Make Unicode data structures global

These structures are read-only, use const C strings, and are truly
global, so no need to have them be interpreter level.  This saves
duplicating and freeing them as threads come and go.

In doing this, I noticed that not every one was properly being
copied/deallocated, so this fixes some potential unreported bugs, and
leaks.

6 files changed, 100 insertions, 88 deletions

diff --git a/embedvar.h b/embedvar.h
index 73c20e7f1b..a3f7fb3dd3 100644
--- a/embedvar.h
+++ b/embedvar.h
@@ -40,8 +40,6 @@
 #    define vTHX	PERL_GET_INTERP
 #  endif
 
-#define PL_AboveLatin1		(vTHX->IAboveLatin1)
-#define PL_Assigned_invlist	(vTHX->IAssigned_invlist)
 #define PL_DBcontrol		(vTHX->IDBcontrol)
 #define PL_DBcv			(vTHX->IDBcv)
 #define PL_DBgv			(vTHX->IDBgv)
@@ -52,25 +50,15 @@
 #define PL_DBtrace		(vTHX->IDBtrace)
 #define PL_Dir			(vTHX->IDir)
 #define PL_Env			(vTHX->IEnv)
-#define PL_GCB_invlist		(vTHX->IGCB_invlist)
-#define PL_HasMultiCharFold	(vTHX->IHasMultiCharFold)
 #define PL_InBitmap		(vTHX->IInBitmap)
-#define PL_LB_invlist		(vTHX->ILB_invlist)
 #define PL_LIO			(vTHX->ILIO)
-#define PL_Latin1		(vTHX->ILatin1)
 #define PL_Mem			(vTHX->IMem)
 #define PL_MemParse		(vTHX->IMemParse)
 #define PL_MemShared		(vTHX->IMemShared)
-#define PL_NonL1NonFinalFold	(vTHX->INonL1NonFinalFold)
 #define PL_Proc			(vTHX->IProc)
-#define PL_SB_invlist		(vTHX->ISB_invlist)
-#define PL_SCX_invlist		(vTHX->ISCX_invlist)
 #define PL_Sock			(vTHX->ISock)
 #define PL_StdIO		(vTHX->IStdIO)
 #define PL_Sv			(vTHX->ISv)
-#define PL_UpperLatin1		(vTHX->IUpperLatin1)
-#define PL_WB_invlist		(vTHX->IWB_invlist)
-#define PL_XPosix_ptrs		(vTHX->IXPosix_ptrs)
 #define PL_Xpv			(vTHX->IXpv)
 #define PL_an			(vTHX->Ian)
 #define PL_argvgv		(vTHX->Iargvgv)
@@ -350,20 +338,13 @@
 #define PL_unsafe		(vTHX->Iunsafe)
 #define PL_utf8_charname_begin	(vTHX->Iutf8_charname_begin)
 #define PL_utf8_charname_continue	(vTHX->Iutf8_charname_continue)
-#define PL_utf8_foldable	(vTHX->Iutf8_foldable)
 #define PL_utf8_foldclosures	(vTHX->Iutf8_foldclosures)
-#define PL_utf8_idcont		(vTHX->Iutf8_idcont)
-#define PL_utf8_idstart		(vTHX->Iutf8_idstart)
 #define PL_utf8_mark		(vTHX->Iutf8_mark)
-#define PL_utf8_perl_idcont	(vTHX->Iutf8_perl_idcont)
-#define PL_utf8_perl_idstart	(vTHX->Iutf8_perl_idstart)
 #define PL_utf8_swash_ptrs	(vTHX->Iutf8_swash_ptrs)
 #define PL_utf8_tofold		(vTHX->Iutf8_tofold)
 #define PL_utf8_tolower		(vTHX->Iutf8_tolower)
 #define PL_utf8_totitle		(vTHX->Iutf8_totitle)
 #define PL_utf8_toupper		(vTHX->Iutf8_toupper)
-#define PL_utf8_xidcont		(vTHX->Iutf8_xidcont)
-#define PL_utf8_xidstart	(vTHX->Iutf8_xidstart)
 #define PL_utf8cache		(vTHX->Iutf8cache)
 #define PL_utf8locale		(vTHX->Iutf8locale)
 #define PL_warn_locale		(vTHX->Iwarn_locale)
@@ -376,8 +357,32 @@
 
 #if defined(PERL_GLOBAL_STRUCT)
 
+#define PL_AboveLatin1		(my_vars->GAboveLatin1)
+#define PL_GAboveLatin1		(my_vars->GAboveLatin1)
+#define PL_Assigned_invlist	(my_vars->GAssigned_invlist)
+#define PL_GAssigned_invlist	(my_vars->GAssigned_invlist)
 #define PL_C_locale_obj		(my_vars->GC_locale_obj)
 #define PL_GC_locale_obj	(my_vars->GC_locale_obj)
+#define PL_GCB_invlist		(my_vars->GGCB_invlist)
+#define PL_GGCB_invlist		(my_vars->GGCB_invlist)
+#define PL_HasMultiCharFold	(my_vars->GHasMultiCharFold)
+#define PL_GHasMultiCharFold	(my_vars->GHasMultiCharFold)
+#define PL_LB_invlist		(my_vars->GLB_invlist)
+#define PL_GLB_invlist		(my_vars->GLB_invlist)
+#define PL_Latin1		(my_vars->GLatin1)
+#define PL_GLatin1		(my_vars->GLatin1)
+#define PL_NonL1NonFinalFold	(my_vars->GNonL1NonFinalFold)
+#define PL_GNonL1NonFinalFold	(my_vars->GNonL1NonFinalFold)
+#define PL_SB_invlist		(my_vars->GSB_invlist)
+#define PL_GSB_invlist		(my_vars->GSB_invlist)
+#define PL_SCX_invlist		(my_vars->GSCX_invlist)
+#define PL_GSCX_invlist		(my_vars->GSCX_invlist)
+#define PL_UpperLatin1		(my_vars->GUpperLatin1)
+#define PL_GUpperLatin1		(my_vars->GUpperLatin1)
+#define PL_WB_invlist		(my_vars->GWB_invlist)
+#define PL_GWB_invlist		(my_vars->GWB_invlist)
+#define PL_XPosix_ptrs		(my_vars->GXPosix_ptrs)
+#define PL_GXPosix_ptrs		(my_vars->GXPosix_ptrs)
 #define PL_appctx		(my_vars->Gappctx)
 #define PL_Gappctx		(my_vars->Gappctx)
 #define PL_check		(my_vars->Gcheck)
@@ -462,6 +467,20 @@
 #define PL_Gtimesbase		(my_vars->Gtimesbase)
 #define PL_use_safe_putenv	(my_vars->Guse_safe_putenv)
 #define PL_Guse_safe_putenv	(my_vars->Guse_safe_putenv)
+#define PL_utf8_foldable	(my_vars->Gutf8_foldable)
+#define PL_Gutf8_foldable	(my_vars->Gutf8_foldable)
+#define PL_utf8_idcont		(my_vars->Gutf8_idcont)
+#define PL_Gutf8_idcont		(my_vars->Gutf8_idcont)
+#define PL_utf8_idstart		(my_vars->Gutf8_idstart)
+#define PL_Gutf8_idstart	(my_vars->Gutf8_idstart)
+#define PL_utf8_perl_idcont	(my_vars->Gutf8_perl_idcont)
+#define PL_Gutf8_perl_idcont	(my_vars->Gutf8_perl_idcont)
+#define PL_utf8_perl_idstart	(my_vars->Gutf8_perl_idstart)
+#define PL_Gutf8_perl_idstart	(my_vars->Gutf8_perl_idstart)
+#define PL_utf8_xidcont		(my_vars->Gutf8_xidcont)
+#define PL_Gutf8_xidcont	(my_vars->Gutf8_xidcont)
+#define PL_utf8_xidstart	(my_vars->Gutf8_xidstart)
+#define PL_Gutf8_xidstart	(my_vars->Gutf8_xidstart)
 #define PL_veto_cleanup		(my_vars->Gveto_cleanup)
 #define PL_Gveto_cleanup	(my_vars->Gveto_cleanup)
 #define PL_watch_pvx		(my_vars->Gwatch_pvx)
diff --git a/intrpvar.h b/intrpvar.h
index 466785b882..33d952682a 100644
--- a/intrpvar.h
+++ b/intrpvar.h
@@ -639,14 +639,8 @@ PERLVARI(I, underlying_numeric_obj, locale_t, NULL)
 #endif /* !USE_LOCALE_NUMERIC */
 
 /* Unicode inversion lists */
-PERLVAR(I, Latin1,	SV *)
-PERLVAR(I, UpperLatin1,	SV *)   /* Code points 128 - 255 */
-PERLVAR(I, AboveLatin1,	SV *)
 PERLVAR(I, InBitmap,	SV *)
 
-PERLVAR(I, NonL1NonFinalFold,   SV *)
-PERLVAR(I, HasMultiCharFold,   SV *)
-
 /* utf8 character class swashes */
 PERLVAR(I, utf8_mark,	SV *)
 PERLVAR(I, utf8_toupper, SV *)
@@ -657,13 +651,6 @@ PERLVAR(I, utf8_charname_begin, SV *)
 PERLVAR(I, utf8_charname_continue, SV *)
 
 PERLVARA(I, utf8_swash_ptrs, POSIX_SWASH_COUNT, SV *)
-PERLVARA(I, XPosix_ptrs, POSIX_CC_COUNT, SV *)
-PERLVAR(I, GCB_invlist, SV *)
-PERLVAR(I, LB_invlist, SV *)
-PERLVAR(I, SB_invlist, SV *)
-PERLVAR(I, SCX_invlist, SV *)
-PERLVAR(I, WB_invlist, SV *)
-PERLVAR(I, Assigned_invlist, SV *)
 PERLVAR(I, seen_deprecated_macro, HV *)
 
 PERLVAR(I, last_swash_hv, HV *)
@@ -733,13 +720,6 @@ PERLVARI(I, known_layers, PerlIO_list_t *, NULL)
 PERLVARI(I, def_layerlist, PerlIO_list_t *, NULL)
 #endif
 
-PERLVAR(I, utf8_idstart, SV *)
-PERLVAR(I, utf8_idcont,	SV *)
-PERLVAR(I, utf8_xidstart, SV *)
-PERLVAR(I, utf8_perl_idstart, SV *)
-PERLVAR(I, utf8_perl_idcont, SV *)
-PERLVAR(I, utf8_xidcont, SV *)
-
 PERLVAR(I, sort_RealCmp, SVCOMPARE_t)
 
 PERLVARI(I, checkav_save, AV *, NULL)	/* save CHECK{}s when compiling */
@@ -783,10 +763,6 @@ PERLVAR(I, blockhooks,	AV *)
  * matching */
 PERLVARI(I, utf8_foldclosures, HV *, NULL)
 
-/* List of characters that participate in folds (except marks, etc in
- * multi-char folds) */
-PERLVARI(I, utf8_foldable, SV *, NULL)
-
 PERLVAR(I, custom_ops,	HV *)		/* custom op registrations */
 
 PERLVAR(I, Xpv,		XPV *)		/* (unused) held temporary value */
diff --git a/perl.c b/perl.c
index f894780a0b..181c597fd2 100644
--- a/perl.c
+++ b/perl.c
@@ -1190,16 +1190,8 @@ perl_destruct(pTHXx)
     SvREFCNT_dec(PL_utf8_totitle);
     SvREFCNT_dec(PL_utf8_tolower);
     SvREFCNT_dec(PL_utf8_tofold);
-    SvREFCNT_dec(PL_utf8_idstart);
-    SvREFCNT_dec(PL_utf8_idcont);
-    SvREFCNT_dec(PL_utf8_foldable);
     SvREFCNT_dec(PL_utf8_foldclosures);
-    SvREFCNT_dec(PL_AboveLatin1);
     SvREFCNT_dec(PL_InBitmap);
-    SvREFCNT_dec(PL_UpperLatin1);
-    SvREFCNT_dec(PL_Latin1);
-    SvREFCNT_dec(PL_NonL1NonFinalFold);
-    SvREFCNT_dec(PL_HasMultiCharFold);
 #ifdef USE_LOCALE_CTYPE
     SvREFCNT_dec(PL_warn_locale);
 #endif
@@ -1208,28 +1200,11 @@ perl_destruct(pTHXx)
     PL_utf8_totitle	= NULL;
     PL_utf8_tolower	= NULL;
     PL_utf8_tofold	= NULL;
-    PL_utf8_idstart	= NULL;
-    PL_utf8_idcont	= NULL;
     PL_utf8_foldclosures = NULL;
-    PL_AboveLatin1       = NULL;
     PL_InBitmap          = NULL;
-    PL_HasMultiCharFold  = NULL;
 #ifdef USE_LOCALE_CTYPE
     PL_warn_locale       = NULL;
 #endif
-    PL_Latin1            = NULL;
-    PL_NonL1NonFinalFold = NULL;
-    PL_UpperLatin1       = NULL;
-    for (i = 0; i < POSIX_CC_COUNT; i++) {
-        SvREFCNT_dec(PL_XPosix_ptrs[i]);
-        PL_XPosix_ptrs[i] = NULL;
-    }
-    PL_GCB_invlist = NULL;
-    PL_LB_invlist = NULL;
-    PL_SB_invlist = NULL;
-    PL_SCX_invlist = NULL;
-    PL_WB_invlist = NULL;
-    PL_Assigned_invlist = NULL;
 
     if (!specialWARN(PL_compiling.cop_warnings))
 	PerlMemShared_free(PL_compiling.cop_warnings);
diff --git a/perlapi.h b/perlapi.h
index b39c8ccd04..f882f6b840 100644
--- a/perlapi.h
+++ b/perlapi.h
@@ -99,8 +99,32 @@ END_EXTERN_C
 
 #else	/* !PERL_CORE */
 
+#undef  PL_AboveLatin1
+#define PL_AboveLatin1		(*Perl_GAboveLatin1_ptr(NULL))
+#undef  PL_Assigned_invlist
+#define PL_Assigned_invlist	(*Perl_GAssigned_invlist_ptr(NULL))
 #undef  PL_C_locale_obj
 #define PL_C_locale_obj		(*Perl_GC_locale_obj_ptr(NULL))
+#undef  PL_GCB_invlist
+#define PL_GCB_invlist		(*Perl_GGCB_invlist_ptr(NULL))
+#undef  PL_HasMultiCharFold
+#define PL_HasMultiCharFold	(*Perl_GHasMultiCharFold_ptr(NULL))
+#undef  PL_LB_invlist
+#define PL_LB_invlist		(*Perl_GLB_invlist_ptr(NULL))
+#undef  PL_Latin1
+#define PL_Latin1		(*Perl_GLatin1_ptr(NULL))
+#undef  PL_NonL1NonFinalFold
+#define PL_NonL1NonFinalFold	(*Perl_GNonL1NonFinalFold_ptr(NULL))
+#undef  PL_SB_invlist
+#define PL_SB_invlist		(*Perl_GSB_invlist_ptr(NULL))
+#undef  PL_SCX_invlist
+#define PL_SCX_invlist		(*Perl_GSCX_invlist_ptr(NULL))
+#undef  PL_UpperLatin1
+#define PL_UpperLatin1		(*Perl_GUpperLatin1_ptr(NULL))
+#undef  PL_WB_invlist
+#define PL_WB_invlist		(*Perl_GWB_invlist_ptr(NULL))
+#undef  PL_XPosix_ptrs
+#define PL_XPosix_ptrs		(*Perl_GXPosix_ptrs_ptr(NULL))
 #undef  PL_appctx
 #define PL_appctx		(*Perl_Gappctx_ptr(NULL))
 #undef  PL_check
@@ -181,6 +205,20 @@ END_EXTERN_C
 #define PL_timesbase		(*Perl_Gtimesbase_ptr(NULL))
 #undef  PL_use_safe_putenv
 #define PL_use_safe_putenv	(*Perl_Guse_safe_putenv_ptr(NULL))
+#undef  PL_utf8_foldable
+#define PL_utf8_foldable	(*Perl_Gutf8_foldable_ptr(NULL))
+#undef  PL_utf8_idcont
+#define PL_utf8_idcont		(*Perl_Gutf8_idcont_ptr(NULL))
+#undef  PL_utf8_idstart
+#define PL_utf8_idstart		(*Perl_Gutf8_idstart_ptr(NULL))
+#undef  PL_utf8_perl_idcont
+#define PL_utf8_perl_idcont	(*Perl_Gutf8_perl_idcont_ptr(NULL))
+#undef  PL_utf8_perl_idstart
+#define PL_utf8_perl_idstart	(*Perl_Gutf8_perl_idstart_ptr(NULL))
+#undef  PL_utf8_xidcont
+#define PL_utf8_xidcont		(*Perl_Gutf8_xidcont_ptr(NULL))
+#undef  PL_utf8_xidstart
+#define PL_utf8_xidstart	(*Perl_Gutf8_xidstart_ptr(NULL))
 #undef  PL_veto_cleanup
 #define PL_veto_cleanup		(*Perl_Gveto_cleanup_ptr(NULL))
 #undef  PL_watch_pvx
diff --git a/perlvars.h b/perlvars.h
index ac97ebcd99..e3ded3c7ef 100644
--- a/perlvars.h
+++ b/perlvars.h
@@ -271,3 +271,27 @@ PERLVARA(G, hash_chars, (1+256) * sizeof(U32), unsigned char) /* perl.c and hv.h
 #ifdef __VMS
 PERLVAR(G, perllib_sep, char)
 #endif
+
+PERLVAR(G, AboveLatin1,	SV *)
+PERLVAR(G, Assigned_invlist, SV *)
+PERLVAR(G, GCB_invlist, SV *)
+PERLVAR(G, HasMultiCharFold,   SV *)
+PERLVAR(G, Latin1,	SV *)
+PERLVAR(G, LB_invlist, SV *)
+PERLVAR(G, NonL1NonFinalFold,   SV *)
+PERLVAR(G, SB_invlist, SV *)
+PERLVAR(G, SCX_invlist, SV *)
+PERLVAR(G, UpperLatin1,	SV *)   /* Code points 128 - 255 */
+
+/* List of characters that participate in folds (except marks, etc in
+ * multi-char folds) */
+PERLVARI(G, utf8_foldable, SV *, NULL)
+
+PERLVAR(G, utf8_idcont,	SV *)
+PERLVAR(G, utf8_idstart, SV *)
+PERLVAR(G, utf8_perl_idcont, SV *)
+PERLVAR(G, utf8_perl_idstart, SV *)
+PERLVAR(G, utf8_xidcont, SV *)
+PERLVAR(G, utf8_xidstart, SV *)
+PERLVAR(G, WB_invlist, SV *)
+PERLVARA(G, XPosix_ptrs, POSIX_CC_COUNT, SV *)
diff --git a/sv.c b/sv.c
index 231215e766..1856ba37c1 100644
--- a/sv.c
+++ b/sv.c
@@ -15582,38 +15582,18 @@ perl_clone_using(PerlInterpreter *proto_perl, UV flags,
     PL_setlocale_bufsize = 0;
 
     /* Unicode inversion lists */
-    PL_Latin1		= sv_dup_inc(proto_perl->ILatin1, param);
-    PL_UpperLatin1	= sv_dup_inc(proto_perl->IUpperLatin1, param);
-    PL_AboveLatin1	= sv_dup_inc(proto_perl->IAboveLatin1, param);
     PL_InBitmap         = sv_dup_inc(proto_perl->IInBitmap, param);
 
-    PL_NonL1NonFinalFold = sv_dup_inc(proto_perl->INonL1NonFinalFold, param);
-    PL_HasMultiCharFold = sv_dup_inc(proto_perl->IHasMultiCharFold, param);
-
     /* utf8 character class swashes */
     for (i = 0; i < POSIX_SWASH_COUNT; i++) {
         PL_utf8_swash_ptrs[i] = sv_dup_inc(proto_perl->Iutf8_swash_ptrs[i], param);
     }
-    for (i = 0; i < POSIX_CC_COUNT; i++) {
-        PL_XPosix_ptrs[i] = sv_dup_inc(proto_perl->IXPosix_ptrs[i], param);
-    }
-    PL_GCB_invlist = sv_dup_inc(proto_perl->IGCB_invlist, param);
-    PL_SB_invlist = sv_dup_inc(proto_perl->ISB_invlist, param);
-    PL_SCX_invlist = sv_dup_inc(proto_perl->ISCX_invlist, param);
-    PL_WB_invlist = sv_dup_inc(proto_perl->IWB_invlist, param);
     PL_seen_deprecated_macro = hv_dup_inc(proto_perl->Iseen_deprecated_macro, param);
     PL_utf8_mark	= sv_dup_inc(proto_perl->Iutf8_mark, param);
     PL_utf8_toupper	= sv_dup_inc(proto_perl->Iutf8_toupper, param);
     PL_utf8_totitle	= sv_dup_inc(proto_perl->Iutf8_totitle, param);
     PL_utf8_tolower	= sv_dup_inc(proto_perl->Iutf8_tolower, param);
     PL_utf8_tofold	= sv_dup_inc(proto_perl->Iutf8_tofold, param);
-    PL_utf8_idstart	= sv_dup_inc(proto_perl->Iutf8_idstart, param);
-    PL_utf8_xidstart	= sv_dup_inc(proto_perl->Iutf8_xidstart, param);
-    PL_utf8_perl_idstart = sv_dup_inc(proto_perl->Iutf8_perl_idstart, param);
-    PL_utf8_perl_idcont = sv_dup_inc(proto_perl->Iutf8_perl_idcont, param);
-    PL_utf8_idcont	= sv_dup_inc(proto_perl->Iutf8_idcont, param);
-    PL_utf8_xidcont	= sv_dup_inc(proto_perl->Iutf8_xidcont, param);
-    PL_utf8_foldable	= sv_dup_inc(proto_perl->Iutf8_foldable, param);
     PL_utf8_charname_begin = sv_dup_inc(proto_perl->Iutf8_charname_begin, param);
     PL_utf8_charname_continue = sv_dup_inc(proto_perl->Iutf8_charname_continue, param);