Use new Svt_INVLIST for inversion lists.

This converts inversion lists to use their own scalar type.

7 files changed, 33 insertions, 22 deletions

diff --git a/embed.fnc b/embed.fnc
index 5441ff54a4..997f42dd1e 100644
--- a/embed.fnc
+++ b/embed.fnc
@@ -1428,7 +1428,7 @@ EsM	|void	|_append_range_to_invlist   |NN SV* const invlist|const UV start|const
 EiMR	|UV*	|_invlist_array_init	|NN SV* const invlist|const bool will_have_0
 EiMR	|UV*	|invlist_array	|NN SV* const invlist
 EsM	|void	|invlist_extend    |NN SV* const invlist|const UV len
-EiMR	|U8*	|get_invlist_offset_addr|NN SV* invlist
+EiMR	|bool*	|get_invlist_offset_addr|NN SV* invlist
 EiMR	|UV	|invlist_max	|NN SV* const invlist
 EiM	|void	|invlist_set_len|NN SV* const invlist|const UV len
 EiMR	|IV*	|get_invlist_previous_index_addr|NN SV* invlist
diff --git a/inline_invlist.c b/inline_invlist.c
index ced42d85c0..a29964524b 100644
--- a/inline_invlist.c
+++ b/inline_invlist.c
@@ -21,7 +21,7 @@ S__get_invlist_len_addr(pTHX_ SV* invlist)
 
     PERL_ARGS_ASSERT__GET_INVLIST_LEN_ADDR;
 
-    return &(LvTARGLEN(invlist));
+    return &(((XINVLIST*) SvANY(invlist))->count);
 }
 
 PERL_STATIC_INLINE UV
diff --git a/perl.h b/perl.h
index 16e339cf94..4bfc8fa21f 100644
--- a/perl.h
+++ b/perl.h
@@ -2278,6 +2278,7 @@ typedef struct xpvuv XPVUV;
 typedef struct xpvnv XPVNV;
 typedef struct xpvmg XPVMG;
 typedef struct xpvlv XPVLV;
+typedef struct xpvinvlist XINVLIST;
 typedef struct xpvav XPVAV;
 typedef struct xpvhv XPVHV;
 typedef struct xpvgv XPVGV;
@@ -5049,7 +5050,7 @@ PL_valid_types_IVX[]    = { 0, 1, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0 };
 EXTCONST bool
 PL_valid_types_NVX[]    = { 0, 0, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0 };
 EXTCONST bool
-PL_valid_types_PVX[]    = { 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1 };
+PL_valid_types_PVX[]    = { 0, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1 };
 EXTCONST bool
 PL_valid_types_RV[]     = { 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1 };
 EXTCONST bool
diff --git a/proto.h b/proto.h
index e782e87da2..865d57747f 100644
--- a/proto.h
+++ b/proto.h
@@ -6486,7 +6486,7 @@ PERL_STATIC_INLINE STRLEN*	S_get_invlist_iter_addr(pTHX_ SV* invlist)
 #define PERL_ARGS_ASSERT_GET_INVLIST_ITER_ADDR	\
 	assert(invlist)
 
-PERL_STATIC_INLINE U8*	S_get_invlist_offset_addr(pTHX_ SV* invlist)
+PERL_STATIC_INLINE bool*	S_get_invlist_offset_addr(pTHX_ SV* invlist)
 			__attribute__warn_unused_result__
 			__attribute__nonnull__(pTHX_1);
 #define PERL_ARGS_ASSERT_GET_INVLIST_OFFSET_ADDR	\
diff --git a/regcomp.c b/regcomp.c
index 160ef717c1..db6f2d445a 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -7025,9 +7025,8 @@ S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
 
 /* This section of code defines the inversion list object and its methods.  The
  * interfaces are highly subject to change, so as much as possible is static to
- * this file.  An inversion list is here implemented as a malloc'd C UV array.
- * Currently it is a SVt_PVLV, with some of the header fields from that
- * repurposed for uses here.
+ * this file.  An inversion list is here implemented as a malloc'd C UV array
+ * as an SVt_INVLIST scalar.
  *
  * An inversion list for Unicode is an array of code points, sorted by ordinal
  * number.  The zeroth element is the first code point in the list.  The 1th
@@ -7087,7 +7086,7 @@ S__invlist_array_init(pTHX_ SV* const invlist, const bool will_have_0)
      * element is either the element reserved for 0, if TRUE, or the element
      * after it, if FALSE */
 
-    U8* offset = get_invlist_offset_addr(invlist);
+    bool* offset = get_invlist_offset_addr(invlist);
     UV* zero_addr = (UV *) SvPVX(invlist);
 
     PERL_ARGS_ASSERT__INVLIST_ARRAY_INIT;
@@ -7143,12 +7142,12 @@ S_invlist_set_len(pTHX_ SV* const invlist, const UV len)
 PERL_STATIC_INLINE IV*
 S_get_invlist_previous_index_addr(pTHX_ SV* invlist)
 {
-    /* Return the address of the UV that is reserved to hold the cached index
+    /* Return the address of the IV that is reserved to hold the cached index
      * */
 
     PERL_ARGS_ASSERT_GET_INVLIST_PREVIOUS_INDEX_ADDR;
 
-    return &(((XPVLV*) SvANY(invlist))->xiv_u.xivu_iv);
+    return &(((XINVLIST*) SvANY(invlist))->prev_index);
 }
 
 PERL_STATIC_INLINE IV
@@ -7186,7 +7185,7 @@ S_invlist_max(pTHX_ SV* const invlist)
            : FROM_INTERNAL_SIZE(SvLEN(invlist));
 }
 
-PERL_STATIC_INLINE U8*
+PERL_STATIC_INLINE bool*
 S_get_invlist_offset_addr(pTHX_ SV* invlist)
 {
     /* Return the address of the field that says whether the inversion list is
@@ -7194,7 +7193,7 @@ S_get_invlist_offset_addr(pTHX_ SV* invlist)
 
     PERL_ARGS_ASSERT_GET_INVLIST_OFFSET_ADDR;
 
-    return (U8*) &(LvFLAGS(invlist));
+    return &(((XINVLIST*) SvANY(invlist))->is_offset);
 }
 
 #ifndef PERL_IN_XSUB_RE
@@ -7207,14 +7206,14 @@ Perl__new_invlist(pTHX_ IV initial_size)
      * system default is used instead */
 
     SV* new_list;
-    U8* offset_addr;
+    bool* offset_addr;
 
     if (initial_size < 0) {
 	initial_size = 10;
     }
 
     /* Allocate the initial space */
-    new_list = newSV_type(SVt_PVLV);
+    new_list = newSV_type(SVt_INVLIST);
     SvGROW(new_list, TO_INTERNAL_SIZE(initial_size) + 1); /* 1 is for trailing
                                                              NUL */
     invlist_set_len(new_list, 0);
@@ -7258,7 +7257,7 @@ S__new_invlist_C_array(pTHX_ const UV* const list)
                                         inversion list of the correct vintage.
                                        */
 
-    SV* invlist = newSV_type(SVt_PVLV);
+    SV* invlist = newSV_type(SVt_INVLIST);
 
     PERL_ARGS_ASSERT__NEW_INVLIST_C_ARRAY;
 
@@ -8167,7 +8166,7 @@ S_get_invlist_iter_addr(pTHX_ SV* invlist)
 
     PERL_ARGS_ASSERT_GET_INVLIST_ITER_ADDR;
 
-    return &(LvTARGOFF(invlist));
+    return &(((XINVLIST*) SvANY(invlist))->iterator);
 }
 
 PERL_STATIC_INLINE void
diff --git a/sv.c b/sv.c
index a3b0bb4a1e..8d23a3a678 100644
--- a/sv.c
+++ b/sv.c
@@ -898,10 +898,11 @@ static const struct body_details bodies_by_type[] = {
       SVt_PV, FALSE, NONV, HASARENA,
       FIT_ARENA(0, sizeof(XPV) - STRUCT_OFFSET(XPV, xpv_cur)) },
 
-    /* The invlist placeholder pretends to be an RV for now.
-       Also it's marked as "can't upgrade" to stop anyone using it before it's
-       implemented.  */
-    { 0, 0, 0, SVt_INVLIST, TRUE, NONV, NOARENA, 0 },
+    { sizeof(XINVLIST) - STRUCT_OFFSET(XPV, xpv_cur),
+      copy_length(XINVLIST, is_offset) - STRUCT_OFFSET(XPV, xpv_cur),
+      + STRUCT_OFFSET(XPV, xpv_cur),
+      SVt_INVLIST, TRUE, NONV, HASARENA,
+      FIT_ARENA(0, sizeof(XINVLIST) - STRUCT_OFFSET(XPV, xpv_cur)) },
 
     { sizeof(XPVIV) - STRUCT_OFFSET(XPV, xpv_cur),
       copy_length(XPVIV, xiv_u) - STRUCT_OFFSET(XPV, xpv_cur),
@@ -1340,6 +1341,7 @@ Perl_sv_upgrade(pTHX_ SV *const sv, svtype new_type)
     case SVt_PVGV:
     case SVt_PVCV:
     case SVt_PVLV:
+    case SVt_INVLIST:
     case SVt_REGEXP:
     case SVt_PVMG:
     case SVt_PVNV:
@@ -4117,7 +4119,7 @@ Perl_sv_setsv_flags(pTHX_ SV *dstr, SV* sstr, const I32 flags)
 	}
 	break;
 
-	/* case SVt_INVLIST: */
+	case SVt_INVLIST:
     case SVt_PVLV:
     case SVt_PVGV:
     case SVt_PVMG:
@@ -6304,6 +6306,7 @@ Perl_sv_clear(pTHX_ SV *const orig_sv)
 	case SVt_PVMG:
 	case SVt_PVNV:
 	case SVt_PVIV:
+	case SVt_INVLIST:
 	case SVt_PV:
 	  freescalar:
 	    /* Don't bother with SvOOK_off(sv); as we're only going to
@@ -12177,7 +12180,6 @@ S_sv_dup_common(pTHX_ const SV *const sstr, CLONE_PARAMS *const param)
 	SvANY(dstr)	= new_XNV();
 	SvNV_set(dstr, SvNVX(sstr));
 	break;
-	/* case SVt_INVLIST: */
     default:
 	{
 	    /* These are all the types that need complex bodies allocating.  */
@@ -12202,6 +12204,7 @@ S_sv_dup_common(pTHX_ const SV *const sstr, CLONE_PARAMS *const param)
 	    case SVt_PVMG:
 	    case SVt_PVNV:
 	    case SVt_PVIV:
+            case SVt_INVLIST:
 	    case SVt_PV:
 		assert(sv_type_details->body_size);
 		if (sv_type_details->arena) {
diff --git a/sv.h b/sv.h
index 475da61b7b..c2be3ba81d 100644
--- a/sv.h
+++ b/sv.h
@@ -528,6 +528,14 @@ struct xpvlv {
     char	xlv_flags;	/* 1 = negative offset  2 = negative len */
 };
 
+struct xpvinvlist {
+    _XPV_HEAD;
+    IV          prev_index;
+    STRLEN	iterator;
+    STRLEN	count;
+    bool	is_offset;	/* */
+};
+
 /* This structure works in 3 ways - regular scalar, GV with GP, or fast
    Boyer-Moore.  */
 struct xpvgv {