| Age | Commit message (Collapse) | Author |
|---|
|
errcheck ./... flagged unchecked HTTP response Close(), file Close(),
os.RemoveAll(), fmt.Scanln(), and fmt.Fprintf() calls across codeberg,
github, release, showcase, and sync. None of these were bugs causing
incorrect behavior today, but leaving them unchecked hid real failure
modes (e.g. a lagging NFS mount failing a file Close() after writes,
which this codebase has hit before per commit 23ecaa2).
- internal/codeberg/codeberg.go, internal/github/github.go,
internal/release/release.go: added a small closeResponseBody(resp)
helper per package and used it for all deferred resp.Body.Close()
calls. The body is always fully read (or abandoned on an earlier
error) by the time Close() runs, so the error is intentionally
discarded - matching the explicit `_ = ...` discard convention
already used elsewhere in this repo (e.g. showcase.go's
os.RemoveAll on the worktree-add failure path).
- internal/release/release.go: the two fmt.Scanln(&response) prompts
now explicitly discard the return values; a Scanln error already
leaves response == "", which the existing y/yes check already
treats as a safe decline, so behavior is unchanged.
- internal/showcase/code_extractor.go, images.go, language_detector.go:
added a shared closeFile(*os.File) helper (package showcase) for the
read-only file Close() calls, matching the same discard rationale.
copyFile's destination Close() is the one write-side case where a
close failure is real data-loss information, so it now uses a named
return to surface it via err without masking any earlier error.
- internal/showcase/showcase.go: the deferred os.RemoveAll(tempRoot)
now explicitly discards its error, matching the sibling `_ =` call
three lines above in the same function.
- internal/sync/branch_analyzer.go: GenerateDeleteScript's per-repository
fmt.Fprintf(file, ...) calls are now checked and wrapped with
fmt.Errorf(...: %w), matching the error-wrapping convention already
used by writeBranchDeletionBlock right below it in the same file
(the repeated repo-header writes were pulled into a new
writeDeleteScriptRepoHeader helper to keep this readable). The
script file's defer file.Close() now uses a named return so a close
failure is reported instead of silently discarded.
Added focused tests for the two behavior-relevant paths: copyFile's
missing-source/success paths after the named-return change, and
PromptConfirmation's empty-input-declines/explicit-yes behavior after
touching the Scanln call. Trivial defer-Close discards elsewhere are
not additionally unit tested per the task's guidance against
overengineering.
Verified: go build ./..., go vet ./..., errcheck ./... (clean), and
go test ./... all pass.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
Auditing a full bidirectional sync to the r0 git-server backup destination
showed a stale/missing-repo pattern that cgit idle times alone did not
explain: several public repos (rampage, ggaze, comicforge, fastforge, gonf,
quicklog, shuriken.sh) were entirely absent from r0 even though they synced
fine to GitHub/Codeberg.
Root cause #1: createSSHBareRepository provisions missing backup repos by
running `git init --bare` over a root SSH session directly on the r0
filesystem. That leaves the new repo directory at mode 0755 (root's
session umask), owner-write only. Pushes into it later go through the
git-server pod's own SSH endpoint as a different, non-root UID (1001,
GID 33/www-data per the git-server helm chart's docker-image/Dockerfile),
which then cannot write new objects into the 0755 tree ("unable to create
temporary object directory"). Fixed by initializing with
`git init --bare --shared=group` (mode 2775, matching the already-working
repos on r0).
Root cause #2: even with correct permissions, the push immediately
following repository creation can still fail transiently (the git-server's
own view of the newly created directory can lag behind, e.g. across an NFS
mount). handlePushError treats any backup push failure as fatal for the
remainder of that sync run and disables the backup destination entirely
(disableBackupForSession) - so one transient failure early in a shuffled
repo order silently skipped backup for every repo processed afterward in
that pass, explaining the fuller set of stale repos. Fixed by retrying the
post-creation push up to 3 times with a short backoff in the new
createAndPushSSHBackupRepo helper, and by capturing/propagating the actual
git stderr on failure (previously swallowed via cmd.Run()) so any future
failure is diagnosable instead of a bare "exit status 128".
Verified against the live r0 git-server (ssh://git@r0:30022/repos,
/data/nfs/k3svolumes/git-server/repos): reinstalled gitsyncer and reran
`sync bidirectional --force --backup --auto-create-releases` end to end
with zero backup-disable events across both passes; all previously
missing/broken repos (plus the pre-existing broken player.git and
irregular.ninja.git, repaired directly via chmod) now match GitHub/Codeberg
HEAD exactly and are browsable on c-git.f3s.buetow.org.
Adds a regression test for the --shared=group init command.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
forcePush/descriptionSync creation support
Backup failures now disable retries only for the failing remote instead of
the whole session, add an opt-in forcePush flag for backup organizations,
and allow repository creation to go through descriptionSyncHost/Root when
configured so the git remote endpoint can stay restricted. Also fixes the
AI release-notes cache being bypassed by --force, which is meant to control
sync scheduling, not cache invalidation.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
0.18.5
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
|
Both the rank-history SVG and the gemtext showcase now use the same rule:
inactive = AvgCommitAge (last 42 commits on HEAD) > 730 days (~2 years)
Changes:
- SVG subtitle: 'no meaningful commits in 1+ years' → 'avg commit age > 2 years'
- SVG struct comment: updated to describe the 730-day/42-commit rule accurately
- gemtext: dropped the secondary 'lastCommit > 365 days' guard (it allowed
projects like muttdelay to escape despite a single stray recent commit)
- gemtext notice text: 'Last meaningful activity was over 2 years ago' →
'The average age of its last 42 commits exceeds 2 years'
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
