+Also I'm going to index \i\c{-output} to ensure that its two
+components are displayed as a joined-up code fragment in the index.
+
+Here are \I{testone}some \I{testtwo}subsections \I{testthree}with
+silly chapter titles and interesting use of Unicode. The Unicode
+oddities are in the titles rather than the body text because that
+way I get to test their handling in the PDF document outline.
+
+\H{app-one} The 1024 \u00D7{x} 768 screen resolution
+
+Err.
+
+\H{app-two} How about Spongletech\u2122{(TM)}?
+
+Umm.
+
+\# I'm going to label one of these with a carefully chosen fragment
+\# name "i1", because I know this will also be generated as an index
+\# fragment name and this allows me to test the fragment name clash
+\# detection.
+\#
+\# To actually run this test you need to configure html-leaf-level
+\# to 0 and html-template-fragment to %k.
+
+\H{i1} Or just Erd\u0151{\u00F6{o}}s?
+
+Ahh.
+