--- parser3/src/main/pa_string.C 2001/02/22 08:16:31 1.28 +++ parser3/src/main/pa_string.C 2001/03/24 15:58:01 1.49 @@ -1,8 +1,14 @@ -/* - $Id: pa_string.C,v 1.28 2001/02/22 08:16:31 paf Exp $ +/** @file + Parser: string class. @see untaint.C. + + Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) + + Author: Alexander Petrosyan (http://design.ru/paf) + + $Id: pa_string.C,v 1.49 2001/03/24 15:58:01 paf Exp $ */ -#include +#include "pa_config_includes.h" #include "pa_pool.h" #include "pa_string.h" @@ -11,7 +17,7 @@ // String -String::String(Pool& apool) : +String::String(Pool& apool, const char *src, bool tainted) : Pooled(apool) { last_chunk=&head; head.count=CR_PREALLOCATED_COUNT; @@ -19,12 +25,18 @@ String::String(Pool& apool) : head.preallocated_link=0; link_row=&head.rows[head.count]; fused_rows=fsize=0; + + if(src) + if(tainted) + APPEND_TAINTED(src, 0, 0, 0); + else + APPEND(src, 0, 0, 0); } void String::expand() { int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; last_chunk=static_cast( - pool().malloc(sizeof(int)+sizeof(Chunk::Row)*new_chunk_count+sizeof(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; @@ -32,11 +44,10 @@ void String::expand() { link_row->link=0; } -String::String(const String& src) : - Pooled(src.pool()) { +String::String(const String& src) : Pooled(src.pool()) { head.count=CR_PREALLOCATED_COUNT; - int src_used_rows=src.used_rows(); + int src_used_rows=src.fused_rows; if(src_used_rows<=head.count) { // all new rows fit into preallocated area int curr_chunk_rows=head.count; @@ -58,7 +69,7 @@ String::String(const String& src) : // remaining rows into new_chunk 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[new_chunk->count]; @@ -90,28 +101,29 @@ String::String(const String& src) : fsize=src.fsize; } -String& String::operator += (const String& src) { - int src_used_rows=src.used_rows(); +String& String::append(const String& src, Untaint_lang lang, bool forced) { + int src_used_rows=src.fused_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, forced, src_used_rows); append_here+=src_used_rows; } else { - // not all new rows don't fit into last chunk: shrinking it to used part, + // 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( - pool().malloc(sizeof(int)+sizeof(Chunk::Row)*src_used_rows+sizeof(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; + const Chunk *old_chunk=&src.head; Chunk::Row *new_rows=last_chunk->rows; int rows_left_to_copy=src_used_rows; while(true) { @@ -121,6 +133,7 @@ String& String::operator += (const Strin // 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; @@ -129,17 +142,45 @@ String& String::operator += (const Strin // 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; } - 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) { + Chunk *chunk=&head; + do { + Chunk::Row *row=chunk->rows; + for(int i=0; icount; i++) { + if(row==append_here) + goto break2; + + row->item.lang=lang; + row++; + } + chunk=row->link; + } while(chunk); +break2: + return; +} +*/ String& String::real_append(STRING_APPEND_PARAMS) { if(!src) return *this; @@ -153,6 +194,7 @@ String& String::real_append(STRING_APPEN append_here->item.ptr=src; fsize+=append_here->item.size=size; + append_here->item.lang=tainted?UL_YES:UL_NO; #ifndef NO_STRING_ORIGIN append_here->item.origin.file=file; append_here->item.origin.line=line; @@ -162,28 +204,6 @@ String& String::real_append(STRING_APPEN return *this; } -char *String::cstr() const { - char *result=static_cast(pool().malloc(size()+1)); - - char *copy_here=result; - const Chunk *chunk=&head; - do { - const Chunk::Row *row=chunk->rows; - for(int i=0; icount; 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 result=0; @@ -203,10 +223,7 @@ 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; @@ -219,33 +236,37 @@ 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; @@ -257,10 +278,16 @@ bool String::operator == (const String& b_countdown=b_chunk->count; } } - return a_break==b_break; + 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::operator == (char* b_ptr) const { +bool String::operator == (const char* b_ptr) const { size_t b_size=b_ptr?strlen(b_ptr):0; if(size() != b_size) return false; @@ -307,3 +334,16 @@ bool String::operator == (char* b_ptr) c } return a_break==b_break; } + +#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 + // first frequently constant. ex: ^load[/file] "document_root" + "/file" + return append_here[-1].item.origin; +} +#endif