remove subtlety whose justification got lost in time -- the new code is possibly less precise but sounder

tuning

handle higher-order occurrences of sets gracefully in model display

prefer explicit version information;

more Nitpick doc updates

reenable Kodkodi in Mira now that Nitpick has been ported to 'a set constructor

reenable Kodkodi in Isatest now that Nitpick has been ported to 'a set constructor

fixed bisimilarity axiom -- avoid "insert" with wrong type

tuning

updated Nitpick docs after "set" reintroduction

no abuse of notation

always treat "unit" as a deep datatype, so that we get a good interaction with the record syntax (2.7 of the Nitpick manual)

more robust destruction of "set Collect" idiom

handle starred predicates correctly w.r.t. "set"

handle "Id" gracefully w.r.t. "set"

reintroduced 'refute' calls taken out after reintroducing the "set" constructor, and use "expect" feature

handle "set" correctly in Refute -- inspired by old code from Isabelle2007

create consts with proper "set" types

tuned Refute

lower cardinality for faster testing

simplify mem Collect

tuning

ported Minipick to "set"

fixed set extensionality code

tuned import

construct correct "set" type for wf goal

fixed Nitpick's typedef handling w.r.t. "set"

fixed type annotations