blanchet [Mon, 04 Aug 2014 12:28:42 +0200] rev 57776
rationalized sorting of facts -- so that preplaying (almost always) coincides with the real thing, preventing odd failures
blanchet [Mon, 04 Aug 2014 11:54:23 +0200] rev 57775
slightly earlier exit from preplaying
blanchet [Mon, 04 Aug 2014 11:43:19 +0200] rev 57774
honor 'dont_minimize' option when preplaying one-liner proof
sultana [Sun, 22 Jun 2014 06:16:57 +0100] rev 57773
Metis is being used to emulate E steps;
sultana [Sun, 22 Jun 2014 06:16:56 +0100] rev 57772
updated application of print_tac to take context parameter;
blanchet [Sat, 02 Aug 2014 00:15:08 +0200] rev 57771
better duplicate detection
blanchet [Fri, 01 Aug 2014 23:58:42 +0200] rev 57770
normalize conjectures vs. negated conjectures when comparing terms
blanchet [Fri, 01 Aug 2014 23:33:43 +0200] rev 57769
tweaked 'clone' formula detection
blanchet [Fri, 01 Aug 2014 23:29:50 +0200] rev 57768
fine-tuned Isar reconstruction, esp. boolean simplifications
blanchet [Fri, 01 Aug 2014 23:29:49 +0200] rev 57767
centralized boolean simplification so that e.g. LEO-II benefits from it
blanchet [Fri, 01 Aug 2014 20:44:51 +0200] rev 57766
careful when compressing 'obtains'
blanchet [Fri, 01 Aug 2014 20:44:29 +0200] rev 57765
better handling of variable names
blanchet [Fri, 01 Aug 2014 20:15:41 +0200] rev 57764
try to get rid of skolems first
blanchet [Fri, 01 Aug 2014 20:08:50 +0200] rev 57763
nicer generated variable names
blanchet [Fri, 01 Aug 2014 19:44:18 +0200] rev 57762
tuning
blanchet [Fri, 01 Aug 2014 19:36:23 +0200] rev 57761
tuning
blanchet [Fri, 01 Aug 2014 19:32:46 +0200] rev 57760
no need to 'obtain' variables not in formula
blanchet [Fri, 01 Aug 2014 19:32:10 +0200] rev 57759
more precise handling of LEO-II skolemization
blanchet [Fri, 01 Aug 2014 16:07:34 +0200] rev 57758
beware of 'skolem' rules that do not skolemize (e.g. LEO-II)
blanchet [Fri, 01 Aug 2014 16:07:33 +0200] rev 57757
tuning
blanchet [Fri, 01 Aug 2014 16:07:33 +0200] rev 57756
peek instead of joining -- is perhaps less risky
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57755
export ML function
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57754
compile
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57753
removed 'metisFT' support in Mirabelle
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57752
removed Mirabelle minimization code
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57751
modernized Mirabelle (a bit) and made it compile
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57750
restored a bit of laziness
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57749
reorder quantifiers to ease Z3 skolemization
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57748
tuned order of arguments
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57747
tuned name context code
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57746
tuned whitespace
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57745
more rational unskolemizing of names
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57744
added appropriate method for skolemization of Z3 steps to the mix
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57743
pushing skolems under 'iff' sometimes breaks things further down the proof (as was to be feared)
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57742
honor 'try0' also for one-liners
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57741
tentatively took out 'fastforce' from the set of tried methods -- it seems to be largely subsumed and is hard to silence
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57740
further minimize one-liner
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57739
tuning
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57738
eliminated needlessly complex message tail
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57737
updated NEWS
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57736
update documentation after removal of 'min' subcommand
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57735
eliminated Sledgehammer's "min" subcommand (and lots of complications in the code)
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57734
rationalized preplaying by eliminating (now superfluous) laziness
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57733
removed proof methods as provers from docs
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57732
simplified minimization logic
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57731
tuning
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57730
remove lambda-lifting related assumptions from generated Isar proofs
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57729
whitespace tuning
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57728
remove YXML formatting when parsing backquoted facts supplied manually to Sledgehammer
blanchet [Fri, 01 Aug 2014 14:43:57 +0200] rev 57727
generate backquotes without markup, since this confuses preplay; bump up spying version identifier;
traytel [Thu, 31 Jul 2014 13:19:57 +0200] rev 57726
simplified tactics slightly
blanchet [Thu, 31 Jul 2014 00:45:55 +0200] rev 57725
cascading timeout in parallel evaluation, to rapidly find optimum
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57724
put faster proof methods first
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57723
use parallel preplay machinery also for one-line proofs
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57722
updated docs
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57721
always minimize Sledgehammer results by default
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57720
tuned ML function name
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57719
reduced preplay timeout to 1 s
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57718
added more proof methods for one-liners
blanchet [Wed, 30 Jul 2014 23:52:56 +0200] rev 57717
unlift before uncombine, because the definition of a lambda-lifted symbol might have an SK combinator in it (in hybrid encodings)
fleury [Wed, 30 Jul 2014 14:03:58 +0200] rev 57716
Improving robustness and indentation corrections.
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57715
Skolemization for tmp_ite_elim rule in the SMT solver veriT.
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57714
Changing the role of rule "tmp_ite_elim" of the SMT solver veriT to Lemma.
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57713
Whitespace and indentation correction.
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57712
Simplifying the labels in the proof of the SMT solver veriT.
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57711
Changing ~ into - for unuary minus (not supported by veriT)
fleury [Wed, 30 Jul 2014 14:03:13 +0200] rev 57710
imported patch satallax_skolemization_in_tree_part
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57709
imported patch hilbert_choice_support
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57708
veriT changes for lifted terms, and ite_elim rules.
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57707
imported patch satallax_proof_support_Sledgehammer
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57706
Basic support for the higher-order ATP Satallax.
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57705
Subproofs for the SMT solver veriT.
fleury [Wed, 30 Jul 2014 14:03:12 +0200] rev 57704
Basic support for the SMT prover veriT.
fleury [Wed, 30 Jul 2014 14:03:11 +0200] rev 57703
removing the '= True' generated by Leo-II.
fleury [Wed, 30 Jul 2014 14:03:11 +0200] rev 57702
Skolemization support for leo-II and Zipperposition.
desharna [Wed, 30 Jul 2014 10:50:30 +0200] rev 57701
document property 'set_induct'
desharna [Wed, 30 Jul 2014 10:50:28 +0200] rev 57700
generate 'set_induct' theorem for codatatypes
blanchet [Wed, 30 Jul 2014 00:50:41 +0200] rev 57699
also try 'metis' with 'full_types'
blanchet [Tue, 29 Jul 2014 23:39:35 +0200] rev 57698
header tuning
blanchet [Mon, 28 Jul 2014 10:57:33 +0200] rev 57697
correctly translate THF functions from terms to types
blanchet [Sun, 27 Jul 2014 21:11:35 +0200] rev 57696
do not embed 'nat' into 'int's in 'smt2' method -- this is highly inefficient and decreases the Sledgehammer success rate significantly
wenzelm [Sun, 27 Jul 2014 15:44:08 +0200] rev 57695
back to post-release mode -- after fork point;
wenzelm [Sun, 27 Jul 2014 15:29:42 +0200] rev 57694
tuned;
wenzelm [Sun, 27 Jul 2014 15:25:00 +0200] rev 57693
tuned;
wenzelm [Sat, 26 Jul 2014 19:19:19 +0200] rev 57692
no -optimise -- produces bad bytecode;
wenzelm [Sat, 26 Jul 2014 14:52:54 +0200] rev 57691
output state first -- avoid fluctuation wrt. warnings, errors, etc.;
wenzelm [Fri, 25 Jul 2014 21:44:03 +0200] rev 57690
tuned comment;
wenzelm [Fri, 25 Jul 2014 21:29:12 +0200] rev 57689
updated to polyml-5.5.2-1 which addresses two hard crashes;
wenzelm [Fri, 25 Jul 2014 20:55:57 +0200] rev 57688
updated to cygwin-20140725, which is presumably close to Cygwin 1.7.31-1;
nipkow [Fri, 25 Jul 2014 18:41:53 +0200] rev 57687
added more functions and lemmas
wenzelm [Fri, 25 Jul 2014 17:13:30 +0200] rev 57686
proper mkdir;
wenzelm [Fri, 25 Jul 2014 16:58:28 +0200] rev 57685
proper option -O;
wenzelm [Fri, 25 Jul 2014 16:50:49 +0200] rev 57684
some reshuffling of Poly/ML version to evade failing tests;
wenzelm [Fri, 25 Jul 2014 15:01:18 +0200] rev 57683
old 'defs' is legacy --- slightly odd side-entry that bypasses regular Local_Theory.define interface;
wenzelm [Fri, 25 Jul 2014 14:47:38 +0200] rev 57682
proper volume name, such that background image is found in /Volumes/Isabelle/.background;
wenzelm [Fri, 25 Jul 2014 14:16:39 +0200] rev 57681
tuned;
wenzelm [Fri, 25 Jul 2014 14:15:02 +0200] rev 57680
tuned message;
wenzelm [Fri, 25 Jul 2014 14:13:19 +0200] rev 57679
setup for drag-and-drop DMG;
blanchet [Fri, 25 Jul 2014 13:22:37 +0200] rev 57678
merge
blanchet [Fri, 25 Jul 2014 13:15:50 +0200] rev 57677
reordered provers
blanchet [Fri, 25 Jul 2014 12:22:18 +0200] rev 57676
compile
blanchet [Fri, 25 Jul 2014 12:20:48 +0200] rev 57675
faster minimization by not adding facts that are already in the simpset
blanchet [Fri, 25 Jul 2014 11:31:20 +0200] rev 57674
added missing facts to proof method
blanchet [Fri, 25 Jul 2014 11:26:23 +0200] rev 57673
don't lose 'minimize' flag before it reaches Isar proof text generation
blanchet [Fri, 25 Jul 2014 11:26:19 +0200] rev 57672
tuning
blanchet [Fri, 25 Jul 2014 11:26:17 +0200] rev 57671
avoid 'eproof' and 'eproof_ram' scripts if possible (i.e. if 'eprover' can produce reasonable enough proofs for one-liner reconstruction)
blanchet [Fri, 25 Jul 2014 11:26:11 +0200] rev 57670
compile
blanchet [Fri, 25 Jul 2014 11:26:11 +0200] rev 57669
more robustness in Isar proof construction
blanchet [Fri, 25 Jul 2014 11:26:10 +0200] rev 57668
tuning
wenzelm [Thu, 24 Jul 2014 23:18:01 +0200] rev 57667
merged
wenzelm [Thu, 24 Jul 2014 23:17:26 +0200] rev 57666
more elementary exception handling: evade hard crash of (Runtime.thread true undefined) on Poly/ML 5.5.1 and 5.5.2;
blanchet [Thu, 24 Jul 2014 23:01:23 +0200] rev 57665
tuned code
kuncar [Thu, 24 Jul 2014 20:21:59 +0200] rev 57664
having extra assumptions (typically from a context) means there is no chance to have a valid code equation => skip decoding and registration of the code equations
kuncar [Thu, 24 Jul 2014 20:21:34 +0200] rev 57663
store explicitly quotient types with no_code => more precise registration of code equations
blanchet [Thu, 24 Jul 2014 19:01:06 +0200] rev 57662
don't needlessly regenerate entire file when the time stamps are equal
blanchet [Thu, 24 Jul 2014 18:53:14 +0200] rev 57661
eliminated source of 'DUP's in MaSh
blanchet [Thu, 24 Jul 2014 18:46:38 +0200] rev 57660
fixed sorting (broken since 9cc802a8ab06)
blanchet [Thu, 24 Jul 2014 18:46:38 +0200] rev 57659
reenabled MaSh for Isabelle2014 release (hopefully)
blanchet [Thu, 24 Jul 2014 18:46:38 +0200] rev 57658
beware of duplicate fact names
blanchet [Thu, 24 Jul 2014 18:46:38 +0200] rev 57657
refined filter for ATP steps to avoid 'have True' steps in E proofs