Re: speed up unicode decomposition and recomposition

+   # Then the second
+   return -1 if $a2 < $b2;
+   return 1 if $a2 > $b2;
Should say "second code point" here?

+       hashkey = pg_hton64(((uint64) start << 32) | (uint64) code);
+       h = recompinfo.hash(&hashkey);
This choice should be documented, and most likely we should have
comments on the perl and C sides to keep track of the relationship
between the two.

<separate headers>