Hi, new patterns to implement overflow arithmetics were added recently (addv4, subv4, mulv4, umulv4, negv3) and they are not documented in the internal manual as far as I can see. Granted, the old patterns (addv3, subv3, mulv3, divv3, negv2) are not documented either but they were not very useful, contrary to the new ones. Any particular reason not to document them?
-- Eric Botcazou
