GitLab

for pi by the best possible bounds in double-precision IEEE-754
floating-point numbers

  + size -> size_bv
  + size_int -> size
  + change two_power_size and max_int definitions
  + add axioms to BVConverter
  + new axiom relating nth and nth_bv
  + some reorganisation
- update coq realisation
- modify in consequence the relevant examples and pull the completed ones out of in_progress

update test suite

add out of bound axiom for nth

test eq_sub

update test suite

New function choose added in bag theory
New axiom Card_nonneg added in bag theory

this is consistent with other, similar theories

new operation set on sequences (with syntax [<-])

Add int.NumOf realization.

- generation in the makefile
- proofs done
Remove complex axioms from Pow2int in bv.why.
Add guards in rotates axioms in bv.why.

this is a first draft (with no Coq realization for the moment)
see the comment at the end of the file for discussion

- Switch from AUFNIRA to AUFBVNIRA logic in cvc4_bare

Module bitvec

no more need for cloning
similar change in array.NumOf and array.NumOfEq
updated proofs

Adding mathematical functions (+ - / *...) and comparison predicates to the theory of bit-vector and cvc4 driver (based on smt2-lib theory of bit-vector).

Coq realization remains to be updated...