diff options
author | Son Ho | 2022-02-13 20:40:31 +0100 |
---|---|---|
committer | Son Ho | 2022-02-13 20:40:31 +0100 |
commit | bd73cf946a2e51e07137a64602479755ca56f35b (patch) | |
tree | 91c4519b21922facaefc788cc2380b109b07a79d | |
parent | a2b032dbcfa1d001e6c908d7bf0e605515f41e0c (diff) |
Add some comments
Diffstat (limited to '')
-rw-r--r-- | tests/hashmap/Hashmap.Properties.fst | 15 |
1 files changed, 12 insertions, 3 deletions
diff --git a/tests/hashmap/Hashmap.Properties.fst b/tests/hashmap/Hashmap.Properties.fst index a6becc96..ad9b39e1 100644 --- a/tests/hashmap/Hashmap.Properties.fst +++ b/tests/hashmap/Hashmap.Properties.fst @@ -151,6 +151,10 @@ open Hashmap.Funs /// directly given by some postconditions which failed for no reasons, or /// "unknown assertion failed" which forced us to manually unfold postconditions... /// +/// 4. As usual, the unstable arithmetic proofs are a lot of fun. We have a few +/// of them because we prove that the table is never over-loaded (it resizes itself +/// in order to respect the max load factor). See [new_max_load_lem] for instance. +/// /// Finally: remember (again) that we are in a pure setting. Imagine having to /// deal with Low*/separation logic at the same time. /// @@ -542,7 +546,10 @@ let slots_s_inv (#t : Type0) (slots : slots_s t{length slots <= usize_max}) : Ty {:pattern index slots i} slot_s_inv (length slots) i (index slots i) -// TODO: write that in terms of slots_s_inv +// At some point we tried to rewrite this in terms of [slots_s_inv]. However it +// made a lot of proofs fail because those proofs relied on the [index_map_lem] +// pattern. We tried writing others lemmas with patterns (like [slots_s_inv] +// implies [slots_t_inv]) but it didn't succeed, so we keep things as they are. let slots_t_inv (#t : Type0) (slots : slots_t t{length slots <= usize_max}) : Type0 = forall(i:nat{i < length slots}). {:pattern index slots i} @@ -2474,7 +2481,6 @@ let times_divid_lem (n m p : pos) : Lemma ((n * m) / p >= n * (m / p)) #pop-options /// The good old arithmetic proofs and their unstability... -/// It seems OK now that I moved an assertion (I ran `quake 100`). #restart-solver #push-options "--z3rlimit 200 --z3cliopt smt.arith.nl=false" let new_max_load_lem @@ -2497,7 +2503,10 @@ let new_max_load_lem let ncapacity = 2 * capacity in let nmax_load = (ncapacity * divid) / divis in FStar.Math.Lemmas.paren_mul_left 2 capacity divid; - assert(2 * (capacity * divid) == 2 * capacity * divid); // Often broke (though given by lemma above): we moved it up + // The following assertion often breaks though it is given by the above + // lemma. I really don't know what to do (I deactivated non-linear + // arithmetic and added the previous lemma, moved the lemma up...). + assert(2 * (capacity * divid) == 2 * capacity * divid); assert(nmax_load = (2 * capacity * divid) / divis); times_divid_lem 2 (capacity * divid) divis; assert((2 * (capacity * divid)) / divis >= 2 * ((capacity * divid) / divis)); |