From 5ba5aa5c5cbd6afa26b44e81f58a0db385b1ad8c Mon Sep 17 00:00:00 2001
From: KMY <tt@kmycode.net>
Date: Thu, 30 May 2024 23:33:59 +0900
Subject: [PATCH] Normalize language code of incoming posts (#30403)

---
 app/lib/activitypub/parser/status_parser.rb   | 11 ++--
 .../activitypub/parser/status_parser_spec.rb  | 50 +++++++++++++++++++
 2 files changed, 57 insertions(+), 4 deletions(-)
 create mode 100644 spec/lib/activitypub/parser/status_parser_spec.rb

diff --git a/app/lib/activitypub/parser/status_parser.rb b/app/lib/activitypub/parser/status_parser.rb
index a77686dcb3..e982a12e63 100644
--- a/app/lib/activitypub/parser/status_parser.rb
+++ b/app/lib/activitypub/parser/status_parser.rb
@@ -3,6 +3,8 @@
 class ActivityPub::Parser::StatusParser
   include JsonLdHelper
 
+  NORMALIZED_LOCALE_NAMES = LanguagesHelper::SUPPORTED_LOCALES.keys.index_by(&:downcase).freeze
+
   # @param [Hash] json
   # @param [Hash] options
   # @option options [String] :followers_collection
@@ -118,10 +120,13 @@ class ActivityPub::Parser::StatusParser
   end
 
   def language
-    @language ||= original_language || (misskey_software? ? 'ja' : nil)
+    lang = raw_language_code || (misskey_software? ? 'ja' : nil)
+    lang.presence && NORMALIZED_LOCALE_NAMES.fetch(lang.downcase.to_sym, lang)
   end
 
-  def original_language
+  private
+
+  def raw_language_code
     if content_language_map?
       @object['contentMap'].keys.first
     elsif name_language_map?
@@ -131,8 +136,6 @@ class ActivityPub::Parser::StatusParser
     end
   end
 
-  private
-
   def audience_to
     as_array(@object['to'] || @json['to']).map { |x| value_or_id(x) }
   end
diff --git a/spec/lib/activitypub/parser/status_parser_spec.rb b/spec/lib/activitypub/parser/status_parser_spec.rb
new file mode 100644
index 0000000000..5d9f008db1
--- /dev/null
+++ b/spec/lib/activitypub/parser/status_parser_spec.rb
@@ -0,0 +1,50 @@
+# frozen_string_literal: true
+
+require 'rails_helper'
+
+RSpec.describe ActivityPub::Parser::StatusParser do
+  subject { described_class.new(json) }
+
+  let(:sender) { Fabricate(:account, followers_url: 'http://example.com/followers', domain: 'example.com', uri: 'https://example.com/actor') }
+  let(:follower) { Fabricate(:account, username: 'bob') }
+
+  let(:json) do
+    {
+      '@context': 'https://www.w3.org/ns/activitystreams',
+      id: [ActivityPub::TagManager.instance.uri_for(sender), '#foo'].join,
+      type: 'Create',
+      actor: ActivityPub::TagManager.instance.uri_for(sender),
+      object: object_json,
+    }.with_indifferent_access
+  end
+
+  let(:object_json) do
+    {
+      id: [ActivityPub::TagManager.instance.uri_for(sender), 'post1'].join('/'),
+      type: 'Note',
+      to: [
+        'https://www.w3.org/ns/activitystreams#Public',
+        ActivityPub::TagManager.instance.uri_for(follower),
+      ],
+      content: '@bob lorem ipsum',
+      contentMap: {
+        EN: '@bob lorem ipsum',
+      },
+      published: 1.hour.ago.utc.iso8601,
+      updated: 1.hour.ago.utc.iso8601,
+      tag: {
+        type: 'Mention',
+        href: ActivityPub::TagManager.instance.uri_for(follower),
+      },
+    }
+  end
+
+  it 'correctly parses status' do
+    expect(subject).to have_attributes(
+      text: '@bob lorem ipsum',
+      uri: [ActivityPub::TagManager.instance.uri_for(sender), 'post1'].join('/'),
+      reply: false,
+      language: :en
+    )
+  end
+end