blanchet [Fri, 20 Aug 2010 16:22:51 +0200] rev 38615
improve "x = A | x = B | x = C"-style axiom detection
blanchet [Fri, 20 Aug 2010 15:56:00 +0200] rev 38614
temporarily disable "fequal" handling in Metis;
one proof fails in "HOL-Auth" because of the new rules
blanchet [Fri, 20 Aug 2010 15:16:27 +0200] rev 38613
use "hypothesis" rather than "conjecture" for hypotheses in TPTP format;
avoids relying on inconsistently implemented feature of TPTP format
blanchet [Fri, 20 Aug 2010 14:18:55 +0200] rev 38612
merged
blanchet [Fri, 20 Aug 2010 14:15:29 +0200] rev 38611
improve "x = A | x = B | x = C"-style fact discovery
blanchet [Fri, 20 Aug 2010 14:09:02 +0200] rev 38610
transform elim theorems before filtering "bool" and "prop" variables out;
another consequence of the transition to FOF
blanchet [Fri, 20 Aug 2010 13:39:47 +0200] rev 38609
more fiddling with Sledgehammer's "add:" option
blanchet [Fri, 20 Aug 2010 10:58:01 +0200] rev 38608
beta eta contract the Sledgehammer conjecture (and also the axioms, although this might not be needed), just like Metis does (implicitly);
another consequence of the transition to FOF
blanchet [Thu, 19 Aug 2010 19:34:11 +0200] rev 38607
generalize the "too general equality" code to handle facts like "x ~= A ==> x = B"
blanchet [Thu, 19 Aug 2010 18:16:47 +0200] rev 38606
encode "fequal" reasoning rules in Metis problem, just as is done for Sledgehammer -- otherwise any proof that relies on "fequal" found by Sledgehammer can't be reconstructed