google-protobuf 4.28.0.rc.1-x86-mingw32 → 4.28.0.rc.2-x86-mingw32

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 95bc4885452d0c487764575c9ee7e92f4251534974e38bd2b0057af4ed3e0837
-  data.tar.gz: 30e4e9e9aee63b7004f2fa0a187a4b9e9a24112613bf7ad23bbadb3c2199ffab
+  metadata.gz: 8d3632fadafba9e7e0ef63f76c813d2c41eea5b126a5b482f3f685ebb6250c6b
+  data.tar.gz: 0d1b58d5c03240a46cad224a807d752e3c919692c977317c874f94542430ca99
 SHA512:
-  metadata.gz: 2e31dcead555699dce5aae4c686d6a27650b65b2de9c176f8696bba8e8f6762c2f65f99ed950147ce8ea2f2a2eb0af9570eab017106e612e823d642afbd7d2fe
-  data.tar.gz: 7944571933b51bedbcbdc3baf0a719699e37fcd27c3c5dbfd1e6746906e121bd8a2532049642486bd12b2eb7b8adad92751a8391bb92edac5e553755e4750d86
+  metadata.gz: 63cd6bfe49ee1a23f56c0e6dcaf29d3877e48d7580a86afac575be249c939ad22b4903ce26bc83992b0054039e014b8734a0d537ee8e21b1e42c00bcaa8ca4a6
+  data.tar.gz: fba0ebf8e3fcdc7b8fb34bcc4c063e0abee05e353ee5ce8e4e1f0485f21f07122785fe2aba6928e2019299d4c0227246d1c2d5f7f4633e9381a1e5223f68a567

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.2
 platform: x86-mingw32
 authors:
 - Protobuf Authors
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-07-11 00:00:00.000000000 Z
+date: 2024-08-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler-dock
@@ -190,7 +190,7 @@ 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-rc2/ruby
 post_install_message: 
 rdoc_options: []
 require_paths: