ingestFromWatch arbitrated by whole-document updatedAt, so concurrent edits
to the same workout on both devices (phone edits exercise A while the watch
completes exercise B) lost one side wholesale — the newer snapshot replaced
the other (H1).
Reconcile per log instead. WorkoutMergePlanner (pure, deterministic) unions
logs by id, resolves each by newest per-log updatedAt, and applies
phone-authored deletion tombstones so an absent log is never ambiguous
between "deleted on the phone" and "just added on the watch". Edits to
different exercises now commute — delivery order and offline gaps stop
mattering. A stale/duplicate push merges back to exactly the cached doc, so
ingest re-pushes authoritative state rather than writing.
The per-log updatedAt scaffolding shipped (unused) in schema v4; it's now
stamped by transition(to:) on status flips and a new touch() at the
content-only edit sites (order, notes, machine settings, adjusted entries,
new logs) on both phone and watch. deletedLogIDs is new: additive on the
wire and cache, phone-authored (deleteLog), pruned after a 30-day grace.
Because an older build rewriting a file would strip the tombstones and
resurrect a deleted exercise, WorkoutDocument schema bumps 4->5 (forward
gate quarantines old builds) and the cache bumps 5->6.
recomputeStatusFromLogs takes an injectable now: so the merge recomputes
status/end deterministically. WorkoutMergePlannerTests pins the decision
table (commute, per-log newer-wins, legacy-nil, watch-add, tombstone
honored/resurrect/union/prune, status recompute, no-op re-push);
WorkoutDocumentMapperTests gains the deletedLogIDs round-trip.
Every completed set now writes a SetEntry (reps/weight or seconds),
pre-filled from the plan by transition(to:) so the list checkbox, both
run flows, and One More all capture for free; reset clears, skip keeps
partials. The rest and finish pages show the just-done set as a pill
that opens a stepper sheet for correcting reps and weight (2.5 lb /
1.25 kg steps). The Weight Progression chart plots the top-set actual
weight and workout volume sums recorded sets, both falling back to the
plan for legacy logs via effectiveSetEntries.
Storage side of UX #3 rides along: plan weights are Double now.
Schema bumps: SplitDocument 2→3, WorkoutDocument 3→4 (a fractional
weight fails an older Int decode, and a rewrite would strip the
irreplaceable actuals), SwiftData cache 4→5. A per-log updatedAt is
reserved for the future cross-device log merge.
Claude-Session: https://claude.ai/code/session_01HJDQQDA9QdP8zByg43H5v3
Editing a workout's start date now removes the file at its old month
bucket so the record no longer duplicates on the next reconcile. Seed
reconcile re-checks the tombstone veto before overwriting an upgraded
seed. The watch applies authoritative-empty pushes so remote deletes
prune, and a re-saved finished workout keeps its original end time.
Adds unit tests for the mappers, path bucketing, and status machine.
Claude-Session: https://claude.ai/code/session_01HJDQQDA9QdP8zByg43H5v3
Machine-based exercises now carry ordered name/value comfort settings
(seat height, back-rest position, ...) on both ExerciseDocument and, as a
plan-time snapshot, WorkoutLogDocument — the optional array doubles as the
machine flag. Editable from the exercise editor's new Machine section and
from the workout row's settings sheet, whose mid-workout edits write back
to the originating split (workout saved first, so seed clone-on-edit
repointing can't clobber the log edit).
Schema tightening rides the same rev: splits bump to v2 (weight-reminder
fields and the unused exercise category removed), workouts to v3 (the
derived `completed` flag removed; status is the single source). Starter
seeds regenerated at v2 with unchanged ULIDs; SwiftData cache schema
bumped to rebuild. SCHEMA.md documents the shapes.
Claude-Session: https://claude.ai/code/session_01LEoff8bXGBS83tK1c55Mf7
Exercise Library/ holds per-exercise reference docs (setup, cues,
mistakes, progressions) with SVG visuals and a Python-rendered motion
pipeline; Workouts/ExerciseFigure renders the bundled *.motion.json
rigs as animated stick figures on the exercise screen. Exercises gain
a warm-up/main-circuit category, timed exercises display hold time via
planSummary, and a completed exercise reopens to a check screen instead
of its timers.
Replaces the app-local copies of the extracted storage core with the
IndieSync 0.1.0 package (pinned from the new tag): ICloudFileManager ->
DocumentFileStore + TombstoneStore, ICloudFileMonitor ->
MetadataObserver (batched events with the content-date churn gate),
and the shared ULID / DocumentCoder / Tombstone / VersionedDocument
pieces. SyncEngine stays app-specific and is rewired onto the package
types; documents rename currentSchema -> currentSchemaVersion to adopt
the package protocol. ULID.make() remains as a shim minting the string
form the app keys on.
Behavior-preserving: identical JSON bytes (same coder config), same
stub wire format (kind encodes as the same strings), same soft-delete
ordering, same 30-day prune. WorkoutDocument keeps its local-calendar
month bucketing rather than adopting TimeBucketedLayout's UTC paths --
changing derivation would strand existing files. The watch target
links the package too (ULID + DocumentCoder via Shared); iOS, watchOS,
and widget targets all build.
Watch sessions now run an HKLiveWorkoutBuilder: heart rate and active
energy are collected live (shown in an in-workout HUD), saved to Health
as a real HKWorkout on completion, and carried back to the phone as
WorkoutMetrics on the workout document. Phone-only workouts get an
estimated Health workout (MET x bodyweight x duration) after a 10s
debounce that a late-arriving watch session cancels; a launch sweep
backfills workouts completed within the last day. Dedupe is keyed on
metrics.healthKitWorkoutUUID plus the metrics source.
Splits gain an activity type (strength, functional, HIIT, core, cardio,
cycling) that categorizes the Health workout and picks the MET value.
A post-workout summary sheet (duration, calories, avg/max HR, HR zones,
total volume) fills in live and is also shown on completed workouts.
Weights can now display in lb or kg (display-only relabel), synced to
the watch over the existing application context.
WorkoutDocument schema 1->2 (metrics are irreplaceable, so older apps
must quarantine rather than strip them); cache schema 2 rebuilds the
SwiftData store with the new metric columns. Deleting a workout in the
app intentionally leaves its Health record in place.
Watch-side follow-through for the End Workout flow:
- The phone now pushes an authoritative set (in-progress, not-started, and
completed within 24h) instead of the 25 most-recent workouts, and the watch
prunes any workout absent from it. So a Discard/Delete (or a completed run aging
out) drops off the watch, empties its active list, and ends the HKWorkoutSession
— fixing the persistent wrist-raise re-foregrounding. The watch never originates
a workout, so pruning can't lose local data; the 24h grace keeps a just-finished
run on screen. The gate pops if the run you're viewing is pruned.
UX tweaks:
- The in-workout ⋯ is now a pull-down Menu (Add Exercise / End Workout) rather than
an action sheet.
- Starting a split while another workout is still active now prompts to end the
current one(s) — keeping their progress — or run in parallel. Wired into both
start paths (the split picker and "Start This Split"), via a shared
WorkoutDocument.endKeepingProgress() helper.
Claude-Session: https://claude.ai/code/session_01SCv7zvGFcKy47KSTnTLxRe
Replace the in-workout "+" toolbar button with an ellipsis menu offering
"Add Exercise" and "End Workout". Ending opens a Save/Discard action sheet:
Save marks the remaining exercises as skipped and resolves the workout to
completed (stamping end), which drops it off the watch's active list and ends
the watch's HealthKit session; Discard soft-deletes it.
Teach the status-from-logs derivation that a skipped log is terminal, and
consolidate the three duplicated copies into a single shared
WorkoutDocument.recomputeStatusFromLogs() so an ended workout stays finished
regardless of which screen the next edit comes from.
Claude-Session: https://claude.ai/code/session_01SCv7zvGFcKy47KSTnTLxRe
Replace Core Data + NSPersistentCloudKitContainer + App-Group store +
WatchConnectivity dictionary sync with the QuickRabbit iCloud-documents
architecture:
- iCloud Drive JSON documents are the sole source of truth (one file per
aggregate: Splits/<ULID>.json, Workouts/YYYY/MM/<ULID>.json), with a
rebuildable SwiftData cache populated only by an NSMetadataQuery observer
and a connect-time reconcile. Soft-delete tombstones; hard iCloud gate.
- Shared model layer (ULID, Codable *Documents + stateless mappers, @Model
cache entities, SwiftData container) compiled into both targets.
- New iPhone<->Watch bridge over WatchConnectivity keyed by ULIDs; the phone
is the sole writer of iCloud Drive, the watch round-trips documents.
- AppServices DI + iCloud-required root gate; Swift 6 strict concurrency.
- Starter splits generated on demand from the bundled YAML catalogs.
- Migrate to XcodeGen (project.yml), iOS 26 / watchOS 26; CloudDocuments
entitlement (drop CloudKit/App Group/aps-environment).
- Duration stored as Int seconds (was a Date epoch hack); fix workout
end-on-create, undismissable delete dialog, toolbar-hiding nav stacks,
and the Settings placeholder.
- Add README/CHANGELOG/LICENSE, .gitignore, refreshed REQUIREMENTS, and the
Scripts/ TestFlight pipeline (release.sh + ASC API scripts).
MARKETING_VERSION 2.0.
rzen