--- parser3/src/include/pa_string.h 2003/02/14 12:59:10 1.144.2.23 +++ parser3/src/include/pa_string.h 2003/03/19 12:22:35 1.144.2.28.2.3 @@ -8,28 +8,31 @@ #ifndef PA_STRING_H #define PA_STRING_H -static const char* IDENT_STRING_H="$Date: 2003/02/14 12:59:10 $"; +static const char* IDENT_STRING_H="$Date: 2003/03/19 12:22:35 $"; // includes -#include "pa_pool.h" #include "pa_types.h" #include "pa_array.h" +#include "cord.h" + // forwards +class Charset; class Table; DECLARE_OBJECT_PTR(Table); class SQL_Connection; class Dictionary; class Request_charsets; -typedef Array ArrayString; DECLARE_OBJECT_PTR(ArrayString); +class String; +typedef Array ArrayString; // helpers #ifndef NO_STRING_ORIGIN # define STRING_APPEND_PARAMS \ const char* src, size_t size, \ - String_UL lang, \ + Untaint_lang lang, \ const char* file, uint line /// appends piece to String @see String::real_append # define APPEND(src, size, lang, file, line) \ @@ -38,7 +41,7 @@ typedef Array ArrayString; DE # define STRING_APPEND_PARAMS \ const char* src, \ size_t size, \ - String_UL lang + Untaint_lang lang /// appends piece to String @see String::real_append # define APPEND(src, size, lang, file, line) \ real_append(src, size, lang) @@ -55,34 +58,16 @@ typedef Array ArrayString; DE /// handy: appends const char* piece to String @see String::real_append #define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) -typedef uint String_UL; - -struct String_fragment { - String_UL/*Untaint_lang*/ lang; ///< untaint flag, later untaint language - const char* ptr; ///< pointer to the start - size_t size; ///< length - -#ifndef NO_STRING_ORIGIN - /// all String pieces hold information of where they come from - struct Origin { - const char* file; ///< macros file name | load file name | sql query text - uint line; ///< file line no | record no - } origin; ///< origin -#endif -}; - /** String which knows the language of all it's fragments. All pieces remember - - the file and its line they are from [can be turned off by NO_STRING_ORIGIN] - whether they are tainted or not, and the language which should be used to detaint them */ -class String: public Array { - size_t fsize; - int count() const; // hiding fro accidental use instead of size() +class String: public PA_Object { + public: /** piece is tainted or not. the language to use when detaint @@ -113,53 +98,69 @@ public: UL_OPTIMIZE_BIT = 0x8000 ///< flag, requiring cstr whitespace optimization }; + struct Fragment { + Untaint_lang lang; ///< untaint flag, later untaint language + size_t size; ///< size + Fragment(Untaint_lang alang, size_t asize): lang(alang), size(asize) {} + }; + + class ArrayFragment: public Array { + void simple_append(element_type src) { + *static_cast *>(this)+=src; + } + public: + ArrayFragment& operator += (element_type src) { + if(count()) { // not empty? + // try to join with last + Fragment& last=get(count()-1); + if(last.lang==src.lang) { + last.size+=src.size; + return *this; + } + } + simple_append(src); + return *this; + } + ArrayFragment& append(const ArrayFragment& src, size_t substr_start, size_t substr_end); + }; + +private: + + CORD body; ///< all characters of string + ArrayFragment fragments; ///< fragment language+size info + public: explicit String(const char* src=0, size_t src_size=0, bool tainted=false); String(const String& src); - bool is_empty() const { return !size(); } - size_t size() const { return fsize; } + operator bool() const { return body!=CORD_EMPTY; } + size_t size() const { return CORD_len(body); } + + struct C { + char *str; + size_t size; + operator char *() { return str; } + C(char *astr, size_t asize): str(astr), size(asize) {} + }; + /// convert to C string. if 'lang' known, forcing 'lang' to it - CharPtr cstr(Untaint_lang lang=UL_AS_IS, + C cstr(Untaint_lang lang=UL_AS_IS, SQL_Connection *connection=0, const Request_charsets *charsets=0) const { - CharPtr result(new char[cstr_bufsize(lang, connection, charsets)]); - char *eol=store_to(result.get(), lang, connection, charsets); - *eol=0; - return result; - } - char *cstr(Pool& pool, Untaint_lang lang=UL_AS_IS, - SQL_Connection *connection=0, - const Request_charsets *charsets=0) const { - char *result=new(pool) char[cstr_bufsize(lang, connection, charsets)]; - char *eol=store_to(result, lang, connection, charsets); + char* cstr=new char[cstr_bufsize(lang, connection, charsets)]; + char* eol=store_to(cstr, lang, connection, charsets); *eol=0; - return result; + return C(cstr, eol-cstr); } - char *cstr_debug_origins(Pool& pool) const; /// puts pieces to buf - void serialize(Pool& pool, size_t prolog_size, char *& buf, size_t& buf_size) const; + void serialize(size_t prolog_size, char *& buf, size_t& buf_size) const; /// appends pieces from buf to self bool deserialize(size_t prolog_size, void *buf, size_t buf_size, const char* file); /** append fragment @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST */ String& real_append(STRING_APPEND_PARAMS); - /// @return <0 ==0 or >0 depending on comparison result - int cmp (int& partial, const String& src, - size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; - bool operator < (const String& src) const { int p; return cmp(p, src)<0; } - bool operator > (const String& src) const { int p; return cmp(p, src)>0; } - bool operator <= (const String& src) const { int p; return cmp(p, src)<=0; } - bool operator >= (const String& src) const { int p; return cmp(p, src)>=0; } - bool operator == (const String& src) const { - if(size()!=src.size()) // can speed up in trivial case - return false; - int p; return cmp(p, src)==0; - } - bool operator != (const String& src) const { int p; return cmp(p, src)!=0; } - /** @param partial returns partial match status. @@ -168,31 +169,30 @@ public: - 1: means @c this starts @c src - 2: means @src starts @this */ - int cmp(int& partial, const char* src_ptr, size_t src_size=0, - size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; - /// this starts with src - bool starts_with(const char* src_ptr, size_t src_size=0) const { - int p; cmp(p, src_ptr, src_size); - return p==0 || p==2; - } - bool operator == (const char* src_ptr) const { - size_t src_size=src_ptr?strlen(src_ptr):0; - if(size() != src_size) - return false; - int partial; // unused - return cmp(partial, src_ptr, src_size)==0; - } - bool operator != (const char* src_ptr) const { - int partial; // unused - return cmp(partial, src_ptr, 0)!=0; + //int cmp(CORD substr, size_t this_offset=0) const; + /// String version of @see cmp(CORD, size_t) + //int cmp(const String& src, size_t this_offset=0) const; + + bool operator < (const String& src) const { return CORD_cmp(body, src.body)<0; } + bool operator > (const String& src) const { return CORD_cmp(body, src.body)>0; } + bool operator <= (const String& src) const { return CORD_cmp(body, src.body)<=0; } + bool operator >= (const String& src) const { return CORD_cmp(body, src.body)>=0; } + bool operator == (const String& src) const { return CORD_cmp(body, src.body)==0; } + bool operator != (const String& src) const { return CORD_cmp(body, src.body)!=0; } + + /// this starts with y + bool starts_with(CORD y) const { + return CORD_ncmp(body, 0/*x_start*/, y, 0/*y_start*/, CORD_len(y))==0; } + bool operator == (CORD y) const { return CORD_cmp(body, y)==0; } + bool operator != (CORD y) const { return CORD_cmp(body, y)!=0; } - String& append_to(String& dest, uchar lang, bool forced) const; - String& append(const String& src, uchar lang, bool forced=false) { + String& append_to(String& dest, Untaint_lang lang, bool forced) const; + String& append(const String& src, Untaint_lang lang, bool forced=false) { return src.append_to(*this, lang, forced); } String& operator << (const String& src) { return append(src, UL_PASS_APPENDED); } - String& operator << (StringPtr src) { return append(*src, UL_PASS_APPENDED); } + //String& operator << (const String* src) { return append(*src, UL_PASS_APPENDED); } String& operator << (const char* src) { return APPEND_CONST(src); } /// simple hash code of string @@ -200,29 +200,36 @@ public: /// extracts first char of a string, if any char first_char() const { - return is_empty()?0:*felements[0].ptr; + return operator bool()?CORD_fetch(body, 0):0; } /// extracts [start, finish) piece of string - StringPtr mid(size_t start, size_t finish) const; + const String& mid(size_t substr_start, size_t substr_end) const; + + /// this is result of pos functions which mean that substr were not found + static const size_t NOT_FOUND; - /// @return position of substr in string, -1 means "not found" [String version] - int pos(const String& substr, - int this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; - /// @return position of substr in string, -1 means "not found" [const char* version] - int pos(const char* substr, size_t substr_size=0, - int this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; + /** + ignore lang if it's UL_UNSPECIFIED + but when specified: look for substring that lies in ONE fragment in THAT lang + @return position of substr in string, -1 means "not found" [const char* version] + */ + size_t pos(CORD substr, + size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; + /// String version of @see pos(const char*, int, Untaint_lang) + size_t pos(const String& substr, + size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; void split(ArrayString& result, - size_t *pos_after_ref, - const char* delim, size_t delim_size, + size_t& pos_after_ref, + const char* delim, Untaint_lang lang=UL_UNSPECIFIED, int limit=-1); void split(ArrayString& result, - size_t *pos_after_ref, + size_t& pos_after_ref, const String& delim, - Untaint_lang lang=UL_UNSPECIFIED, int limit=-1); + Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; - typedef void (*Row_action)(TablePtr table, ArrayStringPtr row, + typedef void (*Row_action)(Table* table, ArrayString* row, int prestart, int prefinish, int poststart, int postfinish, void *info); @@ -233,29 +240,21 @@ public: prematch/match/postmatch/1/2/3/... @endverbatim */ - TablePtr match(Charset& source_charset, - StringPtr aorigin, + Table* match(Charset& source_charset, const String& regexp, - StringPtr options, + const String& options, Row_action row_action, void *info, - bool *was_global=0) const; + bool& just_matched) const; enum Change_case_kind { CC_UPPER, CC_LOWER }; - StringPtr change_case(Pool& pool, Charset& source_charset, + String& change_case(Charset& source_charset, Change_case_kind kind); - StringPtr replace(Pool& pool, Dictionary& dict) const; + const String& replace(const Dictionary& dict) const; double as_double(); int as_int(); - StringPtr join_chains(Pool& pool) const; - -#ifndef NO_STRING_ORIGIN - /// origin of string. calculated by first row - const String_fragment::Origin& origin() const; -#endif - size_t cstr_bufsize(Untaint_lang lang=UL_UNSPECIFIED, SQL_Connection *connection=0, const Request_charsets *charsets=0) const; @@ -266,7 +265,7 @@ public: /* void join_chain( const Chunk*& achunk, const Chunk::Row*& arow, uint& acountdown, - uchar& joined_lang, const char* & joined_ptr, size_t& joined_size) const; + Untaint_lang& joined_lang, const char* & joined_ptr, size_t& joined_size) const; */ private: //disabled @@ -274,18 +273,16 @@ private: //disabled }; +const size_t String::NOT_FOUND=(size_t)-1; /// simple hash code of string. used by Hash inline uint hash_code(const String& self) { return self.hash_code(); } -/// simple hash code of string. used by Hash -inline uint hash_code(StringPtr self) { - return self->hash_code(); -} // helpers +/* #define STRING_PREPARED_FOREACH_FRAGMENT(code) \ for(; fragment