--- parser3/src/main/pa_string.C 2001/02/20 18:45:53 1.26 +++ parser3/src/main/pa_string.C 2001/03/10 12:12:51 1.35 @@ -1,5 +1,5 @@ /* - $Id: pa_string.C,v 1.26 2001/02/20 18:45:53 paf Exp $ + $Id: pa_string.C,v 1.35 2001/03/10 12:12:51 paf Exp $ */ #include @@ -13,21 +13,22 @@ String::String(Pool& apool) : Pooled(apool) { - head.count=curr_chunk_rows=CR_PREALLOCATED_COUNT; + last_chunk=&head; + head.count=CR_PREALLOCATED_COUNT; append_here=head.rows; head.preallocated_link=0; - link_row=&head.rows[curr_chunk_rows]; + link_row=&head.rows[head.count]; fused_rows=fsize=0; } void String::expand() { - curr_chunk_rows+=curr_chunk_rows*CR_GROW_PERCENT/100; - Chunk *chunk=static_cast( - pool().malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); - chunk->count=curr_chunk_rows; - link_row->link=chunk; - append_here=chunk->rows; - link_row=&chunk->rows[curr_chunk_rows]; + int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; + last_chunk=static_cast( + 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; } @@ -38,7 +39,7 @@ String::String(const String& src) : int src_used_rows=src.used_rows(); if(src_used_rows<=head.count) { // all new rows fit into preallocated area - curr_chunk_rows=head.count; + int curr_chunk_rows=head.count; memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows); append_here=&head.rows[src_used_rows]; link_row=&head.rows[curr_chunk_rows]; @@ -55,16 +56,16 @@ String::String(const String& src) : // preallocated chunk src to constructing head memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count); // remaining rows into new_chunk - curr_chunk_rows=src_used_rows-head.count; + int curr_chunk_rows=src_used_rows-head.count; Chunk *new_chunk=static_cast( - pool().malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); + malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *))); new_chunk->count=curr_chunk_rows; head.preallocated_link=new_chunk; - append_here=link_row=&new_chunk->rows[curr_chunk_rows]; + append_here=link_row=&new_chunk->rows[new_chunk->count]; Chunk *old_chunk=src.head.preallocated_link; Chunk::Row *new_rows=new_chunk->rows; - int rows_left_to_copy=curr_chunk_rows; + int rows_left_to_copy=new_chunk->count; while(true) { int old_count=old_chunk->count; Chunk *next_chunk=old_chunk->rows[old_count].link; @@ -88,11 +89,69 @@ String::String(const String& src) : fused_rows=src_used_rows; fsize=src.fsize; } -/* -String(const String_iterator& begin, const String_iterator& end) { - ;//TODO + +String& String::append(const String& src, Untaint_lang lang) { + int src_used_rows=src.used_rows(); + int dst_free_rows=link_row-append_here; + + if(src_used_rows<=dst_free_rows) { + // all new rows fit into last chunk + memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows); + set_lang(append_here, lang, src_used_rows); + append_here+=src_used_rows; + } 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( + 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, 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, 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, size_t size) { + if(lang==PASS_APPENDED) + return; + + while(size--) { + Untaint_lang& item_lang=(row++)->item.lang; + if(item_lang==YES) // tainted? need untaint language assignment + item_lang=lang; // assign untaint language + } } -*/ String& String::real_append(STRING_APPEND_PARAMS) { if(!src) @@ -107,6 +166,7 @@ String& String::real_append(STRING_APPEN append_here->item.ptr=src; fsize+=append_here->item.size=size; + append_here->item.lang=tainted?/*Untaint_lang::*/YES:Untaint_lang::NO; #ifndef NO_STRING_ORIGIN append_here->item.origin.file=file; append_here->item.origin.line=line; @@ -117,7 +177,7 @@ String& String::real_append(STRING_APPEN } char *String::cstr() const { - char *result=static_cast(pool().malloc(size()+1)); + char *result=static_cast(malloc(size()+1)); char *copy_here=result; const Chunk *chunk=&head; @@ -127,7 +187,22 @@ char *String::cstr() const { if(row==append_here) goto break2; - memcpy(copy_here, row->item.ptr, row->item.size); + switch(row->item.lang) { + case NO: + case YES: // for VString.get_double of tainted values + case AS_IS: + memcpy(copy_here, row->item.ptr, row->item.size); + break; + case HTML_TYPO: + memset(copy_here, '?', row->item.size); + break; + default: + THROW(0,0, + this, + "unknown untaint language #%d of %d piece", + static_cast(row->item.lang), + i); + } copy_here+=row->item.size; row++; } @@ -157,10 +232,7 @@ break2: return result; } -bool String::operator == (const String& src) const { - if(size() != src.size()) - return false; - +int String::cmp(const String& src) const { const Chunk *a_chunk=&head; const Chunk *b_chunk=&src.head; const Chunk::Row *a_row=a_chunk->rows; @@ -173,33 +245,37 @@ bool String::operator == (const String& int b_countdown=b_chunk->count; bool a_break=false; bool b_break=false; + int result; while(true) { + a_break=a_row==a_end; + b_break=b_row==b_end; + if(a_break || b_break) + break; + 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; + result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset); + if(result) + return result; 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; + 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 - if(memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset)!=0) - return false; + 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; a_row++; a_countdown--; a_offset=0; } - a_break=a_row==a_end; - b_break=b_row==b_end; - if(a_break || b_break) - break; - if(!a_countdown) { a_chunk=a_row->link; a_row=a_chunk->rows; @@ -211,126 +287,59 @@ bool String::operator == (const String& b_countdown=b_chunk->count; } } - return a_break==b_break; -} -/* -String& String::append(const String_iterator& begin, const String_iterator& end) { - //TODO - return *this; -} - -// Char_types - -Char_types::Char_types() { - memset(types, 0, sizeof(types)); -} - -void Char_types::set(char from, char to, int type) { - memset(&types[static_cast(from)], type, to-from+1); -} - -// String_iterator - -String_iterator::String_iterator(String& astring) : string(astring) { - read_here=string.head.rows; - position=string.size()==0?0:read_here->item.ptr; - link_row=reinterpret_cast(string.head.preallocated_link); -} -/* -String_iterator::String_iterator(String_iterator& asi) { - //TODO -} - -char String_iterator::operator()() const { - return position?*position:0; -} - -void String_iterator::skip() { - if(!position) - return; - - if(++position== - read_here->item.ptr+ - read_here->item.size) { - - // next row - if(++read_here==string.append_here) { - position=0; - return; - } - if(read_here==link_row) { - 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; - link_row=&chunk->rows[chunk->count]; - } - position=read_here->item.ptr; - } + if(a_break==b_break) // ended simultaneously + result=0; + else if(a_break) // first bytes equal, but a ended before b + result=-1; + else + result=+1; + return result; } -bool String_iterator::skip_to(char c) { - if(!position) +bool String::operator == (char* b_ptr) const { + size_t b_size=b_ptr?strlen(b_ptr):0; + if(size() != b_size) return false; + const Chunk *a_chunk=&head; + const Chunk::Row *a_row=a_chunk->rows; + int a_offset=0; + int b_offset=0; + Chunk::Row *a_end=append_here; + int a_countdown=a_chunk->count; + bool a_break=false; + bool b_break=false; while(true) { - if(char *found=static_cast( - memchr(position, c, read_here->item.size-(position-read_here->item.ptr)))) { - position=found; - return true; - } - - // next row - if(++read_here==string.append_here) { - position=0; - return false; - } - if(read_here==link_row) { - String::Chunk *chunk=link_row->link; - if(!chunk) - string.pool().exception().raise(0, 0, - &string, - "String_iterator::skip_to(char) missed " - "read_here==string.append_here check"); + int size_diff= + (a_row->item.size-a_offset)- + (b_size-b_offset); - read_here=chunk->rows; - link_row=&chunk->rows[chunk->count]; + 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; } - position=read_here->item.ptr; - } -} - -int String_iterator::skip_to(Char_types& types) { - if(!position) - return false; - while(true) { - int countdown=read_here->item.size-(position-read_here->item.ptr); - for(; countdown--; position++) - if(int type=types.get(*position)) - return type; - - // next row - if(++read_here==string.append_here) { - position=0; - return -1; - } - if(read_here==link_row) { - String::Chunk *chunk=link_row->link; - if(!chunk) - string.pool().exception().raise(0, 0, - &string, - "String_iterator::skip_to(Char_type) missed " - "read_here==string.append_here check"); + a_break=a_row==a_end; + if(a_break || b_break) + break; - read_here=chunk->rows; - link_row=&chunk->rows[chunk->count]; + if(!a_countdown) { + a_chunk=a_row->link; + a_row=a_chunk->rows; + a_countdown=a_chunk->count; } - position=read_here->item.ptr; } + return a_break==b_break; } -*/ \ No newline at end of file