--- parser3/src/include/pa_string.h 2003/03/20 11:11:55 1.144.2.28.2.8 +++ parser3/src/include/pa_string.h 2003/07/24 11:31:21 1.145 @@ -8,7 +8,7 @@ #ifndef PA_STRING_H #define PA_STRING_H -static const char* IDENT_STRING_H="$Date: 2003/03/20 11:11:55 $"; +static const char* IDENT_STRING_H="$Date: 2003/07/24 11:31:21 $"; // includes @@ -16,6 +16,7 @@ static const char* IDENT_STRING_H="$Date #include "pa_array.h" extern "C" { // cord's author forgot to do that +#define CORD_NO_IO #include "cord.h" }; @@ -27,42 +28,96 @@ class SQL_Connection; class Dictionary; class Request_charsets; class String; -typedef Array ArrayString; - -// helpers - -#ifndef NO_STRING_ORIGIN -# define STRING_APPEND_PARAMS \ - const char* src, size_t size, \ - Untaint_lang lang, \ - const char* file, uint line -/// appends piece to String @see String::real_append -# define APPEND(src, size, lang, file, line) \ - real_append(src, size, lang, file, line) -#else -# define STRING_APPEND_PARAMS \ - const char* src, \ - size_t size, \ - Untaint_lang lang -/// appends piece to String @see String::real_append -# define APPEND(src, size, lang, file, line) \ - real_append(src, size, lang) -#endif -/// appends clean piece to String @see String::real_append -#define APPEND_CLEAN(src, size, file, line) \ - APPEND(src, size, String::UL_CLEAN, file, line) -/// appends piece to String as-is @see String::real_append -#define APPEND_AS_IS(src, size, file, line) \ - APPEND(src, size, String::UL_AS_IS, file, line) -/// appends tainted piece to String @see String::real_append -#define APPEND_TAINTED(src, size, file, line) \ - APPEND(src, size, String::UL_TAINTED, file, line) -/// handy: appends const char* piece to String @see String::real_append -#define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) +typedef Array ArrayString; /// this is result of pos functions which mean that substr were not found #define STRING_NOT_FOUND ((size_t)-1) +class StringBody { + + CORD body; + +public: + + StringBody(): body(CORD_EMPTY) {} + StringBody(CORD abody): body(abody) { + assert(!body // no body + || *body // ordinary string + || body[1]==1 // CONCAT_HDR + || body[1]==4 // FN_HDR + || body[1]==6 // SUBSTR_HDR + ); + } + /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated + StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) { + append_know_length(str, helper_length?helper_length:strlen(str)); + } + static StringBody Format(int value); + + void clear() { body=CORD_EMPTY; } + + bool operator! () const { return is_empty(); } + + uint hash_code() const; + + const char* cstr() const { return CORD_to_const_char_star(body); } + char* cstrm() const { return CORD_to_char_star(body); } + + size_t length() const { return CORD_len(body); } + + bool is_empty() const { return body==CORD_EMPTY; } + + void append_know_length(const char *str, size_t known_length) { + if(known_length) + body=CORD_cat_char_star(body, str, known_length); + } + void append_strdup_know_length(const char* str, size_t known_length) { + if(known_length) + append_know_length(pa_strdup(str, known_length), known_length); + } + void append(char c) { body=CORD_cat_char(body, c); } + StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; } + StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; } + + // could not figure out why this operator is needed [should do this chain: string->simple->==] + bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; } + bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; } + bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; } + bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; } + bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; } + bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; } + + int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const { + return CORD_ncmp(body, x_begin, y.body, y_begin, size); + } + + char fetch(size_t index) const { return CORD_fetch(body, index); } + StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); } + size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); } + size_t pos(const StringBody substr, size_t offset=0) const { + if(!substr.length()) + return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1] + return CORD_str(body, offset, substr.body); + } + size_t pos(char c, + size_t offset=0) const { + return CORD_chr(body, offset, c); + } + + template void for_each(int (*callback)(const char* s, I), I info) const { + CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info); + } + + void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); } + + StringBody normalize() const { + return StringBody(CORD_balance(body)); + } + + void dump() const { + CORD_dump(body); + } +}; /** String which knows the language of all it's fragments. @@ -73,176 +128,210 @@ typedef Array ArrayStrin */ class String: public PA_Object { +// friend class StringBody; + public: /** piece is tainted or not. the language to use when detaint remember to change String_Untaint_lang_name @ untaint.C along */ - enum Untaint_lang { - UL_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum - // these two must go before others, there are checks for >UL_AS_IS - UL_CLEAN, ///< clean - UL_AS_IS, ///< leave all characters intact + enum Language { + L_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum + // these two must go before others, there are checks for >L_AS_IS + L_CLEAN, ///< clean + L_AS_IS, ///< leave all characters intact - UL_PASS_APPENDED, + L_PASS_APPENDED, /**< leave language built into string being appended. just a flag, that value not stored */ - UL_TAINTED, ///< tainted, untaint language as assigned later + L_TAINTED, ///< tainted, untaint language as assigned later // untaint languages. assigned by ^untaint[lang]{...} - UL_FILE_SPEC, ///< file specification - UL_HTTP_HEADER, ///< text in HTTP response header - UL_MAIL_HEADER, ///< text in mail header - UL_URI, ///< text in uri - UL_TABLE, ///< ^table:set body - UL_SQL, ///< ^table:sql body - UL_JS, ///< JavaScript code - UL_XML, ///< ^dom:set xml - UL_HTML, ///< HTML code (for editing) - UL_OPTIMIZE_BIT = 0x8000 ///< flag, requiring cstr whitespace optimization + L_FILE_SPEC, ///< file specification + L_HTTP_HEADER, ///< text in HTTP response header + L_MAIL_HEADER, ///< text in mail header + L_URI, ///< text in uri + L_TABLE, ///< ^table:set body + L_SQL, ///< ^table:sql body + L_JS, ///< JavaScript code + L_XML, ///< ^dom:set xml + L_HTML, ///< HTML code (for editing) + L_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) {} + Language lang; ///< untaint flag, later untaint language + size_t length; ///< length + Fragment(Language alang, size_t asize): lang(alang), length(asize) { + assert(alang!=L_UNSPECIFIED); + assert(asize!=0); + assert(asize!=(size_t)-1); + } }; class ArrayFragment: public Array { - void simple_append(element_type src) { + void append(element_type src) { *static_cast *>(this)+=src; } + /// hiding from accidental USE, use append_positions + void append(const ArrayFragment& src, int offset, int limit) { + static_cast *>(this)->append(src, offset, limit); + } public: ArrayFragment& operator += (element_type src) { - if(count()) { // not empty? + if(size_t lcount=count()) { // not empty? // try to join with last - Fragment& last=get(count()-1); + Fragment& last=get_ref(lcount-1); if(last.lang==src.lang) { - last.size+=src.size; + last.length+=src.length; return *this; } } - simple_append(src); + append(src); return *this; } - ArrayFragment& append(const ArrayFragment& src, size_t substr_start, size_t substr_end); + void append(const ArrayFragment& src) { append(src, 0, ARRAY_OPTION_LIMIT_ALL); } + void append_positions(const ArrayFragment& src, size_t substr_begin, size_t substr_end); + + size_t length() { + size_t result=0; + for(Array_iterator i(*this); i.has_next(); ) { + const Fragment fragment=i.next(); + result+=fragment.length; + } + return result; + } }; struct C { + const char *str; + size_t length; + operator const char *() { return str; } + C(const char *astr, size_t asize): str(astr), length(asize) {} + }; + + struct Cm { char *str; - size_t size; - operator char *() { return str; } - C(char *astr, size_t asize): str(astr), size(asize) {} + size_t length; + //operator char *() { return str; } + Cm(char *astr, size_t asize): str(astr), length(asize) {} }; private: - CORD body; ///< all characters of string - ArrayFragment fragments; ///< fragment language+size info + StringBody body; ///< all characters of string + ArrayFragment fragments; ///< fragment language+length info public: - explicit String(const char* src=0, size_t src_size=0, bool tainted=false); + explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false); + explicit String(const C cstr, bool tainted=false); String(const String& src); - operator bool() const { return body!=CORD_EMPTY; } - size_t size() const { return CORD_len(body); } + String(StringBody abody, Language alang): body(abody) { + fragments+=Fragment(alang, abody.length()); + } + +#define ASSERT_STRING_INVARIANT(string) \ + assert((string).body.length()==(string).fragments.length()) + + /// for convinient hash lookup + operator const StringBody() const { return body; } + + bool is_empty() const { return body.is_empty(); } + size_t length() const { return body.length(); } - /// convert to C string. if 'lang' known, forcing 'lang' to it - CORD cstr_to_cord(Untaint_lang lang=UL_AS_IS, - SQL_Connection *connection=0, + /// convert to CORD. if 'lang' known, forcing 'lang' to it + StringBody cstr_to_string_body(Language lang=L_AS_IS, + SQL_Connection* connection=0, const Request_charsets *charsets=0) const; - /// convert to C string. if 'lang' known, forcing 'lang' to it - const char* cstr(Untaint_lang lang=UL_AS_IS, - SQL_Connection *connection=0, + /// convert to constant C string. if 'lang' known, forcing 'lang' to it + const char* cstr(Language lang=L_AS_IS, + SQL_Connection* connection=0, const Request_charsets *charsets=0) const { - return CORD_to_const_char_star(cstr_to_cord(lang, connection, charsets)); + return cstr_to_string_body(lang, connection, charsets).cstr(); } - /// convert to C string. if 'lang' known, forcing 'lang' to it - char *cstrm(Untaint_lang lang=UL_AS_IS, - SQL_Connection *connection=0, + /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it + char *cstrm(Language lang=L_AS_IS, + SQL_Connection* connection=0, const Request_charsets *charsets=0) const { - return CORD_to_char_star(cstr_to_cord(lang, connection, charsets)); + return cstr_to_string_body(lang, connection, charsets).cstrm(); } /// puts pieces to buf - C serialize(size_t prolog_size) const; + Cm serialize(size_t prolog_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); - /** - @param partial - returns partial match status. - - -1: strings too different - - 0: full match - - 1: means @c this starts @c src - - 2: means @src starts @this - */ - //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; } + bool deserialize(size_t prolog_size, void *buf, size_t buf_size); + /// @see StringBody::append_know_length + String& append_know_length(const char* str, size_t known_length, Language lang); + /// @see StringBody::append_help_length + String& append_help_length(const char* str, size_t helper_length, Language lang); + String& append_strdup(const char* str, size_t helper_length, Language lang); + + bool operator == (const char* y) const { return body==StringBody(y); } + bool operator != (const char* y) const { return body!=StringBody(y); } /// 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 starts_with(const char* y) const { + return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0; } - /// y starts with this - bool this_starts(CORD y) const { - return CORD_ncmp(body, 0/*x_start*/, y, 0/*y_start*/, CORD_len(y))==0; + /// x starts with this + bool this_starts(const char* x) const { + return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==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, Untaint_lang lang, bool forced) const; - String& append(const String& src, Untaint_lang lang, bool forced=false) { + String& append_to(String& dest, Language lang, bool forced) const; + String& append(const String& src, Language lang, bool forced=false) { return src.append_to(*this, lang, forced); } - String& operator << (const String& 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 - uint hash_code() const; + String& operator << (const String& src) { return append(src, L_PASS_APPENDED); } + String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); } + String& operator << (const StringBody src) { + body< (const String& src) const { return body>src.body; } + bool operator <= (const String& src) const { return body<=src.body; } + bool operator >= (const String& src) const { return body>=src.body; } + bool operator != (const String& src) const { return body!=src.body; } + bool operator == (const String& src) const { return body==src.body; } + /// extracts [start, finish) piece of string - const String& mid(size_t substr_start, size_t substr_end) const; + String& mid(size_t substr_begin, size_t substr_end) const; /** - ignore lang if it's UL_UNSPECIFIED + ignore lang if it's L_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 StringBody substr, + size_t this_offset=0, Language lang=L_UNSPECIFIED) const; + /// String version of @see pos(const char*, int, Language) size_t pos(const String& substr, - size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; + size_t this_offset=0, Language lang=L_UNSPECIFIED) const; + size_t pos(char c, + size_t this_offset=0) const { + return body.pos(c, this_offset); + } void split(ArrayString& result, - size_t& pos_after_ref, + size_t& pos_after, const char* delim, - Untaint_lang lang=UL_UNSPECIFIED, int limit=-1); + Language lang=L_UNSPECIFIED, int limit=-1) const; void split(ArrayString& result, - size_t& pos_after_ref, + size_t& pos_after, const String& delim, - Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; + Language lang=L_UNSPECIFIED, int limit=-1) const; - typedef void (*Row_action)(Table* table, ArrayString* row, + typedef void (*Row_action)(Table& table, ArrayString* row, int prestart, int prefinish, int poststart, int postfinish, void *info); @@ -255,7 +344,7 @@ public: */ Table* match(Charset& source_charset, const String& regexp, - const String& options, + const String* options, Row_action row_action, void *info, bool& just_matched) const; enum Change_case_kind { @@ -263,7 +352,7 @@ public: CC_LOWER }; String& change_case(Charset& source_charset, - Change_case_kind kind); + Change_case_kind kind) const; const String& replace(const Dictionary& dict) const; double as_double() const; int as_int() const; @@ -275,7 +364,7 @@ private: //disabled }; /// simple hash code of string. used by Hash -inline uint hash_code(const String& self) { +inline uint hash_code(const StringBody self) { return self.hash_code(); }