More liberal parsing of version numbers.

[perl5.git] / utf8.c

diff --git a/utf8.c b/utf8.c
index 9bb89a4..e86c49f 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -134,6 +134,30 @@ Perl_is_utf8_char(pTHX_ U8 *s)
     return len;
 }
 
+/*
+=for apidoc Am|bool_utf8_string|U8 *s|STRLEN len
+
+Returns true if first C<len> bytes of the given string form valid a UTF8
+string, false otherwise.
+
+=cut
+*/
+
+bool 
+Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
+{
+    U8* x=s;
+    U8* send=s+len;
+    int c;
+    while (x < send) {
+        c = is_utf8_char(x);
+        x += c;
+        if (!c || x > send)
+            return 0;
+    }
+    return 1;
+}
+
 UV
 Perl_utf8_to_uv(pTHX_ U8* s, I32* retlen)
 {
@@ -296,19 +320,26 @@ Perl_bytes_to_utf8(pTHX_ U8* s, STRLEN *len)
     return dst;
 }
 
-/* XXX NOTHING CALLS THE FOLLOWING TWO ROUTINES YET!!! */
 /*
- * Convert native or reversed UTF-16 to UTF-8.
+ * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
  *
  * Destination must be pre-extended to 3/2 source.  Do not use in-place.
  * We optimize for native, for obvious reasons. */
 
 U8*
-Perl_utf16_to_utf8(pTHX_ U16* p, U8* d, I32 bytelen)
+Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
 {
-    U16* pend = p + bytelen / 2;
+    U8* pend;
+    U8* dstart = d;
+
+    if (bytelen & 1)
+       Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
+
+    pend = p + bytelen;
+
     while (p < pend) {
-       UV uv = *p++;
+       UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
+       p += 2;
        if (uv < 0x80) {
            *d++ = uv;
            continue;
@@ -320,13 +351,9 @@ Perl_utf16_to_utf8(pTHX_ U16* p, U8* d, I32 bytelen)
        }
        if (uv >= 0xd800 && uv < 0xdbff) {      /* surrogates */
             dTHR;
-           int low = *p++;
-           if (low < 0xdc00 || low >= 0xdfff) {
-               if (ckWARN_d(WARN_UTF8))     
-                   Perl_warner(aTHX_ WARN_UTF8, "Malformed UTF-16 surrogate");
-               p--;
-               uv = 0xfffd;
-           }
+           UV low = *p++;
+           if (low < 0xdc00 || low >= 0xdfff)
+               Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
            uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
        }
        if (uv < 0x10000) {
@@ -343,13 +370,14 @@ Perl_utf16_to_utf8(pTHX_ U16* p, U8* d, I32 bytelen)
            continue;
        }
     }
+    *newlen = d - dstart;
     return d;
 }
 
 /* Note: this one is slightly destructive of the source. */
 
 U8*
-Perl_utf16_to_utf8_reversed(pTHX_ U16* p, U8* d, I32 bytelen)
+Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
 {
     U8* s = (U8*)p;
     U8* send = s + bytelen;
@@ -359,7 +387,7 @@ Perl_utf16_to_utf8_reversed(pTHX_ U16* p, U8* d, I32 bytelen)
        s[1] = tmp;
        s += 2;
     }
-    return utf16_to_utf8(p, d, bytelen);
+    return utf16_to_utf8(p, d, bytelen, newlen);
 }
 
 /* for now these are all defined (inefficiently) in terms of the utf8 versions */