audit: add rules for non-ascii character and encoding comment

Closes Homebrew/homebrew#40042.

Signed-off-by: Xu Cheng <xucheng@me.com>

1 files changed, 16 insertions, 0 deletions

diff --git a/Library/Homebrew/cmd/audit.rb b/Library/Homebrew/cmd/audit.rb
index 4290bec4c..be6be5c39 100644
--- a/Library/Homebrew/cmd/audit.rb
+++ b/Library/Homebrew/cmd/audit.rb
@@ -70,6 +70,14 @@ class FormulaText
     /\Z\n/ =~ @text
   end
 
+  def has_non_ascii_character?
+    /[^\x00-\x7F]/ =~ @text
+  end
+
+  def has_encoding_comment?
+    /^# (en)?coding: utf-8$/i =~ @text
+  end
+
   def =~ regex
     regex =~ @text
   end
@@ -118,6 +126,14 @@ class FormulaAuditor
       problem "'__END__' was found, but 'DATA' is not used"
     end
 
+    if text.has_non_ascii_character? and not text.has_encoding_comment?
+      problem "Found non-ASCII character: add `# encoding: UTF-8` in the first line"
+    end
+
+    if text.has_encoding_comment? and not text.has_non_ascii_character?
+      problem "Remove the redundant `# encoding: UTF-8`"
+    end
+
     unless text.has_trailing_newline?
       problem "File should end with a newline"
     end