|
|
| version 1.155, 2002/04/19 09:36:51 | version 1.172.2.8, 2003/01/28 12:16:18 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class. @see untasize_t.C. | Parser: string class. @see untasize_t.C. |
| Copyright (c) 2001, 2002 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001, 2003 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| $Id$ | |
| */ | */ |
| static const char* IDENT_STRING_C="$Date$"; | |
| #include "pcre.h" | #include "pcre.h" |
| #include "pa_pool.h" | |
| #include "pa_string.h" | #include "pa_string.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_table.h" | #include "pa_table.h" |
| #include "pa_dictionary.h" | #include "pa_dictionary.h" |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #define DEBUG_STRING_APPENDS_VS_EXPANDS | // helpers |
| /// String::match uses this as replace & global search table columns | |
| #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS | const int MAX_MATCH_GROUPS=100; |
| ulong string_piece_appends=0; | |
| #endif | class String_match_table_template_columns: public Array<ConstStringPtr> { |
| Pool pool; | |
| public: | |
| String_match_table_template_columns() { | |
| *this+=ConstStringPtr(new String("prematch")); | |
| *this+=ConstStringPtr(new String("match")); | |
| *this+=ConstStringPtr(new String("postmatch")); | |
| for(int i=0; i<MAX_MATCH_GROUPS; i++) { | |
| char *cname=new(pool) char[3/*strlen("100")*/+1/*terminating 0*/]; | |
| ConstStringPtr sname(new String(cname, sprintf(cname, "%d", 1+i))); | |
| *this+=sname; | |
| } | |
| } | |
| }; | |
| String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : | Table string_match_table_template( |
| Pooled(apool) { | ConstStringPtrZero, |
| last_chunk=&head.chunk; | Table::columns_type(new String_match_table_template_columns)); |
| head.chunk.count=CR_PREALLOCATED_COUNT; | |
| append_here=head.chunk.rows; | |
| // methods | |
| String::String(const char *src, size_t src_size, bool tainted): Array<String_fragment>(1), fsize(0) { | |
| if(src) | if(src) |
| if(tainted) | if(tainted) |
| APPEND_TAINTED(src, src_size, 0, 0); | APPEND_TAINTED(src, src_size, 0, 0); |
| Line 40 String::String(Pool& apool, const char * | Line 50 String::String(Pool& apool, const char * |
| APPEND_CLEAN(src, src_size, 0, 0); | APPEND_CLEAN(src, src_size, 0, 0); |
| } | } |
| String::String(const String& src) : | String::String(const String& src): Array<String_fragment>(src.count()) { |
| Pooled(src.pool()) { | append(src, UL_PASS_APPENDED); |
| last_chunk=&head.chunk; | |
| head.chunk.count=CR_PREALLOCATED_COUNT; | |
| append_here=head.chunk.rows; | |
| append(src, UL_UNSPECIFIED); | |
| } | |
| size_t String::size() const { | |
| size_t result=0; | |
| STRING_FOREACH_ROW( | |
| result+=row->item.size; | |
| ); | |
| return result; | |
| } | |
| /// @todo not very optimal | |
| uint String::used_rows() const { | |
| uint result=0; | |
| STRING_FOREACH_ROW( | |
| result++; | |
| ); | |
| return result; | |
| } | |
| void String::expand() { | |
| uint new_chunk_count=last_chunk->count+CR_GROW_COUNT; | |
| if(new_chunk_count>max_integral(Chunk::count_type)) | |
| new_chunk_count=max_integral(Chunk::count_type); | |
| Chunk *new_chunk=static_cast<Chunk *>(malloc( | |
| sizeof(Chunk)// count+interpadding(?)+rows[CR_PREALLOCATED_COUNT]+tailpadding(??) | |
| -sizeof(Chunk::rows_type) // PREALLOCATED rows | |
| +sizeof(Chunk::Row)*new_chunk_count // neaded rows | |
| +sizeof(Chunk *) // link size | |
| , 10)); | |
| new_chunk->rows[new_chunk->count=new_chunk_count].link=0; | |
| last_chunk->rows[last_chunk->count].link=new_chunk; | |
| last_chunk=new_chunk; | |
| append_here=last_chunk->rows; | |
| } | } |
| String& String::real_append(STRING_APPEND_PARAMS) { | String& String::real_append(STRING_APPEND_PARAMS) { |
| if(!last_chunk) // growth stopped [we're appended as string to somebody] | |
| throw Exception(0, | |
| this, | |
| "string growth stopped (append cstr)"); | |
| if(!src) | if(!src) |
| return *this; | return *this; |
| if(!size) | if(!size) |
| Line 96 String& String::real_append(STRING_APPEN | Line 62 String& String::real_append(STRING_APPEN |
| if(!size) | if(!size) |
| return *this; | return *this; |
| #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS | if(is_full()) |
| string_piece_appends++; | expand(fdelta); |
| #endif | |
| // manually unrolled to avoid extra check | String_fragment *fragment=&felements[fused++]; |
| while(size>max_integral(Chunk::Row::item_size_type)) { | fragment->ptr=src; |
| if(chunk_is_full()) | fragment->size=size; |
| expand(); | fragment->lang=lang; |
| append_here->item.ptr=src; | |
| append_here->item.size=max_integral(Chunk::Row::item_size_type); | |
| append_here->item.lang=lang; | |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| append_here->item.origin.file=file; | fragment->origin.file=file; |
| append_here->item.origin.line=line; | fragment->origin.line=line; |
| #endif | #endif |
| append_here++; | |
| src+=max_integral(Chunk::Row::item_size_type); | |
| size-=max_integral(Chunk::Row::item_size_type); | |
| } | |
| if(chunk_is_full()) | |
| expand(); | |
| append_here->item.ptr=src; | |
| append_here->item.size=size; | |
| append_here->item.lang=lang; | |
| #ifndef NO_STRING_ORIGIN | |
| append_here->item.origin.file=file; | |
| append_here->item.origin.line=line; | |
| #endif | |
| append_here++; | |
| return *this; | return *this; |
| } | } |
| char String::first_char() const { | |
| if(is_empty()) | |
| throw Exception(0, | |
| this, | |
| "getting first char of empty string"); | |
| return *head.chunk.rows[0].item.ptr; | |
| } | |
| uint String::hash_code() const { | uint String::hash_code() const { |
| uint result=0; | uint result=0; |
| STRING_FOREACH_ROW( | STRING_FOREACH_FRAGMENT( |
| result=Hash::generic_code(result, row->item.ptr, row->item.size); | result=generic_hash_code(result, fragment->ptr, fragment->size); |
| ); | ); |
| return result; | return result; |
| } | } |
| Line 157 int String::cmp(int& partial, const Stri | Line 92 int String::cmp(int& partial, const Stri |
| size_t a_size=size(); | size_t a_size=size(); |
| this_offset=min(this_offset, a_size-1); | this_offset=min(this_offset, a_size-1); |
| const Chunk *a_chunk=&head.chunk; | const String_fragment *a_current=felements; |
| const Chunk *b_chunk=&src.head.chunk; | const String_fragment *b_current=src.felements; |
| const Chunk::Row *a_row=a_chunk->rows; | |
| const Chunk::Row *b_row=b_chunk->rows; | |
| size_t a_offset=this_offset; | size_t a_offset=this_offset; |
| size_t b_offset=0; | size_t b_offset=0; |
| Chunk::Row *a_end=append_here; | String_fragment *a_end=felements+fused; |
| Chunk::Row *b_end=src.append_here; | String_fragment *b_end=src.felements+src.fused; |
| uint a_countdown=a_chunk->count; | |
| uint b_countdown=b_chunk->count; | |
| int result; | int result; |
| size_t pos=0; | size_t pos=0; |
| bool a_break=a_size==0; | bool a_break; |
| bool b_break=src.is_empty(); | bool b_break; |
| if(!(a_break || b_break)) while(true) { | while(!( |
| if(pos+a_row->item.size > this_offset) { | (a_break=(a_current==a_end)) |
| if(lang!=UL_UNSPECIFIED && a_row->item.lang>lang) | || (b_break=(b_current==b_end)))) { |
| if(pos+a_current->size > this_offset) { | |
| if(lang!=UL_UNSPECIFIED && a_current->lang>(String_UL)lang) | |
| return -1; // wrong lang -- bail out | return -1; // wrong lang -- bail out |
| int size_diff= | int size_diff= |
| (a_row->item.size-a_offset)- | (a_current->size-a_offset)- |
| (b_row->item.size-b_offset); | (b_current->size-b_offset); |
| if(size_diff==0) { // a has same size as b | if(size_diff==0) { // a has same size as b |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, | result=memcmp(a_current->ptr+a_offset, b_current->ptr+b_offset, |
| a_row->item.size-a_offset); | a_current->size-a_offset); |
| if(result) | if(result) |
| return result; | return result; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; a_offset=0; | a_current++; a_offset=0; |
| b_row++; b_countdown--; b_offset=0; | b_current++; b_offset=0; |
| } else if (size_diff>0) { // a longer | } else if(size_diff>0) { // a longer |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, | result=memcmp(a_current->ptr+a_offset, b_current->ptr+b_offset, |
| b_row->item.size-b_offset); | b_current->size-b_offset); |
| if(result) | if(result) |
| return result; | return result; |
| a_offset+=b_row->item.size-b_offset; | a_offset+=b_current->size-b_offset; |
| b_row++; b_countdown--; b_offset=0; | b_current++; b_offset=0; |
| } else { // b longer | } else { // b longer |
| result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, | result=memcmp(a_current->ptr+a_offset, b_current->ptr+b_offset, |
| a_row->item.size-a_offset); | a_current->size-a_offset); |
| if(result) | if(result) |
| return result; | return result; |
| b_offset+=a_row->item.size-a_offset; | b_offset+=a_current->size-a_offset; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; a_offset=0; | a_current++; 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 { | } else { |
| a_offset-=a_row->item.size; | a_offset-=a_current->size; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; | a_current++; |
| } | |
| 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; | |
| a_countdown=a_chunk->count; | |
| } | } |
| } | } |
| if(a_break==b_break) { // ended simultaneously | if(a_break==b_break) { // ended simultaneously |
| Line 247 int String::cmp(int& partial, const char | Line 161 int String::cmp(int& partial, const char |
| 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, a_size-1); | this_offset=min(this_offset, a_size-1); |
| const Chunk *a_chunk=&head.chunk; | const String_fragment *a_current=felements; |
| const Chunk::Row *a_row=a_chunk->rows; | |
| size_t a_offset=this_offset; | size_t a_offset=this_offset; |
| size_t b_offset=0; | size_t b_offset=0; |
| Chunk::Row *a_end=append_here; | String_fragment *a_end=felements+fused; |
| uint a_countdown=a_chunk->count; | |
| size_t pos=0; | size_t pos=0; |
| bool a_break=a_size==0; | bool a_break; |
| bool b_break=b_size==0; | bool b_break=b_size==0; |
| if(!(a_break || b_break)) while(true) { | while(!( |
| if(pos+a_row->item.size > this_offset) { | (a_break=(a_current==a_end)) |
| if(lang!=UL_UNSPECIFIED && a_row->item.lang>lang) | || b_break)) { |
| if(pos+a_current->size > this_offset) { | |
| if(lang!=UL_UNSPECIFIED && a_current->lang>(String_UL)lang) | |
| return -1; // wrong lang -- bail out | return -1; // wrong lang -- bail out |
| int size_diff= | int size_diff= |
| (a_row->item.size-a_offset)- | (a_current->size-a_offset)- |
| (b_size-b_offset); | (b_size-b_offset); |
| if(size_diff==0) { // a has same size as b | if(size_diff==0) { // a has same size as b |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_current->ptr+a_offset, b_ptr+b_offset, |
| a_row->item.size-a_offset)!=0) | a_current->size-a_offset)!=0) |
| return result; | return result; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; a_offset=0; | a_current++; a_offset=0; |
| b_break=true; | b_break=true; |
| } else if (size_diff>0) { // a longer | } else if (size_diff>0) { // a longer |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_current->ptr+a_offset, b_ptr+b_offset, |
| b_size-b_offset)!=0) | b_size-b_offset)!=0) |
| return result; | return result; |
| a_offset+=b_size-b_offset; | a_offset+=b_size-b_offset; |
| b_break=true; | b_break=true; |
| } else { // b longer | } else { // b longer |
| if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, | if(int result=memcmp(a_current->ptr+a_offset, b_ptr+b_offset, |
| a_row->item.size-a_offset)!=0) | a_current->size-a_offset)!=0) |
| return result; | return result; |
| b_offset+=a_row->item.size-a_offset; | b_offset+=a_current->size-a_offset; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; a_offset=0; | a_current++; a_offset=0; |
| } | } |
| } else { | } else { |
| a_offset-=a_row->item.size; | a_offset-=a_current->size; |
| pos+=a_row->item.size; | pos+=a_current->size; |
| a_row++; a_countdown--; | a_current++; |
| } | |
| 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; | |
| a_countdown=a_chunk->count; | |
| } | } |
| } | } |
| if(a_break==b_break) { // ended simultaneously | if(a_break==b_break) { // ended simultaneously |
| Line 312 int String::cmp(int& partial, const char | Line 217 int String::cmp(int& partial, const char |
| } | } |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| const Origin& String::origin() const { | const String_fragment::Origin& String::origin() const { |
| if(is_empty()) { | if(is_empty()) { |
| static const Origin empty_origin={"empty string"}; | static const String_fragment::Origin empty_origin={"empty string"}; |
| return empty_origin; | return empty_origin; |
| } | } |
| Line 324 const Origin& String::origin() const { | Line 229 const Origin& String::origin() const { |
| // when last peice is constant, | // when last peice is constant, |
| // ex: parser_root_auto_path{dynamic} / auto.p{const} | // ex: parser_root_auto_path{dynamic} / auto.p{const} |
| // using first piece | // using first piece |
| Origin& first_origin=head.chunk.rows[0].item.origin; | String_fragment::Origin& first_origin=felements[0].origin; |
| return first_origin.file ? first_origin : append_here[-1].item.origin; | return first_origin.file ? first_origin : felements[fused-1].origin; |
| } | } |
| #endif | #endif |
| String& String::mid(size_t start, size_t finish) const { | StringPtr String::mid(size_t start, size_t finish) const { |
| String& result=*NEW String(pool()); | StringPtr result(new String()); |
| start=max(size_t(0), start); | start=min(start, size()); |
| finish=min(size(), finish); | finish=max(start, finish); |
| if(start==finish) | if(start==finish) |
| return result; | return result; |
| size_t pos=0; | size_t pos=0; |
| STRING_FOREACH_ROW( | STRING_FOREACH_FRAGMENT( |
| size_t item_finish=pos+row->item.size; | size_t item_finish=pos+fragment->size; |
| if(item_finish > start) { // started now or already? | if(item_finish > start) { // started now or already? |
| bool started=result.is_empty(); // started now? | bool started=result->is_empty(); // 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:fragment->size; |
| result.APPEND( | result->APPEND( |
| row->item.ptr+offset, size-offset, | fragment->ptr+offset, size-offset, |
| row->item.lang, | fragment->lang, |
| row->item.origin.file, row->item.origin.line); | fragment->origin.file, fragment->origin.line); |
| if(finished) | if(finished) |
| goto break2; | goto break2; |
| } | } |
| pos+=row->item.size; | pos+=fragment->size; |
| ); | ); |
| break2: | break2: |
| // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", | // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'", |
| Line 388 int String::pos(const char *substr, size | Line 293 int String::pos(const char *substr, size |
| return -1; | return -1; |
| } | } |
| void String::split(Array& result, | void String::split(Array<ConstStringPtr>& result, |
| size_t* pos_after_ref, | size_t* pos_after_ref, |
| const char *delim, size_t delim_size, | const char *delim, size_t delim_size, |
| Untaint_lang lang, int limit) const { | Untaint_lang lang, int limit) const { |
| Line 398 void String::split(Array& result, | Line 303 void String::split(Array& result, |
| int pos_before; | int pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) { | for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) { |
| result+=&mid(pos_after, pos_before); | result+=mid(pos_after, pos_before); |
| pos_after=pos_before+delim_size; | pos_after=pos_before+delim_size; |
| } | } |
| // last piece | // last piece |
| if(pos_after<self_size && limit) { | if(pos_after<self_size && limit) { |
| result+=&mid(pos_after, self_size); | result+=mid(pos_after, self_size); |
| pos_after=self_size; | pos_after=self_size; |
| } | } |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref=pos_after; | *pos_after_ref=pos_after; |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=ConstStringPtr(this); |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref+=self_size; | *pos_after_ref+=self_size; |
| } | } |
| } | } |
| void String::split(Array& result, | void String::split(Array<ConstStringPtr>& result, |
| size_t* pos_after_ref, | size_t* pos_after_ref, |
| const String& delim, Untaint_lang lang, | const String& delim, Untaint_lang lang, |
| int limit) const { | int limit) const { |
| Line 424 void String::split(Array& result, | Line 329 void String::split(Array& result, |
| int pos_before; | int pos_before; |
| // while we have 'delim'... | // while we have 'delim'... |
| for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) { | for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) { |
| result+=&mid(pos_after, pos_before); | result+=mid(pos_after, pos_before); |
| pos_after=pos_before+delim.size(); | pos_after=pos_before+delim.size(); |
| } | } |
| // last piece | // last piece |
| if(pos_after<size() && limit) { | if(pos_after<size() && limit) { |
| result+=&mid(pos_after, size()); | result+=mid(pos_after, size()); |
| pos_after=size(); | pos_after=size(); |
| } | } |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref=pos_after; | *pos_after_ref=pos_after; |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=ConstStringPtr(this); |
| if(pos_after_ref) | if(pos_after_ref) |
| *pos_after_ref+=size(); | *pos_after_ref+=size(); |
| } | } |
| } | } |
| static void regex_options(const String *options, int *result, bool& need_pre_post_match){ | static void regex_options(ConstStringPtr options, int *result, bool& need_pre_post_match){ |
| struct Regex_option { | struct Regex_option { |
| const char *keyL; | const char *keyL; |
| const char *keyU; | const char *keyU; |
| Line 457 static void regex_options(const String * | Line 362 static void regex_options(const String * |
| {"'", 0, 0, 0, 0, &need_pre_post_match}, | {"'", 0, 0, 0, 0, &need_pre_post_match}, |
| {0} | {0} |
| }; | }; |
| result[0]=PCRE_EXTRA | PCRE_DOTALL; | result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY; |
| result[1]=0; | result[1]=0; |
| if(options) | if(options) |
| Line 473 static void regex_options(const String * | Line 378 static void regex_options(const String * |
| } | } |
| } | } |
| /// @todo make replacement Table stacked | TablePtr String::match(Charset& source_charset, |
| bool String::match( | ConstStringPtr aorigin, |
| const String *aorigin, | const String& regexp, |
| const String& regexp, | ConstStringPtr options, |
| const String *options, | Row_action row_action, void *info, |
| Table **table, | bool *was_global) const { |
| Row_action row_action, void *info, | |
| bool *was_global) const { | |
| if(regexp.is_empty()) | if(regexp.is_empty()) |
| throw Exception(0, | throw Exception(0, |
| aorigin, | aorigin, |
| "regexp is empty"); | "regexp is empty"); |
| const char *pattern=regexp.cstr(); | CharPtr pattern=regexp.cstr(); |
| const char *errptr; | const char *errptr; |
| int erroffset; | int erroffset; |
| bool need_pre_post_match=false; | bool need_pre_post_match=false; |
| int option_bits[2]; regex_options(options, option_bits, need_pre_post_match); | int option_bits[2]; regex_options(options, option_bits, need_pre_post_match); |
| if(was_global) | if(was_global) |
| *was_global=option_bits[1]!=0; | *was_global=option_bits[1]!=0; |
| pcre *code=pcre_compile(pattern, option_bits[0], | pcre *code=pcre_compile(pattern.get(), option_bits[0], |
| &errptr, &erroffset, | &errptr, &erroffset, |
| pool().get_source_charset().pcre_tables); | source_charset.pcre_tables); |
| if(!code) | if(!code) |
| throw Exception(0, | throw Exception(0, |
| ®exp.mid(erroffset, regexp.size()), | regexp.mid(erroffset, regexp.size()), |
| "regular expression syntax error - %s", errptr); | "regular expression syntax error - %s", errptr); |
| int info_substrings=pcre_info(code, 0, 0); | int info_substrings=pcre_info(code, 0, 0); |
| Line 512 bool String::match( | Line 415 bool String::match( |
| info_substrings); | info_substrings); |
| } | } |
| const char *subject=cstr(); | CharPtr subject=cstr(); |
| int length=strlen(subject); | int length=size(); |
| const int ovecsize=(1/*match*/+MAX_STRING_MATCH_TABLE_COLUMNS)*3; | const int ovecsize=(1/*match*/+MAX_MATCH_GROUPS)*3; |
| int ovector[ovecsize]; | int ovector[ovecsize]; |
| // create table | // create table |
| *table=NEW Table(pool(), 0, string_match_table_columns); | TablePtr result(new Table(string_match_table_template)); |
| int exec_option_bits=0; | int exec_option_bits=0; |
| int prestart=0; | int prestart=0; |
| Line 526 bool String::match( | Line 429 bool String::match( |
| int postfinish=size(); | int postfinish=size(); |
| while(true) { | while(true) { |
| int exec_substrings=pcre_exec(code, 0, | int exec_substrings=pcre_exec(code, 0, |
| subject, length, prestart, | subject.get(), length, prestart, |
| exec_option_bits, ovector, ovecsize); | exec_option_bits, ovector, ovecsize); |
| if(exec_substrings==PCRE_ERROR_NOMATCH) { | if(exec_substrings==PCRE_ERROR_NOMATCH) { |
| pcre_free(code); | pcre_free(code); |
| row_action(**table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); | row_action(*result.get(), 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); |
| return option_bits[1]!=0; // global=true+table, not global=false | if(option_bits[1]) |
| return result; // global=true+result | |
| else | |
| return TablePtr(0);// not global=no result | |
| } | } |
| if(exec_substrings<0) { | if(exec_substrings<0) { |
| Line 545 bool String::match( | Line 451 bool String::match( |
| int prefinish=ovector[0]; | int prefinish=ovector[0]; |
| poststart=ovector[1]; | poststart=ovector[1]; |
| Array& row=*NEW Array(pool()); | object_ptr<Array<ConstStringPtr> > row; |
| row+=need_pre_post_match?&mid(0, prefinish):0; // .prematch column value | if(need_pre_post_match) { |
| row+=need_pre_post_match?&mid(prefinish, poststart):0; // .match | *row+=mid(0, prefinish); // .prematch column value |
| row+=need_pre_post_match?&mid(poststart, postfinish):0; // .postmatch | *row+=mid(prefinish, poststart); // .match |
| *row+=mid(poststart, postfinish); // .postmatch | |
| } else { | |
| *row+=ConstStringPtrZero; // .prematch column value | |
| *row+=ConstStringPtrZero; // .match | |
| *row+=ConstStringPtrZero; // .postmatch | |
| } | |
| for(int i=1; i<exec_substrings; i++) { | for(int i=1; i<exec_substrings; i++) { |
| // -1:-1 case handled peacefully by mid() itself | // -1:-1 case handled peacefully by mid() itself |
| row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value | *row+=mid(ovector[i*2+0], ovector[i*2+1]); // .i column value |
| } | } |
| row_action(**table, &row, prestart, prefinish, poststart, postfinish, info); | row_action(*result, row.get(), prestart, prefinish, poststart, postfinish, info); |
| if(!option_bits[1] || prestart==poststart) { // not global | going to hang | if(!option_bits[1] || prestart==poststart) { // not global | going to hang |
| pcre_free(code); | pcre_free(code); |
| row_action(**table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); | row_action(*result, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); |
| return true; | return result; |
| } | } |
| prestart=poststart; | prestart=poststart; |
| Line 571 bool String::match( | Line 483 bool String::match( |
| } | } |
| } | } |
| String& String::change_case(Pool& pool, | StringPtr String::change_case(Pool& pool, Charset& source_charset, Change_case_kind kind) const { |
| Change_case_kind kind) const { | StringPtr result(new String()); |
| const unsigned char *tables=pool.get_source_charset().pcre_tables; | |
| String& result=*new(pool) String(pool); | const unsigned char *tables=source_charset.pcre_tables; |
| const unsigned char *a; | const unsigned char *a; |
| const unsigned char *b; | const unsigned char *b; |
| Line 589 String& String::change_case(Pool& pool, | Line 501 String& String::change_case(Pool& pool, |
| break; | break; |
| default: | default: |
| throw Exception(0, | throw Exception(0, |
| this, | ConstStringPtr(this), |
| "unknown change case kind #%d", | "unknown change case kind #%d", |
| static_cast<int>(kind)); // never | static_cast<int>(kind)); // never |
| a=b=0; // calm, compiler | a=b=0; // calm, compiler |
| break; // never | break; // never |
| } | } |
| STRING_FOREACH_ROW( | STRING_FOREACH_FRAGMENT( |
| char *new_cstr=(char *)pool.malloc(row->item.size, 12); | char *new_cstr=new(pool) char[fragment->size]; |
| char *dest=new_cstr; | char *dest=new_cstr; |
| const char *src=row->item.ptr; | const char *end=fragment->ptr+fragment->size; |
| for(int size=row->item.size; size--; src++) { | for(const char *current=fragment->ptr; current<end; current++) { |
| unsigned char c=a[(unsigned char)*src]; | unsigned char c=a[(unsigned char)*current]; |
| if(b) | if(b) |
| c=b[c]; | c=b[c]; |
| *dest++=(char)c; | *dest++=(char)c; |
| } | } |
| result.APPEND(new_cstr, row->item.size, | result->APPEND(new_cstr, fragment->size, |
| row->item.lang, | fragment->lang, |
| row->item.origin.file, row->item.origin.line); | fragment->origin.file, fragment->origin.line); |
| ); | ); |
| return result; | return result; |
| } | } |
| void String::join_chain(Pool& pool, | /// @test if in some piece were found no dict words, append it, not it's duplicate |
| const Chunk*& achunk, const Chunk::Row*& arow, uint& acountdown, | StringPtr String::replace(Pool& pool, Dictionary& dict) const { |
| uchar& joined_lang, const char *& joined_ptr, size_t& joined_size) const { | StringPtr result(new String()); |
| joined_lang=arow->item.lang; | CharPtr lcstr(cstr()); |
| const char *current=lcstr.get(); | |
| // calc size | |
| joined_size=0; | |
| { | |
| const Chunk* chunk=achunk; | |
| const Chunk::Row* row=arow; | |
| uint countdown=acountdown; | |
| STRING_PREPARED_FOREACH_ROW(*this, | |
| if(row->item.lang==joined_lang) | |
| joined_size+=row->item.size; | |
| else | |
| break; | |
| ); | |
| } | |
| // if one row, return simply itself | STRING_FOREACH_FRAGMENT( |
| if(joined_size==arow->item.size) { | IFNDEF_NO_STRING_ORIGIN( |
| joined_ptr=arow->item.ptr; | const char *joined_origin_file=fragment->origin.file; |
| } else { | const size_t joined_origin_line=fragment->origin.line; |
| // join adjacent rows | ); |
| char *ptr=(char *)pool.malloc(joined_size,13); | uchar joined_lang=fragment->lang; |
| joined_ptr=ptr; | const char *joined_ptr=current; |
| // calc size | |
| const Chunk* chunk=achunk; | size_t joined_size=0; |
| const Chunk::Row* row=arow; | STRING_PREPARED_FOREACH_FRAGMENT( |
| uint countdown=acountdown; | if(fragment->lang==joined_lang) |
| STRING_PREPARED_FOREACH_ROW(*this, | joined_size+=fragment->size; |
| if(row->item.lang==joined_lang) { | else |
| memcpy(ptr, row->item.ptr, row->item.size); ptr+=row->item.size; | |
| } else | |
| break; // before non-ours | break; // before non-ours |
| ); | ); |
| current+=joined_size; | |
| // set pointers after joined piece | |
| achunk=chunk; arow=row; acountdown=countdown; | |
| // & one step back, see String::reconstruct | |
| --arow; ++acountdown; | |
| } | |
| } | |
| /// @test if in some piece were found no dict words, append it, not it's duplicate | // pointers are after joined piece |
| String& String::replace(Pool& pool, Dictionary& dict) const { | // & one step back, see STRING_PREPARED_FOREACH_FRAGMENT |
| // return reconstruct(pool).replace_in_reconstructed(pool, dict); | --fragment; |
| String& result=*new(pool) String(pool); | |
| STRING_FOREACH_ROW( | |
| uchar joined_lang; | |
| const char *joined_ptr; | |
| size_t joined_size; | |
| IFNDEF_NO_STRING_ORIGIN( | |
| const char *joined_origin_file=row->item.origin.file; | |
| const size_t joined_origin_line=row->item.origin.line; | |
| ); | |
| join_chain(pool, chunk, row, countdown, | |
| joined_lang, joined_ptr, joined_size); | |
| const char *src=joined_ptr; | char *new_cstr=new(pool) char[(size_t)ceil(joined_size*dict.max_ratio())]; |
| size_t src_size=joined_size; | |
| char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14); | |
| char *dest=new_cstr; | char *dest=new_cstr; |
| while(src_size) { | while(joined_size) { |
| // there is a row where first column starts 'src' | // there is a fragment where first column starts 'joined_ptr' |
| if(Table::Item *item=dict.first_that_starts(src, src_size)) { | if(Table::element_type row=dict.first_that_starts(joined_ptr, joined_size)) { |
| // get a=>b values | // get a=>b values |
| const String& a=*static_cast<Array *>(item)->get_string(0); | ConstStringPtr a=row->get(0); |
| const String& b=*static_cast<Array *>(item)->get_string(1); | ConstStringPtr b=row->get(1); |
| // skip 'a' in 'src' && reduce work size | // skip 'a' in 'joined_ptr' && reduce work size |
| src+=a.size(); src_size-=a.size(); | joined_ptr+=a->size(); joined_size-=a->size(); |
| // write 'b' to 'dest' && skip 'b' in 'dest' | // write 'b' to 'dest' && skip 'b' in 'dest' |
| b.store_to(dest); dest+=b.size(); | b->store_to(dest, String::UL_AS_IS); dest+=b->size(); |
| } else { | } else { |
| // write a char to b && reduce work size | // write a char to b && reduce work size |
| *dest++=*src++; src_size--; | *dest++=*joined_ptr++; joined_size--; |
| } | } |
| } | } |
| result.APPEND(new_cstr, dest-new_cstr, joined_lang, | result->APPEND(new_cstr, dest-new_cstr, joined_lang, |
| joined_origin_file, joined_origin_line); | |
| ); | |
| return result; | |
| } | |
| StringPtr String::join_chains(Pool& pool) const { | |
| StringPtr result(new String()); | |
| char *pooled_buf=new(pool) char[cstr_bufsize()]; | |
| store_to(pooled_buf, String::UL_AS_IS); | |
| const char *current=pooled_buf; | |
| STRING_FOREACH_FRAGMENT( | |
| IFNDEF_NO_STRING_ORIGIN( | |
| const char *joined_origin_file=fragment->origin.file; | |
| const size_t joined_origin_line=fragment->origin.line; | |
| ); | |
| uchar joined_lang=fragment->lang; | |
| const char *joined_ptr=current; | |
| // calc size | |
| size_t joined_size=0; | |
| STRING_PREPARED_FOREACH_FRAGMENT( | |
| if(fragment->lang==joined_lang) | |
| joined_size+=fragment->size; | |
| else | |
| break; // before non-ours | |
| ); | |
| current+=joined_size; | |
| // pointers are after joined piece | |
| // & one step back, see STRING_PREPARED_FOREACH_FRAGMENT | |
| --fragment; | |
| result->APPEND(joined_ptr, joined_size, joined_lang, | |
| joined_origin_file, joined_origin_line); | joined_origin_file, joined_origin_line); |
| ); | ); |
| return result; | return result; |
| } | } |
| double String::as_double() const { | double String::as_double() const { |
| double result; | double result; |
| const char *cstr; | char buf[MAX_STRING]; |
| char buf[MAX_NUMBER]; | if(size()>MAX_STRING-1) |
| if(head.chunk.rows+1==append_here) { | throw Exception("number.format", |
| int size=min(head.chunk.rows[0].item.size, MAX_NUMBER-1); | ConstStringPtr(this), |
| memcpy(buf, head.chunk.rows[0].item.ptr, size); | "invalid number too long a string (%u>%u)", size(), MAX_STRING-1); |
| buf[size]=0; | char *eol=store_to(buf, String::UL_AS_IS); *eol=0; |
| cstr=buf; | const char *cstr=buf; |
| } else | |
| cstr=this->cstr(); | while(*cstr && isspace(*cstr)) |
| cstr++; | |
| if(!*cstr) | |
| return 0; | |
| char *error_pos; | char *error_pos; |
| // 0xABC | // 0xABC |
| if(cstr[0]=='0') | if(cstr[0]=='0') |
| Line 723 double String::as_double() const { | Line 640 double String::as_double() const { |
| else | else |
| result=(double)strtod(cstr, &error_pos); | result=(double)strtod(cstr, &error_pos); |
| if(*error_pos/*not EOS*/) | while(char c=*error_pos++) |
| throw Exception("number.format", | if(!isspace(c)) |
| this, | throw Exception("number.format", |
| "invalid number (double)"); | ConstStringPtr(this), |
| "invalid number (double)"); | |
| return result; | return result; |
| } | } |
| int String::as_int() const { | int String::as_int() const { |
| int result; | int result; |
| const char *cstr; | char buf[MAX_STRING]; |
| char buf[MAX_NUMBER]; | if(size()>MAX_STRING-1) |
| if(head.chunk.rows+1==append_here) { | throw Exception("number.format", |
| int size=min(head.chunk.rows[0].item.size, MAX_NUMBER-1); | ConstStringPtr(this), |
| memcpy(buf, head.chunk.rows[0].item.ptr, size); | "invalid number too long a string (%u>%u)", size(), MAX_STRING-1); |
| buf[size]=0; | char *eol=store_to(buf, String::UL_AS_IS); *eol=0; |
| cstr=buf; | const char *cstr=buf; |
| } else | |
| cstr=this->cstr(); | while(*cstr && isspace(*cstr)) |
| cstr++; | |
| if(!*cstr) | |
| return 0; | |
| char *error_pos; | char *error_pos; |
| // 0xABC | // 0xABC |
| if(cstr[0]=='0') | if(cstr[0]=='0') |
| Line 751 int String::as_int() const { | Line 673 int String::as_int() const { |
| else | else |
| result=(int)strtol(cstr, &error_pos, 0); | result=(int)strtol(cstr, &error_pos, 0); |
| if(*error_pos/*not EOS*/) | while(char c=*error_pos++) |
| throw Exception("number.format", | if(!isspace(c)) |
| this, | throw Exception("number.format", |
| "invalid number (int)"); | ConstStringPtr(this), |
| "invalid number (int)"); | |
| return result; | return result; |
| } | } |
| inline void ushort2uchars(ushort word, uchar& byte1, uchar& byte2) { | inline void uint2uchars(uint word, uchar *bytes) { |
| byte1=word&0xFF; | bytes[0]=word&0xFF; |
| byte2=word>>8; | bytes[1]=(word>>8)&0xFF; |
| } | bytes[2]=(word>>16)&0xFF; |
| inline ushort uchars2ushort(uchar byte1, uchar byte2) { | bytes[3]=(word>>24)&0xFF; |
| return (byte2<<8) | byte1; | } |
| } | inline uint uchars2uint(uchar *bytes) { |
| /* @todo maybe network order worth spending some effort? | return bytes[3]<<24 |
| don't bothering myself with network byte order, | | bytes[2]<<16 |
| am not planning to be able to move resulting file across platforms | | bytes[1]<<8 |
| for now | | bytes[0]; |
| */ | } |
| void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const { | |
| void String::serialize(Pool& pool, size_t prolog_size, char *& buf, size_t& buf_size) const { | |
| buf_size= | buf_size= |
| prolog_size | prolog_size |
| +used_rows()*(sizeof(uchar)+sizeof(ushort)) | +fused*(sizeof(uchar)+sizeof(size_t)) |
| +size(); | +size(); |
| buf=malloc(buf_size,15); | buf=new(pool) char[buf_size]; |
| char *cur=(char *)buf+prolog_size; | char *cur=buf+prolog_size; |
| STRING_FOREACH_ROW( | STRING_FOREACH_FRAGMENT( |
| // lang | // lang |
| memcpy(cur, &row->item.lang, sizeof(uchar)); | memcpy(cur, &fragment->lang, sizeof(fragment->lang)); |
| cur+=sizeof(uchar); | cur+=sizeof(fragment->lang); |
| // size | // size |
| uchar byte1; uchar byte2; | // bug on some sparc platform [you can't work with integers on odd pointers] |
| ushort2uchars(row->item.size, byte1, byte2); | // forces us to use byte array instead |
| memcpy(cur, &byte1, sizeof(uchar)); cur+=sizeof(uchar); | uchar bytes[4]; |
| memcpy(cur, &byte2, sizeof(uchar)); cur+=sizeof(uchar); | uint2uchars(fragment->size, bytes); |
| memcpy(cur, &bytes, sizeof(bytes)); cur+=sizeof(bytes); | |
| // bytes | // bytes |
| memcpy(cur, row->item.ptr, row->item.size); | memcpy(cur, fragment->ptr, fragment->size); |
| cur+=row->item.size; | cur+=fragment->size; |
| ); | ); |
| } | } |
| bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) { | bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) { |
| Line 801 bool String::deserialize(size_t prolog_s | Line 726 bool String::deserialize(size_t prolog_s |
| buf_size-=prolog_size; | buf_size-=prolog_size; |
| while(buf_size) { | while(buf_size) { |
| if(sizeof(uchar)+sizeof(ushort)>buf_size) // lang+size | if(sizeof(uchar)+sizeof(size_t)>buf_size) // lang+size |
| return false; | return false; |
| uchar lang=*(uchar *)(cur); | String_UL lang=*(String_UL *)(cur); |
| ushort size=uchars2ushort( | size_t size=uchars2uint((uchar *)cur); |
| *(uchar*)(cur+sizeof(uchar)*1), | |
| *(uchar*)(cur+sizeof(uchar)*2) | |
| ); | |
| size_t piece_size=sizeof(uchar)+sizeof(ushort)+size; | size_t piece_size=sizeof(uchar)+sizeof(size_t)+size; |
| if(piece_size>buf_size) // buffer overrun, can be on incomplete cache files | if(piece_size>buf_size) // buffer overrun, can be on incomplete cache files |
| return false; | return false; |
| const char *ptr=(const char*)(cur+sizeof(uchar)*3); | const char *ptr=(const char*)(cur+sizeof(uchar)+sizeof(size_t)); |
| APPEND(ptr, size, lang, file, 0); | APPEND(ptr, size, lang, file, 0); |
| cur+=piece_size; | cur+=piece_size; |