Skip to content

feat: ship a macos-x86_64 archive and align the packer with rbxoc - #25

Merged
eiei114 merged 2 commits into
mainfrom
feat/macos-intel-and-packer-parity
Oct 23, 2026
Merged

eiei114 merged 2 commits into
mainfrom
feat/macos-intel-and-packer-parity

Conversation

@eiei114

@eiei114 eiei114 commented Oct 23, 2026

Copy link
Copy Markdown
Collaborator

Why

Two follow-ups recorded against #24, both mechanical:

  1. Intel Macs were unsupported. Rokit 1.2.0 checks the OS of a downloaded binary but not its architecture, so an Intel Mac installed macos-aarch64, saw success, and then could not exec rob. The README documented the trap; this removes it.
  2. The packer lagged behind rbxoc's copy. package-release.py still wrote the .sha256 sidecar in text mode, which emits CRLF on Windows and breaks sha256sum -c, and a bad target leaked a raw KeyError tuple.

What

  • Add macos-15-intel to the matrix, emitting macos-x86_64.
  • Pin the arm64 job to macos-15, so both macOS entries name their image explicitly.
  • Allow ("macos", "x86_64") in the packer.
  • Write the .sha256 sidecar with an explicit \n.
  • Report the target name for an out-of-matrix target instead of a KeyError repr.
  • Raise the publish guard from -eq 3 to -eq 4.
  • README: drop the Intel warning, list macos-x86_64.

Cost

macos-15-intel is a Standard GitHub-hosted runner — standard-runners table, no -large / -xlarge suffix — so it is free on a public repository.

Verification

Packer run locally against the built binary; no release or Roblox endpoint touched:

case result
--os macos --arch x86_64 robcli-0.8.0-macos-x86_64.tar.gz with rob, license-mit, license-apache
--os windows --arch x86_64 zip; sidecar ends with a bare \n, sha256sum -c reports OK (it failed before)
--os linux --arch aarch64 target linux/aarch64 is outside the approved archive matrix
--version 0.8.0-rc1 version is not a plain numeric release

Workflow YAML parses; matrix resolves to windows-latest/x86_64, ubuntu-latest/x86_64, macos-15/aarch64, macos-15-intel/x86_64. No Rust source changed.

Not verified until this merges: the macos-15-intel job itself. Backfill onto v0.8.0 follows in a separate dispatch.

Remaining follow-up (not here)

The archive member is rob while the repository is robcli, so Rokit's member lookup uses its scoring fallback instead of an exact filename match. Renaming changes the archive layout, so it needs its own decision.

Rokit validates the OS of a downloaded binary but not its architecture, so on
an Intel Mac it installed macos-aarch64, reported success, and left `rob`
unable to exec. The README worked around that with a warning; this removes the
need for one.

macos-15-intel is a Standard GitHub-hosted runner (no `-large` or `-xlarge`
suffix), so this costs nothing on a public repository. The arm64 job is pinned
to macos-15 for the same reason as in rbxoc: name the image explicitly so a
`-latest` migration cannot silently move one macOS entry.

Two packager fixes ported from rbxoc's copy:

- The `.sha256` sidecar is written with an explicit \n. Python's text mode emits
  CRLF on Windows, which breaks `sha256sum -c`. Verified: the sidecar now ends
  with a bare newline and `sha256sum -c` reports OK.
- A target outside the approved matrix reports the target name. It used to leak
  a raw KeyError tuple, because `str(KeyError)` is the repr of the key.

The publish guard counted exactly three built-commit records; it now expects
four, matching the matrix.

Verification of the packer paths (local, no release touched):
- macos/x86_64 -> robcli-0.8.0-macos-x86_64.tar.gz with rob, license-mit,
  license-apache.
- windows/x86_64 -> zip, sidecar passes `sha256sum -c`.
- linux/aarch64 -> "target linux/aarch64 is outside the approved archive matrix".
- 0.8.0-rc1 -> "version is not a plain numeric release".
@coderabbitai

coderabbitai Bot commented Oct 23, 2026 •

Copy link
Copy Markdown

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 6 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

⚙️ Run configuration
  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: dc6490bc-1024-48da-a2ab-af8fa22a1341
📥 Commits

Reviewing files that changed from the base of the PR and between e87f237 and cbc71e6.

📒 Files selected for processing: 3
  • .github/workflows/release-assets.yml
  • readme.md
  • scripts/package-release.py
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

v0.8.0 does not carry macos-x86_64 yet, and the publish guard blocks a backfill
for that tag until this change is on main. Saying the archive exists without
that caveat would be false at merge time for anyone on exactly 0.8.0.
@eiei114
eiei114 merged commit dcbff8b into main Oct 23, 2026
9 checks passed
@eiei114
eiei114 deleted the feat/macos-intel-and-packer-parity branch October 23, 2026 17:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant