varlena.c

postgresql8.3.4源码,开源数据库

	{
		S1 = Max(S, 1);

		if (length_not_specified)		/* special case - get length to end of
										 * string */
			L1 = -1;
		else
		{
			/* end position */
			int			E = S + length;

			/*
			 * A negative value for L is the only way for the end position to
			 * be before the start. SQL99 says to throw an error.
			 */
			if (E < S)
				ereport(ERROR,
						(errcode(ERRCODE_SUBSTRING_ERROR),
						 errmsg("negative substring length not allowed")));

			/*
			 * A zero or negative value for the end position can happen if the
			 * start was negative or one. SQL99 says to return a zero-length
			 * string.
			 */
			if (E < 1)
				return PG_STR_GET_TEXT("");

			L1 = E - S1;
		}

		/*
		 * If the start position is past the end of the string, SQL99 says to
		 * return a zero-length string -- PG_GETARG_TEXT_P_SLICE() will do
		 * that for us. Convert to zero-based starting position
		 */
		return DatumGetTextPSlice(str, S1 - 1, L1);
	}
	else if (eml > 1)
	{
		/*
		 * When encoding max length is > 1, we can't get LC without
		 * detoasting, so we'll grab a conservatively large slice now and go
		 * back later to do the right thing
		 */
		int32		slice_start;
		int32		slice_size;
		int32		slice_strlen;
		text	   *slice;
		int32		E1;
		int32		i;
		char	   *p;
		char	   *s;
		text	   *ret;

		/*
		 * if S is past the end of the string, the tuple toaster will return a
		 * zero-length string to us
		 */
		S1 = Max(S, 1);

		/*
		 * We need to start at position zero because there is no way to know
		 * in advance which byte offset corresponds to the supplied start
		 * position.
		 */
		slice_start = 0;

		if (length_not_specified)		/* special case - get length to end of
										 * string */
			slice_size = L1 = -1;
		else
		{
			int			E = S + length;

			/*
			 * A negative value for L is the only way for the end position to
			 * be before the start. SQL99 says to throw an error.
			 */
			if (E < S)
				ereport(ERROR,
						(errcode(ERRCODE_SUBSTRING_ERROR),
						 errmsg("negative substring length not allowed")));

			/*
			 * A zero or negative value for the end position can happen if the
			 * start was negative or one. SQL99 says to return a zero-length
			 * string.
			 */
			if (E < 1)
				return PG_STR_GET_TEXT("");

			/*
			 * if E is past the end of the string, the tuple toaster will
			 * truncate the length for us
			 */
			L1 = E - S1;

			/*
			 * Total slice size in bytes can't be any longer than the start
			 * position plus substring length times the encoding max length.
			 */
			slice_size = (S1 + L1) * eml;
		}

		/*
		 * If we're working with an untoasted source, no need to do an extra
		 * copying step.
		 */
		if (VARATT_IS_COMPRESSED(str) || VARATT_IS_EXTERNAL(str))
			slice = DatumGetTextPSlice(str, slice_start, slice_size);
		else
			slice = (text *) DatumGetPointer(str);

		/* see if we got back an empty string */
		if (VARSIZE_ANY_EXHDR(slice) == 0)
		{
			if (slice != (text *) DatumGetPointer(str))
				pfree(slice);
			return PG_STR_GET_TEXT("");
		}

		/* Now we can get the actual length of the slice in MB characters */
		slice_strlen = pg_mbstrlen_with_len(VARDATA_ANY(slice),
											VARSIZE_ANY_EXHDR(slice));

		/*
		 * Check that the start position wasn't > slice_strlen. If so, SQL99
		 * says to return a zero-length string.
		 */
		if (S1 > slice_strlen)
		{
			if (slice != (text *) DatumGetPointer(str))
				pfree(slice);
			return PG_STR_GET_TEXT("");
		}

		/*
		 * Adjust L1 and E1 now that we know the slice string length. Again
		 * remember that S1 is one based, and slice_start is zero based.
		 */
		if (L1 > -1)
			E1 = Min(S1 + L1, slice_start + 1 + slice_strlen);
		else
			E1 = slice_start + 1 + slice_strlen;

		/*
		 * Find the start position in the slice; remember S1 is not zero based
		 */
		p = VARDATA_ANY(slice);
		for (i = 0; i < S1 - 1; i++)
			p += pg_mblen(p);

		/* hang onto a pointer to our start position */
		s = p;

		/*
		 * Count the actual bytes used by the substring of the requested
		 * length.
		 */
		for (i = S1; i < E1; i++)
			p += pg_mblen(p);

		ret = (text *) palloc(VARHDRSZ + (p - s));
		SET_VARSIZE(ret, VARHDRSZ + (p - s));
		memcpy(VARDATA(ret), s, (p - s));

		if (slice != (text *) DatumGetPointer(str))
			pfree(slice);

		return ret;
	}
	else
		elog(ERROR, "invalid backend encoding: encoding max length < 1");

	/* not reached: suppress compiler warning */
	return NULL;
}

/*
 * textpos -
 *	  Return the position of the specified substring.
 *	  Implements the SQL92 POSITION() function.
 *	  Ref: A Guide To The SQL Standard, Date & Darwen, 1997
 * - thomas 1997-07-27
 */
Datum
textpos(PG_FUNCTION_ARGS)
{
	text	   *str = PG_GETARG_TEXT_PP(0);
	text	   *search_str = PG_GETARG_TEXT_PP(1);

	PG_RETURN_INT32((int32) text_position(str, search_str));
}

/*
 * text_position -
 *	Does the real work for textpos()
 *
 * Inputs:
 *		t1 - string to be searched
 *		t2 - pattern to match within t1
 * Result:
 *		Character index of the first matched char, starting from 1,
 *		or 0 if no match.
 *
 *	This is broken out so it can be called directly by other string processing
 *	functions.
 */
static int
text_position(text *t1, text *t2)
{
	TextPositionState state;
	int			result;

	text_position_setup(t1, t2, &state);
	result = text_position_next(1, &state);
	text_position_cleanup(&state);
	return result;
}

/*
 * text_position_setup, text_position_next, text_position_cleanup -
 *	Component steps of text_position()
 *
 * These are broken out so that a string can be efficiently searched for
 * multiple occurrences of the same pattern.  text_position_next may be
 * called multiple times with increasing values of start_pos, which is
 * the 1-based character position to start the search from.  The "state"
 * variable is normally just a local variable in the caller.
 */

static void
text_position_setup(text *t1, text *t2, TextPositionState *state)
{
	int			len1 = VARSIZE_ANY_EXHDR(t1);
	int			len2 = VARSIZE_ANY_EXHDR(t2);

	if (pg_database_encoding_max_length() == 1)
	{
		/* simple case - single byte encoding */
		state->use_wchar = false;
		state->str1 = VARDATA_ANY(t1);
		state->str2 = VARDATA_ANY(t2);
		state->len1 = len1;
		state->len2 = len2;
	}
	else
	{
		/* not as simple - multibyte encoding */
		pg_wchar   *p1,
				   *p2;

		p1 = (pg_wchar *) palloc((len1 + 1) * sizeof(pg_wchar));
		len1 = pg_mb2wchar_with_len(VARDATA_ANY(t1), p1, len1);
		p2 = (pg_wchar *) palloc((len2 + 1) * sizeof(pg_wchar));
		len2 = pg_mb2wchar_with_len(VARDATA_ANY(t2), p2, len2);

		state->use_wchar = true;
		state->wstr1 = p1;
		state->wstr2 = p2;
		state->len1 = len1;
		state->len2 = len2;
	}
}

static int
text_position_next(int start_pos, TextPositionState *state)
{
	int			pos = 0,
				p,
				px;

	Assert(start_pos > 0);		/* else caller error */

	if (state->len2 <= 0)
		return start_pos;		/* result for empty pattern */

	if (!state->use_wchar)
	{
		/* simple case - single byte encoding */
		char	   *p1 = state->str1;
		char	   *p2 = state->str2;

		/* no use in searching str past point where search_str will fit */
		px = (state->len1 - state->len2);

		p1 += start_pos - 1;

		for (p = start_pos - 1; p <= px; p++)
		{
			if ((*p1 == *p2) && (strncmp(p1, p2, state->len2) == 0))
			{
				pos = p + 1;
				break;
			}
			p1++;
		}
	}
	else
	{
		/* not as simple - multibyte encoding */
		pg_wchar   *p1 = state->wstr1;
		pg_wchar   *p2 = state->wstr2;

		/* no use in searching str past point where search_str will fit */
		px = (state->len1 - state->len2);

		p1 += start_pos - 1;

		for (p = start_pos - 1; p <= px; p++)
		{
			if ((*p1 == *p2) && (pg_wchar_strncmp(p1, p2, state->len2) == 0))
			{
				pos = p + 1;
				break;
			}
			p1++;
		}
	}

	return pos;
}

static void
text_position_cleanup(TextPositionState *state)
{
	if (state->use_wchar)
	{
		pfree(state->wstr1);
		pfree(state->wstr2);
	}
}

/* varstr_cmp()
 * Comparison function for text strings with given lengths.
 * Includes locale support, but must copy strings to temporary memory
 *	to allow null-termination for inputs to strcoll().
 * Returns -1, 0 or 1
 */
int
varstr_cmp(char *arg1, int len1, char *arg2, int len2)
{
	int			result;

	/*
	 * Unfortunately, there is no strncoll(), so in the non-C locale case we
	 * have to do some memory copying.	This turns out to be significantly
	 * slower, so we optimize the case where LC_COLLATE is C.  We also try to
	 * optimize relatively-short strings by avoiding palloc/pfree overhead.
	 */
	if (lc_collate_is_c())
	{
		result = strncmp(arg1, arg2, Min(len1, len2));
		if ((result == 0) && (len1 != len2))
			result = (len1 < len2) ? -1 : 1;
	}
	else
	{
#define STACKBUFLEN		1024

		char		a1buf[STACKBUFLEN];
		char		a2buf[STACKBUFLEN];
		char	   *a1p,
				   *a2p;

#ifdef WIN32
		/* Win32 does not have UTF-8, so we need to map to UTF-16 */
		if (GetDatabaseEncoding() == PG_UTF8)
		{
			int			a1len;
			int			a2len;
			int			r;

			if (len1 >= STACKBUFLEN / 2)
			{
				a1len = len1 * 2 + 2;
				a1p = palloc(a1len);
			}
			else
			{
				a1len = STACKBUFLEN;
				a1p = a1buf;
			}
			if (len2 >= STACKBUFLEN / 2)
			{
				a2len = len2 * 2 + 2;
				a2p = palloc(a2len);
			}
			else
			{
				a2len = STACKBUFLEN;
				a2p = a2buf;
			}

			/* stupid Microsloth API does not work for zero-length input */
			if (len1 == 0)
				r = 0;
			else
			{
				r = MultiByteToWideChar(CP_UTF8, 0, arg1, len1,
										(LPWSTR) a1p, a1len / 2);
				if (!r)
					ereport(ERROR,
					 (errmsg("could not convert string to UTF-16: error %lu",
							 GetLastError())));
			}
			((LPWSTR) a1p)[r] = 0;

			if (len2 == 0)
				r = 0;
			else
			{
				r = MultiByteToWideChar(CP_UTF8, 0, arg2, len2,
										(LPWSTR) a2p, a2len / 2);
				if (!r)
					ereport(ERROR,
					 (errmsg("could not convert string to UTF-16: error %lu",
							 GetLastError())));
			}
			((LPWSTR) a2p)[r] = 0;

			errno = 0;
			result = wcscoll((LPWSTR) a1p, (LPWSTR) a2p);
			if (result == 2147483647)	/* _NLSCMPERROR; missing from mingw
										 * headers */
				ereport(ERROR,
						(errmsg("could not compare Unicode strings: %m")));

			/*
			 * In some locales wcscoll() can claim that nonidentical strings
			 * are equal.  Believing that would be bad news for a number of
			 * reasons, so we follow Perl's lead and sort "equal" strings
			 * according to strcmp (on the UTF-8 representation).
			 */
			if (result == 0)
			{
				result = strncmp(arg1, arg2, Min(len1, len2));
				if ((result == 0) && (len1 != len2))
					result = (len1 < len2) ? -1 : 1;
			}

			if (a1p != a1buf)
				pfree(a1p);
			if (a2p != a2buf)
				pfree(a2p);

			return result;
		}
#endif   /* WIN32 */

		if (len1 >= STACKBUFLEN)
			a1p = (char *) palloc(len1 + 1);
		else
			a1p = a1buf;
		if (len2 >= STACKBUFLEN)
			a2p = (char *) palloc(len2 + 1);
		else
			a2p = a2buf;

		memcpy(a1p, arg1, len1);
		a1p[len1] = '\0';
		memcpy(a2p, arg2, len2);
		a2p[len2] = '\0';

		result = strcoll(a1p, a2p);

		/*
		 * In some locales strcoll() can claim that nonidentical strings are
		 * equal.  Believing that would be bad news for a number of reasons,
		 * so we follow Perl's lead and sort "equal" strings according to
		 * strcmp().
		 */
		if (result == 0)
			result = strcmp(a1p, a2p);

		if (a1p != a1buf)
			pfree(a1p);
		if (a2p != a2buf)
			pfree(a2p);
	}

	return result;
}


/* text_cmp()
 * Internal comparison function for text strings.
 * Returns -1, 0 or 1
 */
static int
text_cmp(text *arg1, text *arg2)
{
	char	   *a1p,
			   *a2p;
	int			len1,
				len2;

	a1p = VARDATA_ANY(arg1);
	a2p = VARDATA_ANY(arg2);

	len1 = VARSIZE_ANY_EXHDR(arg1);
	len2 = VARSIZE_ANY_EXHDR(arg2);

	return varstr_cmp(a1p, len1, a2p, len2);
}

/*
 * Comparison functions for text strings.
 *
 * 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
texteq(PG_FUNCTION_ARGS)
{
	text	   *arg1 = PG_GETARG_TEXT_PP(0);
	text	   *arg2 = PG_GETARG_TEXT_PP(1);
	bool		result;

	/*
	 * Since we only care about equality or not-equality, we can avoid all the
	 * expense of strcoll() here, and just do bitwise comparison.
	 */
	if (VARSIZE_ANY_EXHDR(arg1) != VARSIZE_ANY_EXHDR(arg2))
		result = false;
	else
		result = (strncmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2),
						  VARSIZE_ANY_EXHDR(arg1)) == 0);

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

	PG_RETURN_BOOL(result);
}

Datum
textne(PG_FUNCTION_ARGS)
{
	text	   *arg1 = PG_GETARG_TEXT_PP(0);
	text	   *arg2 = PG_GETARG_TEXT_PP(1);
	bool		result;

	/*
	 * Since we only care about equality or not-equality, we can avoid all the
	 * expense of strcoll() here, and just do bitwise comparison.
	 */
	if (VARSIZE_ANY_EXHDR(arg1) != VARSIZE_ANY_EXHDR(arg2))
		result = true;
	else
		result = (strncmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2),
						  VARSIZE_ANY_EXHDR(arg1)) != 0);

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

	PG_RETURN_BOOL(result);
}

Datum
text_lt(PG_FUNCTION_ARGS)
{
	text	   *arg1 = PG_GETARG_TEXT_PP(0);
	text	   *arg2 = PG_GETARG_TEXT_PP(1);
	bool		result;

	result = (text_cmp(arg1, arg2) < 0);

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

	PG_RETURN_BOOL(result);
}