[pull] master from ruby:master

.github/workflows/check_sast.yml

@@ -78,14 +78,14 @@ jobs:
           persist-credentials: false
 
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/init@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         with:
           languages: ${{ matrix.language }}
           build-mode: none
           config-file: .github/codeql/codeql-config.yml
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/analyze@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         with:
           category: '/language:${{ matrix.language }}'
           upload: False
@@ -127,7 +127,7 @@ jobs:
         continue-on-error: true
 
       - name: Upload SARIF
-        uses: github/codeql-action/upload-sarif@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/upload-sarif@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         with:
           sarif_file: sarif-results/${{ matrix.language }}.sarif
         continue-on-error: true

.github/workflows/labeler.yml

@@ -12,4 +12,4 @@ jobs:
       pull-requests: write
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/labeler@f27b608878404679385c85cfa523b85ccb86e213 # v6.1.0
+    - uses: actions/labeler@b8dd2d9be0f68b860e7dae5dae7d772984eacd6d # v6.2.0

.github/workflows/scorecards.yml

@@ -73,6 +73,6 @@ jobs:
       # Upload the results to GitHub's code scanning dashboard (optional).
       # Commenting out will disable upload of results to your repo's Code Scanning dashboard
       - name: "Upload to code-scanning"
-        uses: github/codeql-action/upload-sarif@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/upload-sarif@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         with:
           sarif_file: results.sarif

.github/workflows/zjit-macos.yml

@@ -98,7 +98,7 @@ jobs:
           rustup install ${{ matrix.rust_version }} --profile minimal
           rustup default ${{ matrix.rust_version }}
 
-      - uses: taiki-e/install-action@50414676f9f5d50a65992c6dd2ed02641263226c # v2.82.10
+      - uses: taiki-e/install-action@c7eb1735f09259a5035e8e5d44b1406b1cddc0fb # v2.83.0
         with:
           tool: nextest@0.9
         if: ${{ matrix.test_task == 'zjit-check' }}

.github/workflows/zjit-ubuntu.yml

@@ -141,7 +141,7 @@ jobs:
           ruby-version: '3.1'
           bundler: none
 
-      - uses: taiki-e/install-action@50414676f9f5d50a65992c6dd2ed02641263226c # v2.82.10
+      - uses: taiki-e/install-action@c7eb1735f09259a5035e8e5d44b1406b1cddc0fb # v2.83.0
         with:
           tool: nextest@0.9
         if: ${{ matrix.test_task == 'zjit-check' }}

array.c

@@ -208,7 +208,9 @@ ary_embed_size(long capa)
 static bool
 ary_embeddable_p(long capa)
 {
-    return rb_gc_size_allocatable_p(ary_embed_size(capa));
+    const long embed_len_max = RARRAY_EMBED_LEN_MASK >> RARRAY_EMBED_LEN_SHIFT;
+
+    return capa <= embed_len_max && rb_gc_size_allocatable_p(ary_embed_size(capa));
 }
 
 bool

depend

@@ -16094,6 +16094,7 @@ scheduler.$(OBJEXT): $(top_srcdir)/internal/sanitizers.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/serial.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/set_table.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/static_assert.h
+scheduler.$(OBJEXT): $(top_srcdir)/internal/struct.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/thread.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/variable.h
 scheduler.$(OBJEXT): $(top_srcdir)/internal/vm.h
@@ -16897,6 +16898,7 @@ signal.$(OBJEXT): $(top_srcdir)/internal/set_table.h
 signal.$(OBJEXT): $(top_srcdir)/internal/signal.h
 signal.$(OBJEXT): $(top_srcdir)/internal/static_assert.h
 signal.$(OBJEXT): $(top_srcdir)/internal/string.h
+signal.$(OBJEXT): $(top_srcdir)/internal/struct.h
 signal.$(OBJEXT): $(top_srcdir)/internal/thread.h
 signal.$(OBJEXT): $(top_srcdir)/internal/variable.h
 signal.$(OBJEXT): $(top_srcdir)/internal/vm.h

dir.rb

@@ -213,18 +213,29 @@ def initialize(name, encoding: nil)
   end
 
   # call-seq:
-  #   Dir[patterns, base: nil, sort: true] -> array
+  #   Dir[*patterns, base: nil, sort: true] -> array
+  #
+  # Like Dir.glob, but does not accept keyword argument +flags+,
+  # and may take multiple patterns as arguments:
+  #
+  #   Dir['*.rb', '*.h'].take(3) # => ["KNOWNBUGS.rb", "array.rb", "ast.rb"]
   #
-  # Like Dir.glob, but does not accept keyword argument +flags+.
   def self.[](*args, base: nil, sort: true)
     Primitive.dir_s_aref(args, base, sort)
   end
 
   # call-seq:
   #   Dir.glob(patterns, flags: 0, base: '.', sort: true) -> array_of_entries
+  #   Dir.glob(patterns, flags: 0, base: '.', sort: true) {|entry_name| ... } -> nil
   #
   # Returns an array of filesystem entries;
   # see {Filename Globbing}[rdoc-ref:file/filename_globbing.md].
+  #
+  # With a block given, calls the block with each of the selected entry names
+  # and returns +nil+:
+  #
+  #   Dir.glob('*.rb') {|entry_name| puts entry_name } # => nil
+  #
   def self.glob(pattern, _flags = 0, flags: _flags, base: nil, sort: true)
     Primitive.attr! :use_block
     Primitive.dir_s_glob(pattern, flags, base, sort)

doc/file/filename_globbing.md

@@ -33,7 +33,10 @@ Inputs to the filename-globbing methods:
 
 Their return values:
 
-- Each of the methods `Dir[]` and Dir.glob returns an array of the selected string entries.
+- \Method `Dir[]` returns an array of the selected string entries.
+- \Method Dir.glob with no block returns an array of the selected string entries.
+- \Method Dir.glob with a block calls the block with each selected string entry
+  and returns `nil`.
 - \Method Pathname#glob with no block returns an array of Pathname objects
   each based on a selected string entry.
 - \Method Pathname#glob with a block calls the block with each pathname
@@ -44,8 +47,13 @@ Examples:
 ```ruby
 Dir['*'].take(3)
 # => ["BSDL", "CONTRIBUTING.md", "COPYING"]
+Dir['*.rb', '*.h'].take(3)
+# => ["KNOWNBUGS.rb", "array.rb", "ast.rb"]
 Dir.glob('*').take(3)
 # => ["BSDL", "CONTRIBUTING.md", "COPYING"]
+Dir.glob(['*.rb', '*.h']).take(3)
+# => ["KNOWNBUGS.rb", "array.rb", "ast.rb"]
+Dir.glob('*.rb') {|entry_name| puts entry_name } # => nil
 Pathname('.').glob('*').take(3)
 # => [#<Pathname:BSDL>, #<Pathname:CONTRIBUTING.md>, #<Pathname:COPYING>]
 a = []
@@ -299,6 +307,25 @@ Pathname('.').glob('test/ruby/**/*.rb').take(3)
 #     #<Pathname:test/ruby/box/a.1_1_0.rb>]
 ```
 
+The double-asterisk pattern matches directories recursively
+only when followed by the slash character `'/'`;
+otherwise it is equivalent to pattern `'*'`:
+
+```ruby
+Dir.glob('**') == Dir.glob('*') # => true
+```
+
+A pattern ending with `'/'` matches directory names only;
+each matched name ends with `'/'`:
+
+```ruby
+# Find the directories at the top level.
+Dir.glob('*/').take(3)  # => ["basictest/", "benchmark/", "bin/"]
+
+# Find all directories everywhere.
+Dir.glob('**/').take(3) # => ["basictest/", "benchmark/", "benchmark/gc/"]
+```
+
 The pattern may be escaped:
 
 ```ruby
@@ -325,6 +352,22 @@ Pathname('.').glob('\[k-m][h-j][a-c]') # => []
 Pathname('.').glob('\**/*.rb')         # => []
 ```
 
+Note that the backslash escapes the following character on Windows as well,
+and therefore cannot be used as a path separator in the pattern;
+`Dir.glob('C:\Users\*')` matches nothing.
+Write the pattern with forward slashes instead:
+
+```ruby
+Dir.glob('C:/Users/*')
+```
+
+A Windows path taken from an external source (such as an environment variable)
+may be converted for use as a pattern:
+
+```ruby
+pattern = path.tr('\\', '/')
+```
+
 ## Keyword Arguments
 
 | Keyword           | Value                    | Default | Meaning                                 |
@@ -417,6 +460,11 @@ Pathname('.').glob('\*').size                     # => 0
 Pathname('.').glob('\*', File::FNM_NOESCAPE).size # => 0
 ```
 
+Note that on Windows this flag does not make the backslash usable
+as a path separator in the pattern;
+the backslash is then matched as an ordinary character,
+which cannot occur in an entry name.
+
 #### Constant File::FNM_SHORTNAME
 
 By default, Windows shortname matching is disabled;

gc.c

@@ -1106,6 +1106,16 @@ VALUE class_allocate_complex_instance(VALUE klass, uint32_t capacity)
     return obj;
 }
 
+static inline size_t
+robject_embedded_size(uint32_t fields_count)
+{
+    size_t size = rb_obj_embedded_size(fields_count);
+    if (!rb_gc_size_allocatable_p(size)) {
+        size = sizeof(struct RObject);
+    }
+    return size;
+}
+
 VALUE
 rb_class_allocate_instance(VALUE klass)
 {
@@ -1118,10 +1128,7 @@ rb_class_allocate_instance(VALUE klass)
         obj = class_allocate_complex_instance(klass, index_tbl_num_entries);
     }
     else {
-        size_t size = rb_obj_embedded_size(index_tbl_num_entries);
-        if (!rb_gc_size_allocatable_p(size)) {
-            size = sizeof(struct RObject);
-        }
+        size_t size = robject_embedded_size(index_tbl_num_entries);
 
         // There might be a NEWOBJ tracepoint callback, and it may set fields.
         // So the shape must be passed to `NEWOBJ_OF`.
@@ -1145,6 +1152,25 @@ rb_class_allocate_instance(VALUE klass)
     return obj;
 }
 
+#if USE_ZJIT
+bool
+rb_zjit_class_allocate_instance_fastpath(VALUE klass, size_t *size_out, shape_id_t *shape_id_out)
+{
+    uint32_t index_tbl_num_entries = RCLASS_MAX_IV_COUNT(klass);
+
+    RUBY_ASSERT(rb_shape_max_capacity() > 0);
+    if (RB_UNLIKELY(index_tbl_num_entries > rb_shape_max_capacity())) {
+        return false;
+    }
+
+    size_t size = robject_embedded_size(index_tbl_num_entries);
+    *size_out = size;
+    *shape_id_out = rb_shape_transition_slot_size(rb_shape_transition_robject(0),
+                                                  rb_gc_size_slot_size(size));
+    return true;
+}
+#endif
+
 void
 rb_gc_register_pinning_obj(VALUE obj)
 {
@@ -2159,16 +2185,10 @@ object_id(VALUE obj)
     return object_id0(obj);
 }
 
-static inline void
-obj_free_object_id(VALUE obj)
-{
-}
-
 void
 rb_gc_obj_free_vm_weak_references(VALUE obj)
 {
     ASSUME(!RB_SPECIAL_CONST_P(obj));
-    obj_free_object_id(obj);
 
     if (rb_obj_gen_fields_p(obj)) {
         rb_free_generic_ivar(obj);
@@ -2636,11 +2656,22 @@ ruby_stack_check(void)
 
 /* ==================== Marking ==================== */
 
+/* The traversal mark redirect is per-Ractor, except on a modular GC where
+ * marking can run on worker threads with no current EC and it lives in the VM.
+ * GC_MARK_FUNC_DATA_SLOTP() points at the active slot; a non-modular build pays
+ * nothing extra over a plain GET_VM() (one GET_RACTOR(), no NULL check). */
+#if USE_MODULAR_GC
+#  define GC_MARK_FUNC_DATA_SLOTP()  (&GET_VM()->gc.mark_func_data)
+#else
+#  define GC_MARK_FUNC_DATA_SLOTP()  (&GET_RACTOR()->mark_func_data)
+#endif
+
 #define RB_GC_MARK_OR_TRAVERSE(func, obj_or_ptr, obj, check_obj) do { \
     if (!RB_SPECIAL_CONST_P(obj)) { \
-        rb_vm_t *vm = GET_VM(); \
-        void *objspace = vm->gc.objspace; \
-        if (LIKELY(vm->gc.mark_func_data == NULL)) { \
+        struct gc_mark_func_data_struct **mfdp = GC_MARK_FUNC_DATA_SLOTP(); \
+        struct gc_mark_func_data_struct *mark_func_data = *mfdp; \
+        void *objspace = GET_VM()->gc.objspace; \
+        if (LIKELY(mark_func_data == NULL)) { \
             GC_ASSERT(rb_gc_impl_during_gc_p(objspace)); \
             (func)(objspace, (obj_or_ptr)); \
         } \
@@ -2649,10 +2680,9 @@ ruby_stack_check(void)
                     !rb_gc_impl_garbage_object_p(objspace, obj) : \
                 true) { \
             GC_ASSERT(!rb_gc_impl_during_gc_p(objspace)); \
-            struct gc_mark_func_data_struct *mark_func_data = vm->gc.mark_func_data; \
-            vm->gc.mark_func_data = NULL; \
+            *mfdp = NULL; \
             mark_func_data->mark_func((obj), mark_func_data->data); \
-            vm->gc.mark_func_data = mark_func_data; \
+            *mfdp = mark_func_data; \
         } \
     } \
 } while (0)
@@ -4519,16 +4549,16 @@ rb_objspace_reachable_objects_from(VALUE obj, void (func)(VALUE, void *), void *
         if (rb_gc_impl_during_gc_p(rb_gc_get_objspace())) rb_bug("rb_objspace_reachable_objects_from() is not supported while during GC");
 
         if (!RB_SPECIAL_CONST_P(obj)) {
-            rb_vm_t *vm = GET_VM();
-            struct gc_mark_func_data_struct *prev_mfd = vm->gc.mark_func_data;
+            struct gc_mark_func_data_struct **mfdp = GC_MARK_FUNC_DATA_SLOTP();
+            struct gc_mark_func_data_struct *prev_mfd = *mfdp;
             struct gc_mark_func_data_struct mfd = {
                 .mark_func = func,
                 .data = data,
             };
 
-            vm->gc.mark_func_data = &mfd;
+            *mfdp = &mfd;
             rb_gc_mark_children(rb_gc_get_objspace(), obj);
-            vm->gc.mark_func_data = prev_mfd;
+            *mfdp = prev_mfd;
         }
     }
 }
@@ -4558,16 +4588,17 @@ rb_objspace_reachable_objects_from_root(void (func)(const char *category, VALUE,
         .data = passing_data,
     };
 
-    struct gc_mark_func_data_struct *prev_mfd = vm->gc.mark_func_data;
+    struct gc_mark_func_data_struct **mfdp = GC_MARK_FUNC_DATA_SLOTP();
+    struct gc_mark_func_data_struct *prev_mfd = *mfdp;
     struct gc_mark_func_data_struct mfd = {
         .mark_func = root_objects_from,
         .data = &data,
     };
 
-    vm->gc.mark_func_data = &mfd;
+    *mfdp = &mfd;
     rb_gc_save_machine_context();
     rb_gc_mark_roots(vm->gc.objspace, &data.category);
-    vm->gc.mark_func_data = prev_mfd;
+    *mfdp = prev_mfd;
 }
 
 /*

io.c

@@ -12141,7 +12141,7 @@ io_s_foreach(VALUE v)
  *
  *  - {Open Options}[rdoc-ref:IO@Open+Options].
  *  - {Encoding options}[rdoc-ref:encodings.rdoc@Encoding+Options].
- *  - {Line Options}[rdoc-ref:IO@Line+IO].
+ *  - {Line Input Options}[rdoc-ref:IO@Line+Input+Options].
  *
  *  Returns an Enumerator if no block is given.
  *
@@ -12217,7 +12217,7 @@ io_s_readlines(VALUE v)
  *
  *  - {Open Options}[rdoc-ref:IO@Open+Options].
  *  - {Encoding options}[rdoc-ref:encodings.rdoc@Encoding+Options].
- *  - {Line Options}[rdoc-ref:IO@Line+IO].
+ *  - {Line Input Options}[rdoc-ref:IO@Line+Input+Options].
  *
  */
 
@@ -15253,7 +15253,12 @@ set_LAST_READ_LINE(VALUE val, ID _x, VALUE *_y)
  *  === Line Input
  *
  *  Class \IO supports line-oriented input for
- *  {files}[rdoc-ref:IO@File+Line+Input] and {IO streams}[rdoc-ref:IO@Stream+Line+Input]
+ *  {files}[rdoc-ref:IO@File+Line+Input] and {IO streams}[rdoc-ref:IO@Stream+Line+Input].
+ *
+ *  ==== Line Input Options
+ *
+ *  Optional keyword argument +chomp+ (default: +false+)
+ *  specifies whether line separators are to be excluded from the result of a read.
  *
  *  ==== \File Line Input
  *

jit.c

@@ -41,6 +41,9 @@ enum jit_bindgen_constants {
     // Field offset for the RHash struct
     RUBY_OFFSET_RHASH_IFNONE = offsetof(struct RHash, ifnone),
 
+    // Max pairs an embedded ar_table hash holds before it converts to an st_table
+    RUBY_RHASH_AR_TABLE_MAX_SIZE = RHASH_AR_TABLE_MAX_SIZE,
+
     // Field offsets for the RString struct
     RUBY_OFFSET_RSTRING_LEN = offsetof(struct RString, len),