Skip to content

docs: gourmand's image tag is a packaging version, not a version mismatch - #1

Merged
fatherlinux merged 1 commit into
mainfrom
docs/gourmand-version-wording
Nov 4, 2026
Merged

fatherlinux merged 1 commit into
mainfrom
docs/gourmand-version-wording

Conversation

@fatherlinux

Copy link
Copy Markdown
Member

Summary

I reported crunchtools/gourmand:1.0.0 shipping a binary that calls itself 0.16.5 as drift between an image tag and its build, and wrote that reasoning into four places: the README, the CHANGELOG, the sdlc versions trailer, and the contract test's exemption note for gourmand.

It is not drift. crunchtools/gourmand is a wrapper repo with its own SemVer — 1.0.0 is its first tagged release (2026-09-19, RT #1484) and its Containerfile builds upstream mattdm/gourmand at --tag v0.16.5. Both numbers are correct and they answer different questions.

What changes

sdlc versions still prints both columns, but for the reason it should have given rather than the one I gave:

  • the pinned column is what CI must name, so a local finding is the CI finding;
  • the self-reported column is which upstream is actually running, which a tag alone does not tell you.

That second point is not academic here — Gourmand's Python parser is exactly what crunchtools/gourmand#29 is about, and the upstream version is the one that answers whether a given image has the PEP 758 fix.

The contract test continues to exempt gourmand from the pinned-equals-reported assertion, now because asserting they match would be asserting something untrue, rather than because of a defect somewhere else.

Test plan

  • ruff check / ruff format --check clean over sdlc, validate, tests
  • shellcheck clean over bin/sdlc and the four test scripts
  • tests/contract.sh passes against a locally built image, including the sdlc versions ↔ build-pin assertions for the six tools that are not wrappers
  • tests/test_tiers.sh, tests/test_host_contract.sh, tests/test_codeql_globs.py pass
  • Gourmand passes on this repo

No behaviour change: wording, one docstring, and one test comment.

🤖 Generated with Claude Code

https://claude.ai/code/session_01R7Y1jfy8tZHovMEJUNjjD2

…atch

I reported `crunchtools/gourmand:1.0.0` shipping a binary that calls itself
0.16.5 as drift between a tag and its build, and wrote that into the README,
the CHANGELOG, the `sdlc versions` trailer and the contract test's exemption
note. It is not drift.

`crunchtools/gourmand` is a WRAPPER repo with its own SemVer: 1.0.0 is its first
tagged release (2026-09-19, RT #1484), and its Containerfile builds upstream
`mattdm/gourmand` at `--tag v0.16.5`. Both numbers are correct and they answer
different questions.

So `sdlc versions` keeps printing both columns, for the reason it should have
given in the first place rather than the one I gave: the tag is what CI must
pin, and the self-reported version is which upstream is actually running --
which a tag alone does not tell you, and which matters here because Gourmand's
parser is what crunchtools/gourmand#29 is about.

The contract test still exempts gourmand from the pinned-equals-reported
assertion, now because asserting they match would be asserting something untrue
rather than because of someone else's defect.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R7Y1jfy8tZHovMEJUNjjD2

@github-actions github-actions Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Gatehouse found no issues.

Full review.

@fatherlinux
fatherlinux merged commit e7d8394 into main Nov 4, 2026
9 checks passed
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