[MLton] Unicode / WideChar

Henry Cejtin henry.cejtin@sbcglobal.net
Sun, 20 Nov 2005 23:06:43 -0600

So a total of 2 tests: one to see if the wide char is in 0..128 and, if it
is, a second test on the result of the table lookup.
It is still an extra test (compared to the 1 meg lookup) but probably not
bad at all since that first test (in the case that all the chars are really
ASCII) will always be predicted correctly.