[Bug #22387] Fix termlen in str_shared_replace

The buffer is copied over based on the new string's termlen
so we can just set the new encoding raw instead of recalculating
based on the old encoding's termlen.

Author: rwstauner

string.c

@@ -1853,8 +1853,6 @@ str_shared_replace(VALUE str, VALUE str2)
     if (str_embed_capa(str) >= RSTRING_LEN(str2) + termlen) {
         STR_SET_EMBED(str);
         memcpy(RSTRING_PTR(str), RSTRING_PTR(str2), (size_t)RSTRING_LEN(str2) + termlen);
-        rb_enc_associate(str, enc);
-        ENC_CODERANGE_SET(str, cr);
     }
     else {
         if (STR_EMBED_P(str2)) {
@@ -1886,9 +1884,12 @@ str_shared_replace(VALUE str, VALUE str2)
         STR_SET_EMBED(str2);
         RSTRING_PTR(str2)[0] = 0;
         STR_SET_LEN(str2, 0);
-        rb_enc_associate(str, enc);
-        ENC_CODERANGE_SET(str, cr);
     }
+
+    // We used str2's termlen above so we set enc raw
+    // to avoid adjusting it based on str1's old enc.
+    rb_enc_raw_set(str, enc);
+    ENC_CODERANGE_SET(str, cr);
 }
 
 VALUE

test/-ext-/string/test_capacity.rb

@@ -87,12 +87,34 @@ def test_frozen_root_no_embed_capacity_with_multibyte_terminator
     assert_equal(capacity, capa(s))
   end
 
+  def test_encode_bang_capacity_with_multibyte_terminator
+    s = smallest_slot_utf16le_string
+    assert(Bug::String.cstr_embedded?(s))
+
+    s.encode!("UTF-8")
+
+    # The UTF-8 form no longer fits in the slot so the buffer is on heap.
+    refute(Bug::String.cstr_embedded?(s))
+
+    # and sized precisely
+    assert_equal(s.bytesize, capa(s))
+  end
+
   private
 
   def capa(str)
     Bug::String.capacity(str)
   end
 
+  # A UTF-16LE string that fills the smallest slot, so that its UTF-8 form
+  # (1.5x as long) cannot stay embedded.
+  def smallest_slot_utf16le_string
+    embed_capa = pool_slot_size(0) - embed_header_size
+    str = ("\u{30AF}" * ((embed_capa - 2) / 2)).encode("UTF-16LE").b
+    str.force_encoding("UTF-16LE")
+    str
+  end
+
   def multibyte_terminator_string
     ("\u{30AF}\u{30FC}\u{30DD}\u{30F3}\u{1F381}" * 100).encode("UTF-16LE")
   end