Description:
------------
In ext/standard/html.c , htmlentities() fails to identify BIG5 Chinese characters correctly.

I have checked CVS version 1.87, the bug is still there.

Reproduce code:
---------------
In html.c, look for this piece of code :

case cs_big5:
case cs_gb2312:
case cs_big5hkscs:
    {
	/* check if this is the first of a 2-byte sequence */
	if (this_char >= 0xa1 && this_char <= 0xf9) {
	/* peek at the next char */
	unsigned char next_char = str[pos];
		if ((next_char >= 0x40 && next_char <= 0x73) ||(next_char >= 0xa1 && next_char <= 0xfe)) {
			

Expected result:
----------------
In fact, the first byte should be from 0xa1 to 0xfe, and the second byte should be from 0x40-0x7e and 0xa1-0xfe.

(from page 88, "Understanding Japanese Information Processing" by Ken Lunde , O'Reilly.)

Actual result:
--------------
So it should be :

	if (this_char >= 0xa1 && this_char <= 0xfe) {

and 

		if ((next_char >= 0x40 && next_char <= 0x7e) ||(next_char >= 0xa1 && next_char <= 0xfe)) {