From 7c121692309d0a4d1183ace406b0c4a2848e6da6 Mon Sep 17 00:00:00 2001
From: Jean Boussier <jean.boussier@gmail.com>
Date: Thu, 30 May 2024 18:12:29 +0200
Subject: [PATCH] Eliminate internal uses of `Data_Wrap_Struct`

Ref: https://github.com/ruby/ruby/pull/10872

These should be the last internal uses of the old `Data` API
inside Ruby itself. Some use remain in a couple default gems.
---
 array.c   | 16 +++++++----
 common.mk |  3 ++
 compile.c | 18 ++++++++----
 dir.c     | 11 +++-----
 marshal.c | 82 +++++++++++++++++++++++++++++++++++++------------------
 util.c    | 48 ++++++++++++++++++++++++--------
 6 files changed, 122 insertions(+), 56 deletions(-)

diff --git a/array.c b/array.c
index 8355a45da5..dab933776f 100644
--- a/array.c
+++ b/array.c
@@ -6549,6 +6549,14 @@ rb_ary_shuffle(rb_execution_context_t *ec, VALUE ary, VALUE randgen)
     return ary;
 }
 
+static const rb_data_type_t ary_sample_memo_type = {
+    .wrap_struct_name = "ary_sample_memo",
+    .function = {
+        .dfree = (RUBY_DATA_FUNC)st_free_table,
+    },
+    .flags = RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY
+};
+
 static VALUE
 ary_sample(rb_execution_context_t *ec, VALUE ary, VALUE randgen, VALUE nv, VALUE to_array)
 {
@@ -6630,11 +6638,9 @@ ary_sample(rb_execution_context_t *ec, VALUE ary, VALUE randgen, VALUE nv, VALUE
     }
     else if (n <= memo_threshold / 2) {
         long max_idx = 0;
-#undef RUBY_UNTYPED_DATA_WARNING
-#define RUBY_UNTYPED_DATA_WARNING 0
-        VALUE vmemo = Data_Wrap_Struct(0, 0, st_free_table, 0);
+        VALUE vmemo = TypedData_Wrap_Struct(0, &ary_sample_memo_type, 0);
         st_table *memo = st_init_numtable_with_size(n);
-        DATA_PTR(vmemo) = memo;
+        RTYPEDDATA_DATA(vmemo) = memo;
         result = rb_ary_new_capa(n);
         RARRAY_PTR_USE(result, ptr_result, {
             for (i=0; i<n; i++) {
@@ -6657,7 +6663,7 @@ ary_sample(rb_execution_context_t *ec, VALUE ary, VALUE randgen, VALUE nv, VALUE
                 }
             });
         });
-        DATA_PTR(vmemo) = 0;
+        RTYPEDDATA_DATA(vmemo) = 0;
         st_free_table(memo);
         RB_GC_GUARD(vmemo);
     }
diff --git a/common.mk b/common.mk
index 516481739e..bf820b228e 100644
--- a/common.mk
+++ b/common.mk
@@ -19127,8 +19127,11 @@ transcode.$(OBJEXT): {$(VPATH)}subst.h
 transcode.$(OBJEXT): {$(VPATH)}transcode.c
 transcode.$(OBJEXT): {$(VPATH)}transcode_data.h
 util.$(OBJEXT): $(hdrdir)/ruby/ruby.h
+util.$(OBJEXT): $(top_srcdir)/internal/array.h
 util.$(OBJEXT): $(top_srcdir)/internal/compilers.h
+util.$(OBJEXT): $(top_srcdir)/internal/imemo.h
 util.$(OBJEXT): $(top_srcdir)/internal/sanitizers.h
+util.$(OBJEXT): $(top_srcdir)/internal/static_assert.h
 util.$(OBJEXT): $(top_srcdir)/internal/util.h
 util.$(OBJEXT): $(top_srcdir)/internal/warnings.h
 util.$(OBJEXT): {$(VPATH)}assert.h
diff --git a/compile.c b/compile.c
index 2d52e316a3..a0bbcab54b 100644
--- a/compile.c
+++ b/compile.c
@@ -48,9 +48,6 @@
 #include "insns.inc"
 #include "insns_info.inc"
 
-#undef RUBY_UNTYPED_DATA_WARNING
-#define RUBY_UNTYPED_DATA_WARNING 0
-
 #define FIXNUM_INC(n, i) ((n)+(INT2FIX(i)&~FIXNUM_FLAG))
 #define FIXNUM_OR(n, i) ((n)|INT2FIX(i))
 
@@ -11421,7 +11418,7 @@ iseq_build_from_ary_body(rb_iseq_t *iseq, LINK_ANCHOR *const anchor,
 {
     /* TODO: body should be frozen */
     long i, len = RARRAY_LEN(body);
-    struct st_table *labels_table = DATA_PTR(labels_wrapper);
+    struct st_table *labels_table = RTYPEDDATA_DATA(labels_wrapper);
     int j;
     int line_no = 0, node_id = -1, insn_idx = 0;
     int ret = COMPILE_OK;
@@ -11599,7 +11596,7 @@ iseq_build_from_ary_body(rb_iseq_t *iseq, LINK_ANCHOR *const anchor,
             rb_raise(rb_eTypeError, "unexpected object for instruction");
         }
     }
-    DATA_PTR(labels_wrapper) = 0;
+    RTYPEDDATA_DATA(labels_wrapper) = 0;
     RB_GC_GUARD(labels_wrapper);
     validate_labels(iseq, labels_table);
     if (!ret) return ret;
@@ -11732,6 +11729,15 @@ rb_iseq_mark_and_pin_insn_storage(struct iseq_compile_data_storage *storage)
     }
 }
 
+static const rb_data_type_t labels_wrapper_type = {
+    .wrap_struct_name = "compiler/labels_wrapper",
+    .function = {
+        .dmark = (RUBY_DATA_FUNC)rb_mark_set,
+        .dfree = (RUBY_DATA_FUNC)st_free_table,
+    },
+    .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED,
+};
+
 void
 rb_iseq_build_from_ary(rb_iseq_t *iseq, VALUE misc, VALUE locals, VALUE params,
                          VALUE exception, VALUE body)
@@ -11741,7 +11747,7 @@ rb_iseq_build_from_ary(rb_iseq_t *iseq, VALUE misc, VALUE locals, VALUE params,
     unsigned int arg_size, local_size, stack_max;
     ID *tbl;
     struct st_table *labels_table = st_init_numtable();
-    VALUE labels_wrapper = Data_Wrap_Struct(0, rb_mark_set, st_free_table, labels_table);
+    VALUE labels_wrapper = TypedData_Wrap_Struct(0, &labels_wrapper_type, labels_table);
     VALUE arg_opt_labels = rb_hash_aref(params, SYM(opt));
     VALUE keywords = rb_hash_aref(params, SYM(keyword));
     VALUE sym_arg_rest = ID2SYM(rb_intern_const("#arg_rest"));
diff --git a/dir.c b/dir.c
index 84ef5ee6f5..0a22ad7901 100644
--- a/dir.c
+++ b/dir.c
@@ -113,6 +113,7 @@ char *strchr(char*,char);
 #include "internal/gc.h"
 #include "internal/io.h"
 #include "internal/object.h"
+#include "internal/imemo.h"
 #include "internal/vm.h"
 #include "ruby/encoding.h"
 #include "ruby/ruby.h"
@@ -1389,19 +1390,15 @@ rb_dir_getwd_ospath(void)
     VALUE cwd;
     VALUE path_guard;
 
-#undef RUBY_UNTYPED_DATA_WARNING
-#define RUBY_UNTYPED_DATA_WARNING 0
-    path_guard = Data_Wrap_Struct((VALUE)0, NULL, RUBY_DEFAULT_FREE, NULL);
+    path_guard = rb_imemo_tmpbuf_auto_free_pointer();
     path = ruby_getcwd();
-    DATA_PTR(path_guard) = path;
+    rb_imemo_tmpbuf_set_ptr(path_guard, path);
 #ifdef __APPLE__
     cwd = rb_str_normalize_ospath(path, strlen(path));
 #else
     cwd = rb_str_new2(path);
 #endif
-    DATA_PTR(path_guard) = 0;
-
-    xfree(path);
+    rb_free_tmp_buffer(&path_guard);
     return cwd;
 }
 #endif
diff --git a/marshal.c b/marshal.c
index 83134371fa..c0bc2dd6da 100644
--- a/marshal.c
+++ b/marshal.c
@@ -128,22 +128,6 @@ static VALUE compat_allocator_tbl_wrapper;
 static VALUE rb_marshal_dump_limited(VALUE obj, VALUE port, int limit);
 static VALUE rb_marshal_load_with_proc(VALUE port, VALUE proc, bool freeze);
 
-static int
-mark_marshal_compat_i(st_data_t key, st_data_t value, st_data_t _)
-{
-    marshal_compat_t *p = (marshal_compat_t *)value;
-    rb_gc_mark(p->newclass);
-    rb_gc_mark(p->oldclass);
-    return ST_CONTINUE;
-}
-
-static void
-mark_marshal_compat_t(void *tbl)
-{
-    if (!tbl) return;
-    st_foreach(tbl, mark_marshal_compat_i, 0);
-}
-
 static st_table *compat_allocator_table(void);
 
 void
@@ -156,11 +140,10 @@ rb_marshal_define_compat(VALUE newclass, VALUE oldclass, VALUE (*dumper)(VALUE),
         rb_raise(rb_eTypeError, "no allocator");
     }
 
+    compat_allocator_table();
     compat = ALLOC(marshal_compat_t);
-    compat->newclass = Qnil;
-    compat->oldclass = Qnil;
-    compat->newclass = newclass;
-    compat->oldclass = oldclass;
+    RB_OBJ_WRITE(compat_allocator_tbl_wrapper, &compat->newclass, newclass);
+    RB_OBJ_WRITE(compat_allocator_tbl_wrapper, &compat->oldclass, oldclass);
     compat->dumper = dumper;
     compat->loader = loader;
 
@@ -2525,28 +2508,75 @@ Init_marshal(void)
 }
 
 static int
-free_compat_i(st_data_t key, st_data_t value, st_data_t _)
+marshal_compat_table_mark_i(st_data_t key, st_data_t value, st_data_t _)
+{
+    marshal_compat_t *p = (marshal_compat_t *)value;
+    rb_gc_mark_movable(p->newclass);
+    rb_gc_mark_movable(p->oldclass);
+    return ST_CONTINUE;
+}
+
+static void
+marshal_compat_table_mark(void *tbl)
+{
+    if (!tbl) return;
+    st_foreach(tbl, marshal_compat_table_mark_i, 0);
+}
+
+static int
+marshal_compat_table_free_i(st_data_t key, st_data_t value, st_data_t _)
 {
     xfree((marshal_compat_t *)value);
     return ST_CONTINUE;
 }
 
 static void
-free_compat_allocator_table(void *data)
+marshal_compat_table_free(void *data)
 {
-    st_foreach(data, free_compat_i, 0);
+    st_foreach(data, marshal_compat_table_free_i, 0);
     st_free_table(data);
 }
 
+static size_t
+marshal_compat_table_memsize(const void *data)
+{
+    return st_memsize(data) + sizeof(marshal_compat_t) * st_table_size(data);
+}
+
+static int
+marshal_compat_table_compact_i(st_data_t key, st_data_t value, st_data_t _)
+{
+    marshal_compat_t *p = (marshal_compat_t *)value;
+    p->newclass = rb_gc_location(p->newclass);
+    p->oldclass = rb_gc_location(p->oldclass);
+    return ST_CONTINUE;
+}
+
+static void
+marshal_compat_table_compact(void *tbl)
+{
+    if (!tbl) return;
+    st_foreach(tbl, marshal_compat_table_compact_i, 0);
+}
+
+static const rb_data_type_t marshal_compat_type = {
+    .wrap_struct_name = "marshal_compat_table",
+    .function = {
+        .dmark = marshal_compat_table_mark,
+        .dfree = marshal_compat_table_free,
+        .dsize = marshal_compat_table_memsize,
+        .dcompact = marshal_compat_table_compact,
+    },
+    .flags = RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY,
+};
+
 static st_table *
 compat_allocator_table(void)
 {
     if (compat_allocator_tbl) return compat_allocator_tbl;
     compat_allocator_tbl = st_init_numtable();
-#undef RUBY_UNTYPED_DATA_WARNING
-#define RUBY_UNTYPED_DATA_WARNING 0
     compat_allocator_tbl_wrapper =
-        Data_Wrap_Struct(0, mark_marshal_compat_t, free_compat_allocator_table, compat_allocator_tbl);
+        TypedData_Wrap_Struct(0, &marshal_compat_type, compat_allocator_tbl);
     rb_vm_register_global_object(compat_allocator_tbl_wrapper);
     return compat_allocator_tbl;
 }
diff --git a/util.c b/util.c
index 3c08879ce5..8a18814937 100644
--- a/util.c
+++ b/util.c
@@ -31,6 +31,7 @@
 
 #include "internal.h"
 #include "internal/sanitizers.h"
+#include "internal/imemo.h"
 #include "internal/util.h"
 #include "ruby/util.h"
 #include "ruby_atomic.h"
@@ -543,41 +544,63 @@ ruby_strdup(const char *str)
     return tmp;
 }
 
+#if defined HAVE_GETCWD
+# if defined NO_GETCWD_MALLOC
+
 char *
 ruby_getcwd(void)
 {
-#if defined HAVE_GETCWD
-# undef RUBY_UNTYPED_DATA_WARNING
-# define RUBY_UNTYPED_DATA_WARNING 0
-# if defined NO_GETCWD_MALLOC
-    VALUE guard = Data_Wrap_Struct((VALUE)0, NULL, RUBY_DEFAULT_FREE, NULL);
+    VALUE guard = rb_imemo_tmpbuf_auto_free_pointer();
     int size = 200;
     char *buf = xmalloc(size);
 
     while (!getcwd(buf, size)) {
         int e = errno;
         if (e != ERANGE) {
-            xfree(buf);
-            DATA_PTR(guard) = NULL;
+            rb_free_tmp_buffer(&guard);
             rb_syserr_fail(e, "getcwd");
         }
         size *= 2;
-        DATA_PTR(guard) = buf;
+        rb_imemo_tmpbuf_set_ptr(guard, buf);
         buf = xrealloc(buf, size);
     }
+    rb_free_tmp_buffer(&guard);
+    return buf;
+}
+
 # else
-    VALUE guard = Data_Wrap_Struct((VALUE)0, NULL, free, NULL);
+
+static const rb_data_type_t getcwd_buffer_guard_type = {
+    .wrap_struct_name = "ruby_getcwd_guard",
+    .function = {
+        .dfree = free // not xfree.
+    },
+    .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED
+};
+
+char *
+ruby_getcwd(void)
+{
+    VALUE guard = TypedData_Wrap_Struct((VALUE)0, &getcwd_buffer_guard_type, NULL);
     char *buf, *cwd = getcwd(NULL, 0);
-    DATA_PTR(guard) = cwd;
+    RTYPEDDATA_DATA(guard) = cwd;
     if (!cwd) rb_sys_fail("getcwd");
     buf = ruby_strdup(cwd);	/* allocate by xmalloc */
     free(cwd);
+    RTYPEDDATA_DATA(RB_GC_GUARD(guard)) = NULL;
+    return buf;
+}
+
 # endif
-    DATA_PTR(RB_GC_GUARD(guard)) = NULL;
 #else
+
 # ifndef PATH_MAX
 #  define PATH_MAX 8192
 # endif
+
+char *
+ruby_getcwd(void)
+{
     char *buf = xmalloc(PATH_MAX+1);
 
     if (!getwd(buf)) {
@@ -585,10 +608,11 @@ ruby_getcwd(void)
         xfree(buf);
         rb_syserr_fail(e, "getwd");
     }
-#endif
     return buf;
 }
 
+#endif
+
 void
 ruby_each_words(const char *str, void (*func)(const char*, int, void*), void *arg)
 {