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

Updates to Cryptol Monadification (+ add qsort to saw-core) #1844

Merged
merged 20 commits into from
Mar 28, 2023

Conversation

m-yac
Copy link
Contributor

@m-yac m-yac commented Mar 17, 2023

This PR mainly makes the following changes to Cryptol Monadification:

  • Add a write_coq_cryptol_module_monadic command which monadifies all definitions in a Cryptol module and writes them to a Coq file
  • Always monadify all contained constants, recursively (this was partially implemented before, but it did not always find all contained constants – see 37ba742)
  • Write CryptolM.sawcore to a Coq file in generate_scaffolding.saw
  • Resolve all remaining FIXMEs in CryptolM.sawcore. This mostly involved the mostly mechanical task of translating the Cryptol.sawcore definitions into the SpecM monad (which I did while waiting for things to compile) but also involved adding an uncurryMacro and generalizing the finMacro in Monadify.hs

However, it also adds a new type of sort, qsort, to saw-core. Like the existing isort, it is ignored during saw-core typechecking, but when translated to Coq adds an implicit typeclass argument. For isort this is the Inhabited typeclass, but for qsort this is the QuantType typeclass (needed for the existsS combinator and the CryptolM functions that use it, such as vecMapM and the number functions that in turn use that). This involved:

  • Replacing the Bool in the Sort constructor of FlatTermF (which differentiated sort vs. isort) with a new record type SortFlags, which has two Bools: flagInhabited and flagQuantType.
  • Updating the typed and untyped ASTs as well as the external format to account for replacing this Bool with this new type
  • Updating the Lexer and Grammar to add the new qsort and qisort keywords
  • Updating many places where the default False value for the old Bool flag was used with the new noFlags value of the record where all fields are False

An advantage of making a new SortFlags type instead of just adding another Bool flag is that in the future if we want to add another such flag, it should only involve adding a new field to SortFlags, updating the Coq translation, and updating the Lexer and Grammar. The changes to the external format as well as the many places False was replaced with noFlags should not need to happen again.

@m-yac m-yac added subsystem: saw-core Issues related to the saw-core representation or the saw-core subsystem subsystem: MRSolver Issues related to the Mr. Solver monadic-recursive solver in Heapster labels Mar 17, 2023
@eddywestbrook
Copy link
Contributor

My initial thought is that I don't think we want an exponential explosion of different kinds of sorts. That is, we don't want isort, qsort, and also qisort, because that adds complexity that we don't need. Instead, the purpose of all of these is really just to express that a type comes from Cryptol. So let's have a single name that expresses that, i.e., why not just call it crysort. Let's also have a single typeclass in Coq as well, maybe called CryptolType, that has all the functionality we need for a Cryptol type. For now that's just that it's Inhabited and is a QuantType, but in the future it might be more, and this way we can just add to the CryptolType class and not change SAW core at all rather than having to add yet another Bool flag to sort.

@m-yac
Copy link
Contributor Author

m-yac commented Mar 20, 2023

@eddywestbrook Oh, sorry for misunderstanding you! The worry I have with always bundling QuantType and Inhabited is that sometimes we might want to quantify over a Prop which we don't know is inhabited. If we weren't able to separate the two, this would force us to only use existsS over Props which we can prove are Inhabited. I feel like I can imagine a situation where the Inhabited-ness of a Prop we're quantifying over depends on the value of variables from earlier in the monadic term. Maybe this isn't a problem in practice, though? What are your thoughts on this

@eddywestbrook
Copy link
Contributor

@m-yac I see your concern, and we certainly wouldn't want to existsS on a Prop that must be Inhabited because the Inhabited proof is a proof of that Prop! However, we already expose a separate assertS combinator in SAW, that does exactly what you are talking about, using existsS over Props. But it does that "under the hood", on the Coq side. So the solution would be to use existsS when quantifying over "actual data", in the form of Cryptol types, and to use assertP to quantify over Props (which always trivially satisfy QuantType). How does that sound?

@m-yac
Copy link
Contributor Author

m-yac commented Mar 21, 2023

Update: We realized what we really want here is some sort of typeclass system in saw-core (#1845). At least, that was the best solution we could think of that was equally as easy to use as the current solution of having special sorts. Based on Eddy's comment we could simplify the hierarchy a little bit by making qsort imply isort, but we're just going to leave this as-is.

Copy link
Contributor

@eddywestbrook eddywestbrook left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This mostly looks great! Please perform the small doc changes I requested. I'll mark as "approve" assuming that you'll make those changes and merge yourself.

saw-core/src/Verifier/SAW/Term/Functor.hs Show resolved Hide resolved
saw-core/src/Verifier/SAW/Term/Functor.hs Outdated Show resolved Hide resolved
src/SAWScript/Prover/Exporter.hs Show resolved Hide resolved
@m-yac m-yac added PR: ready to merge Magic flag for pull requests to ask Mergify to merge given an approval and a successful CI run and removed PR: ready to merge Magic flag for pull requests to ask Mergify to merge given an approval and a successful CI run labels Mar 23, 2023
@KarmaHaze

This comment was marked as abuse.

KarmaHaze

This comment was marked as spam.

@m-yac m-yac added the PR: ready to merge Magic flag for pull requests to ask Mergify to merge given an approval and a successful CI run label Mar 28, 2023
@m-yac m-yac merged commit ecab2cd into master Mar 28, 2023
@m-yac m-yac deleted the monadify-coq-module branch March 28, 2023 16:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
PR: ready to merge Magic flag for pull requests to ask Mergify to merge given an approval and a successful CI run subsystem: MRSolver Issues related to the Mr. Solver monadic-recursive solver in Heapster subsystem: saw-core Issues related to the saw-core representation or the saw-core subsystem
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants