Skip to content

Merging latest master into feature/p-token - #1058

Merged
mariaKt merged 6 commits into
feature/p-tokenfrom
mk/merge-feature-p-token
Apr 10, 2026
Merged

Merging latest master into feature/p-token#1058
mariaKt merged 6 commits into
feature/p-tokenfrom
mk/merge-feature-p-token

Conversation

@mariaKt

@mariaKt mariaKt commented Apr 10, 2026

Copy link
Copy Markdown
Contributor

Stevengre and others added 6 commits April 7, 2026 04:35
## Summary

- make the manual `test-stable-mir-ui` workflow runnable on the
self-hosted runner
- derive the Rust toolchain from repo-tracked `rust-toolchain.toml`
instead of hardcoding it in the workflow
- derive the `rust-lang/rust` checkout commit from the active
toolchain's `rustc -vV` output instead of maintaining a separate fixed
Rust source pin
- upload the real failure outputs from inside the Docker container,
including updated `skip.txt` when requested

## Context

- Stacked on #968 / `codex/import-stable-mir-ui-tests`
- The base PR adds the workflow; this PR makes that workflow usable
end-to-end in the containerized setup while keeping both the installed
toolchain and the checked-out Rust source tree aligned to the same
nightly

## Testing

- Manual validation on `zhaoji`: rebuilt the workflow image after
deriving `RUST_TOOLCHAIN=nightly-2024-11-29` from `rust-toolchain.toml`
- Manual validation on `zhaoji`: verified `make stable-mir-json`
succeeds before the UI run
- Manual validation on `zhaoji`: verified `test-filter='foo or bar'` is
passed through correctly
- Manual validation on `zhaoji`: verified proof artifacts and updated
`skip.txt` are copied back from the container
- Local container validation: verified `rustc -vV` in the workflow image
resolves `commit-hash: a2545fd6fc66b4323f555223a860c451885d1d2b` for
`nightly-2024-11-29`, and the workflow's shallow `rust-lang/rust`
checkout can `fetch` and `checkout` that commit successfully
- Local container validation: verified `make stable-mir-json` and `make
build` succeed after aligning the Rust checkout from the active
toolchain
- Local container validation: ran a single-case smoke path
(`tests/ui/issues/issue-11205.rs`) through `RUST_DIR_ROOT=rust make
test-stable-mir-ui`; workflow plumbing succeeded and the test reached
the proof stage, while the proof itself still failed with
`AssertionError: Proof did not pass`
Our PRs currently mostly fail CI due to the `Haskell Exec SMIR` CI test
timing out. I increased it from 20 to 30 minutes.
The timeout increase
[here](#1021)
is not enough (depending on the state of the GH runners, this is still a
bit too tight). I measured locally, and it took around 40 minutes, so I
added another 10 as a leeway for a total of 50 min.
The timeout limits for our CI tests are tight. Depending on the state of
the github runners, our CI often fails with just a few more minutes
needed. I have noticed test runs being cancelled while making progress,
at 90% + (even just after the test have been completed). It would be
better to have them be more generous, as having to rerun jobs cancelled
due to this is actually leading to more time on the runners.
#1057)

This reverts commit b694d02 (merge
commit of PR
[#995](#995) in
master).

It caused a regression in the spl-token initialize_multisig proofs,
where the Haskell backend tried to evaluate Int2Float (the
FLOAT.int2float hook).
@mariaKt
mariaKt marked this pull request as ready for review April 10, 2026 15:25
@mariaKt
mariaKt requested a review from dkcumming April 10, 2026 15:25

@Stevengre Stevengre left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

LGTM. My understanding is that this change only removes the float casting logic that was causing verification issues—please correct me if that’s not accurate. It might also be worth waiting for Daniel’s review before merging.

Could you share more details about the underlying issue? If possible, please also provide a minimal reproducer in a separate PR targeting master.

@mariaKt

mariaKt commented Apr 10, 2026

Copy link
Copy Markdown
Contributor Author

Here is the kaas job that shows the issue: initialize_multisig, initialize_multisig2

@mariaKt
mariaKt merged commit 9017fee into feature/p-token Apr 10, 2026
21 of 22 checks passed
@mariaKt
mariaKt deleted the mk/merge-feature-p-token branch April 10, 2026 17:23
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.

3 participants