chore(ci): replace the reproducibility double build with a build-configuration check

Open
#251 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
1-2 days
Newbie friendliness
25/100
Issue type
Refactor
Clarity
Mostly clear
Activity status
Active
Tech stack
github-actions, go

Research direction

Start with the reproducibility job in .github/workflows/ci.yml and the check:reproducibility target it runs, then read .goreleaser.yaml for the -trimpath, -buildid= and mod_timestamp settings. Check docs/RELEASE.md section 3 and openspec/config.yaml, since their reproducibility claims must match what the chosen gate proves. Done means one of the four options is picked, the claims are reconciled, and the work follows or merges with open PR #242.

Written by the indexing model from the issue text.

Description

area:ci chore priority:low
What kind of chore?

CI / workflow

What needs doing?

The reproducibility required check in .github/workflows/ci.yml runs task check:reproducibility. That target builds codegraph-linux-amd64 twice and compares the SHA-256 digests. The check costs CI time and mostly tests tools this repository does not own. Replace it with a check of the build configuration this repository owns. The options are:

  • A static check of .goreleaser.yaml. It asserts that each build sets -trimpath and -buildid=, that mod_timestamp comes from {{ .CommitTimestamp }}, and that no build-clock or random value reaches the ldflags. It needs no second build.
  • Keep the double build and share the Go and zig caches between the two builds. This still tests Go, zig and GoReleaser determinism on one runner.
  • Move the bit-for-bit proof to .github/workflows/post-release-verify.yml. It would rebuild the tag and compare the result with the published linux/amd64 binary. The check then runs after release, off the pull-request path.
  • Drop the required check. This needs an edit to the protect-main ruleset, where the job name is pinned.

Whether the static check alone satisfies the reproducibility claim is open. openspec/config.yaml promises reproducible releases, and docs/RELEASE.md section 3 says the linux/amd64 binary is reproducible bit for bit. Whichever option lands, those two claims must match what the gate proves.

Why now?

PR #242 (open, branch feat/221-rework-ci-test-suite) changes build 2 of check:reproducibility. Build 2 now uses an empty GOCACHE and empty zig cache directories. The suite review logged this as QUAL-16, and it reports that the reproducibility job rose from 62 s to 191 s.

The repository owns the reproducibility settings in .goreleaser.yaml: -trimpath, -buildid=, the Date ldflag from {{ .CommitDate }}, and mod_timestamp. It also owns the pinned toolchain in mise.toml and mise.lock. The release posture is documented in docs/RELEASE.md section 3.

PR #242 edits this job and this target too, so this work must follow that PR or merge with it. Issue #212 tracks the GNU date -d failure on macOS in the same target and its arm64 leg. A rewrite of the target also settles #212.

This issue reads .goreleaser.yaml, mise.toml and mise.lock without editing them. The first load-bearing box stays unticked for that reason.

Does this touch anything load-bearing?
  • The release pipeline (release.yml, .goreleaser.yaml, signing)
  • The upgrade identity (internal/upgrade/verify.go)
  • The perf baseline (tools/bench/baseline.json)
  • The graph schema / meta-key layout
  • None of these
Before submitting
  • I searched existing issues and this is not a duplicate
Dominant language
Go
Stars
0
Forks
0
Avg merge
5h 16m
Merged PRs (30d)
107

Getting set up

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 seanb4t/codegraph-go

All issues in seanb4t/codegraph-go

Similar issues

More Go issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.