RESENT - [PATCH] utf8_heavy.pl

[p5sagit/p5-mst-13.2.git] / utf8.c

diff --git a/utf8.c b/utf8.c
index 747b78c..f21b13c 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -68,6 +68,8 @@ Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
        return d;
     }
     if (uv < 0x10000) {
+        if (UNICODE_IS_SURROGATE(uv))
+           Perl_croak(aTHX_ "UTF-16 surrogate 0x%04"UVxf, uv);
        *d++ = (( uv >> 12)         | 0xe0);
        *d++ = (((uv >>  6) & 0x3f) | 0x80);
        *d++ = (( uv        & 0x3f) | 0x80);
@@ -903,7 +905,7 @@ Perl_is_uni_punct(pTHX_ UV c)
 bool
 Perl_is_uni_xdigit(pTHX_ UV c)
 {
-    U8 tmpbuf[UTF8_MAXLEN*2+1];
+    U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
     uvchr_to_utf8(tmpbuf, (UV)c);
     return is_utf8_xdigit(tmpbuf);
 }
@@ -911,7 +913,7 @@ Perl_is_uni_xdigit(pTHX_ UV c)
 UV
 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
 {
-    U8 tmpbuf[UTF8_MAXLEN*2+1];
+    U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
     uvchr_to_utf8(tmpbuf, (UV)c);
     return to_utf8_upper(tmpbuf, p, lenp);
 }
@@ -919,7 +921,7 @@ Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
 UV
 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
 {
-    U8 tmpbuf[UTF8_MAXLEN*2+1];
+    U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
     uvchr_to_utf8(tmpbuf, (UV)c);
     return to_utf8_title(tmpbuf, p, lenp);
 }
@@ -927,7 +929,7 @@ Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
 UV
 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
 {
-    U8 tmpbuf[UTF8_MAXLEN+1];
+    U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
     uvchr_to_utf8(tmpbuf, (UV)c);
     return to_utf8_lower(tmpbuf, p, lenp);
 }
@@ -935,7 +937,7 @@ Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
 UV
 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
 {
-    U8 tmpbuf[UTF8_MAXLEN+1];
+    U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
     uvchr_to_utf8(tmpbuf, (UV)c);
     return to_utf8_fold(tmpbuf, p, lenp);
 }
@@ -1247,7 +1249,8 @@ Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp,char *normal
              return 0;
         }
     }
-    *lenp = UNISKIP(uv);
+    if (lenp)
+       *lenp = UNISKIP(uv);
     uvuni_to_utf8(ustrp, uv);
     return uv;
 }
@@ -1596,8 +1599,8 @@ Perl_ibcmp_utf8(pTHX_ const char *s1, bool u1, register I32 len1, const char *s2
      STRLEN la, lb;
      UV ca, cb;
      STRLEN ulen1, ulen2;
-     U8 tmpbuf1[UTF8_MAXLEN*3+1];
-     U8 tmpbuf2[UTF8_MAXLEN*3+1];
+     U8 tmpbuf1[UTF8_MAXLEN_FOLD+1];
+     U8 tmpbuf2[UTF8_MAXLEN_FOLD+1];
      
      while (a < ae && b < be) {
          if (u1) {