|
|
| version 1.30, 2001/02/22 10:43:46 | version 1.60, 2001/04/03 05:23:41 |
|---|---|
| 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://design.ru/paf) | |
| $Id$ | |
| */ | */ |
| #include <string.h> | #include "pa_config_includes.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_sapi.h" | |
| // String | // String |
| String::String(Pool& apool) : | String::String(Pool& apool, const char *src, bool tasize_ted) : |
| Pooled(apool) { | Pooled(apool) { |
| last_chunk=&head; | last_chunk=&head; |
| head.count=CR_PREALLOCATED_COUNT; | head.count=CR_PREALLOCATED_COUNT; |
| Line 19 String::String(Pool& apool) : | Line 30 String::String(Pool& apool) : |
| head.preallocated_link=0; | head.preallocated_link=0; |
| link_row=&head.rows[head.count]; | link_row=&head.rows[head.count]; |
| fused_rows=fsize=0; | fused_rows=fsize=0; |
| if(src) | |
| if(tasize_ted) | |
| APPEND_TAINTED(src, 0, 0, 0); | |
| else | |
| APPEND_CONST(src); | |
| } | } |
| void String::expand() { | void String::expand() { |
| int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; | size_t new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; |
| last_chunk=static_cast<Chunk *>( | last_chunk=static_cast<Chunk *>( |
| malloc(sizeof(int)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *))); | malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *))); |
| last_chunk->count=new_chunk_count; | last_chunk->count=new_chunk_count; |
| link_row->link=last_chunk; | link_row->link=last_chunk; |
| append_here=last_chunk->rows; | append_here=last_chunk->rows; |
| Line 32 void String::expand() { | Line 49 void String::expand() { |
| link_row->link=0; | 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(); | 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; | 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 52 String::String(const String& src) : | Line 68 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 |
| 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 *>( | Chunk *new_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; | new_chunk->count=curr_chunk_rows; |
| head.preallocated_link=new_chunk; | head.preallocated_link=new_chunk; |
| append_here=link_row=&new_chunk->rows[new_chunk->count]; | append_here=link_row=&new_chunk->rows[new_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=new_chunk->rows; |
| int rows_left_to_copy=new_chunk->count; | size_t rows_left_to_copy=new_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 90 String::String(const String& src) : | Line 106 String::String(const String& src) : |
| fsize=src.fsize; | fsize=src.fsize; |
| } | } |
| String& String::operator += (const String& src) { | String& String::append(const String& src, Untaint_lang lang, bool forced) { |
| int src_used_rows=src.used_rows(); | const Chunk *chunk=&src.head; |
| int dst_free_rows=link_row-append_here; | do { |
| const Chunk::Row *row=chunk->rows; | |
| if(src_used_rows<=dst_free_rows) { | for(size_t i=0; i<chunk->count; i++, row++) { |
| // all new rows fit into last chunk | if(row==src.append_here) |
| memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows); | goto break2; |
| append_here+=src_used_rows; | |
| } else { | APPEND(row->item.ptr, row->item.size, |
| // not all new rows don't fit into last chunk: shrinking it to used part, | (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang, |
| int used_rows=last_chunk->count-dst_free_rows; | row->item.origin.file, row->item.origin.line); |
| //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]; | |
| Chunk *old_chunk=src.head.preallocated_link; | |
| 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); | |
| 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); | |
| break; | |
| } | |
| } | } |
| link_row->link=0; | chunk=row->link; |
| } | } while(chunk); |
| fused_rows+=src_used_rows; | break2: |
| fsize+=src.fsize; | |
| return *this; | return *this; |
| } | } |
| Line 153 String& String::real_append(STRING_APPEN | Line 137 String& String::real_append(STRING_APPEN |
| append_here->item.ptr=src; | append_here->item.ptr=src; |
| fsize+=append_here->item.size=size; | fsize+=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; |
| Line 162 String& String::real_append(STRING_APPEN | Line 147 String& String::real_append(STRING_APPEN |
| return *this; | return *this; |
| } | } |
| char *String::cstr() const { | |
| char *result=static_cast<char *>(malloc(size()+1)); | |
| char *copy_here=result; | |
| 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; |
| 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 203 break2: | Line 166 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; | |
| 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; | bool a_break=false; |
| bool b_break=false; | bool b_break=false; |
| size_t result; | |
| size_t pos=0; | |
| while(true) { | while(true) { |
| int size_diff= | |
| (a_row->item.size-a_offset)- | |
| (b_row->item.size-b_offset); | |
| if(size_diff==0) { // a has same size as b | |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset)!=0) | |
| return false; | |
| a_row++; a_countdown--; a_offset=0; | |
| b_row++; b_countdown--; b_offset=0; | |
| } else if (size_diff>0) { // a longer | |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset)!=0) | |
| return false; | |
| a_offset+=b_row->item.size-b_offset; | |
| b_row++; b_countdown--; b_offset=0; | |
| } else { // b longer | |
| if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset)!=0) | |
| return false; | |
| b_offset+=a_row->item.size-a_offset; | |
| a_row++; a_countdown--; a_offset=0; | |
| } | |
| a_break=a_row==a_end; | a_break=a_row==a_end; |
| b_break=b_row==b_end; | b_break=b_row==b_end; |
| if(a_break || b_break) | if(a_break || b_break) |
| break; | break; |
| if(pos+a_row->item.size > this_offset) { | |
| if(lang!=UL_UNKNOWN && a_row->item.lang!=lang) | |
| return -1; // wrong lang -- bail out | |
| int size_diff= | |
| (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(result) | |
| return result; | |
| pos+=a_row->item.size; | |
| 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_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_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; | |
| } | |
| } | } |
| bool String::operator == (char* b_ptr) const { | /// @todo move 'lang' skipping to pos |
| size_t b_size=b_ptr?strlen(b_ptr):0; | int String::cmp(int& partial, const char* b_ptr, size_t src_size, |
| if(size() != b_size) | size_t this_offset, Untaint_lang lang) const { |
| return false; | partial=-1; |
| size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0; | |
| this_offset=min(this_offset, 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; |
| 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; | bool a_break=false; |
| bool b_break=false; | bool b_break=false; |
| size_t pos=0; | |
| while(true) { | while(true) { |
| int size_diff= | |
| (a_row->item.size-a_offset)- | |
| (b_size-b_offset); | |
| if(size_diff==0) { // a has same size as b | |
| if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0) | |
| return false; | |
| a_row++; a_countdown--; a_offset=0; | |
| b_break=true; | |
| } else if (size_diff>0) { // a longer | |
| if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0) | |
| return false; | |
| a_offset+=b_size-b_offset; | |
| b_break=true; | |
| } else { // b longer | |
| if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0) | |
| return false; | |
| b_offset+=a_row->item.size-a_offset; | |
| a_row++; a_countdown--; a_offset=0; | |
| } | |
| a_break=a_row==a_end; | a_break=a_row==a_end; |
| if(a_break || b_break) | if(a_break || b_break) |
| break; | break; |
| if(pos+a_row->item.size > this_offset) { | |
| if(lang!=UL_UNKNOWN && 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(size_t 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(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--; | |
| } | |
| 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; |
| } | } |
| } | } |
| 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; | |
| } | |
| } | |
| #ifndef NO_STRING_ORIGIN | |
| const Origin& String::origin() const { | |
| if(!fused_rows) | |
| THROW(0, 0, | |
| 0, | |
| "String::origin() of empty string called"); | |
| // determining origin by last appended piece | |
| // because first one frequently constant. | |
| // ex: ^load[/file] "document_root" + "/file" | |
| return append_here[-1].item.origin; | |
| } | |
| #endif | |
| String& String::piece(size_t start, size_t finish) const { | |
| start=max(0, start); | |
| finish=min(size(), finish); | |
| if(start==finish) | |
| return *empty_string; | |
| String& result=*NEW String(pool()); | |
| size_t pos=0; | |
| const Chunk *chunk=&head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) { | |
| if(row==append_here) | |
| 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; | |
| } while(chunk); | |
| 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+=&piece(pos_after, pos_before); | |
| pos_after=pos_before+delim_size; | |
| } | |
| // last piece | |
| if(pos_after<size() && limit) { | |
| result+=&piece(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+=&piece(pos_after, pos_before); | |
| pos_after=pos_before+delim.size(); | |
| } | |
| // last piece | |
| if(pos_after<size() && limit) { | |
| result+=&piece(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(); | |
| } | |
| } | |