Sun, 23 Nov 2008 18:37:56 +0100 wenzelm tuned;
Sun, 23 Nov 2008 17:27:15 +0100 wenzelm eliminated finish_proof, keep pre/post normalization of results separate;
Sun, 23 Nov 2008 17:25:56 +0100 wenzelm future: norm_proof after make_result reduces memory requirements but increases runtime (factor 1.5 for both for HOL with proof terms);
Fri, 21 Nov 2008 18:02:19 +0100 ballarin Regression tests for new locale implementation.
Fri, 21 Nov 2008 18:01:39 +0100 ballarin add_locale functional.
Fri, 21 Nov 2008 15:54:53 +0100 paulson Added a line that was missing from the definition
Fri, 21 Nov 2008 14:21:42 +0100 krauss added binary logarithm
Fri, 21 Nov 2008 13:17:43 +0100 paulson Strange. The proof worked in the 2008 release. In order to make it work now, the last line of the proof must be moved up two places. In other words, the first proof step is now returning its subgoals in a different order from before.
(0) -10000 -3000 -1000 -300 -100 -30 -10 -8 +8 +10 +30 +100 +300 +1000 +3000 +10000 +30000 tip