Skip to content

docs: a pull request opens as a draft and the hardware suite runs before it is marked ready - #54

Merged
jonathanstokes merged 3 commits into
mainfrom
docs/pr-policy
Sep 8, 2026
Merged

jonathanstokes merged 3 commits into
mainfrom
docs/pr-policy

Conversation

@jonathanstokes

@jonathanstokes jonathanstokes commented Sep 8, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Writes down the pull request lifecycle this repo already follows in practice and has broken twice this week:

  • A pull request opens as a draft.
  • The hardware suite runs on its final commit; the description records the commit hash, the unit's CorOS version, the operations on ... block and pytest's last line.
  • A contributor with no unit says so and marks ready anyway; a maintainer runs the suite before merging. A maintainer may waive the run, in the description, for a change that cannot reach the wire.
  • Only then is it marked ready for review.

contributing.md gets "Before you mark a pull request ready" and the contribution flow says "draft". The PR template gets a Hardware section with the three states. CLAUDE.md gets the same rule for the agent, with the reason. STEERING's pattern row and section 10 stop saying hardware verification is manual, and its Change Log records this.

Docs only. No code changes. Rebased onto main after #53 merged (2026-09-08); #53 is the evidence this PR cites for the rule.

Triage

Full triage-pr pass on the draft (two reviewers): 9 findings, all fixed in 82a5c81 (the template still carried the old optional-hardware rule; "totals line" named output only #53 prints; the two files disagreed on the no-unit state; nothing tied the run to a commit; the parent intro still said "if you want to"; prerequisites were elsewhere; "only evidence" overclaimed; STEERING's pattern row was stale). Re-triage of that commit found 4 more, fixed in bc45ab8 (--profile is for a unit the registry refuses, not "other than 4.0.1"; the report's two parts named in the order pytest prints them; the agent's rule dropped the CorOS version; the template's no-unit line did not say "mark ready anyway").

Hardware

Waived by the maintainer. Stokes, 2026-09-07 evening, in the session that wrote this PR: "the unit isn't available this evening, but we can open PRs without the hardware run. I approve it." The change is docs only and cannot reach the wire.

For the record, the suite did run once on this branch, on its first commit f802350 (the code under test was main's before #53), CorOS 4.0.1:

============== operations on QuadCortex (CorOS 4.0.1, MAINTAINER) ==============
passed:                                     ['active_scene', 'clear_expression', 'read_current_preset', 'set_chain_input', 'set_expression', 'set_global_eq', 'set_hold_timing', 'set_input_port', 'set_param', 'set_scene_color', 'set_scene_label', 'set_tempo_mode', 'switch_scene', 'tempo_mode', 'update_settings']
failed or skipped:                          ['set_bypass']
passed, not VERIFIED:                       []   <- candidates to add
VERIFIED and claimed by a test, not passed: ['set_bypass']   <- regressions by name
2 failed, 93 passed, 4 skipped in 188.82s (0:03:08)

The two failures were the two stale burst records that #53 corrected, and #53 is now in this branch's base. The unit dropped off USB before the later commits could be run, hence the waiver.

…ore it is marked ready

contributing.md gains "Before you mark a pull request ready": the suite runs on the
final commit, its totals go in the description, a contributor with no unit says so
and a maintainer runs it, and a maintainer may waive it for a change that cannot
reach the wire. CLAUDE.md carries the same rule for the agent, with why: the
2026-09-07 post-merge run found two stale records a green offline suite had passed.
…ed as pytest prints it, and the rule's evidence is PR #53

Triage of the draft found: the template still carried the old optional-hardware
rule; "totals line" named output only PR #53's branch prints; the two files
disagreed on the no-unit state; nothing tied the run to a commit; the section's
parent intro still said "if you want to"; the prerequisites lived elsewhere;
"only evidence" overclaimed and dropped the manual read-back ask; and STEERING's
pattern row still said hardware verification is manual via examples/.
…ed in the order pytest prints it, and the agent's rule asks for the CorOS version too
@jonathanstokes
jonathanstokes marked this pull request as ready for review September 8, 2026 02:19
@jonathanstokes
jonathanstokes merged commit b2331cc into main Sep 8, 2026
5 checks passed
@jonathanstokes
jonathanstokes deleted the docs/pr-policy branch September 8, 2026 02:54
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