Rust adapter: selective runs on real repositories

Open
#131 1 comment 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
30/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
github-actions, rust

Research direction

Start with the Cargo adapter registration and the rust-lang/glob replay pinned in #95, and read #91's out-of-workspace manifest agreement plus #92–#95 for the adapter contract. Compare how existing adapters classify non-code paths and runtime-input tokens, then use those paths as the reference. Done means at least one historical revision in the pinned replay narrows instead of staying FULL, while the selected tests still catch the same failures as the full baseline.

Written by the indexing model from the issue text.

Description

Follow-up to #62. Once #92–#95 merge, the Rust (Cargo) adapter is registered and safe, but it cannot narrow a real repository yet: the pinned rust-lang/glob replay stayed FULL on every revision. This issue tracks the work needed for real-repository selectivity.

Scope

  • External dependency inputs. The blockers for non-workspace packages, dependency features or default-features = false, build scripts and proc macros currently run over every resolved package. Apply them only to the crates a change reaches, wherever the dependency graph can show that.
  • Runtime-input tokens. fs, unsafe, exists, metadata, process and similar words in a source file force a repo-wide full run. Narrow this to the owning crate and its dependents.
  • Out-of-workspace Cargo.toml. As agreed in #91: treat unmodeled manifests as ownership boundaries and require full validation only when the changed set or its dependents reach them. Keep the repo-wide fallback when metadata can't tell.
  • CI invocation evidence. Accept common toolchain setup and cache steps (for example dtolnay/rust-toolchain, Swatinem/rust-cache) when they don't change how cargo test runs.
  • Non-code paths. Classify README.md, docs/** and similar files the way the other adapters do, instead of forcing a full run.
  • Workspace scope. Handle default-members and a root [package], where plain cargo test covers only some members.

Done when a pinned public Rust replay narrows at least one historical revision, and the selected tests catch the same failures as the full baseline.

Maintainer: @FredApps.

Dominant language
TypeScript
Stars
0
Forks
4
Avg merge
18h 41m
Merged PRs (30d)
48

Getting set up

Open in Codespaces

Starts the project's dev container in your browser, under your own GitHub account.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from DiffCI/core

All issues in DiffCI/core

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.