-/* Flags for node->flags of ANYOF. These are in short supply, so some games
- * are done to share them, as described below. Already, the ANYOF_LOCALE and
- * ANYOF_CLASS bits are shared, making a space penalty for all locale nodes.
- * An option would be to push them into new nodes. E.g. there could be an
- * ANYOF_LOCALE node that would be in place of the flag of the same name. But
- * there are better options. The UNICODE_ALL bit could be freed up by
- * resorting to creating a swash containing everything above 255. This
- * introduces a performance penalty. Better would be to split it off into a
- * separate node, which actually would improve performance by allowing adding a
- * case statement to regexec.c use the bit map for code points under 256, and
- * to match everything above. If flags need to be added that are applicable to
- * the synthetic start class only, with some work, they could be put in the
- * next-node field, or in an unused bit of the classflags field. This could be
- * done with the current EOS flag, and a new node type created that is just for
- * the scc, freeing up that bit */
+/* Flags for node->flags of ANYOF. These are in short supply, but there is one
+ * currently available. If more than this are needed, the ANYOF_LOCALE and
+ * ANYOF_CLASS bits could be shared, making a space penalty for all locale nodes.
+ * Also, the UNICODE_ALL bit could be freed up by resorting to creating a swash
+ * containing everything above 255. This introduces a performance penalty.
+ * Better would be to split it off into a separate node, which actually would
+ * improve performance a bit by allowing regexec.c to test for a UTF-8
+ * character being above 255 without having to call a function nor calculate
+ * its code point value. However, this solution might need to have a second
+ * node type, ANYOF_SYNTHETIC_ABOVE_LATIN1_ALL */