Repository navigation
chore(sync): drop the hidden --profile alias and name the store in every message - #407
Merged
Merged
Conversation
…ery message dot sync accepted --profile as a deprecated alias of --store for one release after #380. That release is out (v2.80.1 knows --store), so the alias is removed: `dot sync --profile` now fails with "unknown flag" instead of warning on stderr, and the fallback that read the flag, the conflict refusal between the two names, and syncStoreFrom are gone. Everything dot prints and everything it runs on the peer moves to --store too: the owner-mismatch and fence refusals, the NFD fix that peer doctor and peer inventory report, the handover messages, the conflicts hint at the end of `dot peer sync`, the launchd plist and systemd unit the scheduler renders, and the spawn-failed reload line. remotePeerNormalizeCommand now sends --store=peer to the peer; the comment that hedged for a peer whose dot knew only --profile is no longer true, because the peer runs the same release as this Mac. --profile on apply, diff, config, status, init, watchdog setup|notify and container setup is untouched: that flag picks a dotfiles profile and is unrelated to the sync store. Part of #394
Codex Review Summary: ✅ CompletedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7c007cd029
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
A unit stores the whole command line and the service manager replays it on every tick, and neither `dot update` nor `dot apply` rewrites an installed unit. A scheduler installed for a non-default store therefore kept invoking `dot sync push --profile=<store>` after the alias removal of #394 and failed with "unknown flag" on every run, silently. `dot sync resume` is the command that re-arms installed units, so it now rewrites the retired argument to --store before the reload that follows, which is what makes the service manager read the new line. The rewrite changes only the flag: store, mode, interval, --home and log paths stay exactly as the operator installed them, because a fresh render would re-apply setup's current settings and could silently re-time or re-enable a scheduler. `dot sync status` stays read-only and names the installed unit plus the `dot sync --store=<store> pause && ... resume` pair that heals it. The peer store is exempt: com.dotfiles.peer.plist belongs to `dot peer setup` (it runs `dot peer sync` and renders no store argument), and `dot sync --store=peer` is refused, so rewriting that file would leave a command that cannot run. Verified: this Mac's com.dotfiles.sync.plist carries no store argument (default store). The peer Mac was unreachable over WARP at review time, so its units are unverified. The migration was exercised against real template renders spelled the old way, for both the launchd plist and the systemd service. Part of #394
entelecheia
added a commit
that referenced
this pull request
Nov 9, 2026
entelecheia
added a commit
that referenced
this pull request
Nov 9, 2026
* chore(release): add v2.81.0 release notes Covers #399, #400, #401, #402, #404, #405, #407 and #409. #406 (spec docs) and #408 (test isolation) are left out per the release-note rules. Claude-Session: https://claude.ai/code/session_01Vb48Vt5iGMfc7MJWPTCWCL * chore(release): v2.81.0 notes, apply review findings Name the resume reload fix (#411), the peer pair upgrade, the exact anchor scope, the dot sync target change and the dot admit wait. Claude-Session: https://claude.ai/code/session_01Vb48Vt5iGMfc7MJWPTCWCL
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Part of #394 (row: hidden
dot sync --profile, from #380).Removes the hidden persistent
--profileflag ondot sync, the fallback that read it (syncStoreFrom), the stderr deprecation warning, and thePersistentPreRunEthat refused two conflicting values.dot sync --profilenow fails withunknown flag, which is the end of the one-release alias window opened in #380.Every command dot prints or runs on the peer moves to
--store: the owner-mismatch refusal (dot sync owner --store=<p> --set-self,dot sync pull --store=<p>), the NFD fix thatdot peer doctorand the peer inventory report, the handover and fence messages,noLocalOwnerFix, the "list them with" hint at the end ofdot peer sync, the launchd plist and systemd unit the scheduler renders, and the launchd spawn-failed reload line.remotePeerNormalizeCommandnow executessync names normalize --store=peer --yeson the peer; the comment that hedged for a peer running a dot that knew only--profileno longer holds.--profileonapply,diff,config,status,init,watchdog setup|notifyandcontainer setupis untouched: that flag selects a dotfiles profile and has nothing to do with the sync store. The codex--profilelaunch argument ininternal/aipolicyis also untouched.Installed scheduler units
A unit stores the whole command line and the service manager replays it on every tick, and neither
dot updatenordot applyrewrites an installed unit, so a scheduler installed for a non-default store kept runningdot sync push --profile=<store>and failed withunknown flagon every tick after this change. Both halves of the fix are here:dot sync resume(the command documented as the one that re-arms installed units) rewrites a stale unit from--profile=to--store=before the reload it already performs, and prints✓ rewrote the installed scheduler unit to call dot sync --store=<store>. The rewrite changes only the flag, so store, mode, interval,--homeand log paths stay exactly as installed; a fresh render would re-apply setup's current settings and could silently re-time or re-enable a scheduler. Under--dry-runit prints the rewrite it would make.dot sync statusstays read-only and names an installed unit that still calls the retired flag, together with thedot sync --store=<store> pause && dot sync --store=<store> resumepair that heals it.The peer store is exempt from both:
com.dotfiles.peer.plistbelongs todot peer setup(it runsdot peer syncand renders no store argument), anddot sync --store=peeris refused, so rewriting that file would leave a command that cannot run.Scheduler units checked: every dotfiles unit installed on this Mac (
~/Library/LaunchAgents/com.dotfiles.sync.plist,com.dotfiles.peer.plist,com.dotfiles.noindex.plist) carries no store argument, so none of them is affected and none can exercise the migration against a live unit; the new tests do it against real template renders spelled the old way. The peer Mac was unreachable over WARP at review time, so its units are unverified.Known ceiling: the detection is train-only (a unit whose store argument was written some other way than
--profile=/--store=is left alone), and a hand-edited unit that names no store keeps working as before.Evidence and checks
On 2026-11-09 both Macs run dot 2.80.1 (
dot versionlocally and over ssh on the peer), which knows--store. No peer can be left on a release that only accepts--profile.make docs: regenerates nothing (the flag was hidden, sodocs/commandsnever listed it; no overlap left with refactor(cli)!: remove hidden --include-secrets flag #402).go test ./internal/syncer/...: pass, real-rsync parity tests included (303s).go test ./internal/cli/...: pass, 25s. TheTestJSONGoldens/ai_policy_*failure from the previous round was an exportedCLAUDE_CONFIG_DIRin the shell; withenv -u CLAUDE_CONFIG_DIRthe whole suite is green.make lint: 0 issues.make build: ok.TestSchedulerMigratesUnitsInstalledWithTheRetiredStoreFlag(old-style fixture rendered from the current templates with--store=spelled--profile=, for both the launchd plist and the systemd service, byte-identical to a fresh render after migrating and a no-op on a second pass),TestSchedulerLeavesUnitsWithoutAStoreArgumentAlone,TestSchedulerLeavesThePeerSchedulerUnitAlone,TestRetiredStoreInReadsArgumentBoundaries,TestSyncResumeMigratesAUnitInstalledWithTheRetiredStoreFlag(resume rewrites then reloads, service-manager argv is exactly the state query plus the resume), andTestSyncReportsAUnitInstalledWithTheRetiredStoreFlag(status warns, stays read-only, resume heals it, status is quiet afterwards).Release note:
dot syncno longer accepts the deprecated--profilealias, and the guidance it prints and the jobs it installs use--store;dot sync statusnow names an installed scheduler unit still running the retired flag anddot sync resumerewrites it in place. Action needed: ifdot sync status --store=<store>warns about the installed unit, rundot sync resume --store=<store>(ordot sync --store=<store> pause && dot sync --store=<store> resume), or re-install it withdot sync setup --store=<store> --push-interval=<DUR>/--pull-interval=<DUR>.