varlena.c

PostgreSQL 8.1.4的源码 适用于Linux下的开源数据库系统

	/* now null pad to full length... */
	while (len < NAMEDATALEN)
	{
		*(NameStr(*result) + len) = '\0';
		len++;
	}

	PG_RETURN_NAME(result);
}

/* name_text()
 * Converts a Name type to a text type.
 */
Datum
name_text(PG_FUNCTION_ARGS)
{
	Name		s = PG_GETARG_NAME(0);
	text	   *result;
	int			len;

	len = strlen(NameStr(*s));

#ifdef STRINGDEBUG
	printf("text- convert string length %d (%d) ->%d\n",
		   VARSIZE(s) - VARHDRSZ, VARSIZE(s), len);
#endif

	result = palloc(VARHDRSZ + len);
	VARATT_SIZEP(result) = VARHDRSZ + len;
	memcpy(VARDATA(result), NameStr(*s), len);

	PG_RETURN_TEXT_P(result);
}


/*
 * textToQualifiedNameList - convert a text object to list of names
 *
 * This implements the input parsing needed by nextval() and other
 * functions that take a text parameter representing a qualified name.
 * We split the name at dots, downcase if not double-quoted, and
 * truncate names if they're too long.
 */
List *
textToQualifiedNameList(text *textval)
{
	char	   *rawname;
	List	   *result = NIL;
	List	   *namelist;
	ListCell   *l;

	/* Convert to C string (handles possible detoasting). */
	/* Note we rely on being able to modify rawname below. */
	rawname = DatumGetCString(DirectFunctionCall1(textout,
												  PointerGetDatum(textval)));

	if (!SplitIdentifierString(rawname, '.', &namelist))
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_NAME),
				 errmsg("invalid name syntax")));

	if (namelist == NIL)
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_NAME),
				 errmsg("invalid name syntax")));

	foreach(l, namelist)
	{
		char	   *curname = (char *) lfirst(l);

		result = lappend(result, makeString(pstrdup(curname)));
	}

	pfree(rawname);
	list_free(namelist);

	return result;
}

/*
 * SplitIdentifierString --- parse a string containing identifiers
 *
 * This is the guts of textToQualifiedNameList, and is exported for use in
 * other situations such as parsing GUC variables.	In the GUC case, it's
 * important to avoid memory leaks, so the API is designed to minimize the
 * amount of stuff that needs to be allocated and freed.
 *
 * Inputs:
 *	rawstring: the input string; must be overwritable!	On return, it's
 *			   been modified to contain the separated identifiers.
 *	separator: the separator punctuation expected between identifiers
 *			   (typically '.' or ',').	Whitespace may also appear around
 *			   identifiers.
 * Outputs:
 *	namelist: filled with a palloc'd list of pointers to identifiers within
 *			  rawstring.  Caller should list_free() this even on error return.
 *
 * Returns TRUE if okay, FALSE if there is a syntax error in the string.
 *
 * Note that an empty string is considered okay here, though not in
 * textToQualifiedNameList.
 */
bool
SplitIdentifierString(char *rawstring, char separator,
					  List **namelist)
{
	char	   *nextp = rawstring;
	bool		done = false;

	*namelist = NIL;

	while (isspace((unsigned char) *nextp))
		nextp++;				/* skip leading whitespace */

	if (*nextp == '\0')
		return true;			/* allow empty string */

	/* At the top of the loop, we are at start of a new identifier. */
	do
	{
		char	   *curname;
		char	   *endp;

		if (*nextp == '\"')
		{
			/* Quoted name --- collapse quote-quote pairs, no downcasing */
			curname = nextp + 1;
			for (;;)
			{
				endp = strchr(nextp + 1, '\"');
				if (endp == NULL)
					return false;		/* mismatched quotes */
				if (endp[1] != '\"')
					break;		/* found end of quoted name */
				/* Collapse adjacent quotes into one quote, and look again */
				memmove(endp, endp + 1, strlen(endp));
				nextp = endp;
			}
			/* endp now points at the terminating quote */
			nextp = endp + 1;
		}
		else
		{
			/* Unquoted name --- extends to separator or whitespace */
			char	   *downname;
			int			len;

			curname = nextp;
			while (*nextp && *nextp != separator &&
				   !isspace((unsigned char) *nextp))
				nextp++;
			endp = nextp;
			if (curname == nextp)
				return false;	/* empty unquoted name not allowed */

			/*
			 * Downcase the identifier, using same code as main lexer does.
			 *
			 * XXX because we want to overwrite the input in-place, we cannot
			 * support a downcasing transformation that increases the string
			 * length.	This is not a problem given the current implementation
			 * of downcase_truncate_identifier, but we'll probably have to do
			 * something about this someday.
			 */
			len = endp - curname;
			downname = downcase_truncate_identifier(curname, len, false);
			Assert(strlen(downname) <= len);
			strncpy(curname, downname, len);
			pfree(downname);
		}

		while (isspace((unsigned char) *nextp))
			nextp++;			/* skip trailing whitespace */

		if (*nextp == separator)
		{
			nextp++;
			while (isspace((unsigned char) *nextp))
				nextp++;		/* skip leading whitespace for next */
			/* we expect another name, so done remains false */
		}
		else if (*nextp == '\0')
			done = true;
		else
			return false;		/* invalid syntax */

		/* Now safe to overwrite separator with a null */
		*endp = '\0';

		/* Truncate name if it's overlength */
		truncate_identifier(curname, strlen(curname), false);

		/*
		 * Finished isolating current name --- add it to list
		 */
		*namelist = lappend(*namelist, curname);

		/* Loop back if we didn't reach end of string */
	} while (!done);

	return true;
}


/*****************************************************************************
 *	Comparison Functions used for bytea
 *
 * Note: btree indexes need these routines not to leak memory; therefore,
 * be careful to free working copies of toasted datums.  Most places don't
 * need to be so careful.
 *****************************************************************************/

Datum
byteaeq(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	bool		result;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	/* fast path for different-length inputs */
	if (len1 != len2)
		result = false;
	else
		result = (memcmp(VARDATA(arg1), VARDATA(arg2), len1) == 0);

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL(result);
}

Datum
byteane(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	bool		result;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	/* fast path for different-length inputs */
	if (len1 != len2)
		result = true;
	else
		result = (memcmp(VARDATA(arg1), VARDATA(arg2), len1) != 0);

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL(result);
}

Datum
bytealt(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	int			cmp;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	cmp = memcmp(VARDATA(arg1), VARDATA(arg2), Min(len1, len2));

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 < len2)));
}

Datum
byteale(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	int			cmp;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	cmp = memcmp(VARDATA(arg1), VARDATA(arg2), Min(len1, len2));

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 <= len2)));
}

Datum
byteagt(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	int			cmp;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	cmp = memcmp(VARDATA(arg1), VARDATA(arg2), Min(len1, len2));

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 > len2)));
}

Datum
byteage(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	int			cmp;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	cmp = memcmp(VARDATA(arg1), VARDATA(arg2), Min(len1, len2));

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 >= len2)));
}

Datum
byteacmp(PG_FUNCTION_ARGS)
{
	bytea	   *arg1 = PG_GETARG_BYTEA_P(0);
	bytea	   *arg2 = PG_GETARG_BYTEA_P(1);
	int			len1,
				len2;
	int			cmp;

	len1 = VARSIZE(arg1) - VARHDRSZ;
	len2 = VARSIZE(arg2) - VARHDRSZ;

	cmp = memcmp(VARDATA(arg1), VARDATA(arg2), Min(len1, len2));
	if ((cmp == 0) && (len1 != len2))
		cmp = (len1 < len2) ? -1 : 1;

	PG_FREE_IF_COPY(arg1, 0);
	PG_FREE_IF_COPY(arg2, 1);

	PG_RETURN_INT32(cmp);
}

/*
 * appendStringInfoText
 *
 * Append a text to str.
 * Like appendStringInfoString(str, PG_TEXT_GET_STR(s)) but faster.
 */
static void
appendStringInfoText(StringInfo str, const text *t)
{
	appendBinaryStringInfo(str, VARDATA(t), VARSIZE(t) - VARHDRSZ);
}

/*
 * replace_text
 * replace all occurrences of 'old_sub_str' in 'orig_str'
 * with 'new_sub_str' to form 'new_str'
 *
 * returns 'orig_str' if 'old_sub_str' == '' or 'orig_str' == ''
 * otherwise returns 'new_str'
 */
Datum
replace_text(PG_FUNCTION_ARGS)
{
	text	   *src_text = PG_GETARG_TEXT_P(0);
	text	   *from_sub_text = PG_GETARG_TEXT_P(1);
	text	   *to_sub_text = PG_GETARG_TEXT_P(2);
	int			src_text_len = TEXTLEN(src_text);
	int			from_sub_text_len = TEXTLEN(from_sub_text);
	text	   *left_text;
	text	   *right_text;
	text	   *buf_text;
	text	   *ret_text;
	int			curr_posn;
	StringInfo	str;

	if (src_text_len == 0 || from_sub_text_len == 0)
		PG_RETURN_TEXT_P(src_text);

	curr_posn = TEXTPOS(src_text, from_sub_text);

	/* When the from_sub_text is not found, there is nothing to do. */
	if (curr_posn == 0)
		PG_RETURN_TEXT_P(src_text);

	str = makeStringInfo();
	buf_text = src_text;

	while (curr_posn > 0)
	{
		left_text = text_substring(PointerGetDatum(buf_text),
								   1, curr_posn - 1, false);
		right_text = text_substring(PointerGetDatum(buf_text),
									curr_posn + from_sub_text_len, -1, true);

		appendStringInfoText(str, left_text);
		appendStringInfoText(str, to_sub_text);

		if (buf_text != src_text)
			pfree(buf_text);
		pfree(left_text);
		buf_text = right_text;
		curr_posn = TEXTPOS(buf_text, from_sub_text);
	}

	appendStringInfoText(str, buf_text);
	if (buf_text != src_text)
		pfree(buf_text);

	ret_text = PG_STR_GET_TEXT(str->data);
	pfree(str->data);
	pfree(str);

	PG_RETURN_TEXT_P(ret_text);
}

/*
 * check_replace_text_has_escape_char
 *
 * check whether replace_text contains escape char.
 */
static bool
check_replace_text_has_escape_char(const text *replace_text)
{
	const char *p = VARDATA(replace_text);
	const char *p_end = p + (VARSIZE(replace_text) - VARHDRSZ);

	if (pg_database_encoding_max_length() == 1)
	{
		for (; p < p_end; p++)
		{
			if (*p == '\\')
				return true;
		}
	}
	else
	{
		for (; p < p_end; p += pg_mblen(p))
		{
			if (*p == '\\')
				return true;
		}
	}

	return false;
}

/*
 * appendStringInfoRegexpSubstr
 *
 * Append replace_text to str, substituting regexp back references for
 * \n escapes.
 */
static void
appendStringInfoRegexpSubstr(StringInfo str, text *replace_text,
							 regmatch_t *pmatch, text *src_text)
{
	const char *p = VARDATA(replace_text);
	const char *p_end = p + (VARSIZE(replace_text) - VARHDRSZ);
	int			eml = pg_database_encoding_max_length();

	for (;;)
	{
		const char *chunk_start = p;
		int			so;
		int			eo;

		/* Find next escape char. */
		if (eml == 1)
		{
			for (; p < p_end && *p != '\\'; p++)
				 /* nothing */ ;
		}
		else
		{
			for (; p < p_end && *p != '\\'; p += pg_mblen(p))
				 /* nothing */ ;
		}

		/* Copy the text we just scanned over, if any. */
		if (p > chunk_start)
			appendBinaryStringInfo(str, chunk_start, p - chunk_start);

		/* Done if at end of string, else advance over escape char. */
		if (p >= p_end)
			break;
		p++;

		if (p >= p_end)
		{
			/* Escape at very end of input.  Treat same as unexpected char */
			appendStringInfoChar(str, '\\');
			break;
		}

		if (*p >= '1' && *p <= '9')
		{
			/* Use the back reference of regexp. */
			int			idx = *p - '0';

			so = pmatch[idx].rm_so;
			eo = pmatch[idx].rm_eo;
			p++;
		}
		else if (*p == '&')
		{
			/* Use the entire matched string. */
			so = pmatch[0].rm_so;
			eo = pmatch[0].rm_eo;
			p++;
		}
		else if (*p == '\\')
		{
			/* \\ means transfer one \ to output. */
			appendStringInfoChar(str, '\\');
			p++;
			continue;
		}
		else
		{
			/*
			 * If escape char is not followed by any expected char, just treat
			 * it as ordinary data to copy.  (XXX would it be better to throw
			 * an error?)