X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=utf8.c;h=ff113f9fdfbc3542f32000cd1de7d537e205115d;hb=f63f8e2fd0961bccc25f9a9f7fffef07b3c2a65a;hp=b27774af7f9aeb4f08a6aa1d9c6de0493ec8ccd7;hpb=6b8eaf932222db04db65aff99717b9c1dbd0a692;p=p5sagit%2Fp5-mst-13.2.git

diff --git a/utf8.c b/utf8.c
index b27774a..ff113f9 100644
--- a/utf8.c
+++ b/utf8.c
@@ -1,6 +1,6 @@
 /*    utf8.c
  *
- *    Copyright (c) 1998-1999, Larry Wall
+ *    Copyright (c) 1998-2000, Larry Wall
  *
  *    You may distribute under the terms of either the GNU General Public
  *    License or the Artistic License, as specified in the README file.
@@ -69,7 +69,7 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv)
 	return d;
     }
 #ifdef HAS_QUAD
-    if (uv < 0x2000000000)
+    if (uv < 0x1000000000LL)
 #endif
     {
 	*d++ =                        0xfe;	/* Can't match U+FEFF! */
@@ -84,6 +84,11 @@ Perl_uv_to_utf8(pTHX_ U8 *d, UV uv)
 #ifdef HAS_QUAD
     {
 	*d++ =                        0xff;	/* Can't match U+FFFE! */
+	*d++ =                        0x80;	/* 6 Reserved bits */
+	*d++ = (((uv >> 60) & 0x0f) | 0x80);	/* 2 Reserved bits */
+	*d++ = (((uv >> 54) & 0x3f) | 0x80);
+	*d++ = (((uv >> 48) & 0x3f) | 0x80);
+	*d++ = (((uv >> 42) & 0x3f) | 0x80);
 	*d++ = (((uv >> 36) & 0x3f) | 0x80);
 	*d++ = (((uv >> 30) & 0x3f) | 0x80);
 	*d++ = (((uv >> 24) & 0x3f) | 0x80);
@@ -120,8 +125,8 @@ Perl_utf8_to_uv(pTHX_ U8* s, I32* retlen)
     else if (!(uv & 0x08))	{ len = 4; uv &= 0x07; }
     else if (!(uv & 0x04))	{ len = 5; uv &= 0x03; }
     else if (!(uv & 0x02))	{ len = 6; uv &= 0x01; }
-    else if (!(uv & 0x01))	{ len = 7; uv &= 0x00; }
-    else 			  len = 8;	/* whoa! */
+    else if (!(uv & 0x01))	{ len = 7;  uv = 0; }
+    else 			{ len = 13; uv = 0; } /* whoa! */
 
     if (retlen)
 	*retlen = len;
@@ -255,7 +260,7 @@ Perl_utf16_to_utf8_reversed(pTHX_ U16* p, U8* d, I32 bytelen)
 bool
 Perl_is_uni_alnum(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_alnum(tmpbuf);
 }
@@ -263,7 +268,7 @@ Perl_is_uni_alnum(pTHX_ U32 c)
 bool
 Perl_is_uni_alnumc(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_alnumc(tmpbuf);
 }
@@ -271,7 +276,7 @@ Perl_is_uni_alnumc(pTHX_ U32 c)
 bool
 Perl_is_uni_idfirst(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_idfirst(tmpbuf);
 }
@@ -279,7 +284,7 @@ Perl_is_uni_idfirst(pTHX_ U32 c)
 bool
 Perl_is_uni_alpha(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_alpha(tmpbuf);
 }
@@ -287,7 +292,7 @@ Perl_is_uni_alpha(pTHX_ U32 c)
 bool
 Perl_is_uni_ascii(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_ascii(tmpbuf);
 }
@@ -295,7 +300,7 @@ Perl_is_uni_ascii(pTHX_ U32 c)
 bool
 Perl_is_uni_space(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_space(tmpbuf);
 }
@@ -303,7 +308,7 @@ Perl_is_uni_space(pTHX_ U32 c)
 bool
 Perl_is_uni_digit(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_digit(tmpbuf);
 }
@@ -311,7 +316,7 @@ Perl_is_uni_digit(pTHX_ U32 c)
 bool
 Perl_is_uni_upper(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_upper(tmpbuf);
 }
@@ -319,7 +324,7 @@ Perl_is_uni_upper(pTHX_ U32 c)
 bool
 Perl_is_uni_lower(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_lower(tmpbuf);
 }
@@ -327,7 +332,7 @@ Perl_is_uni_lower(pTHX_ U32 c)
 bool
 Perl_is_uni_cntrl(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_cntrl(tmpbuf);
 }
@@ -335,7 +340,7 @@ Perl_is_uni_cntrl(pTHX_ U32 c)
 bool
 Perl_is_uni_graph(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_graph(tmpbuf);
 }
@@ -343,7 +348,7 @@ Perl_is_uni_graph(pTHX_ U32 c)
 bool
 Perl_is_uni_print(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_print(tmpbuf);
 }
@@ -351,7 +356,7 @@ Perl_is_uni_print(pTHX_ U32 c)
 bool
 Perl_is_uni_punct(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_punct(tmpbuf);
 }
@@ -359,7 +364,7 @@ Perl_is_uni_punct(pTHX_ U32 c)
 bool
 Perl_is_uni_xdigit(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return is_utf8_xdigit(tmpbuf);
 }
@@ -367,7 +372,7 @@ Perl_is_uni_xdigit(pTHX_ U32 c)
 U32
 Perl_to_uni_upper(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return to_utf8_upper(tmpbuf);
 }
@@ -375,7 +380,7 @@ Perl_to_uni_upper(pTHX_ U32 c)
 U32
 Perl_to_uni_title(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return to_utf8_title(tmpbuf);
 }
@@ -383,7 +388,7 @@ Perl_to_uni_title(pTHX_ U32 c)
 U32
 Perl_to_uni_lower(pTHX_ U32 c)
 {
-    U8 tmpbuf[10];
+    U8 tmpbuf[UTF8_MAXLEN];
     uv_to_utf8(tmpbuf, (UV)c);
     return to_utf8_lower(tmpbuf);
 }