Repository navigation
chore!: delete the UniFFI fork - #690
Merged
Merged
Conversation
Quest-Claim: f971856dfc57254d
moq decided never to build on it, so its only consumers here were the binding benchmark's baseline and nightly jobs. Removes uniffi/, the root Cargo workspace, the flake's uniffi* packages and .#uniffi shell, the nightly uniffi-* jobs, and the binding benchmark's harness and workflow. The benchmark's README keeps the BoltFFI comparison and the fork's sizes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…deletion # Conflicts: # justfile
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.

Deletes the UniFFI fork and everything that exists only for it, finishing
quest/m0/delete-uniffi.md. Decided with the maintainer on 2026-11-09 in the quest audit: moq never builds on the fork, so its only consumers here were the binding benchmark's baseline and CI.Changes
uniffi/(1,426 files, ~118k lines), the rootCargo.tomlandCargo.lock(the workspace held only the fork), andjust uniffi.sh/dispatch.sh: theuniffimodule and its rootcargo fmt/clippy;sh/dispatch.test.pyfollows.just fixdrops its rootcargo fmt --all.flake.nix: theuniffi,uniffi-bindgen*packages, thegeneratorhelper, the.#uniffidev shell, andflock(onlybench/ffi/run.shused it).nightly.yml: theuniffi-fixtures,uniffi-cpp(MSVC, Apple Clang), anduniffi-installjobs.bench/ffi/{run.sh,browser.sh,browser.ts,browser-page.ts,pins.toml,report.jq},bench-ffi.yml, andjust bench-ffi.bench/ffi/README.mdkeeps the BoltFFI comparison and the fork's generated sizes, retargeted at rust2-ffi's generators..taplo.toml'suniffi/**exclude.README.md's fork bullet,.#unifficommand, and license note;AGENTS.md's first line;ffi/README.md's first paragraph.quest mv quest/m2/uniffi-swift-arm.md quest/m2/swift-arm.md, its Goal rewritten around rust2-ffi's Swift skips on aarch64-linux (the paths inffi/README.md,ffi/justfile, andffi/tests/fixtures.rsfollow).quest/m1/ffi-export/bench.mdandquest/m0/nightly-green.mdlose this quest fromRequiredand are now unblocked.quest/README.md's m0 line andquest/m1/README.md's decision drop the link.Cost before and after
From the 2026-11-09 nightly (run 42655190735) and Bench FFI runs on
main:just gate(nightly Gate job), uniffi modulecheck uniffi3m38s +test uniffi3m52s of a ~20 min runjust cion a PR touchinguniffi/or a shared inputUniFFI fixturesUniFFI C and C++(MSVC, Apple Clang)UniFFI installBench FFI(Linux, Swift)Not re-timed after: the remaining modules are unchanged, so the gate loses exactly the uniffi steps.
Public API
Breaking: the flake's
uniffi,uniffi-bindgen,uniffi-bindgen-go,uniffi-bindgen-dart, anduniffi-bindgen-cpppackages and the.#uniffidev shell are removed. No consumer pins them (moq decided on 2026-11-02 to skip the fork). No wire impact.Choices made without asking
rust-toolchain.toml(the flake's default shell andjust benchread it) andrustfmt.toml(every project'scargo fmtfinds it by walking up).quest/m2/swift-arm.md; its check after a nixpkgs bump is now that the default shell'snativeBuildInputson aarch64-linux lists Swift, since.#uniffiis gone.quest/m0/checks/cache.md, whose Goal cites "root clippy" and "uniffi's test build" numbers that this makes stale: moq-dev/rust2#426 holds that quest and also editsuniffi/uniffi_macrosand the fork's trybuild tests. Whichever lands second resolves the conflict by deleting those paths and re-measuring the warm gate without the fork.For moq's tree
moq's
quest/m3/ffi-exporter.mdstill describes the fork era: it links rust2 quests long deleted (quest/m1/uniffi.md,ffi-bytes.md,kt-ffi-calls.md,ffi-poll-protocol.md), plans moq-ffi onrust2-uniffi*crates, and citesuniffi/cpp/bindgen/.../obj.hpp. Its link tobench/ffi/README.md#boltffi-comparisonstill resolves. Its next audit should drop the fork-era requirements.Follow-ups
None recommended: the benchmark's rebuild on rust2-ffi is already
quest/m1/ffi-export/bench.md, and the remaining red nightly jobs arequest/m0/nightly-green.md.Checks
just cipassed locally on the merge withmainat 25bf348: it ran every module (lint, quest, driver, ffi, go, ts checks; sh, driver, ffi, go, ts tests), sinceflake.nixandjustfileare shared inputs. Also passed:actionlint,shellcheck,shfmt,taplo format --check,nixfmt --check,sh/dispatch.test.py,quest check --base origin/main,just --list, and evaluating the flake's packages and default dev shell.Merged
origin/mainfor ajustfileconflict:maindeletedbench-nixand this PR deletedbench-ffiandjust fix'scargo fmt --all; the merge drops all three.Models
Size: M
Implementer: Claude Code, claude-opus-5-5, medium effort
Implementer: Claude Code, claude-opus-5-5, medium effort (iterate, M)
🤖 Generated with Claude Code
(Written by Opus 5.5 with medium effort)
Review
Rewritten by
quest reviewafter each round.01a1c45b-7bec-72e0-8a3d-4e0875760c36, round 1The deletion consistently removes the UniFFI workspace, tooling, CI jobs, and benchmark harness without disrupting the remaining independent workspaces; code inspection found no actionable defects, though builds and tests were not run as instructed.