wenzelm [Thu, 15 Apr 2010 21:24:00 +0200] rev 36159
more robust record syntax: use Type.raw_match to ignore sort constraints as in regular abbreviations (also note that constraints only affect operations, not types);
wenzelm [Thu, 15 Apr 2010 20:56:04 +0200] rev 36158
HOL record: explicitly allow sort constraints;
wenzelm [Thu, 15 Apr 2010 20:37:27 +0200] rev 36157
misc tuning and simplification;
wenzelm [Thu, 15 Apr 2010 20:31:21 +0200] rev 36156
explicit ProofContext.check_tfree;
wenzelm [Thu, 15 Apr 2010 18:13:25 +0200] rev 36155
merged
Cezary Kaliszyk <kaliszyk@in.tum.de> [Thu, 15 Apr 2010 16:55:12 +0200] rev 36154
Respectfullness and preservation of list_rel
wenzelm [Thu, 15 Apr 2010 18:09:22 +0200] rev 36153
replaced slightly odd Typedecl.predeclare_constraints by plain declaration of type arguments -- also avoid "recursive" declaration of type constructor, which can cause problems with sequential definitions B.foo = A.foo;
simplified via ProofContext.check_tfree;
wenzelm [Thu, 15 Apr 2010 18:00:21 +0200] rev 36152
get_sort: suppress dummyS from input;
added check_tvar, check_tfree convenience;
tuned;
wenzelm [Thu, 15 Apr 2010 16:58:12 +0200] rev 36151
modernized treatment of sort constraints in specification;
pass-through type variables as usual as (string * sort) internally -- recovers proper sort handling;
wenzelm [Thu, 15 Apr 2010 16:55:49 +0200] rev 36150
typecopy: observe given sort constraints more precisely;