Two gaps left over from the v6.0.0 release.
The v6.0.0 archives shipped unstripped: 3.63 MB vs 4.53 MB locally, 168
symbols instead of 9483. Setting it on the profile rather than in the
workflow covers every target at once and `cargo install` along with them.
Nothing gated the crates.io publish. A published version can be yanked but
never withdrawn, so the release job now waits for the CI run on the same
commit and refuses to publish unless it is green. release-pr deliberately
does not wait; opening the release PR is harmless.
This is not hypothetical: CI failed on eef7bf7, the v6.0.0 release commit.
It was a flaky e2e timeout rather than a regression - the very next commit
touched no code and passed - but the gate would have stopped that release,
which is the point. The gate reads the newest CI run for the commit, so
re-running a flaky job and then re-running Release clears it.
upload-rust-binary-action reaches for `cross` on any *-musl target. Every
entry in the matrix builds on a runner of its own architecture — only the
libc differs — so cross buys nothing, and on the native arm64 runner it
hard-failed:
error: toolchain 'stable-x86_64-unknown-linux-gnu' may not be able to
run on this system
That cost v6.0.0 its aarch64-unknown-linux-musl archive; the other five
targets uploaded because the matrix is fail-fast: false.
Re-running the failed job from the Actions UI replays the old workflow file,
which is useless when the workflow is the bug. So the Release workflow now
takes an optional `tag` input that rebuilds and re-uploads binaries for an
existing tag. That value reaches actions/checkout as `ref:` and is typed by
whoever dispatches, so it is checked against a literal version tag first.
The push and pull_request triggers both matched every branch, so each PR
kicked off two identical runs of lint, the three-OS test matrix and the
soulfind e2e job. Push now only covers master; branches are covered by the
pull_request trigger. Adds a concurrency group so a follow-up push cancels
the superseded run instead of queueing behind it.
Replace the 326-line semantic-release/npm workflow with release-plz, which
handles the whole chain the old one only half did:
- versioning and CHANGELOG from the conventional commits already in use
- a `chore: release vX.Y.Z` PR as the gate; merging it tags and releases
- publishes both crates to crates.io in dependency order, waiting for the
index between them (RELEASE.md used to ask you to do this by hand)
- release-plz.toml keeps the two crates on one version, one tag, one changelog
Binaries now cover six targets instead of four. Linux switches from gnu to
musl so the downloads are statically linked and stop caring about the host's
glibc version; linux-aarch64 and windows-aarch64 are new. Archives carry
sha256 checksums plus README and LICENSE.
CI drops twelve hand-rolled actions/cache steps for Swatinem/rust-cache.
Extend the existing CI workflow with an ubuntu-only "End-to-end (soulfind)"
job that builds soulfind from source (LDC + `dub build :server`), points
SOULFIND_BIN at it, and runs the e2e suite with SOULSEEK_E2E_REQUIRED=1.
Also widen the clippy job to the whole workspace.
The e2e harness gains SOULSEEK_E2E_REQUIRED: when set, a missing/unstartable
server is a hard failure instead of a silent skip, so CI genuinely exercises
the client end-to-end. Add e2e coverage for two existing behaviours a single
server can verify: wrong-password rejection and concurrent logins.
- Add check-release job to determine if a release is needed
- Build binaries in parallel for all platforms before release
- Create release and attach binaries in final step
- Start with version 0.1.0
Michel de Graaf