--- parser3/src/main/pa_string.C 2001/04/05 13:27:14 1.71 +++ parser3/src/main/pa_string.C 2001/08/01 12:08:40 1.97 @@ -4,15 +4,13 @@ Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) Author: Alexander Petrosyan (http://design.ru/paf) - - $Id: pa_string.C,v 1.71 2001/04/05 13:27:14 paf Exp $ */ +static const char *RCSId="$Id: pa_string.C,v 1.97 2001/08/01 12:08:40 parser Exp $"; #include "pa_config_includes.h" -#include - #include "pcre.h" +#include "internal.h" #include "pa_pool.h" #include "pa_string.h" @@ -22,14 +20,10 @@ #include "pa_array.h" #include "pa_globals.h" #include "pa_table.h" -#include "pa_threads.h" - -//#include "pa_sapi.h" -// String - -String::String(Pool& apool, const char *src, bool tasize_ted) : - Pooled(apool) { +String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : + Pooled(apool), + forigins_mode(false) { last_chunk=&head; head.count=CR_PREALLOCATED_COUNT; append_here=head.rows; @@ -38,24 +32,15 @@ String::String(Pool& apool, const char * fused_rows=fsize=0; if(src) - if(tasize_ted) - APPEND_TAINTED(src, 0, 0, 0); + if(tainted) + APPEND_TAINTED(src, src_size, 0, 0); else - APPEND_CONST(src); + APPEND_CLEAN(src, src_size, 0, 0); } -void String::expand() { - size_t new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; - last_chunk=static_cast( - malloc(sizeof(size_t)+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; size_t src_used_rows=src.fused_rows; @@ -112,6 +97,17 @@ String::String(const String& src) : Pool fsize=src.fsize; } +void String::expand() { + size_t new_chunk_count=last_chunk->count+CR_GROW_COUNT; + last_chunk=static_cast( + malloc(sizeof(size_t)+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::append(const String& src, Untaint_lang lang, bool forced) { const Chunk *chunk=&src.head; do { @@ -153,6 +149,15 @@ String& String::real_append(STRING_APPEN 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 result=0; @@ -188,16 +193,12 @@ int String::cmp(int& partial, const Stri Chunk::Row *b_end=src.append_here; size_t a_countdown=a_chunk->count; size_t b_countdown=b_chunk->count; - bool a_break=false; - bool b_break=false; size_t result; size_t pos=0; - while(true) { - a_break=a_row==a_end; - b_break=b_row==b_end; - if(a_break || b_break) - break; + bool a_break=size()==0; + bool b_break=src.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 @@ -230,7 +231,10 @@ int String::cmp(int& partial, const Stri 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; @@ -242,6 +246,10 @@ int String::cmp(int& partial, const Stri 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; @@ -270,14 +278,11 @@ int String::cmp(int& partial, const char size_t b_offset=0; Chunk::Row *a_end=append_here; size_t a_countdown=a_chunk->count; - bool a_break=false; - bool b_break=false; size_t pos=0; - while(true) { - a_break=a_row==a_end; - if(a_break || b_break) - break; + bool a_break=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 @@ -313,6 +318,9 @@ int String::cmp(int& partial, const char a_row++; a_countdown--; } + a_break=a_row==a_end; + if(a_break || b_break) + break; if(!a_countdown) { a_chunk=a_row->link; a_row=a_chunk->rows; @@ -330,15 +338,19 @@ int String::cmp(int& partial, const char #ifndef NO_STRING_ORIGIN const Origin& String::origin() const { - if(!fused_rows) - THROW(0, 0, - 0, - "String::origin() of empty string called"); + if(!fused_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" - 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 @@ -458,7 +470,6 @@ void String::split(Array& result, } } -/// @test really @b test: s x m [tested: i & g ] static void regex_options(char *options, int *result){ struct Regex_option { char key; @@ -466,7 +477,7 @@ static void regex_options(char *options, int *result; } regex_option[]={ {'i', 0, PCRE_CASELESS, result}, // a=A - {'s', 0, PCRE_DOTALL, result}, // \n\n$ + {'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 @@ -485,46 +496,41 @@ static void regex_options(char *options, } } -/** - returns true if fills table. - table format is defined and fixed[can be used by others]: - @verbatim - pre-match/match/post-match/1/2/3/... - @endverbatim - @test setlocale param to auto.p -*/ -bool String::match(const String *aorigin, +/// @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) const { + Row_action row_action, void *info, + bool *was_global) const { - static const unsigned char *tables=0; { SYNCHRONIZED(true); - if(!tables) { - setlocale(LC_CTYPE, "ru"); - tables=pcre_maketables(); - } - } + 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, - tables); + pcre_tables); if(!code) THROW(0, 0, ®exp.mid(erroffset, regexp.size()), - "match error - %s", errptr); + "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); + aorigin, + "pcre_info error (%d)", + info_substrings); } int startoffset=0; @@ -563,14 +569,14 @@ bool String::match(const String *aorigin (*pcre_free)(code); THROW(0, 0, aorigin, - "pcre_exec error #%d", + "regular expression execute error (%d)", exec_substrings); } Array& row=*NEW Array(pool()); - row+=&mid(0, ovector[0]); // .pre-match column value + row+=&mid(0, ovector[0]); // .prematch column value row+=&mid(ovector[0], ovector[1]); // .match - row+=&mid(ovector[1], size()); // .post-match + row+=&mid(ovector[1], size()); // .postmatch for(int i=1; i(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; icount; 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; +} + +double String::as_double() const { + double result; + const char *cstr=this->cstr(); + char *error_pos=0; + // 0xABC + if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X')) + result=(double)(unsigned long)strtol(cstr, &error_pos, 0); + else + result=strtod(cstr, &error_pos); + + if(error_pos && *error_pos) + THROW(0, 0, + this, + "invalid number (double)"); + + return result; +} +int String::as_int() const { + int result; + const char *cstr=this->cstr(); + char *error_pos=0; + // 0xABC + if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X')) + result=(int)(unsigned long)strtol(cstr, &error_pos, 0); + else + result=(int)strtol(cstr, &error_pos, 0); + + if(error_pos && *error_pos) + THROW(0, 0, + this, + "invalid number (int)"); + + return result; +}