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

Add saw-script functions term_eval and term_eval_unint. #927

Open
wants to merge 6 commits into
base: master
Choose a base branch
from

Conversation

brianhuffman
Copy link
Contributor

These are analogous to the goal_eval and goal_eval_unint proof
tactics, but instead of proof goals they work on arbitrary terms.

@brianhuffman
Copy link
Contributor Author

We might want to reconsider the naming of these operations.

Brian Huffman and others added 3 commits May 29, 2022 13:45
Currently this only works for first-order values; function types
are not supported yet.
These are analogous to the `goal_eval` and `goal_eval_unint` proof
tactics, but instead of proof goals they work on arbitrary terms.
@RyanGlScott
Copy link
Contributor

Landing this would address one aspect of #1788.

@sauclovian-g
Copy link
Contributor

I think the conclusion is that we want this functionality, and I see no reason to start over rather than taking this code. So I'm going to see if it's still buildable.

@sauclovian-g
Copy link
Contributor

Answer: it wasn't, but was easily fixed. I'm going to un-draft it, I think, at least until we identify any reasons it's not ready. @RyanGlScott do you have any examples we could add into the test suite? I am pretty sure I understand what it's doing and why but concocting a simple example involves too many things I'm not yet quite on top of.

@sauclovian-g sauclovian-g marked this pull request as ready for review November 14, 2024 23:04
@sauclovian-g sauclovian-g assigned sauclovian-g and unassigned chameco Nov 14, 2024
@RyanGlScott
Copy link
Contributor

Two potential test cases that come to mind:

  1. The test case from Bitblasting of multipliers differs wildly between Cryptol and SAW #1788. But upon closer inspection, the original problem from Bitblasting of multipliers differs wildly between Cryptol and SAW #1788 can no longer be reproduced in modern SAW, as this was fixed in a different way in write_aig: Normalize symmetry in multiplication #1833. As such, this is no longer a good test case...

  2. The test case from muxBVal: VInt (0, 1) #1436. This can still be reproduced in modern SAW:

    // test.cry
    hasRepeat : {a, b} (fin a, a >= 2, Eq b) => [a]b -> b -> Integer -> Bit
    hasRepeat xs b num = out == num
      where
        out = foldl (\a x -> if a == num then num
                              | x == b   then a+1
                             else 0) 0 xs
    
    // test.saw
    import "test.cry";
    write_aig "test.aig" {{ \(x : [10]) -> hasRepeat x True 2 }};
    
    $ ./bin/saw test.saw
    [14:39:36.840] Loading file "/home/ryanscott/Documents/Hacking/Haskell/saw-script/test.saw"
    [14:39:36.890] Stack trace:
    "write_aig" (/home/ryanscott/Documents/Hacking/Haskell/saw-script/test.saw:3:1-3:10)
    muxBVal: VInt (1,0)
    
    

    In principle, one ought to be able to fix this by adding a use of term_eval, i.e.,

    // test.saw
    import "test.cry";
    write_aig "test.aig" (term_eval {{ \(x : [10]) -> hasRepeat x True 2 }});
    

    Unfortunately, that doesn't work either:

    $ ./bin/saw test.saw
    [14:40:24.864] Loading file "/home/ryanscott/Documents/Hacking/Haskell/saw-script/test.saw"
    [14:40:24.916] Stack trace:
    "term_eval" (/home/ryanscott/Documents/Hacking/Haskell/saw-script/test.saw:3:23-3:32)
    rebuildTerm VFun
    
    

    Ideally, we would teach rebuildTerm how to look underneath lambdas, as I imagine that that will be a pretty common use case for this feature. I don't have time to look at this right now, but I can't imagine it would be too difficult, given that other Term-manipulation primitives also look underneath lambdas (e.g., normalize_term).

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

Successfully merging this pull request may close these issues.

4 participants