|
|
| version 1.75, 2001/04/09 09:48:20 | version 1.175.2.7, 2003/09/25 09:06:35 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class. @see untasize_t.C. | Parser: string class. @see untalength_t.C. |
| Copyright (c) 2001 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: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | |
| $Id$ | |
| */ | */ |
| #include "pa_config_includes.h" | 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_threads.h" | #include "pa_dictionary.h" |
| #include "pa_charset.h" | |
| //#include "pa_sapi.h" | // cord lib extension |
| // String | #ifndef DOXYGEN |
| typedef struct { | |
| String::String(Pool& apool, const char *src, size_t src_size, bool tainted) : | ssize_t countdown; |
| Pooled(apool) { | char target; /* Character we're looking for */ |
| last_chunk=&head; | } chr_data; |
| head.count=CR_PREALLOCATED_COUNT; | #endif |
| append_here=head.rows; | static int CORD_range_contains_chr_greater_then_proc(char c, size_t size, void* client_data) |
| head.preallocated_link=0; | { |
| link_row=&head.rows[head.count]; | register chr_data * d = (chr_data *)client_data; |
| fused_rows=fsize=0; | |
| if (d -> countdown<=0) return(2); | |
| if(src) | d -> countdown -= size; |
| if(tainted) | if (c > d -> target) return(1); |
| APPEND_TAINTED(src, src_size, 0, 0); | return(0); |
| else | |
| APPEND_CLEAN(src, src_size, 0, 0); | |
| } | } |
| int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c) | |
| { | |
| chr_data d; | |
| void String::expand() { | d.countdown = n; |
| size_t new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100; | d.target = c; |
| last_chunk=static_cast<Chunk *>( | return(CORD_block_iter(x, i, CORD_range_contains_chr_greater_then_proc, &d) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/); |
| 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()) { | // helpers |
| head.count=CR_PREALLOCATED_COUNT; | |
| /// String::match uses this as replace & global search table columns | |
| size_t src_used_rows=src.fused_rows; | |
| if(src_used_rows<=head.count) { | const int MAX_MATCH_GROUPS=100; |
| // all new rows fit size_to preallocated area | |
| size_t curr_chunk_rows=head.count; | class String_match_table_template_columns: public ArrayString { |
| memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows); | public: |
| append_here=&head.rows[src_used_rows]; | String_match_table_template_columns() { |
| link_row=&head.rows[curr_chunk_rows]; | *this+=new String("prematch"); |
| } else { | *this+=new String("match"); |
| // warning: | *this+=new String("postmatch"); |
| // heavily relies on the fact | for(int i=0; i<MAX_MATCH_GROUPS; i++) { |
| // "preallocated area is the same for all strings" | *this+=new String(String::Body::Format(1+i), String::L_CLEAN); |
| // | |
| // info: | |
| // allocating only enough mem to fit src string rows | |
| // next append would allocate a new chunk | |
| // | |
| // new rows don't fit size_to preallocated area: splitting size_to two chunks | |
| // preallocated chunk src to constructing head | |
| memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count); | |
| // remaining rows size_to new_chunk | |
| size_t curr_chunk_rows=src_used_rows-head.count; | |
| Chunk *new_chunk=static_cast<Chunk *>( | |
| malloc(sizeof(size_t)+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]; | |
| Chunk *old_chunk=src.head.preallocated_link; | |
| Chunk::Row *new_rows=new_chunk->rows; | |
| size_t rows_left_to_copy=new_chunk->count; | |
| while(true) { | |
| size_t 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; | }; |
| fused_rows=src_used_rows; | |
| fsize=src.fsize; | |
| } | |
| String& String::append(const String& src, Untaint_lang lang, bool forced) { | |
| const Chunk *chunk=&src.head; | |
| do { | |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; i++, row++) { | |
| if(row==src.append_here) | |
| goto break2; | |
| APPEND(row->item.ptr, row->item.size, | |
| (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | |
| return *this; | |
| } | |
| String& String::real_append(STRING_APPEND_PARAMS) { | |
| if(!src) | |
| return *this; | |
| if(!size) | |
| size=strlen(src); | |
| if(!size) | |
| return *this; | |
| if(chunk_is_full()) | Table string_match_table_template(new String_match_table_template_columns); |
| expand(); | |
| append_here->item.ptr=src; | // String::Body methods |
| fsize+=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++; fused_rows++; | |
| return *this; | String::Body String::Body::Format(int value) { |
| char local[MAX_NUMBER]; | |
| size_t length=snprintf(local, MAX_NUMBER, "%d", value); | |
| return String::Body(pa_strdup(local, length), length); | |
| } | } |
| uint String::hash_code() const { | static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) { |
| uint& result=*static_cast<uint*>(client_data); | |
| generic_hash_code(result, c); | |
| return 0; | |
| } | |
| static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) { | |
| uint& result=*static_cast<uint*>(client_data); | |
| generic_hash_code(result, s); | |
| return 0; | |
| }; | |
| uint String::Body::hash_code() const { | |
| uint result=0; | uint result=0; |
| CORD_iter5(body, 0, | |
| const Chunk *chunk=&head; | CORD_batched_iter_fn_generic_hash_code, |
| do { | CORD_batched_iter_fn_generic_hash_code, &result); |
| const Chunk::Row *row=chunk->rows; | |
| for(size_t i=0; i<chunk->count; i++) { | |
| if(row==append_here) | |
| goto break2; | |
| result=Hash::generic_code(result, row->item.ptr, row->item.size); | |
| row++; | |
| } | |
| chunk=row->link; | |
| } while(chunk); | |
| break2: | |
| return result; | return result; |
| } | } |
| /// @todo move 'lang' skipping to pos | // String methods |
| int String::cmp(int& partial, const String& src, | |
| size_t this_offset, Untaint_lang lang) const { | |
| partial=-1; | |
| this_offset=min(this_offset, size()-1); | |
| const Chunk *a_chunk=&head; | |
| const Chunk *b_chunk=&src.head; | |
| const Chunk::Row *a_row=a_chunk->rows; | |
| const Chunk::Row *b_row=b_chunk->rows; | |
| size_t a_offset=this_offset; | |
| size_t b_offset=0; | |
| Chunk::Row *a_end=append_here; | |
| 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; | |
| if(pos+a_row->item.size > this_offset) { | |
| if(lang!=UL_UNSPECIFIED && 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) { | String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) { |
| a_chunk=a_row->link; | append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN); |
| a_row=a_chunk->rows; | } |
| a_countdown=a_chunk->count; | String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) { |
| } | append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN); |
| } | |
| 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; | |
| } | |
| } | } |
| /// @todo move 'lang' skipping to pos | String& String::append_know_length(const char* str, size_t known_length, Language lang) { |
| int String::cmp(int& partial, const char* b_ptr, size_t src_size, | if(!known_length) |
| size_t this_offset, Untaint_lang lang) const { | return *this; |
| 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::Row *a_row=a_chunk->rows; | |
| size_t a_offset=this_offset; | |
| 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; | |
| if(pos+a_row->item.size > this_offset) { | |
| if(lang!=UL_UNSPECIFIED && 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) { | // first: langs |
| a_chunk=a_row->link; | langs.append(body, lang, known_length); |
| a_row=a_chunk->rows; | // next: letters themselves |
| a_countdown=a_chunk->count; | body.append_know_length(str, known_length); |
| } | |
| } | ASSERT_STRING_INVARIANT(*this); |
| if(a_break==b_break) { // ended simultaneously | return *this; |
| 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; | |
| } | |
| } | } |
| String& String::append_help_length(const char* str, size_t helper_length, Language lang) { | |
| if(!str) | |
| return *this; | |
| size_t known_length=helper_length?helper_length:strlen(str); | |
| if(!known_length) | |
| return *this; | |
| #ifndef NO_STRING_ORIGIN | return append_know_length(str, known_length, lang); |
| 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::append_strdup(const char* str, size_t helper_length, Language lang) { |
| size_t known_length=helper_length?helper_length:strlen(str); | |
| if(!known_length) | |
| return *this; | |
| String& String::mid(size_t start, size_t finish) const { | // first: langs |
| start=max(0, start); | langs.append(body, lang, known_length); |
| finish=min(size(), finish); | // next: letters themselves |
| if(start==finish) | body.append_strdup_know_length(str, known_length); |
| return *empty_string; | |
| ASSERT_STRING_INVARIANT(*this); | |
| String& result=*NEW String(pool()); | return *this; |
| } | |
| size_t pos=0; | |
| const Chunk *chunk=&head; | /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab" |
| do { | String& String::mid(size_t substr_begin, size_t substr_end) const { |
| const Chunk::Row *row=chunk->rows; | String& result=*new String; |
| for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) { | |
| if(row==append_here) | size_t self_length=length(); |
| goto break2; | substr_begin=min(substr_begin, self_length); |
| substr_end=min(max(substr_end, substr_begin), self_length); | |
| size_t item_finish=pos+row->item.size; | size_t substr_length=substr_end-substr_begin; |
| if(item_finish > start) { // started now or already? | if(!substr_length) |
| bool started=result.size()==0; // started now? | return result; |
| bool finished=finish <= item_finish; // finished now? | |
| size_t offset=started?start-pos:0; | // first: their langs |
| size_t size=finished?finish-pos:row->item.size; | result.langs.append(result.body, langs, substr_begin, substr_length); |
| result.APPEND( | // next: letters themselves |
| row->item.ptr+offset, size-offset, | result.body=body.mid(substr_begin, substr_length); |
| row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | // SAPI::log("piece of '%s' from %d to %d is '%s'", |
| if(finished) | //cstr(), substr_begin, substr_end, result.cstr()); |
| goto break2; | ASSERT_STRING_INVARIANT(result); |
| } | |
| } | |
| 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; | return result; |
| } | } |
| int String::pos(const String& substr, | size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const { |
| size_t result, Untaint_lang lang) const { | size_t substr_begin=body.pos(substr, this_offset); |
| for(; result<size(); result++) { | if(substr_begin==CORD_NOT_FOUND || !langs.check_lang(lang, substr_begin, substr.length())) |
| int partial; cmp(partial, substr, result, lang); | return STRING_NOT_FOUND; |
| if( | |
| partial==0 || // full match | return substr_begin; |
| partial==2) // 'substr' starts 'this'+'result' | |
| return result; | |
| } | |
| return -1; | |
| } | } |
| int String::pos(const char *substr, size_t substr_size, | size_t String::pos(const String& substr, |
| size_t result, Untaint_lang lang) const { | size_t this_offset, Language lang) const { |
| for(; result<size(); result++) { | return pos(substr.body, this_offset, lang); |
| 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, | void String::split(ArrayString& result, |
| size_t* pos_after_ref, | size_t& pos_after, |
| const char *delim, size_t delim_size, | const char* delim, |
| Untaint_lang lang, int limit) const { | Language lang, int limit) const { |
| if(delim_size) { | size_t self_length=length(); |
| size_t pos_after=pos_after_ref?*pos_after_ref:0; | if(size_t delim_length=strlen(delim)) { |
| 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, pos_after, lang))!=STRING_NOT_FOUND && 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_length; |
| } | } |
| // last piece | // last piece |
| if(pos_after<size() && limit) { | if(pos_after<self_length && limit) { |
| result+=&mid(pos_after, size()); | result+=&mid(pos_after, self_length); |
| pos_after=size(); | pos_after=self_length; |
| } | } |
| if(pos_after_ref) | |
| *pos_after_ref=pos_after; | |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=this; |
| if(pos_after_ref) | pos_after+=self_length; |
| *pos_after_ref+=size(); | |
| } | } |
| } | } |
| void String::split(Array& result, | void String::split(ArrayString& result, |
| size_t* pos_after_ref, | size_t& pos_after, |
| const String& delim, Untaint_lang lang, | const String& delim, Language lang, |
| int limit) const { | int limit) const { |
| if(delim.size()) { | if(!delim.is_empty()) { |
| size_t pos_after=pos_after_ref?*pos_after_ref:0; | |
| 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))!=STRING_NOT_FOUND && 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.length(); |
| } | } |
| // last piece | // last piece |
| if(pos_after<size() && limit) { | if(pos_after<length() && limit) { |
| result+=&mid(pos_after, size()); | result+=&mid(pos_after, length()); |
| pos_after=size(); | pos_after=length(); |
| } | } |
| if(pos_after_ref) | |
| *pos_after_ref=pos_after; | |
| } else { // empty delim | } else { // empty delim |
| result+=this; | result+=this; |
| if(pos_after_ref) | pos_after+=length(); |
| *pos_after_ref+=size(); | |
| } | } |
| } | } |
| /// @test really @b test: s x m [tested: i & g ] | static void regex_options(const String* options, int *result, bool& need_pre_post_match){ |
| static void regex_options(char *options, int *result){ | |
| struct Regex_option { | struct Regex_option { |
| char key; | const char* keyL; |
| const char* keyU; | |
| int clear, set; | int clear, set; |
| int *result; | int *result; |
| bool *flag; | |
| } regex_option[]={ | } regex_option[]={ |
| {'i', 0, PCRE_CASELESS, result}, // a=A | {"i", "I", 0, PCRE_CASELESS, result}, // a=A |
| {'s', 0, PCRE_DOTALL, result}, // \n\n$ | {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default] |
| {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored | {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored |
| {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$ | {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$ |
| {'g', 0, true, result+1}, // many rows | {"g", "G", 0, true, result+1}, // many rows |
| {0}, | {"'", 0, 0, 0, 0, &need_pre_post_match}, |
| {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 && !options->is_empty()) |
| for(Regex_option *o=regex_option; o->key; o++) | for(Regex_option *o=regex_option; o->keyL; o++) |
| if( | if(options->pos(o->keyL)!=STRING_NOT_FOUND |
| strchr(options, o->key) || | || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) { |
| strchr(options, toupper(o->key))) { | if(o->flag) |
| *(o->result)&=~o->clear; | *o->flag=true; |
| *(o->result)|=o->set; | else { // result |
| *o->result &= ~o->clear; | |
| *o->result |= o->set; | |
| } | |
| } | } |
| } | } |
| /** | Table* String::match(Charset& source_charset, |
| returns true if fills table. | const String& regexp, |
| table format is defined and fixed[can be used by others]: | const String* options, |
| @verbatim | Row_action row_action, void *info, |
| pre-match/match/post-match/1/2/3/... | bool& just_matched) const { |
| @endverbatim | if(regexp.is_empty()) |
| */ | throw Exception(0, |
| bool String::match(const String *aorigin, | 0, |
| const String& regexp, | |
| const String *options, | |
| Table **table, | |
| Row_action row_action, void *info) const { | |
| if(!regexp.size()) | |
| THROW(0, 0, | |
| aorigin, | |
| "regexp is empty"); | "regexp is empty"); |
| const char *pattern=regexp.cstr(UL_AS_IS); | |
| const char *errptr; | const char* pattern=regexp.cstr(); |
| const char* errptr; | |
| int erroffset; | int erroffset; |
| int option_bits[2]; regex_options(options?options->cstr():0, option_bits); | bool need_pre_post_match=false; |
| int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match); | |
| bool global=option_bits[1]!=0; | |
| pcre *code=pcre_compile(pattern, option_bits[0], | pcre *code=pcre_compile(pattern, option_bits[0], |
| &errptr, &erroffset, | &errptr, &erroffset, |
| pcre_tables); | source_charset.pcre_tables); |
| if(!code) | if(!code) |
| THROW(0, 0, | throw Exception(0, |
| ®exp.mid(erroffset, regexp.size()), | ®exp.mid(erroffset, regexp.length()), |
| "regular expression syntax error - %s", errptr); | "regular expression syntax error - %s", errptr); |
| int info_substrings=pcre_info(code, 0, 0); | int subpatterns=pcre_info(code, 0, 0); |
| if(info_substrings<0) { | if(subpatterns<0) { |
| (*pcre_free)(code); | pcre_free(code); |
| THROW(0, 0, | throw Exception(0, |
| aorigin, | ®exp, |
| "pcre_info error #%d", | "pcre_info error (%d)", |
| info_substrings); | subpatterns); |
| } | } |
| int startoffset=0; | const char* subject=cstr(); |
| const char *subject=cstr(UL_AS_IS); | size_t subject_length=strlen(subject); |
| int length=strlen(subject); | const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3; |
| int ovecsize; | int ovector[oveclength]; |
| int *ovector=(int *)malloc(sizeof(int)* | |
| (ovecsize=(1/*match*/+info_substrings)*3)); | // create table |
| Table::Action_options table_options; | |
| { // create table | Table& table=*new Table(string_match_table_template, table_options); |
| Array& columns=*NEW Array(pool()); | |
| columns+=string_pre_match_name; | |
| columns+=string_match_name; | |
| columns+=string_post_match_name; | |
| for(int i=1; i<=info_substrings; i++) { | |
| char *column=(char *)malloc(MAX_NUMBER); | |
| snprintf(column, MAX_NUMBER, "%d", i); | |
| columns+=NEW String(pool(), column); // .i column name | |
| } | |
| *table=NEW Table(pool(), aorigin, &columns); | |
| } | |
| int exec_option_bits=0; | int exec_option_bits=0; |
| int prestart=0; | |
| int poststart=0; | |
| int postfinish=length(); | |
| while(true) { | while(true) { |
| int exec_substrings=pcre_exec(code, 0, | int exec_substrings=pcre_exec(code, 0, |
| subject, length, startoffset, | subject, subject_length, prestart, |
| exec_option_bits, ovector, ovecsize); | exec_option_bits, ovector, oveclength); |
| 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, info); | row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info); |
| return option_bits[1]!=0; // global=true+table, not global=false | if(global || subpatterns) |
| return &table; // global or with subpatterns=true+result | |
| else { | |
| just_matched=false; return 0; // not global=no result | |
| } | |
| } | } |
| if(exec_substrings<0) { | if(exec_substrings<0) { |
| (*pcre_free)(code); | pcre_free(code); |
| THROW(0, 0, | throw Exception(0, |
| aorigin, | ®exp, |
| "regular expression execute error #%d", | "regular expression execute error (%d)", |
| exec_substrings); | exec_substrings); |
| } | } |
| Array& row=*NEW Array(pool()); | int prefinish=ovector[0]; |
| row+=&mid(0, ovector[0]); // .pre-match column value | poststart=ovector[1]; |
| row+=&mid(ovector[0], ovector[1]); // .match | ArrayString* row=new ArrayString; |
| row+=&mid(ovector[1], size()); // .post-match | if(need_pre_post_match) { |
| *row+=&mid(0, prefinish); // .prematch column value | |
| *row+=&mid(prefinish, poststart); // .match | |
| *row+=&mid(poststart, postfinish); // .postmatch | |
| } else { | |
| *row+=0; // .prematch column value | |
| *row+=0; // .match | |
| *row+=0; // .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, startoffset, ovector[0], info); | row_action(table, row, prestart, prefinish, poststart, postfinish, info); |
| if(!option_bits[1] || !(startoffset=ovector[1])) { // not global | going to hang | if(!global || prestart==poststart) { // not global | going to hang |
| (*pcre_free)(code); | pcre_free(code); |
| (*row_action)(**table, 0/*last time, no row*/, 0, 0, info); | row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info); |
| return true; | return &table; |
| } | } |
| prestart=poststart; | |
| /* | /* |
| if(option_bits[0] & PCRE_MULTILINE) | if(option_bits[0] & PCRE_MULTILINE) |
| Line 586 bool String::match(const String *aorigin | Line 346 bool String::match(const String *aorigin |
| */ | */ |
| } | } |
| } | } |
| String& String::change_case(Charset& source_charset, Change_case_kind kind) const { | |
| String& result=*new String(); | |
| if(is_empty()) | |
| return result; | |
| const unsigned char *tables=source_charset.pcre_tables; | |
| const unsigned char *a; | |
| const unsigned char *b; | |
| switch(kind) { | |
| case CC_UPPER: | |
| a=tables+lcc_offset; | |
| b=tables+fcc_offset; | |
| break; | |
| case CC_LOWER: | |
| a=tables+lcc_offset; | |
| b=0; | |
| break; | |
| default: | |
| throw Exception(0, | |
| this, | |
| "unknown change case kind #%d", | |
| static_cast<int>(kind)); // never | |
| a=b=0; // calm, compiler | |
| break; // never | |
| } | |
| char* new_cstr=cstrm(); | |
| char *dest=new_cstr; | |
| unsigned char index; | |
| for(const char* current=new_cstr; index=(unsigned char)*current; current++) { | |
| unsigned char c=a[index]; | |
| if(b) | |
| c=b[c]; | |
| *dest++=(char)c; | |
| } | |
| result.langs=langs; | |
| result.body=new_cstr; | |
| return result; | |
| } | |
| const String& String::replace(const Dictionary& dict) const { | |
| String& result=*new String(); | |
| const char* old_cstr=cstr(); | |
| const char* prematch_begin=old_cstr; | |
| const char* current=old_cstr; | |
| while(*current) { | |
| if(Table::element_type row=dict.first_that_begins(current)) { | |
| // prematch | |
| if(size_t prematch_length=current-prematch_begin) { | |
| result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr); | |
| result.body.append_strdup_know_length(prematch_begin, prematch_length); | |
| } | |
| // match | |
| const String* a=row->get(0); | |
| // skip 'a' in 'current'; move prematch_begin | |
| current+=a->length(); prematch_begin=current; | |
| if(row->count()>1) { // are there any b? | |
| const String* b=row->get(1); | |
| result<<*b; | |
| } | |
| } else // simply advance | |
| current++; | |
| } | |
| // postmatch | |
| if(size_t postmatch_length=current-prematch_begin) { | |
| result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr); | |
| result.body.append_strdup_know_length(prematch_begin, postmatch_length); | |
| } | |
| ASSERT_STRING_INVARIANT(result); | |
| return result; | |
| } | |
| double String::as_double() const { | |
| double result; | |
| const char *str=cstr(); | |
| while(*str && isspace(*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| char *error_pos; | |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(double)(unsigned long)strtol(str, &error_pos, 0); | |
| else | |
| result=(double)strtod(str+1/*skip leading 0*/, &error_pos); | |
| else | |
| result=(double)strtod(str, &error_pos); | |
| while(char c=*error_pos++) | |
| if(!isspace(c)) | |
| throw Exception("number.format", | |
| this, | |
| "invalid number (double)"); | |
| return result; | |
| } | |
| int String::as_int() const { | |
| int result; | |
| const char *str=cstr(); | |
| while(*str && isspace(*str)) | |
| str++; | |
| if(!*str) | |
| return 0; | |
| char *error_pos; | |
| // 0xABC | |
| if(str[0]=='0') | |
| if(str[1]=='x' || str[1]=='X') | |
| result=(int)(unsigned long)strtol(str, &error_pos, 0); | |
| else | |
| result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0); | |
| else | |
| result=(int)strtol(str, &error_pos, 0); | |
| while(char c=*error_pos++) | |
| if(!isspace(c)) | |
| throw Exception("number.format", | |
| this, | |
| "invalid number (int)"); | |
| return result; | |
| } | |
| inline void uint2uchars(uint word, uchar *bytes) { | |
| bytes[0]=word&0xFF; | |
| bytes[1]=(word>>8)&0xFF; | |
| bytes[2]=(word>>16)&0xFF; | |
| bytes[3]=(word>>24)&0xFF; | |
| } | |
| inline uint uchars2uint(uchar *bytes) { | |
| return bytes[3]<<24 | |
| | bytes[2]<<16 | |
| | bytes[1]<<8 | |
| | bytes[0]; | |
| } | |
| static int serialize_body_piece(const char* s, char** cur) { | |
| size_t length=strlen(s); | |
| memcpy(*cur, s, length); *cur+=length; | |
| return 0; | |
| }; | |
| String::Cm String::serialize(size_t prolog_length) const { | |
| #if TODO | |
| //_asm int 3; | |
| size_t buf_length= | |
| prolog_length | |
| +sizeof(size_t) | |
| +langs.count()*(sizeof(Language)+sizeof(size_t)) | |
| +length(); | |
| String::Cm result(new(PointerFreeGC) char[buf_length], buf_length); | |
| // 1: prolog | |
| char *cur=result.str+prolog_length; | |
| // 2: langs.count | |
| size_t fragments_count=langs.count(); | |
| memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count); | |
| // 3: lang info | |
| for(Array_iterator<ArrayFragment::element_type> i(langs); i.has_next(); ) { | |
| const Fragment fragment=i.next(); | |
| // lang | |
| memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang); | |
| // length | |
| memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length); | |
| } | |
| // 4: letters | |
| body.for_each(serialize_body_piece, &cur); | |
| return result; | |
| #endif | |
| return String::Cm(0, 0); | |
| } | |
| bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) { | |
| #if TODO | |
| if(buf_length<=prolog_length) | |
| return false; | |
| buf_length-=prolog_length; | |
| // 1: prolog | |
| const char* cur=(const char* )buf+prolog_length; | |
| // 2: langs.count | |
| if(buf_length<sizeof(size_t)) // langs.count don't fit? | |
| return false; | |
| size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t); | |
| buf_length-=sizeof(size_t); | |
| if(fragments_count) { | |
| // 3: lang info | |
| size_t total_length=0; | |
| for(size_t f=0; f<fragments_count; f++) { | |
| size_t piece_length=sizeof(Language)+sizeof(size_t); | |
| if(buf_length<piece_length) // lang+length | |
| return false; | |
| Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language); | |
| size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t); | |
| langs+=Fragment(lang, fragment_length); | |
| total_length+=fragment_length; | |
| buf_length-=piece_length; | |
| } | |
| // 4: letters | |
| if(buf_length!=total_length) | |
| return false; | |
| body=String::Body(cur, buf_length); | |
| } | |
| ASSERT_STRING_INVARIANT(*this); | |
| return true; | |
| #endif | |
| return false; | |
| } | |
| const char* String::Body::v() const { | |
| return CORD_to_const_char_star(body); | |
| } | |
| const char* String::Languages::v() const { | |
| if(is_not_just_lang) | |
| return CORD_to_const_char_star(langs); | |
| else | |
| return (const char*)&langs; | |
| } | |
| const char* String::v() const { | |
| #define LIMIT_VIEW 20 | |
| char* buf=(char*)malloc(MAX_STRING); | |
| const char*body_view=body.v(); | |
| const char*langs_view=langs.v(); | |
| snprintf(buf, MAX_STRING, | |
| "%.*s%s} " | |
| "{%d:%s", | |
| LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"", | |
| strlen(body_view), body_view | |
| ); | |
| return buf; | |
| #undef LIMIT_VIEW | |
| } |