-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
SmallStep/SmallStepDepRec: interpreter + compiler via explicit recurs…
…ion.
- Loading branch information
1 parent
2393c56
commit a7d8ddd
Showing
1 changed file
with
106 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,106 @@ | ||
-- | ||
-- Small-step operational semantics | ||
-- making use of dependent types | ||
-- | ||
|
||
{- | ||
This (slightly modified) code is from | ||
Proof by Smugness | ||
by Conor on August 7, 2007. | ||
https://fplab.bitbucket.org/posts/2007-08-07-proof-by-smugness.html | ||
The same stuff as in SmallStepDep, but via explicit recursion. | ||
-} | ||
|
||
module SmallStepDepRec where | ||
|
||
open import Data.Nat | ||
using (ℕ; _+_) | ||
open import Data.Vec | ||
using (Vec; []; _∷_) | ||
open import Function using (_∘_) | ||
open import Relation.Binary.PropositionalEquality | ||
|
||
-- | ||
-- A Toy Language | ||
-- | ||
|
||
infixl 6 _⊕_ | ||
|
||
data Tm : Set where | ||
val : (n : ℕ) → Tm | ||
_⊕_ : (t1 t2 : Tm) → Tm | ||
|
||
-- Big-step evaluator | ||
|
||
eval : Tm → ℕ | ||
eval (val n) = n | ||
eval (t1 ⊕ t2) = eval t1 + eval t2 | ||
|
||
-- | ||
-- Virtual machine | ||
-- | ||
|
||
-- Program | ||
|
||
-- The idea is to index code by initial and final stack depth | ||
-- in order to ensure stack safety. | ||
|
||
data Code : (i j : ℕ) → Set where | ||
seq : ∀ {i j k} (c1 : Code i j) (c2 : Code j k) → Code i k | ||
push : ∀ {i} (n : ℕ) → Code i (1 + i) | ||
add : ∀ {i} → Code (2 + i) (1 + i) | ||
|
||
-- State | ||
|
||
Stack : ℕ → Set | ||
Stack i = Vec ℕ i | ||
|
||
-- Interpreter | ||
|
||
exec : ∀ {i j} (c : Code i j) (s : Stack i) → Stack j | ||
exec (seq c1 c2) s = exec c2 (exec c1 s) | ||
exec (push n) s = n ∷ s | ||
exec add (n2 ∷ n1 ∷ s) = (n1 + n2) ∷ s | ||
|
||
-- Compiler | ||
|
||
compile : ∀ {i} (t : Tm) → Code i (1 + i) | ||
compile (val n) = push n | ||
compile (t1 ⊕ t2) = seq (seq (compile t1) (compile t2)) add | ||
|
||
-- `compile` is correct with respect to `exec` | ||
|
||
correct : ∀ {i} (t : Tm) (s : Stack i) → | ||
exec {i} (compile t) s ≡ eval t ∷ s | ||
correct (val n) s = refl | ||
correct (t1 ⊕ t2) s = | ||
exec (compile (t1 ⊕ t2)) s | ||
≡⟨⟩ | ||
exec (seq (seq c1 c2) add) s | ||
≡⟨⟩ | ||
exec add (exec c2 (exec c1 s)) | ||
≡⟨ cong (exec add ∘ exec c2) (correct t1 s) ⟩ | ||
exec add (exec c2 (n1 ∷ s)) | ||
≡⟨ cong (exec add) (correct t2 (n1 ∷ s)) ⟩ | ||
exec add (n2 ∷ n1 ∷ s) | ||
≡⟨⟩ | ||
n1 + n2 ∷ s | ||
≡⟨⟩ | ||
eval (t1 ⊕ t2) ∷ s | ||
∎ | ||
where | ||
open ≡-Reasoning | ||
n1 = eval t1; n2 = eval t2 | ||
c1 = compile t1; c2 = compile t2 | ||
|
||
-- Here is another proof, which is shorter, but is more mysterious. | ||
|
||
correct′ : ∀ {i} (t : Tm) (s : Stack i) → | ||
exec {i} (compile t) s ≡ eval t ∷ s | ||
correct′ (val n) s = refl | ||
correct′ {i} (t1 ⊕ t2) s | ||
rewrite correct t1 s | correct t2 (eval t1 ∷ s) | ||
= refl |