Fixed: doc generation crash due to bad UTF.

Not clear what's causing the problem, but the ± in CPDate.j get prefixed with some garbage bytes in the Doxygen output even that the input encoding looks fine.

The work-around for now is to just discard the bad bytes in the Python step. The output comes out correct (with the ± preserved).

Tools/Documentation/preprocess/003.make_headers.rb

@@ -5,6 +5,9 @@
 # $1 Cappuccino documentation directory
 
 
+Encoding.default_external = 'UTF-8'
+
+
 ACCESSOR_GET_TEMPLATE = <<EOS
 /*!
 Synthesized accessor method.
@@ -39,7 +42,7 @@ DUMMY_IVAR = "    id __doxygen__;"
 
 def makeHeaderFileFrom(fileName)
     # Grab the entire file (text)
-    sourceFile = File.new(fileName, "r")
+    sourceFile = File.new(fileName, "r", :encoding => 'UTF-8')
     source = sourceFile.read
     sourceFile.close()
 
@@ -51,7 +54,7 @@ def makeHeaderFileFrom(fileName)
     source.gsub!(/^\s*(@implementation \s*\w+(?:\s*:\s*\w+)?)\n(\s*[^{])/, "\\1\n{\n#{DUMMY_IVAR}\n}\n\\2")
     source.gsub!(/^\s*(@implementation \s*\w+(?:\s*:\s*\w+)?)\n\s*\{\s*\}/, "\\1\n{\n#{DUMMY_IVAR}\n}")
 
-    sourceFile = File.new(fileName, "w")
+    sourceFile = File.new(fileName, "w", :encoding => 'UTF-8')
 
     # Remove @accessor declarations from ivars before writing the source file
     sourceFile.write(source.gsub(/(\s*\w+\s+\w+)\s+@accessors(\(.+?\))?;/m, "\\1;"))

Tools/Documentation/support/massage_text.py

@@ -28,10 +28,15 @@ transforms = [
 html = glob.glob(os.path.join(sys.argv[1], "*.html"))
 
 for count, filename in enumerate(html):
-    f = open(filename, "r+")
-    text = f.read()
+    try:
+        # For some reason we get some UTF errors in the output HTML. Ignoring them should be fine.
+        f = open(filename, "r+", encoding='UTF-8', errors='ignore')
+        text = f.read()
+    except:
+        print("Failed to read %s." % filename)
+        raise
+        
     i = 0
-
     while i < len(transforms):
         text = transforms[i].sub(transforms[i + 1], text)
         i += 2