
| Line 657 double String::as_double() const { | Line 828 double String::as_double() const { |
| } | } |
| int String::as_int() const { | int String::as_int() const { |
| int result; | int result; |
| const char *cstr=this->cstr(); | const char *cstr; |
| char *error_pos=0; | char buf[MAX_NUMBER]; |
| if(head.rows+1==append_here) { | |
| int size=min(head.rows[0].item.size, MAX_NUMBER-1); | |
| memcpy(buf, head.rows[0].item.ptr, size); | |
| buf[size]=0; | |
| cstr=buf; | |
| } else | |
| cstr=this->cstr(); | |
| char *error_pos; | |
| // 0xABC | // 0xABC |
| if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X')) | if(cstr[0]=='0') |
| result=(int)(unsigned long)strtol(cstr, &error_pos, 0); | if(cstr[1]=='x' || cstr[1]=='X') |
| result=(int)(unsigned long)strtol(cstr, &error_pos, 0); | |
| else | |
| result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0); | |
| else | else |
| result=(int)strtol(cstr, &error_pos, 0); | result=(int)strtol(cstr, &error_pos, 0); |
| if(error_pos && *error_pos) | if(*error_pos/*not EOS*/) |
| THROW(0, 0, | throw Exception(0, 0, |
| this, | this, |
| "invalid number (int)"); | "invalid number (int)"); |
| return result; | return result; |
| } | } |
| inline void ushort2uchars(ushort word, uchar& byte1, uchar& byte2) { | |
| byte1=word&0xFF; | |
| byte2=word>>8; | |
| } | |
| inline ushort uchars2ushort(uchar byte1, uchar byte2) { | |
| return (byte2<<8) | byte1; | |
| } | |
| /* @todo maybe network order worth spending some effort? | |
| don't bothering myself with network byte order, | |
| am not planning to be able to move resulting file across platforms | |
| for now | |
| */ | |
| void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const { | |
| buf_size= | |
| prolog_size | |
| +used_rows()*(sizeof(uchar)+sizeof(ushort)) | |
| +size(); | |
| buf=malloc(buf_size,15); | |
| char *cur=(char *)buf+prolog_size; | |
| STRING_FOREACH_ROW( | |
| // lang | |
| memcpy(cur, &row->item.lang, sizeof(uchar)); | |
| cur+=sizeof(uchar); | |
| // size | |
| uchar byte1; uchar byte2; | |
| ushort2uchars(row->item.size, byte1, byte2); | |
| memcpy(cur, &byte1, sizeof(uchar)); cur+=sizeof(uchar); | |
| memcpy(cur, &byte2, sizeof(uchar)); cur+=sizeof(uchar); | |
| // bytes | |
| memcpy(cur, row->item.ptr, row->item.size); | |
| cur+=row->item.size; | |
| ); | |
| break2: | |
| ; | |
| } | |
| void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) { | |
| char *cur=(char *)buf+prolog_size; | |
| buf_size-=prolog_size; | |
| while(buf_size) { | |
| uchar lang=*(uchar *)(cur); | |
| ushort size=uchars2ushort( | |
| *(uchar*)(cur+sizeof(uchar)*1), | |
| *(uchar*)(cur+sizeof(uchar)*2) | |
| ); | |
| const char *ptr=(const char*)(cur+sizeof(uchar)*3); | |
| APPEND(ptr, size, lang, file, 0); | |
| size_t piece_size=sizeof(uchar)+sizeof(ushort)+size; | |
| cur+=piece_size; | |
| buf_size-=piece_size; | |
| } | |
| } |