]> matita.cs.unibo.it Git - helm.git/commit
Back-porting from new Matita:
authorClaudio Sacerdoti Coen <claudio.sacerdoticoen@unibo.it>
Mon, 29 Nov 2010 12:54:09 +0000 (12:54 +0000)
committerClaudio Sacerdoti Coen <claudio.sacerdoticoen@unibo.it>
Mon, 29 Nov 2010 12:54:09 +0000 (12:54 +0000)
commit8a218c35f9bde765a73f14867a294c874f9dc15c
treea36ab7dfc57996a13c90fa07feb25366db1d421c
parent9078eadc955930e7e37cfc2451a40416512c95fa
Back-porting from new Matita:

- GREAT: when unifying ?1 : Type[i]  with ?2: Type[j] the unifier
  randomly tried to instantiate ?1 with ?2 even when j > i, yielding to
  an unification error later on. In turn, this created that horrible
  behaviour of rewriting that failed if you did not pass enough types
  to the lemma.
- in rewriting the argument is now automatically saturated

These two fixes together allow most of the time to write simply >f as we
did in the old system.
helm/software/components/ng_refiner/nCicUnification.ml