Skip to content

doctor shows CI's pins next to what is installed, and never fails on them - #114

Merged
lex00 merged 1 commit into
mainfrom
doctor-pins
Aug 7, 2026
Merged

doctor shows CI's pins next to what is installed, and never fails on them#114
lex00 merged 1 commit into
mainfrom
doctor-pins

Conversation

@lex00

@lex00 lex00 commented Aug 7, 2026

Copy link
Copy Markdown
Contributor

Closes #63, in its own words: print the pinned version next to the installed one where a pin exists, without failing.

  • K3D_VERSION consolidates into workflows/shared.ts (it was restated in both k3d-using pipelines); the rendered workflows are byte-identical after the move, and just ci-check proves it.
  • doctor reads the pins from that one file (pin() is three lines of sed, in the repo's awk-the-artifact tradition) and shows them beside the local k3d and node versions — plus a just line, which was never presence-checked because you are running it, but whose version can still silently disagree with CI's.
  • Never fails on a mismatch, per the issue's own reasoning: that is how preflights get ignored.

First run on this machine immediately surfaced two real divergences nobody had seen: k3d v5.8.3 (CI pins v5.9.0) and just 1.46.0 (CI pins 1.36.0).

  ✓ k3d      k3d version v5.8.3  (CI pins v5.9.0)
  ✓ node     v24.13.1  (CI pins 24)
  ✓ just     just 1.46.0  (CI pins 1.36.0)

just check green.

🤖 Generated with Claude Code

https://claude.ai/code/session_017BQGewVRzp9FS1jXn1H6A6

…them

K3D_VERSION moves to workflows/shared.ts — it was restated in both
k3d-using pipelines, and the pin display only stays honest while there is
exactly one place the pin lives. The rendered workflows are byte-identical
after the move.

doctor reads the pins off that file and prints them beside the local
version for k3d and node, plus a line for just itself — which was never
presence-checked, you are running it, but whose version can still disagree
with the one CI installs. Informational only: a working 5.8.3 against a
pinned 5.9.0 is information, and failing a preflight over a patch
difference is how a preflight gets ignored — the reasoning the recipe
already applied to sops and age.

First run on this machine promptly showed two real divergences (k3d
v5.8.3 vs v5.9.0, just 1.46.0 vs 1.36.0), neither of which had ever been
visible anywhere an operator looks.

Closes #63.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017BQGewVRzp9FS1jXn1H6A6
@lex00
lex00 merged commit 01f0e30 into main Aug 7, 2026
2 checks passed
@lex00
lex00 deleted the doctor-pins branch August 7, 2026 20:35
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.

just doctor reports tool versions but never compares them to the pins

1 participant