--- parser3/src/include/pa_string.h	2003/03/24 08:27:45	1.144.2.28.2.21
+++ parser3/src/include/pa_string.h	2003/04/08 12:47:30	1.144.2.28.2.47
@@ -8,7 +8,7 @@
 #ifndef PA_STRING_H
 #define PA_STRING_H
 
-static const char* IDENT_STRING_H="$Date: 2003/03/24 08:27:45 $";
+static const char* IDENT_STRING_H="$Date: 2003/04/08 12:47:30 $";
 
 // 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"
 };
 
@@ -29,17 +30,6 @@ class Request_charsets;
 class String;
 typedef Array<const String*> ArrayString;
 
-// helpers
-/*
-/// appends clean piece to String @see String::append
-#define APPEND_CLEAN(src, length) append(src, length, String::L_CLEAN)
-/// appends piece to String as-is @see String::append
-#define APPEND_AS_IS(src, length) append(src, length, String::L_AS_IS)
-/// appends tainted piece to String  @see String::append
-#define APPEND_TAINTED(src, length) append(src, length, String::L_TAINTED)
-/// handy: appends const char* piece to String  @see String::append
-#define	APPEND_CONST(src) APPEND_AS_IS(src, 0)
-*/
 /// this is result of pos functions which mean that substr were not found
 #define STRING_NOT_FOUND ((size_t)-1)
 
@@ -50,10 +40,24 @@ class StringBody {
 public:
 
 	StringBody(): body(CORD_EMPTY) {}
-	StringBody(CORD abody): body(abody) {}
+	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); }
@@ -63,19 +67,23 @@ public:
 
 	bool is_empty() const { return body==CORD_EMPTY; }
 
-	void append(const char *str, size_t length) {
-		CORD_cat_char_star(body, str, length);
+	void append_know_length(const char *str, size_t known_length) {
+		body=CORD_cat_char_star(body, str, known_length);
 	}
-	void append(char c) { CORD_cat_char(body, c); }
-	void append(StringBody src) { CORD_cat(body, src.body); }
+	void append_strdup_know_length(const char* str, size_t 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; }
+	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);
@@ -83,8 +91,16 @@ public:
 
 	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(size_t offset, const char* substr) const { return CORD_str(body, offset, substr); }
-	size_t pos(size_t offset, const StringBody substr) const { return CORD_str(body, offset, substr.body); }
+	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<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
 		CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
@@ -92,6 +108,13 @@ public:
 
 	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);
+	}
 };
 
 /** 
@@ -103,7 +126,7 @@ public:
 */
 class String: public PA_Object {
 
-	friend class StringBody;
+//	friend class StringBody;
 
 public:
 
@@ -138,7 +161,12 @@ public:
 	struct Fragment { 
 		Language lang; ///< untaint flag, later untaint language
 		size_t length;  ///< length
-		Fragment(Language alang, size_t asize): lang(alang), length(asize) {}
+		Fragment(Language alang, size_t asize): lang(alang), length(asize) {
+			if(alang==L_UNSPECIFIED)
+				_asm int 3;
+			assert(alang!=L_UNSPECIFIED);
+			assert(asize!=(size_t)-1);
+		}
 	};
 
 	class ArrayFragment: public Array<Fragment> {
@@ -151,9 +179,9 @@ public:
 		}
 	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.length+=src.length;
 					return *this;
@@ -164,13 +192,29 @@ public:
 		}
 		void append(const ArrayFragment& src) { append(src, 0, 0); }
 		void append_positions(const ArrayFragment& src, size_t substr_begin, size_t substr_end);
+
+		size_t length() {
+			size_t result=0;
+			for(Array_iterator<element_type> 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 length;
-		operator char *() { return str; }
-		C(char *astr, size_t asize): str(astr), length(asize) {}
+		//operator char *() { return str; }
+		Cm(char *astr, size_t asize): str(astr), length(asize) {}
 	};
 
 private:
@@ -180,15 +224,22 @@ private:
 
 public:
 
-	/// @todo check all places for length to be only HELPER length, and never be !=strlen(cstr)
 	explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
+	explicit String(C cstr, bool tainted=false);
 	String(const String& src);
 	String(StringBody abody, Language alang): body(abody) {
 		fragments+=Fragment(alang, abody.length());
 	}
+#ifndef NDEBUG
+	void invariant() {
+		if(body.length()!=fragments.length())
+			_asm int 3;
+		//assert(body.length()==fragments.length());
+	}
+#endif
 
 	/// for convinient hash lookup
-	operator StringBody() const { return body; }
+	operator const StringBody() const { return body; }
 
 	bool is_empty() const { return body.is_empty(); }
 	size_t length() const { return body.length(); }
@@ -211,14 +262,14 @@ public:
 		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, 
-		@c length can be zero, it will be autocalced
-		when length<strlen(str) makes heap copy of @c str piece
-	*/
-	String& append(const char* str, size_t length, Language lang);
+	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); }
@@ -227,9 +278,9 @@ public:
 	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(const char* y) const {
-		return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(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;
 	}
 
 	String& append_to(String& dest, Language lang, bool forced) const;
@@ -237,9 +288,9 @@ public:
 		return src.append_to(*this, lang, forced);
 	}
 	String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
-	String& operator << (const char* src) { return append(src, 0, L_AS_IS); }
-	String& operator << (const StringBody& src) { 
-		body.append(src);
+	String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
+	String& operator << (const StringBody src) { 
+		body<<src;
 		fragments+=Fragment(L_AS_IS, src.length());
 		return *this;
 	}
@@ -257,7 +308,7 @@ public:
 	bool operator == (const String& src) const { return body==src.body; }
 
 	/// extracts [start, finish) piece of string
-	const String& mid(size_t substr_begin, size_t substr_end) const;
+	String& mid(size_t substr_begin, size_t substr_end) const;
 
 	/** 
 		ignore lang if it's L_UNSPECIFIED
@@ -269,17 +320,21 @@ public:
 	/// String version of @see pos(const char*, int, Language)
 	size_t pos(const String& substr, 
 		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, 
+		size_t& pos_after,
 		const char* delim, 
-		Language lang=L_UNSPECIFIED, int limit=-1);
+		Language lang=L_UNSPECIFIED, int limit=-1) const;
 	void split(ArrayString& result, 
 		size_t& pos_after, 
 		const String& delim, 
 		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);
@@ -292,7 +347,7 @@ 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 {
@@ -300,7 +355,7 @@ 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;
@@ -312,7 +367,7 @@ private: //disabled
 };
 
 /// simple hash code of string. used by Hash
-inline uint hash_code(const StringBody& self) {
+inline uint hash_code(const StringBody self) {
 	return self.hash_code();
 }