This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Re: Scalar leaked in 'local $0' under ithreads + taint mode
[perl5.git] / regexec.c
index fae084e..6a644d9 100644 (file)
--- a/regexec.c
+++ b/regexec.c
@@ -55,7 +55,6 @@
 #  define PERL_NO_GET_CONTEXT
 #endif
 
-/*SUPPRESS 112*/
 /*
  * pregcomp and pregexec -- regsub and regerror are not used in perl
  *
 #define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim))
 #define HOPMAYBE3c(pos,off,lim) ((char*)HOPMAYBE3(pos,off,lim))
 
-#define LOAD_UTF8_CHARCLASS(a,b) STMT_START { if (!CAT2(PL_utf8_,a)) { ENTER; save_re_context(); (void)CAT2(is_utf8_, a)((const U8*)b); LEAVE; } } STMT_END
+#define LOAD_UTF8_CHARCLASS(class,str) STMT_START { \
+    if (!CAT2(PL_utf8_,class)) { bool ok; ENTER; save_re_context(); ok=CAT2(is_utf8_,class)((const U8*)str); assert(ok); LEAVE; } } STMT_END
+#define LOAD_UTF8_CHARCLASS_ALNUM() LOAD_UTF8_CHARCLASS(alnum,"a")
+#define LOAD_UTF8_CHARCLASS_DIGIT() LOAD_UTF8_CHARCLASS(digit,"0")
+#define LOAD_UTF8_CHARCLASS_SPACE() LOAD_UTF8_CHARCLASS(space," ")
+#define LOAD_UTF8_CHARCLASS_MARK()  LOAD_UTF8_CHARCLASS(mark, "\xcd\x86")
 
 /* for use after a quantifier and before an EXACT-like node -- japhy */
 #define JUMPABLE(rn) ( \
@@ -412,8 +416,8 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
     char *check_at = Nullch;           /* check substr found at this pos */
     const I32 multiline = prog->reganch & PMf_MULTILINE;
 #ifdef DEBUGGING
-    char *i_strpos = strpos;
-    SV *dsv = PERL_DEBUG_PAD_ZERO(0);
+    const char * const i_strpos = strpos;
+    SV * const dsv = PERL_DEBUG_PAD_ZERO(0);
 #endif
 
     GET_RE_DEBUG_FLAGS_DECL;
@@ -611,7 +615,8 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
        if (check == (do_utf8 ? prog->float_utf8 : prog->float_substr)) {
          do_other_anchored:
            {
-               char *last = HOP3c(s, -start_shift, strbeg), *last1, *last2;
+               char * const last = HOP3c(s, -start_shift, strbeg);
+               char *last1, *last2;
                char *s1 = s;
                SV* must;
 
@@ -857,7 +862,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
           regstclass does not come from lookahead...  */
        /* If regstclass takes bytelength more than 1: If charlength==1, OK.
           This leaves EXACTF only, which is dealt with in find_byclass().  */
-        const U8* str = (U8*)STRING(prog->regstclass);
+        const U8* const str = (U8*)STRING(prog->regstclass);
         const int cl_l = (PL_regkind[(U8)OP(prog->regstclass)] == EXACT
                    ? CHR_DIST(str+STR_LEN(prog->regstclass), str)
                    : 1);
@@ -1189,7 +1194,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
                }
                tmp = ((OP(c) == BOUND ?
                        isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (tmp == !(OP(c) == BOUND ?
                                 swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
@@ -1232,7 +1237,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
                }
                tmp = ((OP(c) == NBOUND ?
                        isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (tmp == !(OP(c) == NBOUND ?
                                 swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
@@ -1261,7 +1266,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case ALNUM:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1319,7 +1324,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case NALNUM:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (!swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1377,7 +1382,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case SPACE:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(space," ");
+               LOAD_UTF8_CHARCLASS_SPACE();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8)) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1435,7 +1440,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case NSPACE:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(space," ");
+               LOAD_UTF8_CHARCLASS_SPACE();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8))) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1493,7 +1498,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case DIGIT:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(digit,"0");
+               LOAD_UTF8_CHARCLASS_DIGIT();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1551,7 +1556,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32
            break;
        case NDIGIT:
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(digit,"0");
+               LOAD_UTF8_CHARCLASS_DIGIT();
                while (s + (uskip = UTF8SKIP(s)) <= strend) {
                    if (!swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
                        if (tmp && (norun || regtry(prog, s)))
@@ -1646,7 +1651,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
 
     GET_RE_DEBUG_FLAGS_DECL;
 
-    (void)data; /* Currently unused */
+    PERL_UNUSED_ARG(data);
     RX_MATCH_UTF8_set(prog,do_utf8);
 
     PL_regcc = 0;
@@ -1844,7 +1849,6 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
                                   "Did not find anchored character...\n")
                );
     }
-    /*SUPPRESS 560*/
     else if (prog->anchored_substr != Nullsv
              || prog->anchored_utf8 != Nullsv
              || ((prog->float_substr != Nullsv || prog->float_utf8 != Nullsv)
@@ -1944,8 +1948,8 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
        }
        DEBUG_EXECUTE_r({
            SV *prop = sv_newmortal();
-           char *s0;
-           char *s1;
+           const char *s0;
+           const char *s1;
            int len0;
            int len1;
 
@@ -1953,7 +1957,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
            s0 = UTF ?
              pv_uni_display(dsv0, (U8*)SvPVX_const(prop), SvCUR(prop), 60,
                             UNI_DISPLAY_REGEX) :
-             SvPVX(prop);
+             SvPVX_const(prop);
            len0 = UTF ? SvCUR(dsv0) : SvCUR(prop);
            s1 = UTF ?
              sv_uni_display(dsv1, sv, 60, UNI_DISPLAY_REGEX) : s;
@@ -1989,7 +1993,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *
            }
            else {
                STRLEN len;
-                const char * const little = SvPV(float_real, len);
+                const char * const little = SvPV_const(float_real, len);
 
                if (SvTAIL(float_real)) {
                    if (memEQ(strend - len + 1, little, len - 1))
@@ -2056,7 +2060,7 @@ got_it:
        RX_MATCH_COPY_FREE(prog);
        if (flags & REXEC_COPY_STR) {
            I32 i = PL_regeol - startpos + (stringarg - strbeg);
-#ifdef PERL_COPY_ON_WRITE
+#ifdef PERL_OLD_COPY_ON_WRITE
            if ((SvIsCOW(sv)
                 || (SvFLAGS(sv) & CAN_COW_MASK) == CAN_COW_FLAGS)) {
                if (DEBUG_C_TEST) {
@@ -2065,7 +2069,7 @@ got_it:
                                  (int) SvTYPE(sv));
                }
                prog->saved_copy = sv_setsv_cow(prog->saved_copy, sv);
-               prog->subbeg = SvPVX(prog->saved_copy);
+               prog->subbeg = (char *)SvPVX_const(prog->saved_copy);
                assert (SvPOKp(prog->saved_copy));
            } else
 #endif
@@ -2143,7 +2147,7 @@ S_regtry(pTHX_ regexp *prog, char *startpos)
            SAVEDESTRUCTOR_X(restore_pos, 0);
         }
         if (!PL_reg_curpm) {
-           Newz(22, PL_reg_curpm, 1, PMOP);
+           Newxz(PL_reg_curpm, 1, PMOP);
 #ifdef USE_ITHREADS
             {
                 SV* repointer = newSViv(0);
@@ -2164,7 +2168,7 @@ S_regtry(pTHX_ regexp *prog, char *startpos)
                $` inside (?{}) could fail... */
            PL_reg_oldsaved = prog->subbeg;
            PL_reg_oldsavedlen = prog->sublen;
-#ifdef PERL_COPY_ON_WRITE
+#ifdef PERL_OLD_COPY_ON_WRITE
            PL_nrs = prog->saved_copy;
 #endif
            RX_MATCH_COPIED_off(prog);
@@ -2189,7 +2193,7 @@ S_regtry(pTHX_ regexp *prog, char *startpos)
         if(PL_reg_start_tmp)
             Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
         else
-            New(22, PL_reg_start_tmp, PL_reg_start_tmpl, char*);
+            Newx(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
     }
 
     /* XXXX What this code is doing here?!!!  There should be no need
@@ -2609,7 +2613,7 @@ S_regmatch(pTHX_ regnode *prog)
 
                    if ( base ) {
 
-                       if ( do_utf8 || UTF ) {
+                       if ( do_utf8 ) {
                            if ( foldlen>0 ) {
                                uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags );
                                foldlen -= len;
@@ -2675,7 +2679,7 @@ S_regmatch(pTHX_ regnode *prog)
 
                    if ( base ) {
 
-                       if ( do_utf8 || UTF ) {
+                       if ( do_utf8 ) {
                            uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags );
                        } else {
                            uvc = (U32)*uc;
@@ -2722,7 +2726,7 @@ S_regmatch(pTHX_ regnode *prog)
                            "%*s  %sonly one match : #%d <%s>%s\n",
                            REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],
                            accept_buff[ 0 ].wordnum,
-                           tmp ? SvPV_nolen( *tmp ) : "not compiled under -Dr",
+                           tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",
                            PL_colors[5] );
                    });
                    PL_reginput = (char *)accept_buff[ 0 ].endpos;
@@ -2757,7 +2761,7 @@ S_regmatch(pTHX_ regnode *prog)
                            PerlIO_printf( Perl_debug_log, "%*s  %strying alternation #%d <%s> at 0x%p%s\n",
                                REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],
                                accept_buff[best].wordnum,
-                               tmp ? SvPV_nolen( *tmp ) : "not compiled under -Dr",scan,
+                               tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",scan,
                                PL_colors[5] );
                        });
                        if ( best<accepted ) {
@@ -2929,7 +2933,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (!nextchr)
                sayNO;
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                if (!(OP(scan) == ALNUM
                      ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
                      : isALNUM_LC_utf8((U8*)locinput)))
@@ -2952,7 +2956,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (!nextchr && locinput >= PL_regeol)
                sayNO;
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(alnum,"a");
+               LOAD_UTF8_CHARCLASS_ALNUM();
                if (OP(scan) == NALNUM
                    ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
                    : isALNUM_LC_utf8((U8*)locinput))
@@ -2985,7 +2989,7 @@ S_regmatch(pTHX_ regnode *prog)
                }
                if (OP(scan) == BOUND || OP(scan) == NBOUND) {
                    ln = isALNUM_uni(ln);
-                   LOAD_UTF8_CHARCLASS(alnum,"a");
+                   LOAD_UTF8_CHARCLASS_ALNUM();
                    n = swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8);
                }
                else {
@@ -3017,7 +3021,7 @@ S_regmatch(pTHX_ regnode *prog)
                sayNO;
            if (do_utf8) {
                if (UTF8_IS_CONTINUED(nextchr)) {
-                   LOAD_UTF8_CHARCLASS(space," ");
+                   LOAD_UTF8_CHARCLASS_SPACE();
                    if (!(OP(scan) == SPACE
                          ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
                          : isSPACE_LC_utf8((U8*)locinput)))
@@ -3047,7 +3051,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (!nextchr && locinput >= PL_regeol)
                sayNO;
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(space," ");
+               LOAD_UTF8_CHARCLASS_SPACE();
                if (OP(scan) == NSPACE
                    ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
                    : isSPACE_LC_utf8((U8*)locinput))
@@ -3070,7 +3074,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (!nextchr)
                sayNO;
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(digit,"0");
+               LOAD_UTF8_CHARCLASS_DIGIT();
                if (!(OP(scan) == DIGIT
                      ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
                      : isDIGIT_LC_utf8((U8*)locinput)))
@@ -3093,7 +3097,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (!nextchr && locinput >= PL_regeol)
                sayNO;
            if (do_utf8) {
-               LOAD_UTF8_CHARCLASS(digit,"0");
+               LOAD_UTF8_CHARCLASS_DIGIT();
                if (OP(scan) == NDIGIT
                    ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
                    : isDIGIT_LC_utf8((U8*)locinput))
@@ -3113,7 +3117,7 @@ S_regmatch(pTHX_ regnode *prog)
            if (locinput >= PL_regeol)
                sayNO;
            if  (do_utf8) {
-               LOAD_UTF8_CHARCLASS(mark,"~");
+               LOAD_UTF8_CHARCLASS_MARK();
                if (swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8))
                    sayNO;
                locinput += PL_utf8skip[nextchr];
@@ -3248,7 +3252,7 @@ S_regmatch(pTHX_ regnode *prog)
                    }
                    else {
                        STRLEN len;
-                       char *t = SvPV(ret, len);
+                       const char *t = SvPV_const(ret, len);
                        PMOP pm;
                        char * const oprecomp = PL_regprecomp;
                        const I32 osize = PL_regsize;
@@ -3256,7 +3260,7 @@ S_regmatch(pTHX_ regnode *prog)
 
                        Zero(&pm, 1, PMOP);
                         if (DO_UTF8(ret)) pm.op_pmdynflags |= PMdf_DYN_UTF8;
-                       re = CALLREGCOMP(aTHX_ t, t + len, &pm);
+                       re = CALLREGCOMP(aTHX_ (char*)t, (char*)t + len, &pm);
                        if (!(SvFLAGS(ret)
                              & (SVs_TEMP | SVs_PADTMP | SVf_READONLY
                                | SVs_GMG)))
@@ -3551,7 +3555,7 @@ S_regmatch(pTHX_ regnode *prog)
                    }
                    else {
                        PL_reg_poscache_size = size;
-                       Newz(29, PL_reg_poscache, size, char);
+                       Newxz(PL_reg_poscache, size, char);
                    }
                    DEBUG_EXECUTE_r(
                        PerlIO_printf(Perl_debug_log,
@@ -4323,7 +4327,6 @@ do_no:
                goto do_no;
            }
            /* Have more choice yet.  Reuse the same uwb.  */
-           /*SUPPRESS 560*/
            if ((n = (uwb->type == RE_UNWIND_BRANCH
                      ? NEXT_OFF(next) : ARG(next))))
                next += n;
@@ -4437,7 +4440,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case ALNUM:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(alnum,"a");
+           LOAD_UTF8_CHARCLASS_ALNUM();
            while (hardcount < max && scan < loceol &&
                   swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
                scan += UTF8SKIP(scan);
@@ -4465,7 +4468,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case NALNUM:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(alnum,"a");
+           LOAD_UTF8_CHARCLASS_ALNUM();
            while (hardcount < max && scan < loceol &&
                   !swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
                scan += UTF8SKIP(scan);
@@ -4493,7 +4496,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case SPACE:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(space," ");
+           LOAD_UTF8_CHARCLASS_SPACE();
            while (hardcount < max && scan < loceol &&
                   (*scan == ' ' ||
                    swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
@@ -4522,7 +4525,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case NSPACE:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(space," ");
+           LOAD_UTF8_CHARCLASS_SPACE();
            while (hardcount < max && scan < loceol &&
                   !(*scan == ' ' ||
                     swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
@@ -4551,7 +4554,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case DIGIT:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(digit,"0");
+           LOAD_UTF8_CHARCLASS_DIGIT();
            while (hardcount < max && scan < loceol &&
                   swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
                scan += UTF8SKIP(scan);
@@ -4565,7 +4568,7 @@ S_regrepeat(pTHX_ const regnode *p, I32 max)
     case NDIGIT:
        if (do_utf8) {
            loceol = PL_regeol;
-           LOAD_UTF8_CHARCLASS(digit,"0");
+           LOAD_UTF8_CHARCLASS_DIGIT();
            while (hardcount < max && scan < loceol &&
                   !swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
                scan += UTF8SKIP(scan);
@@ -4668,8 +4671,8 @@ Perl_regclass_swash(pTHX_ register const regnode* node, bool doinit, SV** listsv
        const U32 n = ARG(node);
 
        if (PL_regdata->what[n] == 's') {
-           SV *rv = (SV*)PL_regdata->data[n];
-           AV *av = (AV*)SvRV((SV*)rv);
+           SV * const rv = (SV*)PL_regdata->data[n];
+           AV * const av = (AV*)SvRV((SV*)rv);
            SV **ary = AvARRAY(av);
            SV **a, **b;
        
@@ -4735,7 +4738,7 @@ S_reginclass(pTHX_ register const regnode *n, register const U8* p, STRLEN* lenp
            match = TRUE;
        if (!match) {
            AV *av;
-           SV *sw = regclass_swash(n, TRUE, 0, (SV**)&av);
+           SV * const sw = regclass_swash(n, TRUE, 0, (SV**)&av);
        
            if (sw) {
                if (swash_fetch(sw, p, do_utf8))
@@ -4743,11 +4746,10 @@ S_reginclass(pTHX_ register const regnode *n, register const U8* p, STRLEN* lenp
                else if (flags & ANYOF_FOLD) {
                    if (!match && lenp && av) {
                        I32 i;
-                     
                        for (i = 0; i <= av_len(av); i++) {
-                           SV* sv = *av_fetch(av, i, FALSE);
+                           SV* const sv = *av_fetch(av, i, FALSE);
                            STRLEN len;
-                           const char *s = SvPV(sv, len);
+                           const char * const s = SvPV_const(sv, len);
                        
                            if (len <= plen && memEQ(s, (char*)p, len)) {
                                *lenp = len;
@@ -4898,12 +4900,12 @@ S_reghopmaybe3(pTHX_ U8* s, I32 off, U8* lim)
 static void
 restore_pos(pTHX_ void *arg)
 {
-    (void)arg; /* unused */
+    PERL_UNUSED_ARG(arg);
     if (PL_reg_eval_set) {
        if (PL_reg_oldsaved) {
            PL_reg_re->subbeg = PL_reg_oldsaved;
            PL_reg_re->sublen = PL_reg_oldsavedlen;
-#ifdef PERL_COPY_ON_WRITE
+#ifdef PERL_OLD_COPY_ON_WRITE
            PL_reg_re->saved_copy = PL_nrs;
 #endif
            RX_MATCH_COPIED_on(PL_reg_re);
@@ -4917,8 +4919,8 @@ restore_pos(pTHX_ void *arg)
 STATIC void
 S_to_utf8_substr(pTHX_ register regexp *prog)
 {
-    SV* sv;
     if (prog->float_substr && !prog->float_utf8) {
+       SV* sv;
        prog->float_utf8 = sv = newSVsv(prog->float_substr);
        sv_utf8_upgrade(sv);
        if (SvTAIL(prog->float_substr))
@@ -4927,6 +4929,7 @@ S_to_utf8_substr(pTHX_ register regexp *prog)
            prog->check_utf8 = sv;
     }
     if (prog->anchored_substr && !prog->anchored_utf8) {
+       SV* sv;
        prog->anchored_utf8 = sv = newSVsv(prog->anchored_substr);
        sv_utf8_upgrade(sv);
        if (SvTAIL(prog->anchored_substr))
@@ -4939,8 +4942,8 @@ S_to_utf8_substr(pTHX_ register regexp *prog)
 STATIC void
 S_to_byte_substr(pTHX_ register regexp *prog)
 {
-    SV* sv;
     if (prog->float_utf8 && !prog->float_substr) {
+       SV* sv;
        prog->float_substr = sv = newSVsv(prog->float_utf8);
        if (sv_utf8_downgrade(sv, TRUE)) {
            if (SvTAIL(prog->float_utf8))
@@ -4953,6 +4956,7 @@ S_to_byte_substr(pTHX_ register regexp *prog)
            prog->check_substr = sv;
     }
     if (prog->anchored_utf8 && !prog->anchored_substr) {
+       SV* sv;
        prog->anchored_substr = sv = newSVsv(prog->anchored_utf8);
        if (sv_utf8_downgrade(sv, TRUE)) {
            if (SvTAIL(prog->anchored_utf8))