|
|
| version 1.26, 2001/02/20 18:45:53 | version 1.130, 2001/12/13 10:24:12 |
|---|---|
| Line 1 | Line 1 |
| /* | /** @file |
| $Id$ | Parser: string class. @see untasize_t.C. |
| Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) | |
| Author: Alexander Petrosyan <paf@design.ru> (http://paf.design.ru) | |
| $Id$ | |
| */ | */ |
| #include <string.h> | #include "pa_config_includes.h" |
| #include "pcre.h" | |
| #include "internal.h" | |
| #include "pa_pool.h" | #include "pa_pool.h" |
| #include "pa_string.h" | #include "pa_string.h" |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_common.h" | |
| #include "pa_array.h" | |
| #include "pa_globals.h" | |
| #include "pa_table.h" | |
| #include "pa_dictionary.h" | |
| // String | String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : |
| String::String(Pool& apool) : | |
| Pooled(apool) { | Pooled(apool) { |
| head.count=curr_chunk_rows=CR_PREALLOCATED_COUNT; | last_chunk=&head; |
| head.count=CR_PREALLOCATED_COUNT; | |
| append_here=head.rows; | append_here=head.rows; |
| head.preallocated_link=0; | head_link=0; |
| link_row=&head.rows[curr_chunk_rows]; | link_row=&head.rows[head.count]; |
| fused_rows=fsize=0; | |
| } | |
| void String::expand() { | if(src) |
| curr_chunk_rows+=curr_chunk_rows*CR_GROW_PERCENT/100; | if(tainted) |
| Chunk *chunk=static_cast<Chunk *>( | APPEND_TAINTED(src, src_size, 0, 0); |
| pool().malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); | else |
| chunk->count=curr_chunk_rows; | APPEND_CLEAN(src, src_size, 0, 0); |
| link_row->link=chunk; | |
| append_here=chunk->rows; | |
| link_row=&chunk->rows[curr_chunk_rows]; | |
| link_row->link=0; | |
| } | } |
| String::String(const String& src) : | String::String(const String& src) : |
| Pooled(src.pool()) { | Pooled(src.pool()) { |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| int src_used_rows=src.used_rows(); | uint src_used_rows=src.used_rows(); |
| if(src_used_rows<=head.count) { | if(src_used_rows<=head.count) { |
| // all new rows fit into preallocated area | // all new rows fit size_to preallocated area |
| 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 51 String::String(const String& src) : | Line 58 String::String(const String& src) : |
| // allocating only enough mem to fit src string rows | // allocating only enough mem to fit src string rows |
| // next append would allocate a new chunk | // next append would allocate a new chunk |
| // | // |
| // new rows don't fit into preallocated area: splitting into two chunks | // new rows don't fit size_to preallocated area: splitting size_to two chunks |
| // 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 into new_chunk | // remaining rows size_to new_chunk |
| 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 *>( |
| pool().malloc(sizeof(int)+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[curr_chunk_rows]; | 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; |
| int rows_left_to_copy=curr_chunk_rows; | uint rows_left_to_copy=last_chunk->count; |
| while(true) { | while(true) { |
| int 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 85 String::String(const String& src) : | Line 92 String::String(const String& src) : |
| } | } |
| } | } |
| link_row->link=0; | link_row->link=0; |
| fused_rows=src_used_rows; | src_used_rows; |
| fsize=src.fsize; | |
| } | } |
| /* | |
| String(const String_iterator& begin, const String_iterator& end) { | size_t String::size() const { |
| ;//TODO | size_t result=0; |
| STRING_FOREACH_ROW( | |
| result+=row->item.size; | |
| ); | |
| break2: | |
| 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) { |
| if(!src) | if(!src) |
| Line 102 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; | |
| #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; |
| } | } |
| char *String::cstr() const { | char String::first_char() const { |
| char *result=static_cast<char *>(pool().malloc(size()+1)); | if(!used_rows()) |
| throw Exception(0, 0, | |
| this, | |
| "getting first char of empty string"); | |
| char *copy_here=result; | return *head.rows[0].item.ptr; |
| const Chunk *chunk=&head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(int i=0; i<chunk->count; i++) { | |
| if(row==append_here) | |
| goto break2; | |
| memcpy(copy_here, row->item.ptr, row->item.size); | |
| copy_here+=row->item.size; | |
| row++; | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | |
| *copy_here=0; | |
| return result; | |
| } | } |
| uint String::hash_code() const { | uint String::hash_code() const { |
| uint result=0; | uint result=0; |
| STRING_FOREACH_ROW( | |
| const Chunk *chunk=&head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(int i=0; i<chunk->count; i++) { | |
| if(row==append_here) | |
| goto break2; | |
| 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; |
| } | } |
| bool String::operator == (const String& src) const { | /// @todo move 'lang' skipping to pos |
| if(size() != src.size()) | int String::cmp(int& partial, const String& src, |
| return false; | size_t this_offset, Untaint_lang lang) const { |
| partial=-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; |
| const Chunk::Row *a_row=a_chunk->rows; | const Chunk::Row *a_row=a_chunk->rows; |
| const Chunk::Row *b_row=b_chunk->rows; | const Chunk::Row *b_row=b_chunk->rows; |
| int a_offset=0; | size_t a_offset=this_offset; |
| int 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; |
| int a_countdown=a_chunk->count; | uint a_countdown=a_chunk->count; |
| int b_countdown=b_chunk->count; | uint b_countdown=b_chunk->count; |
| bool a_break=false; | int result; |
| bool b_break=false; | size_t pos=0; |
| while(true) { | |
| int size_diff= | bool a_break=a_size==0; |
| (a_row->item.size-a_offset)- | bool b_break=src.size()==0; |
| (b_row->item.size-b_offset); | if(!(a_break || b_break)) while(true) { |
| if(pos+a_row->item.size > this_offset) { | |
| if(size_diff==0) { // a has same size as b | if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang) |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset)!=0) | return -1; // wrong lang -- bail out |
| return false; | |
| a_row++; a_countdown--; a_offset=0; | int size_diff= |
| b_row++; b_countdown--; b_offset=0; | (a_row->item.size-a_offset)- |
| } else if (size_diff>0) { // a longer | (b_row->item.size-b_offset); |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset)!=0) | |
| return false; | if(size_diff==0) { // a has same size as b |
| a_offset+=b_row->item.size-b_offset; | result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, |
| b_row++; b_countdown--; b_offset=0; | a_row->item.size-a_offset); |
| } else { // b longer | if(result) |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset)!=0) | return result; |
| return false; | pos+=a_row->item.size; |
| b_offset+=a_row->item.size-a_offset; | a_row++; a_countdown--; a_offset=0; |
| a_row++; a_countdown--; a_offset=0; | b_row++; b_countdown--; b_offset=0; |
| } else if (size_diff>0) { // a longer | |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, | |
| b_row->item.size-b_offset); | |
| if(result) | |
| return result; | |
| a_offset+=b_row->item.size-b_offset; | |
| b_row++; b_countdown--; b_offset=0; | |
| } else { // b longer | |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, | |
| a_row->item.size-a_offset); | |
| if(result) | |
| return result; | |
| b_offset+=a_row->item.size-a_offset; | |
| pos+=a_row->item.size; | |
| a_row++; a_countdown--; a_offset=0; | |
| } | |
| if(b_break=b_row==b_end) { | |
| a_break=a_row==a_end; | |
| break; | |
| } | |
| if(!b_countdown) { | |
| b_chunk=b_row->link; | |
| b_row=b_chunk->rows; | |
| b_countdown=b_chunk->count; | |
| } | |
| } else { | |
| a_offset-=a_row->item.size; | |
| pos+=a_row->item.size; | |
| a_row++; a_countdown--; | |
| } | |
| if(a_break=a_row==a_end) { | |
| b_break=b_row==b_end; | |
| break; | |
| } | |
| if(!a_countdown) { | |
| a_chunk=a_row->link; | |
| a_row=a_chunk->rows; | |
| a_countdown=a_chunk->count; | |
| } | |
| } | |
| if(a_break==b_break) { // ended simultaneously | |
| partial=0; return 0; | |
| } else if(a_break) { // first bytes equal, but a ended before b | |
| partial=1; return -1; | |
| } else { | |
| partial=2; return +1; | |
| } | |
| } | |
| /// @todo move 'lang' skipping to pos | |
| int String::cmp(int& partial, const char* b_ptr, size_t src_size, | |
| size_t this_offset, Untaint_lang lang) const { | |
| partial=-1; | |
| size_t a_size=size(); | |
| size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0; | |
| this_offset=min(this_offset, a_size-1); | |
| const Chunk *a_chunk=&head; | |
| const Chunk::Row *a_row=a_chunk->rows; | |
| size_t a_offset=this_offset; | |
| size_t b_offset=0; | |
| Chunk::Row *a_end=append_here; | |
| uint a_countdown=a_chunk->count; | |
| size_t pos=0; | |
| bool a_break=a_size==0; | |
| bool b_break=b_size==0; | |
| if(!(a_break || b_break)) while(true) { | |
| if(pos+a_row->item.size > this_offset) { | |
| if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang) | |
| return -1; // wrong lang -- bail out | |
| int size_diff= | |
| (a_row->item.size-a_offset)- | |
| (b_size-b_offset); | |
| if(size_diff==0) { // a has same size as b | |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | |
| a_row->item.size-a_offset)!=0) | |
| return result; | |
| pos+=a_row->item.size; | |
| a_row++; a_countdown--; a_offset=0; | |
| b_break=true; | |
| } else if (size_diff>0) { // a longer | |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | |
| b_size-b_offset)!=0) | |
| return result; | |
| a_offset+=b_size-b_offset; | |
| b_break=true; | |
| } else { // b longer | |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | |
| a_row->item.size-a_offset)!=0) | |
| return result; | |
| b_offset+=a_row->item.size-a_offset; | |
| pos+=a_row->item.size; | |
| a_row++; a_countdown--; a_offset=0; | |
| } | |
| } else { | |
| a_offset-=a_row->item.size; | |
| pos+=a_row->item.size; | |
| a_row++; a_countdown--; | |
| } | } |
| a_break=a_row==a_end; | a_break=a_row==a_end; |
| b_break=b_row==b_end; | |
| if(a_break || b_break) | if(a_break || b_break) |
| break; | break; |
| if(!a_countdown) { | if(!a_countdown) { |
| a_chunk=a_row->link; | a_chunk=a_row->link; |
| a_row=a_chunk->rows; | a_row=a_chunk->rows; |
| a_countdown=a_chunk->count; | a_countdown=a_chunk->count; |
| } | } |
| if(!b_countdown) { | |
| b_chunk=b_row->link; | |
| b_row=b_chunk->rows; | |
| b_countdown=b_chunk->count; | |
| } | |
| } | } |
| return a_break==b_break; | if(a_break==b_break) { // ended simultaneously |
| partial=0; return 0; | |
| } else if(a_break) { // first bytes equal, but a ended before b | |
| partial=1; return -1; | |
| } else { | |
| partial=2; return +1; | |
| } | |
| } | } |
| /* | |
| String& String::append(const String_iterator& begin, const String_iterator& end) { | #ifndef NO_STRING_ORIGIN |
| //TODO | const Origin& String::origin() const { |
| return *this; | if(!used_rows()) { |
| static const Origin empty_origin={"empty string"}; | |
| return empty_origin; | |
| } | |
| // determining origin by last appended piece | |
| // because first one frequently constant. | |
| // ex: ^load[/file] "document_root" + "/file" | |
| // when last peice is constant, | |
| // ex: parser_root_auto_path{dynamic} / auto.p{const} | |
| // using first piece | |
| Origin& last_origin=append_here[-1].item.origin; | |
| return last_origin.file ? last_origin : head.rows[0].item.origin; | |
| } | } |
| #endif | |
| // Char_types | String& String::mid(size_t start, size_t finish) const { |
| String& result=*NEW String(pool()); | |
| Char_types::Char_types() { | start=max(0, start); |
| memset(types, 0, sizeof(types)); | finish=min(size(), finish); |
| if(start==finish) | |
| return result; | |
| size_t pos=0; | |
| STRING_FOREACH_ROW( | |
| 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; | |
| } | |
| pos+=row->item.size; | |
| ); | |
| break2: | |
| // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", | |
| //cstr(), start, finish, result.cstr()); | |
| return result; | |
| } | } |
| void Char_types::set(char from, char to, int type) { | int String::pos(const String& substr, |
| memset(&types[static_cast<unsigned int>(from)], type, to-from+1); | int result, Untaint_lang lang) const { |
| size_t self_size=size(); | |
| for(; result<self_size; result++) { | |
| int partial; cmp(partial, substr, result, lang); | |
| if( | |
| partial==0 || // full match | |
| partial==2) // 'substr' starts 'this'+'result' | |
| return result; | |
| } | |
| return -1; | |
| } | } |
| // String_iterator | int String::pos(const char *substr, size_t substr_size, |
| int result, Untaint_lang lang) const { | |
| String_iterator::String_iterator(String& astring) : string(astring) { | size_t self_size=size(); |
| read_here=string.head.rows; | for(; result<self_size; result++) { |
| position=string.size()==0?0:read_here->item.ptr; | int partial; cmp(partial, substr, substr_size, result, lang); |
| link_row=reinterpret_cast<String::Chunk::Row*>(string.head.preallocated_link); | if( |
| partial==0 || // full match | |
| partial==2) // 'substr' starts 'this'+'result' | |
| return result; | |
| } | |
| return -1; | |
| } | } |
| /* | |
| String_iterator::String_iterator(String_iterator& asi) { | void String::split(Array& result, |
| //TODO | size_t* pos_after_ref, |
| const char *delim, size_t delim_size, | |
| Untaint_lang lang, int limit) const { | |
| size_t self_size=size(); | |
| if(delim_size) { | |
| size_t pos_after=pos_after_ref?*pos_after_ref:0; | |
| int pos_before; | |
| // while we have 'delim'... | |
| for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) { | |
| result+=&mid(pos_after, pos_before); | |
| pos_after=pos_before+delim_size; | |
| } | |
| // last piece | |
| if(pos_after<self_size && limit) { | |
| result+=&mid(pos_after, self_size); | |
| pos_after=self_size; | |
| } | |
| if(pos_after_ref) | |
| *pos_after_ref=pos_after; | |
| } else { // empty delim | |
| result+=this; | |
| if(pos_after_ref) | |
| *pos_after_ref+=self_size; | |
| } | |
| } | } |
| char String_iterator::operator()() const { | void String::split(Array& result, |
| return position?*position:0; | size_t* pos_after_ref, |
| const String& delim, Untaint_lang lang, | |
| int limit) const { | |
| if(delim.size()) { | |
| size_t pos_after=pos_after_ref?*pos_after_ref:0; | |
| int pos_before; | |
| // while we have 'delim'... | |
| for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) { | |
| result+=&mid(pos_after, pos_before); | |
| pos_after=pos_before+delim.size(); | |
| } | |
| // last piece | |
| if(pos_after<size() && limit) { | |
| result+=&mid(pos_after, size()); | |
| pos_after=size(); | |
| } | |
| if(pos_after_ref) | |
| *pos_after_ref=pos_after; | |
| } else { // empty delim | |
| result+=this; | |
| if(pos_after_ref) | |
| *pos_after_ref+=size(); | |
| } | |
| } | } |
| void String_iterator::skip() { | static void regex_options(char *options, int *result){ |
| if(!position) | struct Regex_option { |
| return; | char key; |
| int clear, set; | |
| int *result; | |
| } regex_option[]={ | |
| {'i', 0, PCRE_CASELESS, result}, // a=A | |
| {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default] | |
| {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored | |
| {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$ | |
| {'g', 0, true, result+1}, // many rows | |
| {0}, | |
| }; | |
| result[0]=PCRE_EXTRA | PCRE_DOTALL; | |
| result[1]=0; | |
| if(options) | |
| for(Regex_option *o=regex_option; o->key; o++) | |
| if( | |
| strchr(options, o->key) || | |
| strchr(options, toupper(o->key))) { | |
| *(o->result)&=~o->clear; | |
| *(o->result)|=o->set; | |
| } | |
| } | |
| if(++position== | /// @todo maybe need speedup: some option to remove pre/match/post string generation |
| read_here->item.ptr+ | bool String::match(const unsigned char *pcre_tables, |
| read_here->item.size) { | const String *aorigin, |
| const String& regexp, | |
| const String *options, | |
| Table **table, | |
| Row_action row_action, void *info, | |
| bool *was_global) const { | |
| if(!regexp.size()) | |
| throw Exception(0, 0, | |
| aorigin, | |
| "regexp is empty"); | |
| const char *pattern=regexp.cstr(); | |
| const char *errptr; | |
| int erroffset; | |
| 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], | |
| &errptr, &erroffset, | |
| pcre_tables); | |
| if(!code) | |
| throw Exception(0, 0, | |
| ®exp.mid(erroffset, regexp.size()), | |
| "regular expression syntax error - %s", errptr); | |
| int info_substrings=pcre_info(code, 0, 0); | |
| if(info_substrings<0) { | |
| pcre_free(code); | |
| throw Exception(0, 0, | |
| aorigin, | |
| "pcre_info error (%d)", | |
| info_substrings); | |
| } | |
| // next row | int startoffset=0; |
| if(++read_here==string.append_here) { | const char *subject=cstr(); |
| position=0; | int length=strlen(subject); |
| return; | int ovecsize; |
| int *ovector=(int *)malloc(sizeof(int)* | |
| (ovecsize=(1/*match*/+info_substrings)*3), 11); | |
| { // create table | |
| Array& columns=*NEW Array(pool()); | |
| columns+=string_pre_match_name; | |
| columns+=string_match_name; | |
| columns+=string_post_match_name; | |
| for(int i=1; i<=info_substrings; i++) { | |
| char *column=(char *)malloc(MAX_NUMBER); | |
| snprintf(column, MAX_NUMBER, "%d", i); | |
| columns+=NEW String(pool(), column); // .i column name | |
| } | } |
| if(read_here==link_row) { | *table=NEW Table(pool(), aorigin, &columns); |
| String::Chunk *chunk=link_row->link; | } |
| if(!chunk) | |
| string.pool().exception().raise(0, 0, | |
| &string, | |
| "String_iterator::skip() missed " | |
| "read_here==string.append_here check"); | |
| read_here=chunk->rows; | int exec_option_bits=0; |
| link_row=&chunk->rows[chunk->count]; | while(true) { |
| int exec_substrings=pcre_exec(code, 0, | |
| subject, length, startoffset, | |
| exec_option_bits, ovector, ovecsize); | |
| if(exec_substrings==PCRE_ERROR_NOMATCH) { | |
| pcre_free(code); | |
| row_action(**table, 0/*last time, no row*/, 0, 0, info); | |
| return option_bits[1]!=0; // global=true+table, not global=false | |
| } | |
| if(exec_substrings<0) { | |
| pcre_free(code); | |
| throw Exception(0, 0, | |
| aorigin, | |
| "regular expression execute error (%d)", | |
| exec_substrings); | |
| } | |
| Array& row=*NEW Array(pool()); | |
| row+=&mid(0, ovector[0]); // .prematch column value | |
| row+=&mid(ovector[0], ovector[1]); // .match | |
| row+=&mid(ovector[1], size()); // .postmatch | |
| for(int i=1; i<exec_substrings; i++) { | |
| // -1:-1 case handled peacefully by mid() itself | |
| row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value | |
| } | |
| row_action(**table, &row, startoffset, ovector[0], info); | |
| if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang | |
| pcre_free(code); | |
| row_action(**table, 0/*last time, no row*/, 0, 0, info); | |
| return true; | |
| } | } |
| position=read_here->item.ptr; | startoffset=ovector[1]; |
| /* | |
| if(option_bits[0] & PCRE_MULTILINE) | |
| exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL | |
| */ | |
| } | } |
| } | } |
| bool String_iterator::skip_to(char c) { | String& String::change_case(Pool& pool, const unsigned char *tables, |
| if(!position) | Change_case_kind kind) const { |
| return false; | String& result=*new(pool) String(pool); |
| 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, 0, | |
| this, | |
| "unknown change case kind #%d", | |
| static_cast<int>(kind)); // never | |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| while(true) { | const Chunk *chunk=&head; |
| if(char *found=static_cast<char *>( | do { |
| memchr(position, c, read_here->item.size-(position-read_here->item.ptr)))) { | const Chunk::Row *row=chunk->rows; |
| position=found; | for(uint i=0; i<chunk->count; i++, row++) { |
| return true; | if(row==append_here) |
| } | goto break2; |
| // next row | char *new_cstr=(char *)pool.malloc(row->item.size, 12); |
| if(++read_here==string.append_here) { | char *dest=new_cstr; |
| position=0; | const char *src=row->item.ptr; |
| return false; | for(int size=row->item.size; size--; src++) { |
| } | unsigned char c=a[(unsigned char)*src]; |
| if(read_here==link_row) { | if(b) |
| String::Chunk *chunk=link_row->link; | c=b[c]; |
| if(!chunk) | |
| string.pool().exception().raise(0, 0, | |
| &string, | |
| "String_iterator::skip_to(char) missed " | |
| "read_here==string.append_here check"); | |
| read_here=chunk->rows; | *dest++=(char)c; |
| link_row=&chunk->rows[chunk->count]; | } |
| result.APPEND(new_cstr, row->item.size, | |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| } | } |
| position=read_here->item.ptr; | chunk=row->link; |
| } | } while(chunk); |
| break2: | |
| return result; | |
| } | } |
| int String_iterator::skip_to(Char_types& types) { | void String::join_chain(Pool& pool, |
| if(!position) | uint& ai, const Chunk*& achunk, const Chunk::Row*& arow, |
| return false; | 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) { | while(true) { |
| int countdown=read_here->item.size-(position-read_here->item.ptr); | const Chunk::Row *row=chunk->rows; |
| for(; countdown--; position++) | for(uint i=0; i<chunk->count; ) { |
| if(int type=types.get(*position)) | if(row==append_here) |
| return type; | goto break2; |
| // next row | uchar joined_lang; |
| if(++read_here==string.append_here) { | const char *joined_ptr; |
| position=0; | size_t joined_size; |
| return -1; | #ifndef NO_STRING_ORIGIN |
| } | const char *joined_origin_file=row->item.origin.file; |
| if(read_here==link_row) { | const size_t joined_origin_line=row->item.origin.line; |
| String::Chunk *chunk=link_row->link; | #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) | if(!chunk) |
| string.pool().exception().raise(0, 0, | goto break2; |
| &string, | } |
| "String_iterator::skip_to(Char_type) missed " | } |
| "read_here==string.append_here check"); | break2: |
| read_here=chunk->rows; | return result; |
| link_row=&chunk->rows[chunk->count]; | }; |
| 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--; | |
| } | |
| } | } |
| position=read_here->item.ptr; | |
| 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 result; | |
| const char *cstr; | |
| 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 | |
| if(cstr[0]=='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 | |
| result=(double)strtod(cstr, &error_pos); | |
| if(*error_pos/*not EOS*/) | |
| throw Exception(0, 0, | |
| this, | |
| "invalid number (double)"); | |
| return result; | |
| } | |
| int String::as_int() const { | |
| int result; | |
| const char *cstr; | |
| 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 | |
| if(cstr[0]=='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 | |
| result=(int)strtol(cstr, &error_pos, 0); | |
| if(*error_pos/*not EOS*/) | |
| throw Exception(0, 0, | |
| this, | |
| "invalid number (int)"); | |
| 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; | |
| } | } |
| } | } |
| */ | |