|
|
| version 1.54, 2001/03/29 15:36:16 | version 1.106, 2001/10/05 16:12:40 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class. @see untaint.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://design.ru/paf) | Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) |
| $Id$ | $Id$ |
| Line 10 | Line 9 |
| #include "pa_config_includes.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_common.h" |
| #include "pa_array.h" | |
| // String | #include "pa_globals.h" |
| #include "pa_table.h" | |
| String::String(Pool& apool, const char *src, bool tainted) : | #include "pa_dictionary.h" |
| Pooled(apool) { | |
| String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : | |
| Pooled(apool), | |
| forigins_mode(false) { | |
| last_chunk=&head; | last_chunk=&head; |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| append_here=head.rows; | append_here=head.rows; |
| Line 29 String::String(Pool& apool, const char * | Line 34 String::String(Pool& apool, const char * |
| if(src) | if(src) |
| if(tainted) | if(tainted) |
| APPEND_TAINTED(src, 0, 0, 0); | APPEND_TAINTED(src, src_size, 0, 0); |
| else | else |
| APPEND_CONST(src); | APPEND_CLEAN(src, src_size, 0, 0); |
| } | |
| void String::expand() { | |
| int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; | |
| last_chunk=static_cast<Chunk *>( | |
| malloc(sizeof(int)+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()) { | String::String(const String& src) : |
| Pooled(src.pool()), | |
| forigins_mode(false) { | |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| int src_used_rows=src.fused_rows; | size_t src_used_rows=src.fused_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 |
| int curr_chunk_rows=head.count; | last_chunk=&head; |
| size_t 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 64 String::String(const String& src) : Pool | Line 61 String::String(const String& src) : Pool |
| // 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 |
| int curr_chunk_rows=src_used_rows-head.count; | size_t curr_chunk_rows=src_used_rows-head.count; |
| Chunk *new_chunk=static_cast<Chunk *>( | last_chunk=static_cast<Chunk *>( |
| malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); | malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); |
| new_chunk->count=curr_chunk_rows; | last_chunk->count=curr_chunk_rows; |
| head.preallocated_link=new_chunk; | head.preallocated_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.preallocated_link; |
| Chunk::Row *new_rows=new_chunk->rows; | Chunk::Row *new_rows=last_chunk->rows; |
| int rows_left_to_copy=new_chunk->count; | size_t rows_left_to_copy=last_chunk->count; |
| while(true) { | while(true) { |
| int old_count=old_chunk->count; | size_t 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 102 String::String(const String& src) : Pool | Line 99 String::String(const String& src) : Pool |
| fsize=src.fsize; | fsize=src.fsize; |
| } | } |
| String& String::append(const String& src, Untaint_lang lang, bool forced) { | void String::expand() { |
| int src_used_rows=src.fused_rows; | size_t new_chunk_count=last_chunk->count+CR_GROW_COUNT; |
| int dst_free_rows=link_row-append_here; | last_chunk=static_cast<Chunk *>( |
| malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *))); | |
| if(src_used_rows<=dst_free_rows) { | last_chunk->count=new_chunk_count; |
| // all new rows fit into last chunk | link_row->link=last_chunk; |
| memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows); | append_here=last_chunk->rows; |
| set_lang(append_here, lang, forced, src_used_rows); | link_row=&last_chunk->rows[last_chunk->count]; |
| append_here+=src_used_rows; | link_row->link=0; |
| } else { | |
| // not all new rows fit into last chunk: shrinking it to used part, | |
| int used_rows=last_chunk->count-dst_free_rows; | |
| //int *countp=append_here | |
| link_row=&last_chunk->rows[last_chunk->count=used_rows]; | |
| // allocating only enough mem to fit src string rows | |
| // next append would allocate a new chunk | |
| last_chunk=static_cast<Chunk *>( | |
| malloc(sizeof(int)+sizeof(Chunk::Row)*src_used_rows+sizeof(Chunk *))); | |
| last_chunk->count=src_used_rows; | |
| link_row->link=last_chunk; | |
| append_here=link_row=&last_chunk->rows[src_used_rows]; | |
| const Chunk *old_chunk=&src.head; | |
| Chunk::Row *new_rows=last_chunk->rows; | |
| int rows_left_to_copy=src_used_rows; | |
| while(true) { | |
| int old_count=old_chunk->count; | |
| Chunk *next_chunk=old_chunk->rows[old_count].link; | |
| if(next_chunk) { | |
| // not last source chunk | |
| // taking it all | |
| memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count); | |
| set_lang(new_rows, lang, forced, old_count); | |
| new_rows+=old_count; | |
| rows_left_to_copy-=old_count; | |
| old_chunk=next_chunk; | |
| } else { | |
| // the last source chunk | |
| // taking only those rows of chunk that _left_to_copy | |
| memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy); | |
| set_lang(new_rows, lang, forced, rows_left_to_copy); | |
| break; | |
| } | |
| } | |
| link_row->link=0; | |
| } | |
| fused_rows+=src_used_rows; | |
| fsize+=src.fsize; | |
| return *this; | |
| } | |
| void String::set_lang(Chunk::Row *row, Untaint_lang lang, bool forced, size_t size) { | |
| if(lang==UL_PASS_APPENDED) | |
| return; | |
| while(size--) { | |
| Untaint_lang& item_lang=(row++)->item.lang; | |
| if(item_lang==UL_YES || forced) // tainted? need untaint language assignment | |
| item_lang=lang; // assign untaint language | |
| } | |
| } | } |
| /*void String::change_lang(Untaint_lang lang) { | String& String::append(const String& src, Untaint_lang lang, bool forced) { |
| Chunk *chunk=&head; | const Chunk *chunk=&src.head; |
| do { | do { |
| Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(int i=0; i<chunk->count; i++) { | for(size_t i=0; i<chunk->count; i++, row++) { |
| if(row==append_here) | if(row==src.append_here) |
| goto break2; | goto break2; |
| row->item.lang=lang; | APPEND(row->item.ptr, row->item.size, |
| row++; | (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; | chunk=row->link; |
| } while(chunk); | } while(chunk); |
| break2: | break2: |
| return; | return *this; |
| } | } |
| */ | |
| String& String::real_append(STRING_APPEND_PARAMS) { | String& String::real_append(STRING_APPEND_PARAMS) { |
| if(!src) | if(!src) |
| return *this; | return *this; |
| Line 205 String& String::real_append(STRING_APPEN | Line 151 String& String::real_append(STRING_APPEN |
| return *this; | return *this; |
| } | } |
| char String::first_char() const { | |
| if(!fused_rows) | |
| THROW(0, 0, | |
| this, | |
| "getting first char of empty string"); | |
| return *head.rows[0].item.ptr; | |
| } | |
| uint String::hash_code() const { | uint String::hash_code() const { |
| uint result=0; | uint result=0; |
| const Chunk *chunk=&head; | const Chunk *chunk=&head; |
| do { | do { |
| const Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(int i=0; i<chunk->count; i++) { | for(size_t i=0; i<chunk->count; i++) { |
| if(row==append_here) | if(row==append_here) |
| goto break2; | goto break2; |
| Line 224 break2: | Line 179 break2: |
| return result; | return result; |
| } | } |
| int String::cmp(const String& src) const { | /// @todo move 'lang' skipping to pos |
| int String::cmp(int& partial, const String& src, | |
| size_t this_offset, Untaint_lang lang) const { | |
| partial=-1; | |
| this_offset=min(this_offset, 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; | size_t a_countdown=a_chunk->count; |
| int b_countdown=b_chunk->count; | size_t b_countdown=b_chunk->count; |
| bool a_break=false; | size_t result; |
| bool b_break=false; | size_t pos=0; |
| int result; | |
| while(true) { | bool a_break=size()==0; |
| a_break=a_row==a_end; | bool b_break=src.size()==0; |
| b_break=b_row==b_end; | if(!(a_break || b_break)) while(true) { |
| if(a_break || b_break) | if(pos+a_row->item.size > this_offset) { |
| break; | if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang) |
| return -1; // wrong lang -- bail out | |
| int size_diff= | |
| (a_row->item.size-a_offset)- | int size_diff= |
| (b_row->item.size-b_offset); | (a_row->item.size-a_offset)- |
| (b_row->item.size-b_offset); | |
| if(size_diff==0) { // a has same size as b | |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset); | if(size_diff==0) { // a has same size as b |
| if(result) | result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, |
| return result; | a_row->item.size-a_offset); |
| a_row++; a_countdown--; a_offset=0; | if(result) |
| b_row++; b_countdown--; b_offset=0; | return result; |
| } else if (size_diff>0) { // a longer | pos+=a_row->item.size; |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset); | a_row++; a_countdown--; a_offset=0; |
| if(result) | b_row++; b_countdown--; b_offset=0; |
| return result; | } else if (size_diff>0) { // a longer |
| 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; | b_row->item.size-b_offset); |
| } else { // b longer | if(result) |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset); | return result; |
| if(result) | a_offset+=b_row->item.size-b_offset; |
| return result; | b_row++; b_countdown--; b_offset=0; |
| b_offset+=a_row->item.size-a_offset; | } else { // b longer |
| a_row++; a_countdown--; a_offset=0; | 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) { | 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; | |
| } | |
| } | } |
| if(a_break==b_break) // ended simultaneously | if(a_break==b_break) { // ended simultaneously |
| result=0; | partial=0; return 0; |
| else if(a_break) // first bytes equal, but a ended before b | } else if(a_break) { // first bytes equal, but a ended before b |
| result=-1; | partial=1; return -1; |
| else | } else { |
| result=+1; | partial=2; return +1; |
| return result; | } |
| } | } |
| int String::cmp(const char* b_ptr, int& partial, size_t src_size) const { | /// @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 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); | |
| partial=-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; |
| 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; |
| int a_countdown=a_chunk->count; | size_t a_countdown=a_chunk->count; |
| bool a_break=false; | size_t pos=0; |
| bool b_break=false; | |
| while(true) { | |
| a_break=a_row==a_end; | |
| if(a_break || b_break) | |
| break; | |
| int size_diff= | bool a_break=size()==0; |
| (a_row->item.size-a_offset)- | bool b_break=b_size==0; |
| (b_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(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0) | return -1; // wrong lang -- bail out |
| return result; | |
| a_row++; a_countdown--; a_offset=0; | int size_diff= |
| b_break=true; | (a_row->item.size-a_offset)- |
| } else if (size_diff>0) { // a longer | (b_size-b_offset); |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0) | |
| return result; | if(size_diff==0) { // a has same size as b |
| a_offset+=b_size-b_offset; | if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, |
| b_break=true; | a_row->item.size-a_offset)!=0) |
| } else { // b longer | return result; |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0) | pos+=a_row->item.size; |
| return result; | a_row++; a_countdown--; a_offset=0; |
| b_offset+=a_row->item.size-a_offset; | b_break=true; |
| a_row++; a_countdown--; a_offset=0; | } else if (size_diff>0) { // a longer |
| if(size_t 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(size_t 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; | |
| if(a_break || b_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(a_break==b_break) // ended simultaneously | if(a_break==b_break) { // ended simultaneously |
| return partial=0; | partial=0; return 0; |
| else if(a_break) // first bytes equal, but a ended before b | } else if(a_break) { // first bytes equal, but a ended before b |
| return partial=1; | partial=1; return -1; |
| else | } else { |
| return partial=2; | partial=2; return +1; |
| } | |
| } | } |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| const Origin& String::origin() const { | const Origin& String::origin() const { |
| if(!fused_rows) | if(!fused_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. |
| // ex: ^load[/file] "document_root" + "/file" | // ex: ^load[/file] "document_root" + "/file" |
| return append_here[-1].item.origin; | // 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 | #endif |
| String& String::piece(size_t start, size_t finish) const { | String& String::mid(size_t start, size_t finish) const { |
| start=max(0, start); | start=max(0, start); |
| finish=min(size(), finish); | finish=min(size(), finish); |
| if(start==finish) | |
| return *empty_string; | |
| String& result=*NEW String(pool()); | String& result=*NEW String(pool()); |
| Line 364 String& String::piece(size_t start, size | Line 368 String& String::piece(size_t start, size |
| const Chunk *chunk=&head; | const Chunk *chunk=&head; |
| do { | do { |
| const Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(int i=0; i<chunk->count; pos+=row->item.size, i++, row++) { | for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) { |
| if(row==append_here) | if(row==append_here) |
| goto break2; | goto break2; |
| if(start>=pos) { // started now or already? | size_t item_finish=pos+row->item.size; |
| size_t item_finish=pos+row->item.size; | if(item_finish > start) { // started now or already? |
| bool started=start < item_finish; // started now? | bool started=result.size()==0; // started now? |
| bool finished=finish < item_finish; // finished now? | bool finished=finish <= item_finish; // finished now? |
| size_t offset=started?start-pos:0; | size_t offset=started?start-pos:0; |
| size_t size=finished?finish-pos:row->item.size; | size_t size=finished?finish-pos:row->item.size; |
| result.APPEND( | result.APPEND( |
| Line 385 String& String::piece(size_t start, size | Line 389 String& String::piece(size_t start, size |
| chunk=row->link; | chunk=row->link; |
| } while(chunk); | } while(chunk); |
| break2: | break2: |
| // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", | |
| //cstr(), start, finish, result.cstr()); | |
| return result; | |
| } | |
| int String::pos(const String& substr, | |
| size_t result, Untaint_lang lang) const { | |
| for(; result<size(); result++) { | |
| int partial; cmp(partial, substr, result, lang); | |
| if( | |
| partial==0 || // full match | |
| partial==2) // 'substr' starts 'this'+'result' | |
| return result; | |
| } | |
| return -1; | |
| } | |
| int String::pos(const char *substr, size_t substr_size, | |
| size_t result, Untaint_lang lang) const { | |
| for(; result<size(); result++) { | |
| int partial; cmp(partial, substr, substr_size, result, lang); | |
| if( | |
| partial==0 || // full match | |
| partial==2) // 'substr' starts 'this'+'result' | |
| return result; | |
| } | |
| return -1; | |
| } | |
| void String::split(Array& result, | |
| size_t* pos_after_ref, | |
| const char *delim, size_t delim_size, | |
| 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, 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<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::split(Array& result, | |
| 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(); | |
| } | |
| } | |
| static void regex_options(char *options, int *result){ | |
| struct Regex_option { | |
| 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; | |
| } | |
| } | |
| /// @todo maybe need speedup: some option to remove pre/match/post string generation | |
| bool String::match(const unsigned char *pcre_tables, | |
| 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(0, 0, | |
| aorigin, | |
| "regexp is empty"); | |
| const char *pattern=regexp.cstr(UL_AS_IS); | |
| 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(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(0, 0, | |
| aorigin, | |
| "pcre_info error (%d)", | |
| info_substrings); | |
| } | |
| int startoffset=0; | |
| const char *subject=cstr(UL_AS_IS); | |
| int length=strlen(subject); | |
| int ovecsize; | |
| int *ovector=(int *)malloc(sizeof(int)* | |
| (ovecsize=(1/*match*/+info_substrings)*3)); | |
| { // 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 | |
| } | |
| *table=NEW Table(pool(), aorigin, &columns); | |
| } | |
| int exec_option_bits=0; | |
| 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(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; | |
| } | |
| startoffset=ovector[1]; | |
| /* | |
| if(option_bits[0] & PCRE_MULTILINE) | |
| exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL | |
| */ | |
| } | |
| } | |
| String& String::change_case(Pool& pool, const unsigned char *tables, | |
| Change_case_kind kind) const { | |
| 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: | |
| PTHROW(0, 0, | |
| this, | |
| "unknown change case kind #%d", | |
| static_cast<int>(kind)); // never | |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| const Chunk *chunk=&head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; i++, row++) { | |
| if(row==append_here) | |
| goto break2; | |
| char *new_cstr=(char *)pool.malloc(row->item.size); | |
| char *dest=new_cstr; | |
| const char *src=row->item.ptr; | |
| for(int size=row->item.size; size--; src++) { | |
| unsigned char c=a[(unsigned char)*src]; | |
| if(b) | |
| c=b[c]; | |
| *dest++=(char)c; | |
| } | |
| result.APPEND(new_cstr, row->item.size, | |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | |
| return result; | |
| } | |
| /// @test restructure string to link pieces of same language together prior to this. | |
| String& String::replace(Pool& pool, Dictionary& dict) const { | |
| //_asm int 3; | |
| String& result=*new(pool) String(pool); | |
| const Chunk *chunk=&head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; i++, row++) { | |
| if(row==append_here) | |
| goto break2; | |
| 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())); | |
| 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); | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | |
| return result; | return result; |
| } | } |
| size_t String::pos(const String& substr) const { | |
| return 0; | |
| } | |
| double String::as_double() const { | |
| double result; | |
| const char *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(0, 0, | |
| this, | |
| "invalid number (double)"); | |
| return result; | |
| } | |
| int String::as_int() const { | |
| int result; | |
| const char *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(0, 0, | |
| this, | |
| "invalid number (int)"); | |
| return result; | |
| } |