docs: document single-binary rollout #133
No reviewers
Labels
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
Slipstream/auto_review!133
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "issue-120-single-binary-docs-tests"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
Verification
Closes #120
The PR introduces documentation updates and a new test to ensure the single-binary OCI rollout is accurately described. The changes appear safe to merge, with minor notes on improving clarity and error diagnostics.
@ -0,0 +1,130 @@const OPERATIONS: &str = include_str!("../../../docs/OPERATIONS.md");These kinds of tests with deterministic checks for language in documentation are completely worthless. Remove this and create a follow-up issue to remove all such tests from the repository and add the necessary guardrails and review guidance to prevent this in the future. We write tests to cover production code and sometimes to cover developer utility scripts if they are essential to have correct. Otherwise...use better judgement.
Reflection: I treated documentation wording as something that needed deterministic test pressure instead of recognizing this was docs-only prose that should be reviewed directly. Classification: guardrail-gap; the existing TDD guidance exempted docs-only work but did not explicitly forbid docs wording tests.
Remediation: removed
single_binary_rollout_docs_contract.rs, removed its now-unneeded flake source inclusions, tightened the TDD guidance to say docs-only wording should not get deterministic tests, and opened follow-up #134 to inventory/remove any remaining docs wording tests.@ -5,12 +5,19 @@ For self-hosters who run the gateway directly on a Linux host[`docker-compose.yml`](../docker-compose.yml) as the thirddeploy option.The Docker/OCI image is still the recommended production deployment because itFor both this and the bare command without systemd, it would be worth noting that the reason these are provided is for people who want to build the program into their own, custom container images or VMs.
Also, we should provide capability and instructions for using this program in Nix/NixOS via our flake.nix if we don't already. It should support deploying containerized, deploying via systemd, or just installing the program.
Reflection: I documented bare/systemd mechanics without explaining the intended operator use case for those paths, and I did not separate existing flake capabilities from missing NixOS deployment support. Classification: one-off for the wording gap; the NixOS module support is new scope tracked separately.
Remediation: clarified that bare/systemd direct-binary paths exist for operators building custom VM/container images, added the same context to the bare quickstart path, documented the existing flake package/image outputs, and opened follow-up #135 for first-class NixOS module/deployment support.
New commits pushed, approval review dismissed automatically according to repository settings
This PR updates documentation to clarify deployment options and testing guidance, particularly around the use of deterministic tests for documentation wording. The changes align with the project's focus on container-first deployment and human review for docs-only changes.
Walkthrough
Δ since
88f0c0d:QUICKSTART.mdandOPERATIONS.md.systemd/README.mdto specify its use for custom deployments.