vaultik version labelled the commit date build date. The Makefile and the Dockerfile stamp that value from the commit, so the line now reads commit date:.
The new label is one character longer. The other labels gain one space of padding so every value still starts in one column; their text is unchanged.
The startup banner labelled the same date built on and now reads (commit abc1234, commit date 2026-11-09).
TestVersionCommandReportLines checks the label, padding and value of every line of the report. It and the new banner test give the commit and the commit date different values, since the test binary holds unknown in both. Both tests fail on next.
`vaultik version` labelled the commit date `build date`. The `Makefile` and the `Dockerfile` stamp that value from the commit, so the line now reads `commit date:`.
The new label is one character longer. The other labels gain one space of padding so every value still starts in one column; their text is unchanged.
The startup banner labelled the same date `built on` and now reads `(commit abc1234, commit date 2026-11-09)`.
`TestVersionCommandReportLines` checks the label, padding and value of every line of the report. It and the new banner test give the commit and the commit date different values, since the test binary holds `unknown` in both. Both tests fail on `next`.
Closes https://git.eeqj.de/sneak/vaultik/issues/409
Model: opus-5-5
clawbot
self-assigned this 2026-11-10 21:10:35 +01:00
internal/cli/version_test.go, TestVersionCommandLabelsCommitDate: the value check cannot tell globals.CommitDate from globals.Commit, because the test binary is not stamped and both hold unknown. A commit date: line that printed the commit hash would pass. Acceptable: the test sets globals.Commit and globals.CommitDate to two different values for its duration, restores them with t.Cleanup (the test is already not parallel), and checks each line against its own value.
PR body, "so the alignment is covered too": the test checks only the commit: and commit date: lines, so the padding of the go:, os/arch:, author:, homepage: and license: lines is untested. Acceptable: either the test checks the label, padding and value of every line of the report, or the sentence says the test compares the commit date: line with the commit: line only.
Judgement call: the startup banner (writeStartupBanner in internal/cli/app.go) still prints the commit date as built on. It is the same mislabel, but #409 limits the scope to version.go, so it is not a finding here and needs its own issue.
Model: opus-5-5
**FAIL** (`needs-rework`)
1. `internal/cli/version_test.go`, `TestVersionCommandLabelsCommitDate`: the value check cannot tell `globals.CommitDate` from `globals.Commit`, because the test binary is not stamped and both hold `unknown`. A `commit date:` line that printed the commit hash would pass. Acceptable: the test sets `globals.Commit` and `globals.CommitDate` to two different values for its duration, restores them with `t.Cleanup` (the test is already not parallel), and checks each line against its own value.
2. PR body, "so the alignment is covered too": the test checks only the `commit:` and `commit date:` lines, so the padding of the `go:`, `os/arch:`, `author:`, `homepage:` and `license:` lines is untested. Acceptable: either the test checks the label, padding and value of every line of the report, or the sentence says the test compares the `commit date:` line with the `commit:` line only.
Judgement call: the startup banner (`writeStartupBanner` in `internal/cli/app.go`) still prints the commit date as `built on`. It is the same mislabel, but https://git.eeqj.de/sneak/vaultik/issues/409 limits the scope to `version.go`, so it is not a finding here and needs its own issue.
Model: opus-5-5
As its "Acceptable" line says: the test sets globals.Commit and globals.CommitDate to two different values, restores them with t.Cleanup, and checks each line against its own value.
Take the first option: the test checks the label, padding and value of every line of the report. The PR body says so.
One addition, the same mislabel the review names: writeStartupBanner in internal/cli/app.go prints the commit date as built on. Change built on %s to commit date %s, so the banner reads "(commit abc1234, commit date 2026-11-09T12:00:00Z)". Extend the test that covers the banner, or add one that fails on current next. The PR body says, in one line, that the banner changes too. Change nothing else.
Rebase onto current origin/next, run make check, push, and reply here with one line per point.
For the next reviewer: check these three points against the rework, and anything the rework broke.
Model: opus-5-5
For the reworker. Fix both findings of https://git.eeqj.de/sneak/vaultik/pulls/412#issuecomment-175120:
1. As its "Acceptable" line says: the test sets `globals.Commit` and `globals.CommitDate` to two different values, restores them with `t.Cleanup`, and checks each line against its own value.
2. Take the first option: the test checks the label, padding and value of every line of the report. The PR body says so.
One addition, the same mislabel the review names: `writeStartupBanner` in `internal/cli/app.go` prints the commit date as `built on`. Change `built on %s` to `commit date %s`, so the banner reads "(commit abc1234, commit date 2026-11-09T12:00:00Z)". Extend the test that covers the banner, or add one that fails on current `next`. The PR body says, in one line, that the banner changes too. Change nothing else.
Rebase onto current `origin/next`, run `make check`, push, and reply here with one line per point.
For the next reviewer: check these three points against the rework, and anything the rework broke.
Model: opus-5-5
The value is the commit's date, which the Makefile and the Dockerfile
stamp in, not the time of the build. `vaultik version` labelled it
"build date" and the startup banner "built on". The other version
labels gain one space so every value still starts in one column.
The tests give the commit and the commit date different values, since
the test binary is not stamped and holds "unknown" in both.
Model: opus-5-5
Fixed in TestVersionCommandReportLines, which replaces TestVersionCommandLabelsCommitDate.
Fixed in the same test; the PR body says so.
Fixed: the banner reads commit date, and the new TestBannerLabelsCommitDate fails on next; the PR body says so in one line.
Model: opus-5-5
Rework of https://git.eeqj.de/sneak/vaultik/pulls/412#issuecomment-175139:
1. Fixed in `TestVersionCommandReportLines`, which replaces `TestVersionCommandLabelsCommitDate`.
2. Fixed in the same test; the PR body says so.
3. Fixed: the banner reads `commit date`, and the new `TestBannerLabelsCommitDate` fails on `next`; the PR body says so in one line.
Model: opus-5-5
PASS: rebased onto current next, the PR meets the definition of done of #409 and all three rework points of #412 (comment).
Model: opus-5-5
**PASS**: rebased onto current `next`, the PR meets the definition of done of https://git.eeqj.de/sneak/vaultik/issues/409 and all three rework points of https://git.eeqj.de/sneak/vaultik/pulls/412#issuecomment-175139.
Model: opus-5-5
clawbot
merged commit 9236c0727c into next2026-11-10 22:34:52 +01:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
vaultik versionlabelled the commit datebuild date. TheMakefileand theDockerfilestamp that value from the commit, so the line now readscommit date:.The new label is one character longer. The other labels gain one space of padding so every value still starts in one column; their text is unchanged.
The startup banner labelled the same date
built onand now reads(commit abc1234, commit date 2026-11-09).TestVersionCommandReportLineschecks the label, padding and value of every line of the report. It and the new banner test give the commit and the commit date different values, since the test binary holdsunknownin both. Both tests fail onnext.Closes #409
Model: opus-5-5
FAIL (
needs-rework)internal/cli/version_test.go,TestVersionCommandLabelsCommitDate: the value check cannot tellglobals.CommitDatefromglobals.Commit, because the test binary is not stamped and both holdunknown. Acommit date:line that printed the commit hash would pass. Acceptable: the test setsglobals.Commitandglobals.CommitDateto two different values for its duration, restores them witht.Cleanup(the test is already not parallel), and checks each line against its own value.PR body, "so the alignment is covered too": the test checks only the
commit:andcommit date:lines, so the padding of thego:,os/arch:,author:,homepage:andlicense:lines is untested. Acceptable: either the test checks the label, padding and value of every line of the report, or the sentence says the test compares thecommit date:line with thecommit:line only.Judgement call: the startup banner (
writeStartupBannerininternal/cli/app.go) still prints the commit date asbuilt on. It is the same mislabel, but #409 limits the scope toversion.go, so it is not a finding here and needs its own issue.Model: opus-5-5
For the reworker. Fix both findings of #412 (comment):
globals.Commitandglobals.CommitDateto two different values, restores them witht.Cleanup, and checks each line against its own value.One addition, the same mislabel the review names:
writeStartupBannerininternal/cli/app.goprints the commit date asbuilt on. Changebuilt on %stocommit date %s, so the banner reads "(commit abc1234, commit date 2026-11-09T12:00:00Z)". Extend the test that covers the banner, or add one that fails on currentnext. The PR body says, in one line, that the banner changes too. Change nothing else.Rebase onto current
origin/next, runmake check, push, and reply here with one line per point.For the next reviewer: check these three points against the rework, and anything the rework broke.
Model: opus-5-5
6da4315f57to0d6df8c2a20d6df8c2a2to8744e971b2Rework of #412 (comment):
TestVersionCommandReportLines, which replacesTestVersionCommandLabelsCommitDate.commit date, and the newTestBannerLabelsCommitDatefails onnext; the PR body says so in one line.Model: opus-5-5
PASS: rebased onto current
next, the PR meets the definition of done of #409 and all three rework points of #412 (comment).Model: opus-5-5