Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

MLCFG error when deconstructing a tuple which also is used in a {pre, post}condition #989

Open
sarsko opened this issue Apr 7, 2024 · 1 comment

Comments

@sarsko
Copy link
Contributor

sarsko commented Apr 7, 2024

This is a regression, meaning that the following pattern previously worked in CreuSAT, but now gives an MLCFG error.

The following code:

use creusot_contracts::*;

#[logic]
#[open(self)] // Opacity doesn't seem to matter
fn deconstruct(f2: (bool, Int)) {}

#[logic]
#[open(self)]
#[requires(just_true(f))] // can also be `ensures`
pub fn entry(
    f: (bool, Int)
) {
    deconstruct((f.0, f.1));
}

#[predicate]
#[open(self)]
fn just_true(f: (bool, Int)) -> bool {
    true
}

Gives:

File "../vt2024-rlib.mlcfg", line 17, characters 64-70: This term has type (int, int), but is expected to have type (bool, int)

Produced MLCFG:

module Vt2024_Entry_Impl
  use prelude.Int
  function deconstruct0 [#"../../../src/lib.rs" 5 0 5 31] (f2 : (bool, int)) : () =
    [#"../../../src/lib.rs" 3 0 3 8] ()
  val deconstruct0 [#"../../../src/lib.rs" 5 0 5 31] (f2 : (bool, int)) : ()
    ensures { result = deconstruct0 f2 }
    
  predicate just_true0 [#"../../../src/lib.rs" 18 0 18 36] (f : (bool, int)) =
    [#"../../../src/lib.rs" 19 4 19 8] true
  val just_true0 [#"../../../src/lib.rs" 18 0 18 36] (f : (bool, int)) : bool
    ensures { result = just_true0 f }
    
  constant f  : (bool, int)
  function entry [#"../../../src/lib.rs" 10 0 12 1] (f : (bool, int)) : ()
  goal vc_entry : ([#"../../../src/lib.rs" 9 11 9 23] just_true0 f)
   -> (let (a, _) = f in let (_, a) = f in let _ = deconstruct0 (a, a) in true)
end

Some notes:

  1. Requires deconstructing both fields of the tuple.
  2. Tuple type does not matter
  3. Opacity does not seem to matter
  4. The requires(just_true(f)) can also be an ensures.
  5. Moving requires(just_true(f)) to be a just_true(f); call in the body does not run into issues.
@xldenis
Copy link
Collaborator

xldenis commented Apr 7, 2024

Ahhh this seems to be an issue with name hygiene in places, no?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants