 30 Mar, 2016 1 commit


Robbert Krebbers authored
This seems to shorten type class search.

 29 Mar, 2016 1 commit


Robbert Krebbers authored

 23 Mar, 2016 2 commits


Ralf Jung authored

Robbert Krebbers authored

 22 Mar, 2016 1 commit


Ralf Jung authored

 21 Mar, 2016 2 commits


Robbert Krebbers authored
Also, slightly reorganize.

Robbert Krebbers authored

 11 Mar, 2016 1 commit


Robbert Krebbers authored

 10 Mar, 2016 1 commit


Robbert Krebbers authored
Thanks to Amin Timany for the suggestion.

 05 Mar, 2016 2 commits
 04 Mar, 2016 3 commits


Robbert Krebbers authored

Robbert Krebbers authored

Robbert Krebbers authored

 03 Mar, 2016 3 commits


Robbert Krebbers authored

Robbert Krebbers authored
Contrary to destruct_conj from Program.

Robbert Krebbers authored

 02 Mar, 2016 1 commit


Ralf Jung authored

 27 Feb, 2016 1 commit


Robbert Krebbers authored

 26 Feb, 2016 1 commit


Robbert Krebbers authored

 25 Feb, 2016 11 commits


Ralf Jung authored

Robbert Krebbers authored
It now turns setoid equalities into Leibniz equalities when possible, and substitutes those.

Robbert Krebbers authored

Robbert Krebbers authored

Ralf Jung authored
In principle, we could now unseal heap_mapsto, saved_prop_own etc., and mark them as "Typeclass Opaque", and ecancel would still work just as fast as it does now. Thanks to Matthieu for pointing me to this unify feature.

Ralf Jung authored

Ralf Jung authored

Ralf Jung authored

Robbert Krebbers authored
Conflicts: prelude/tactics.v

Robbert Krebbers authored

Ralf Jung authored
This replaces f_equiv and solve_proper with our own, hopefully better, versions

 24 Feb, 2016 5 commits


Robbert Krebbers authored
It now traverses terms at most once, whereas the setoid_rewrite approach was travering terms many times. Also, the tactic can now be extended by defining type class instances.

Robbert Krebbers authored

Robbert Krebbers authored

Ralf Jung authored

Robbert Krebbers authored
Also, use "set_solver by tac" to specify a tactic.

 23 Feb, 2016 1 commit


Ralf Jung authored

 22 Feb, 2016 3 commits


Robbert Krebbers authored
And now the part that I forgot to commit.

Robbert Krebbers authored

Robbert Krebbers authored
In most cases there is a lot of duplicate proof search performed by both naive_solver and eauto. Especially since naive_solver calls its tactic (in the case of set_solver this used to be eauto) quite eagerly this made it very slow. Note that set_solver is this too slow and should be improved.
