0x5C is not a Yen sign in CP932 (or CP51932)

When Microsoft created CP932 (their version of Shift-JIS), they explicitly
used bytes 0-0x7F to represent ASCII characters rather than JIS X 0201
characters.

So when converting Unicode to CP932, it is not correct to convert U+00A5
to CP932 0x5C. Fortunately, CP932 does have a multi-byte FULLWIDTH YEN SIGN
character which we can use instead.

CP51932 uses the same extended character set as CP932; while CP932 is
MicroSoft's extended version of Shift-JIS, CP51932 is their extended version
of EUC-JP. So the same reasoning applies to CP51932.

diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c b/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c
index d9fecc9d4d65a52d12fad7fc394398bce1d24644..aa52d05481c4e9316dc468aa9db8b7431d119c1a 100644 (file)
--- a/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c
+++ b/ext/mbstring/libmbfl/filters/mbfilter_cp51932.c
@@ -214,8 +214,8 @@ mbfl_filt_conv_wchar_cp51932(int c, mbfl_convert_filter *filter)
        }
        if (s1 >= 0x8080) s1 = -1; /* we don't support JIS X0213 */
        if (s1 <= 0) {
-               if (c == 0xa5) {                /* YEN SIGN */
-                       s1 = 0x005c;                    /* YEN SIGN */
+               if (c == 0xa5) { /* YEN SIGN */
+                       s1 = 0x216F; /* FULLWIDTH YEN SIGN */
                } else if (c == 0x203e) {       /* OVER LINE */
                        s1 = 0x007e;                    /* FULLWIDTH MACRON */
                } else if (c == 0xff3c) {       /* FULLWIDTH REVERSE SOLIDUS */

diff --git a/ext/mbstring/libmbfl/filters/mbfilter_cp932.c b/ext/mbstring/libmbfl/filters/mbfilter_cp932.c
index 6246600de8248b4b332d68426f0bb5ebc04984de..ec192faa2b1aa3eacf130eedd3b1359876b877f5 100644 (file)
--- a/ext/mbstring/libmbfl/filters/mbfilter_cp932.c
+++ b/ext/mbstring/libmbfl/filters/mbfilter_cp932.c
@@ -251,8 +251,8 @@ mbfl_filt_conv_wchar_cp932(int c, mbfl_convert_filter *filter)
                s2 = 1;
        }
        if (s1 <= 0) {
-               if (c == 0xa5) {                /* YEN SIGN */
-                       s1 = 0x005c;    /* YEN SIGN */
+               if (c == 0xa5) { /* YEN SIGN */
+                       s1 = 0x216F; /* FULLWIDTH YEN SIGN */
                } else if (c == 0x203e) {       /* OVER LINE */
                        s1 = 0x007e;    /* FULLWIDTH MACRON */
                } else if (c == 0xff3c) {       /* FULLWIDTH REVERSE SOLIDUS */

diff --git a/ext/mbstring/tests/cp51932_encoding.phpt b/ext/mbstring/tests/cp51932_encoding.phpt
index 8dbbeb85a023cb1e00e6775203ce3d12a4c9e9eb..2fc25fd3c8d1b4f52706799f18b4e6b6ac310449 100644 (file)
--- a/ext/mbstring/tests/cp51932_encoding.phpt
+++ b/ext/mbstring/tests/cp51932_encoding.phpt
@@ -84,6 +84,9 @@ unset($fromUnicode["\x30\x94"]); // Don't map hiragana vu to katakana vu
 for ($i = 0; $i <= 0x7F; $i++)
   $validChars[chr($i)] = "\x00" . chr($i);
 
+/* U+00A5 is YEN SIGN; convert to FULLWIDTH YEN SIGN */
+$fromUnicode["\x00\xA5"] = "\xA1\xEF";
+
 testAllValidChars($validChars, 'CP51932', 'UTF-16BE', false);
 testAllValidChars($fromUnicode, 'UTF-16BE', 'CP51932', false);
 echo "CP51932 verification and conversion works on all valid characters\n";

diff --git a/ext/mbstring/tests/cp932_encoding.phpt b/ext/mbstring/tests/cp932_encoding.phpt
index ddcf8a449b181c2fb9be9108fbd81855c247efcc..ec9e76f1f6651a154bd3f7f8150653f982d732c1 100644 (file)
--- a/ext/mbstring/tests/cp932_encoding.phpt
+++ b/ext/mbstring/tests/cp932_encoding.phpt
@@ -30,6 +30,8 @@ for ($i = 0xF0; $i <= 0xF9; $i++) {
 $fromUnicode["\x00\xA2"] = "\x81\x91";
 /* U+00A3 is POUND SIGN; convert to FULLWIDTH POUND SIGN */
 $fromUnicode["\x00\xA3"] = "\x81\x92";
+/* U+00A5 is YEN SIGN; convert to FULLWIDTH YEN SIGN */
+$fromUnicode["\x00\xA5"] = "\x81\x8F";
 
 /* We map the JIS X 0208 FULLWIDTH TILDE to U+FF5E (FULLWIDTH TILDE)
  * But when converting Unicode to CP932, we also accept U+301C (WAVE DASH) */