Skip to content

Latest commit

 

History

History
29 lines (20 loc) · 2.61 KB

roadmap.md

File metadata and controls

29 lines (20 loc) · 2.61 KB

Public roadmap

This is an evaluation roadmap, not a delivery guarantee. Existing support limits are in release-boundary.md.

Milestone Completion evidence
Make local trials easy One-command CLI trial, clear fallback interpretation, contribution setup
Recruit five assisted pilots Maintainer consent, exact revision pairs, declared measurement protocol
Publish two reproducible pilot stories Raw reports, repeated timings, analysis overhead, limitations
Expand regression fixtures Bounded issues with positive and fallback cases
Publish a benchmark roundup Denominator includes failures, refusals and regressions; distinguish static selection from runtime

Language support

JS/TS, Vue, Go, Maven and Python (pytest) are supported today, and Rust (Cargo) experimentally: it has not yet narrowed a real project's history, so expect full runs there for now (see Rust adapter safety). Django, PHP and Gradle are next and open for contributors. Each tracking issue lists small steps that are one PR each:

Go is supported, but many real Go repositories still fall back to the full suite (mixed Go/TypeScript, go:generate, go.work, build tags and more). The Go tracking issue lists the steps to close those fallbacks, each one PR. The Maven tracking issue does the same for custom source roots, external parent POMs and unresolved ${property} versions.

Want another language? 👍 an existing tracking issue or open one with the adapter label. Reactions decide what comes next.

Watch → Custom → Releases and Discussions to follow published results and support changes. Release notes should state changed support boundaries, new fallback cases, and measured improvements only when artifacts exist. Aim to publish a monthly benchmark report; a month with no new measurements should say so explicitly.

Monthly reports should start from the benchmark report template, which keeps unsuccessful attempts in the denominator and separates static selection, measured runtime, and modeled impacts.