merge revision(s) 7e4b1f8e1935a10df3c41ee60ca0987d73281126: [Backport #20322]

	[Bug #20322] Fix rb_enc_interned_str_cstr null encoding

	The documentation for `rb_enc_interned_str_cstr` notes that `enc` can be
	a null pointer, but this currently causes a segmentation fault when
	trying to autoload the encoding. This commit fixes the issue by checking
	for NULL before calling `rb_enc_autoload`.

6 files changed, 86 insertions, 4 deletions

diff --git a/ext/-test-/string/fstring.c b/ext/-test-/string/fstring.c
index 2374319fe3..64f079251d 100644
--- a/ext/-test-/string/fstring.c
+++ b/ext/-test-/string/fstring.c
@@ -12,13 +12,13 @@ bug_s_fstring(VALUE self, VALUE str)
 VALUE
 bug_s_rb_enc_interned_str(VALUE self, VALUE encoding)
 {
-    return rb_enc_interned_str("foo", 3, RDATA(encoding)->data);
+    return rb_enc_interned_str("foo", 3, NIL_P(encoding) ? NULL : RDATA(encoding)->data);
 }
 
 VALUE
 bug_s_rb_enc_str_new(VALUE self, VALUE encoding)
 {
-    return rb_enc_str_new("foo", 3, RDATA(encoding)->data);
+    return rb_enc_str_new("foo", 3, NIL_P(encoding) ? NULL : RDATA(encoding)->data);
 }
 
 void
diff --git a/spec/ruby/optional/capi/ext/string_spec.c b/spec/ruby/optional/capi/ext/string_spec.c
index 9cbb50484d..73b5aba278 100644
--- a/spec/ruby/optional/capi/ext/string_spec.c
+++ b/spec/ruby/optional/capi/ext/string_spec.c
@@ -597,6 +597,15 @@ static VALUE string_spec_rb_str_unlocktmp(VALUE self, VALUE str) {
   return rb_str_unlocktmp(str);
 }
 
+static VALUE string_spec_rb_enc_interned_str_cstr(VALUE self, VALUE str, VALUE enc) {
+  rb_encoding *e = NIL_P(enc) ? 0 : rb_to_encoding(enc);
+  return rb_enc_interned_str_cstr(RSTRING_PTR(str), e);
+}
+
+static VALUE string_spec_rb_str_to_interned_str(VALUE self, VALUE str) {
+  return rb_str_to_interned_str(str);
+}
+
 void Init_string_spec(void) {
   VALUE cls = rb_define_class("CApiStringSpecs", rb_cObject);
   rb_define_method(cls, "rb_cstr2inum", string_spec_rb_cstr2inum, 2);
@@ -698,6 +707,8 @@ void Init_string_spec(void) {
   rb_define_method(cls, "rb_str_catf", string_spec_rb_str_catf, 1);
   rb_define_method(cls, "rb_str_locktmp", string_spec_rb_str_locktmp, 1);
   rb_define_method(cls, "rb_str_unlocktmp", string_spec_rb_str_unlocktmp, 1);
+  rb_define_method(cls, "rb_enc_interned_str_cstr", string_spec_rb_enc_interned_str_cstr, 2);
+  rb_define_method(cls, "rb_str_to_interned_str", string_spec_rb_str_to_interned_str, 1);
 }
 
 #ifdef __cplusplus
diff --git a/spec/ruby/optional/capi/string_spec.rb b/spec/ruby/optional/capi/string_spec.rb
index 3a47fa9762..c0fbf3d1ba 100644
--- a/spec/ruby/optional/capi/string_spec.rb
+++ b/spec/ruby/optional/capi/string_spec.rb
@@ -1227,4 +1227,67 @@ end
       -> { @s.rb_str_unlocktmp("test") }.should raise_error(RuntimeError, 'temporal unlocking already unlocked string')
     end
   end
+
+  describe "rb_enc_interned_str_cstr" do
+    it "returns a frozen string" do
+      str = "hello"
+      val = @s.rb_enc_interned_str_cstr(str, Encoding::US_ASCII)
+
+      val.should.is_a?(String)
+      val.encoding.should == Encoding::US_ASCII
+      val.should.frozen?
+    end
+
+    it "returns the same frozen string" do
+      str = "hello"
+      result1 = @s.rb_enc_interned_str_cstr(str, Encoding::US_ASCII)
+      result2 = @s.rb_enc_interned_str_cstr(str, Encoding::US_ASCII)
+      result1.should.equal?(result2)
+    end
+
+    it "returns different frozen strings for different encodings" do
+      str = "hello"
+      result1 = @s.rb_enc_interned_str_cstr(str, Encoding::US_ASCII)
+      result2 = @s.rb_enc_interned_str_cstr(str, Encoding::UTF_8)
+      result1.should_not.equal?(result2)
+    end
+
+    it "returns the same string as String#-@" do
+      @s.rb_enc_interned_str_cstr("hello", Encoding::UTF_8).should.equal?(-"hello")
+    end
+
+    ruby_bug "#20322", ""..."3.4" do
+      it "uses the default encoding if encoding is null" do
+        str = "hello"
+        val = @s.rb_enc_interned_str_cstr(str, nil)
+        val.encoding.should == Encoding::ASCII_8BIT
+      end
+    end
+  end
+
+  describe "rb_str_to_interned_str" do
+    it "returns a frozen string" do
+      str = "hello"
+      result = @s.rb_str_to_interned_str(str)
+      result.should.is_a?(String)
+      result.should.frozen?
+    end
+
+    it "returns the same frozen string" do
+      str = "hello"
+      result1 = @s.rb_str_to_interned_str(str)
+      result2 = @s.rb_str_to_interned_str(str)
+      result1.should.equal?(result2)
+    end
+
+    it "returns different frozen strings for different encodings" do
+      result1 = @s.rb_str_to_interned_str("hello".force_encoding(Encoding::US_ASCII))
+      result2 = @s.rb_str_to_interned_str("hello".force_encoding(Encoding::UTF_8))
+      result1.should_not.equal?(result2)
+    end
+
+    it "returns the same string as String#-@" do
+      @s.rb_str_to_interned_str("hello").should.equal?(-"hello")
+    end
+  end
 end
diff --git a/string.c b/string.c
index ed51fa4af7..40b46eee57 100644
--- a/string.c
+++ b/string.c
@@ -11969,7 +11969,7 @@ rb_interned_str_cstr(const char *ptr)
 VALUE
 rb_enc_interned_str(const char *ptr, long len, rb_encoding *enc)
 {
-    if (UNLIKELY(rb_enc_autoload_p(enc))) {
+    if (enc != NULL && UNLIKELY(rb_enc_autoload_p(enc))) {
         rb_enc_autoload(enc);
     }
 
diff --git a/test/-ext-/string/test_fstring.rb b/test/-ext-/string/test_fstring.rb
index 5a3456c566..92982a315e 100644
--- a/test/-ext-/string/test_fstring.rb
+++ b/test/-ext-/string/test_fstring.rb
@@ -20,6 +20,10 @@ class Test_String_Fstring < Test::Unit::TestCase
     RUBY
   end
 
+  def test_rb_enc_interned_str_null_encoding
+    assert_equal Encoding::ASCII_8BIT, Bug::String.rb_enc_interned_str(nil).encoding
+  end
+
   def test_rb_enc_str_new_autoloaded_encoding
     assert_separately([], <<~RUBY)
       require '-test-/string'
@@ -28,6 +32,10 @@ class Test_String_Fstring < Test::Unit::TestCase
     RUBY
   end
 
+  def test_rb_enc_str_new_null_encoding
+    assert_equal Encoding::ASCII_8BIT, Bug::String.rb_enc_str_new(nil).encoding
+  end
+
   def test_instance_variable
     str = __method__.to_s * 3
     str.instance_variable_set(:@test, 42)
diff --git a/version.h b/version.h
index 2aab0f2935..ae323455f8 100644
--- a/version.h
+++ b/version.h
@@ -11,7 +11,7 @@
 # define RUBY_VERSION_MINOR RUBY_API_VERSION_MINOR
 #define RUBY_VERSION_TEENY 4
 #define RUBY_RELEASE_DATE RUBY_RELEASE_YEAR_STR"-"RUBY_RELEASE_MONTH_STR"-"RUBY_RELEASE_DAY_STR
-#define RUBY_PATCHLEVEL 194
+#define RUBY_PATCHLEVEL 195
 
 #include "ruby/version.h"
 #include "ruby/internal/abi.h"