fix unicode split /\s+/

[perl5.git] / pp.c

diff --git a/pp.c b/pp.c
index 645ba62..4b021c0 100644 (file)
--- a/pp.c
+++ b/pp.c
@@ -1,7 +1,7 @@
 /*    pp.c
  *
  *    Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 /*    pp.c
  *
  *    Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,

- *    2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
+ *    2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 by Larry Wall and others

  *
  *    You may distribute under the terms of either the GNU General Public
  *    License or the Artistic License, as specified in the README file.
  *
  *    You may distribute under the terms of either the GNU General Public
  *    License or the Artistic License, as specified in the README file.


@@ -4035,7 +4035,7 @@ PP(pp_hslice)
 
         if (lval) {
             if (!svp || *svp == &PL_sv_undef) {
 
         if (lval) {
             if (!svp || *svp == &PL_sv_undef) {

-                DIE(aTHX_ PL_no_helem_sv, (void*)keysv);
+                DIE(aTHX_ PL_no_helem_sv, SVfARG(keysv));

             }
             if (localizing) {
                if (HvNAME_get(hv) && isGV(*svp))
             }
             if (localizing) {
                if (HvNAME_get(hv) && isGV(*svp))


@@ -4606,12 +4606,29 @@ PP(pp_split)
     if (!limit)
        limit = maxiters + 2;
     if (pm->op_pmflags & PMf_WHITE) {
     if (!limit)
        limit = maxiters + 2;
     if (pm->op_pmflags & PMf_WHITE) {

+        if (do_utf8 && !PL_utf8_space) { 
+            /* force PL_utf8_space to be loaded */
+            bool ok; 
+            ENTER; 
+            ok = is_utf8_space((const U8*)" "); 
+            assert(ok); 
+            LEAVE; 
+        } 

        while (--limit) {
            m = s;
        while (--limit) {
            m = s;

-           while (m < strend &&
-                  !((pm->op_pmflags & PMf_LOCALE)
-                    ? isSPACE_LC(*m) : isSPACE(*m)))
-               ++m;
+           /* this one uses 'm' and is a negative test */
+           if (do_utf8) {
+               STRLEN uskip;
+                while (m < strend &&
+                       !( *m == ' ' || swash_fetch(PL_utf8_space,(U8*)m, do_utf8) ))
+                   m +=  UTF8SKIP(m);
+            } else if (pm->op_pmflags & PMf_LOCALE) {
+               while (m < strend && !isSPACE_LC(*m))
+                   ++m;
+            } else {
+                while (m < strend && !isSPACE(*m))
+                    ++m;
+            }  

            if (m >= strend)
                break;
 
            if (m >= strend)
                break;
 


@@ -4623,13 +4640,21 @@ PP(pp_split)
            XPUSHs(dstr);
 
            s = m + 1;
            XPUSHs(dstr);
 
            s = m + 1;

-           while (s < strend &&
-                  ((pm->op_pmflags & PMf_LOCALE)
-                   ? isSPACE_LC(*s) : isSPACE(*s)))
-               ++s;
+           /* this one uses 's' and is a positive test */
+           if (do_utf8) {
+                while (s < strend &&
+                       ( *s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8) ))
+                   s +=  UTF8SKIP(s);
+            } else if (pm->op_pmflags & PMf_LOCALE) {
+               while (s < strend && isSPACE_LC(*s))
+                   ++s;
+            } else {
+                while (s < strend && isSPACE(*s))
+                    ++s;
+            }      

        }
     }
        }
     }

-    else if (rx->precomp[0] == '^' && rx->precomp[1] == '\0') {
+    else if (rx->extflags & RXf_START_ONLY) {

        while (--limit) {
            for (m = s; m < strend && *m != '\n'; m++)
                ;
        while (--limit) {
            for (m = s; m < strend && *m != '\n'; m++)
                ;