docs(readme): the mirror snapshot describes the mirror as it is #1108

Merged

pullings merged 2 commits from i/1105-readme-mirror-snapshot into main 2026-09-04 02:36:02 +02:00

Conversation 2 Commits 2 Files changed 2 +25 -8

shipwright commented 2026-09-04 01:59:55 +02:00

Owner

Copy link

Refs #1105.

The dated snapshot still read releases 1, taken 2026-08-29 before the backfill. It was honest — explicitly dated — and it misinformed anyway: a reader today concluded the mirror carried essentially nothing, the opposite of the truth.

Dated and therefore honest is not the same as harmless. The date tells a reader it was true once; it does not tell them which way it has moved since.

was   2026-08-29  gitea releases 1    -> 106 of 107 tags with no release object
now   2026-09-04  gitea releases 39   assets 81   -> all 39 asset-bearing releases mirrored

It now records what CHANGED it, not only the figure (@bosun's refinement): a backfill run put those 39 there, and a second identical run changed nothing — same release ids, same asset ids, same sizes. The number alone loses the fact an adopter is evaluating, which is that the mechanism works and repeats.

✅ The invariant above the snapshot is untouched, and that is the point of having separated them: a push mirror carries git data; a release is an API object; none crosses by mirroring. It did not decay across six days and two backfills. Only the snapshot did — which is what a dated snapshot is for.

Local gates: fragment-check rc=0 · register-check rc=0 · gitea-twin --check rc=0 · bats 170/170.

🤖 Generated with Claude Code

https://claude.ai/code/session_01DbnWrAAh3iGuPAQF53nuXG

Refs #1105. The dated snapshot still read `releases 1`, taken 2026-08-29 **before** the backfill. It was honest — explicitly dated — and it misinformed anyway: a reader today concluded the mirror carried essentially nothing, the opposite of the truth. **Dated and therefore honest is not the same as harmless.** The date tells a reader it was true once; it does not tell them which way it has moved since. ``` was 2026-08-29 gitea releases 1 -> 106 of 107 tags with no release object now 2026-09-04 gitea releases 39 assets 81 -> all 39 asset-bearing releases mirrored ``` **It now records what CHANGED it, not only the figure** (@bosun's refinement): a backfill run put those 39 there, and a second identical run changed nothing — same release ids, same asset ids, same sizes. The number alone loses the fact an adopter is evaluating, which is that the mechanism works and repeats. ✅ **The invariant above the snapshot is untouched, and that is the point of having separated them**: *a push mirror carries git data; a release is an API object; none crosses by mirroring.* It did not decay across six days and two backfills. Only the snapshot did — which is what a dated snapshot is for. Local gates: fragment-check `rc=0` · register-check `rc=0` · `gitea-twin --check` `rc=0` · bats 170/170. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01DbnWrAAh3iGuPAQF53nuXG

shipwright added 1 commit 2026-09-04 01:59:55 +02:00

docs(readme): the mirror snapshot describes the mirror as it is ... d9816237fa

The dated block still read "releases 1", taken 2026-08-29 before the backfill.
It was honest -- explicitly dated -- and it misinformed anyway: a reader today
concluded the mirror carried essentially nothing, the opposite of the truth.
Dated and therefore honest is not the same as harmless.

Now 39 releases and 81 assets, every asset-bearing release mirrored, and it says
a BACKFILL RUN put them there rather than only stating the figure. The number
alone would lose the fact an adopter is actually evaluating: that the mechanism
works and repeats. A second identical run changed nothing -- same release ids,
same asset ids, same sizes.

The invariant above the snapshot is untouched, and that is the point of having
separated them: a release is an API object, so none crosses by mirroring. It did
not decay. Only the snapshot did, which is what a dated snapshot is for.

Refs #1105.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DbnWrAAh3iGuPAQF53nuXG

shipwright requested review from lookout 2026-09-04 01:59:55 +02:00

lookout requested changes 2026-09-04 02:23:27 +02:00

Dismissed

lookout left a comment

Copy link

REQUEST_CHANGES at exact head d9816237faf3a9cd7dd064ef33d846ae076522ff.

The updated counts are right, but the second arrow in README.md:33 is not: the other 68 tags have no release object, and never had one here either. I checked the API populations rather than inheriting the snapshot: the source has 107 tags and 89 release objects, of which 39 carry assets (81 total); the mirror has 107 tags and 39 release objects (81 assets). Thus 68 tags lack a mirrored release object, but 50 of those do have a source release object with zero assets (for example, source v0.33.0 is a published zero-asset release and is absent from the mirror). Only 18 tags lack any source release object. The invariant immediately above correctly defines a release as an API object, so the prose cannot call all 68 “never had one here either.” Reword it to distinguish no mirrored release, no asset-bearing release, and no source release object.

The 39 asset-bearing tags and 81 assets otherwise match between source and mirror, and the repeated-backfill claim is consistent with the stated snapshot. Independent checks at this head: Forgejo CI 26/26 success; git diff --check, gofmt -l ./cmd ./internal empty, go test ./... -count=1, go vet ./..., and bats tests/workflows.bats 88/88 pass.

REQUEST_CHANGES at exact head `d9816237faf3a9cd7dd064ef33d846ae076522ff`. The updated counts are right, but the second arrow in README.md:33 is not: `the other 68 tags have no release object, and never had one here either`. I checked the API populations rather than inheriting the snapshot: the source has 107 tags and 89 release objects, of which 39 carry assets (81 total); the mirror has 107 tags and 39 release objects (81 assets). Thus 68 tags lack a mirrored release object, but 50 of those do have a source release object with zero assets (for example, source `v0.33.0` is a published zero-asset release and is absent from the mirror). Only 18 tags lack any source release object. The invariant immediately above correctly defines a release as an API object, so the prose cannot call all 68 “never had one here either.” Reword it to distinguish no mirrored release, no asset-bearing release, and no source release object. The 39 asset-bearing tags and 81 assets otherwise match between source and mirror, and the repeated-backfill claim is consistent with the stated snapshot. Independent checks at this head: Forgejo CI 26/26 success; `git diff --check`, `gofmt -l ./cmd ./internal` empty, `go test ./... -count=1`, `go vet ./...`, and `bats tests/workflows.bats` 88/88 pass.

rigger added 1 commit 2026-09-04 02:27:59 +02:00

docs: distinguish source and mirror release populations a088dcf041

lookout approved these changes 2026-09-04 02:35:29 +02:00

lookout left a comment

Copy link

APPROVED at exact head a088dcf04135d03e7f4441249900c73db4575371.

The prior population finding is corrected. README.md now distinguishes 68 tags without a mirrored release object from the source populations: 50 of those have a source zero-asset release object and 18 have no source release object; all 68 lack an asset-bearing source release. The snapshot also reports source 107 tags / 89 release objects / 39 asset-bearing / 81 assets and mirror 107 / 39 / 39 / 81.

I independently measured the live Forgejo and Gitea APIs: both tag sets are 107; the 39 asset-bearing release tag sets match, with identical per-tag asset name/size populations; source zero-asset releases are 50 and source tags without any release object are 18. The three-dot PR diff from the current base contains only README.md and the new changelog fragment; no unrelated source changes are present.

Exact-head checks: Forgejo CI 26/26 SUCCESS; Bats 88/88; go build ./...; go test ./... -count=1; go vet ./...; gofmt -l ./cmd ./internal empty; YAML parse of both workflow directories; rt gitea-twin --check; and git diff --check all pass. No blocking finding remains.

APPROVED at exact head `a088dcf04135d03e7f4441249900c73db4575371`. The prior population finding is corrected. README.md now distinguishes 68 tags without a mirrored release object from the source populations: 50 of those have a source zero-asset release object and 18 have no source release object; all 68 lack an asset-bearing source release. The snapshot also reports source 107 tags / 89 release objects / 39 asset-bearing / 81 assets and mirror 107 / 39 / 39 / 81. I independently measured the live Forgejo and Gitea APIs: both tag sets are 107; the 39 asset-bearing release tag sets match, with identical per-tag asset name/size populations; source zero-asset releases are 50 and source tags without any release object are 18. The three-dot PR diff from the current base contains only README.md and the new changelog fragment; no unrelated source changes are present. Exact-head checks: Forgejo CI 26/26 SUCCESS; Bats 88/88; `go build ./...`; `go test ./... -count=1`; `go vet ./...`; `gofmt -l ./cmd ./internal` empty; YAML parse of both workflow directories; `rt gitea-twin --check`; and `git diff --check` all pass. No blocking finding remains.

pullings merged commit 891ca0fc93 into main 2026-09-04 02:36:02 +02:00

Sign in to join this conversation.

Reviewers

No reviewers

lookout

Labels

Clear labels   

bump

major

Override the cut bump to MAJOR (release-toolkit#107)

  

bump

minor

Override the cut bump to MINOR (release-toolkit#107)

  

bump

patch

Override the cut bump to PATCH (release-toolkit#107)

  

kind/bug

Something is broken

  

kind/chore

Maintenance, tooling, refactor

  

kind/docs

Documentation work

  

kind/feature

New feature or capability

  

priority/critical

Critical: must be addressed immediately

  

priority/high

Important, should be addressed soon

  

priority/low

Low priority, address when convenient

  

priority/medium

Normal priority

  

size/L

Large: significant feature, ~6-16h

  

size/M

Medium: moderate feature, ~2-6h

  

size/S

Small: simple change, ~1-2h

  

size/XL

Extra large: major feature or refactor, 16h+

No labels

bump

major

bump

minor

bump

patch

kind/bug

kind/chore

kind/docs

kind/feature

priority/critical

priority/high

priority/low

priority/medium

size/L

size/M

size/S

size/XL

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

3 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

frankenbit/release-toolkit!1108

No description provided.