google-protobuf 4.28.0.rc.1 → 4.28.0.rc.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3f8ae5732a2289222f4aecad3236ff2efe38dabec9671f5c063bfa176fa3f8fc
-  data.tar.gz: 9acafac876b4dc73ba2870c7c958ebd64ee5c9c748521d13246fc9eb61ee351d
+  metadata.gz: a734f713e56feb26f832325a750f6f900da2e6ee8d5caddb5c7fada9c018ec97
+  data.tar.gz: ba69aaf96db26dc1ba721f2155aa5b2811b9167e4e76d984c7fe4a873ace1124
 SHA512:
-  metadata.gz: fb116cf42a8f1d4df406a823cff60c37ab9da85131c045f74b8fda39205c8ca92c9ed7a11279cad0e53f3dcf88a5dfd0adacb05cf74fb8e1e5b305dec3071caf
-  data.tar.gz: d00f7e15815217b04f1480023aa1d7bc340860adaa5b5efe32b9897e165876a4d5e53e5f4f05f7b005862e2bec4a47159d35710ba88c9a464885078cf99e1529
+  metadata.gz: b6a54fa9509710a695d076f9db12d0eb4a621d2dfdfa34418888a8a1811524fd345d745227eb64e00af67144cacab5f5c9a22f7936801fdb311be4fec57020b0
+  data.tar.gz: 658cb8ab707d4a0c770034be7fcea78b36d307e5d708e47c9198e8460502ed66f12dd23a2587176ee16bcdd041de5e31604eafc35fb66100d0bf883a205146fe

data/ext/google/protobuf_c/convert.c

@@ -104,6 +104,41 @@ unknownval:
   rb_raise(rb_eRangeError, "Unknown symbol value for enum field '%s'.", name);
 }
 
+VALUE Convert_CheckStringUtf8(VALUE str) {
+  VALUE utf8 = rb_enc_from_encoding(rb_utf8_encoding());
+
+  if (rb_obj_encoding(str) == utf8) {
+    // Note: Just because a string is marked as having UTF-8 encoding does
+    // not mean that it is *valid* UTF-8.  We have to check separately
+    // whether it is valid.
+    if (rb_enc_str_coderange(str) == ENC_CODERANGE_BROKEN) {
+      // TODO: For now
+      // we only warn for this case.  We will remove the warning and throw an
+      // exception below in the 30.x release
+
+      rb_warn(
+          "String is invalid UTF-8. This will be an error in a future "
+          "version.");
+      // VALUE exc = rb_const_get_at(
+      //     rb_cEncoding, rb_intern("InvalidByteSequenceError"));
+      // rb_raise(exc, "String is invalid UTF-8");
+    }
+  } else {
+    // Note: this will not duplicate underlying string data unless
+    // necessary.
+    //
+    // This will throw an exception if the conversion cannot be performed:
+    // - Encoding::UndefinedConversionError if certain characters cannot be
+    //   converted to UTF-8.
+    // - Encoding::InvalidByteSequenceError if certain characters were invalid
+    //   in the source encoding.
+    str = rb_str_encode(str, utf8, 0, Qnil);
+    PBRUBY_ASSERT(rb_enc_str_coderange(str) != ENC_CODERANGE_BROKEN);
+  }
+
+  return str;
+}
+
 upb_MessageValue Convert_RubyToUpb(VALUE value, const char* name,
                                    TypeInfo type_info, upb_Arena* arena) {
   upb_MessageValue ret;
@@ -137,8 +172,7 @@ upb_MessageValue Convert_RubyToUpb(VALUE value, const char* name,
       }
       break;
     }
-    case kUpb_CType_String: {
-      VALUE utf8 = rb_enc_from_encoding(rb_utf8_encoding());
+    case kUpb_CType_String:
       if (rb_obj_class(value) == rb_cSymbol) {
         value = rb_funcall(value, rb_intern("to_s"), 0);
       } else if (!rb_obj_is_kind_of(value, rb_cString)) {
@@ -147,19 +181,9 @@ upb_MessageValue Convert_RubyToUpb(VALUE value, const char* name,
                  rb_class2name(CLASS_OF(value)));
       }
 
-      if (rb_obj_encoding(value) != utf8) {
-        // Note: this will not duplicate underlying string data unless
-        // necessary.
-        value = rb_str_encode(value, utf8, 0, Qnil);
-
-        if (rb_enc_str_coderange(value) == ENC_CODERANGE_BROKEN) {
-          rb_raise(rb_eEncodingError, "String is invalid UTF-8");
-        }
-      }
-
+      value = Convert_CheckStringUtf8(value);
       ret.str_val = Convert_StringData(value, arena);
       break;
-    }
     case kUpb_CType_Bytes: {
       VALUE bytes = rb_enc_from_encoding(rb_ascii8bit_encoding());
       if (rb_obj_class(value) != rb_cString) {

data/lib/google/protobuf/ffi/internal/convert.rb

@@ -33,11 +33,18 @@ module Google
             return_value[:bool_val] = value
           when :string
             raise TypeError.new "Invalid argument for string field '#{name}' (given #{value.class})." unless value.is_a?(String) or value.is_a?(Symbol)
-            begin
+            value = value.to_s if value.is_a?(Symbol)
+            if value.encoding == Encoding::UTF_8
+              unless value.valid_encoding?
+                # TODO:
+                # For now we only warn for this case.  We will remove the
+                # warning and throw an exception below in the 30.x release
+                warn "String is invalid UTF-8. This will be an error in a future version."
+                # raise Encoding::InvalidByteSequenceError.new "String is invalid UTF-8"
+              end
+              string_value = value
+            else
               string_value = value.to_s.encode("UTF-8")
-            rescue Encoding::UndefinedConversionError
-              # TODO - why not include the field name here?
-              raise Encoding::UndefinedConversionError.new "String is invalid UTF-8"
             end
             return_value[:str_val][:size] = string_value.bytesize
             return_value[:str_val][:data] = Google::Protobuf::FFI.arena_malloc(arena, string_value.bytesize)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: google-protobuf
 version: !ruby/object:Gem::Version
-  version: 4.28.0.rc.1
+  version: 4.28.0.rc.3
 platform: ruby
 authors:
 - Protobuf Authors
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-07-11 00:00:00.000000000 Z
+date: 2024-08-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler-dock
@@ -188,7 +188,7 @@ homepage: https://developers.google.com/protocol-buffers
 licenses:
 - BSD-3-Clause
 metadata:
-  source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v4.28.0-rc1/ruby
+  source_code_uri: https://github.com/protocolbuffers/protobuf/tree/v4.28.0-rc3/ruby
 post_install_message: 
 rdoc_options: []
 require_paths: