[feaLib] Change semantics of markClass to also define a glyphclass

While not really documented in the OpenType Feature File specification,
the AFDKO makeotf tool handles the `markClass` statement this way.
Also, some examples in the specification seem to imply this semantics.

Lib/fontTools/feaLib/parser.py

@@ -273,6 +273,9 @@ class Parser(object):
         for glyph in glyphs:
             markClass.anchors[glyph] = anchor
             markClass.glyphLocations[glyph] = location
+        glyphclass = ast.GlyphClassDefinition(location, name, glyphs)
+        self.glyphclasses_.define(name, glyphclass)
+        return glyphclass
 
     def is_next_glyphclass_(self):
         return (self.next_token_ == "[" or

Lib/fontTools/feaLib/parser_test.py

@@ -142,6 +142,15 @@ class ParserTest(unittest.TestCase):
         self.assertEqual(foo.glyphs, {"a", "b"})
         self.assertEqual(bar.glyphs, {"a", "b"})
 
+    def test_glyphclass_from_markClass(self):
+        doc = self.parse(
+            "markClass [acute grave] <anchor 500 800> @TOP_MARKS;"
+            "markClass cedilla <anchor 500 -100> @BOTTOM_MARKS;"
+            "@MARKS = [@TOP_MARKS @BOTTOM_MARKS ogonek];"
+            "@ALL = @MARKS;")
+        self.assertEqual(doc.statements[-1].glyphs,
+                         {"acute", "cedilla", "grave", "ogonek"})
+
     def test_glyphclass_range_uppercase(self):
         [gc] = self.parse("@swashes = [X.swash-Z.swash];").statements
         self.assertEqual(gc.name, "swashes")