Avoid "will not stay shared" warnings.

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

diff --git a/utf8.c b/utf8.c
index 0c09469..d73809c 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -1180,46 +1180,95 @@ Perl_is_utf8_mark(pTHX_ U8 *p)
     return swash_fetch(PL_utf8_mark, p, TRUE);
 }
 
+/*
+=for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
+
+The "p" contains the pointer to the UTF-8 string encoding
+the character that is being converted.
+
+The "ustrp" is a pointer to the character buffer to put the
+conversion result to.  The "lenp" is a pointer to the length
+of the result.
+
+The "swash" is a pointer to the swash to use.
+
+The "normal" is a string like "ToLower" which means the swash
+$utf8::ToLower, which is stored in lib/unicore/To/Lower.pl,
+and loaded by SWASHGET, using lib/utf8_heavy.pl.
+
+The "special" is a string like "utf8::ToSpecLower", which means
+the hash %utf8::ToSpecLower, which is stored in the same file,
+lib/unicore/To/Lower.pl, and also loaded by SWASHGET.  The access
+to the hash is by Perl_to_utf8_case().
+
+=cut
+ */
+
 UV
-Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
+Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp,char *normal, char *special)
 {
     UV uv;
 
-    if (!PL_utf8_toupper)
-       PL_utf8_toupper = swash_init("utf8", "ToUpper", &PL_sv_undef, 4, 0);
-    uv = swash_fetch(PL_utf8_toupper, p, TRUE);
-    uv = uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p, 0);
+    if (!*swashp)
+        *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
+    uv = swash_fetch(*swashp, p, TRUE);
+    if (uv)
+        uv = UNI_TO_NATIVE(uv);
+    else {
+        HV *hv;
+        SV *keysv;
+        HE *he;
+
+        uv = utf8_to_uvchr(p, 0);
+
+        if ((hv    = get_hv(special, FALSE)) &&
+            (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv))) &&
+            (he    = hv_fetch_ent(hv, keysv, FALSE, 0))) {
+             SV *val = HeVAL(he);
+             char *s = SvPV(val, *lenp);
+             U8 c = *(U8*)s;
+             if (*lenp > 1 || UNI_IS_INVARIANT(c))
+                  Copy(s, ustrp, *lenp, U8);
+             else {
+                  /* something in the 0x80..0xFF range */
+                  ustrp[0] = UTF8_EIGHT_BIT_HI(c);
+                  ustrp[1] = UTF8_EIGHT_BIT_LO(c);
+                  *lenp = 2;
+             }
+             return 0;
+        }
+    }
     *lenp = UNISKIP(uv);
     uvuni_to_utf8(ustrp, uv);
     return uv;
 }
 
 UV
-Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
+Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
 {
-    UV uv;
+    return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
+                             &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
+}
 
-    if (!PL_utf8_totitle)
-       PL_utf8_totitle = swash_init("utf8", "ToTitle", &PL_sv_undef, 4, 0);
-    uv = swash_fetch(PL_utf8_totitle, p, TRUE);
-    uv = uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p, 0);
-    *lenp = UNISKIP(uv);
-    uvuni_to_utf8(ustrp, uv);
-    return uv;
+UV
+Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
+{
+    return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
+                             &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
 }
 
 UV
 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
 {
-    UV uv;
+    return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
+                             &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
+}
 
-    if (!PL_utf8_tolower)
-       PL_utf8_tolower = swash_init("utf8", "ToLower", &PL_sv_undef, 4, 0);
-    uv = swash_fetch(PL_utf8_tolower, p, TRUE);
-    uv = uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p, 0);
-    *lenp = UNISKIP(uv);
-    uvuni_to_utf8(ustrp, uv);
-    return uv;
+UV
+Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
+{
+    return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
+                             &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
 }
 
 /* a "swash" is a swatch hash */
@@ -1295,7 +1344,7 @@ Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
     U32 off;
     STRLEN slen;
     STRLEN needents;
-    U8 *tmps;
+    U8 *tmps = NULL;
     U32 bit;
     SV *retval;
     U8 tmputf8[2];
@@ -1466,4 +1515,31 @@ Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
     return UNI_TO_NATIVE(uv);
 }
 
+char *
+Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
+{
+    int truncated = 0;
+    char *s, *e;
 
+    sv_setpvn(dsv, "", 0);
+    for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
+        UV u;
+        if (pvlim && SvCUR(dsv) >= pvlim) {
+             truncated++;
+             break;
+        }
+        u = utf8_to_uvchr((U8*)s, 0);
+        Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
+    }
+    if (truncated)
+        sv_catpvn(dsv, "...", 3);
+    
+    return SvPVX(dsv);
+}
+
+char *
+Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
+{
+  return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
+                            pvlim, flags);
+}