Backed-up photos row in clean sheet now toggles on/off like other rows.
Page 1 storage stats (USED/FREE/JUNK) pushed to bottom of screen via Spacer,
matching the alignment of the stats strip on the main backup page.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Pages 1 and 2 both show PULL/RELEASE TO CLEAN. Only page 0
triggers backup; all others trigger the clean confirmation.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Moved pullGesture and pageSwitchGesture from the ring ZStack to
the outermost ZStack via simultaneousGesture. Both gestures now
fire from anywhere on the screen without blocking scroll.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Applies to both compactStat (NAS/iPhone/Pending/Safe) and
storageStat (USED/FREE/JUNK).
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Removed gray card background/clipShape. storageStat now uses
30pt light monospacedDigit with statDividers, identical to
compactStat on page 0. Consistent design across all pages.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
36pt semibold rounded → 36pt light monospacedDigit, matching
the thin numeral design language used across all widgets.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Semibold rounded was wrong. Values are now 30pt light matching
the thin numeral style shown in the design screenshot.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Pull zone now shown on pages 0 and 2. Labels route to
PULL/RELEASE TO CLEAN on page 2. Releasing triggers
triggerCleanViaPull() which fires the clean confirmation dialog
with the same ripple effect as the backup pull.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
TimelineView drives a 2.4s cycle: two sin-bell chevron pulses
(primary + ghost trailing 7pt below) then a pause. Mirrors the
cascadingArrows pattern — suggests pull-down on all three pages.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
All pages now share circleSection. Page 2 shows a green arc
(freeable/backed-up fractions) with a leaf icon and GB FREEABLE
label at center, matching the anchored position of pages 0 and 1.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Three amber/gray capsule pills show the current page; a small
rounded-square button with an amber border advances to the next page.
Positioned below the status message (between heroView and the scroll
content), replacing the old left/right chevron bar at the bottom.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
9 files were present on disk but not referenced by NASBackup.xcodeproj,
causing 18 "cannot find type in scope" errors. Added to their respective
groups (Services, Models, Browse, Components) in the main app target.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Pages 0 (backup) and 1 (storage) now share a single KisaniRingView
instance. Arc values (progress, color) swap in-place via computed
properties so the ring never re-mounts or shifts position during
horizontal swipes. Page 2 (cleanup) retains its own hero layout.
Pull gesture and page-switch gesture coexist via simultaneousGesture.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Pull zone is now an .overlay(alignment: .top) offset upward by its own
height, so it appears above the ring without pushing it down. Removed the
ring's offset(y:) so it stays perfectly fixed while pulling.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
BackupView:
- Move circle hero outside ScrollView so DragGesture no longer
competes with the ScrollView's pull-to-refresh capture
- Pull-to-refresh (native iOS gesture) now starts/resumes backup
instead of only refreshing status
- Ring diameter 240 → 270pt, ripple circle 258 → 288pt
- Restructure: heroView (fixed) + ScrollView(belowContent) + pageNav
- page1/page2 content split into heroView + belowContent dispatch
BackupStatusService:
- refreshAfterBackup now bumps nasArchiveTotal by newlyOnNAS count
(fileSize > 0 entries only), preventing the NAS stat from
flickering back to the pre-backup value while writeManifest runs
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
- Replace paginated ring center with full WidgetState machine (offline/checking/pending/running/safe/paused)
- Pull-down gesture on circle triggers backup with haptic + ripple animation
- Cascading TimelineView arrows, status dot pulse, live sub-label
- Add innerColor param to KisaniRingView for per-state color transitions
- Add pageSwitchGesture for horizontal swipe on pages 1 & 2
- Remove all action buttons; pull gesture is the sole backup trigger
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Adds a blue-tinted shadow glow to the outer NAS progress arc for
visibility. Darkens track circles (0.92 → 0.76) and the middle
static ring (0.78 → 0.62). Bumps outer arc lineWidth to 1.5.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Translates the React KisaniCircle design to SwiftUI: three concentric
thin rings (outer slate progress, middle static gray, inner green
progress) replace the old single ProgressRing. All ring center text
updated to ultraLight 46pt numbers with 8pt monospaced uppercase labels.
Adds ringCheckmark (22px circle + tiny checkmark) for completed/all-safe
states. Updates checking, disconnected, failed center views to match.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Replace ZStack-based cell with Color.clear.aspectRatio(1,.fit) + .overlay.
The overlay constrains all children (image, video badge, sync dot) to the
exact square frame, so scaledToFill cannot inflate the cell height in
LazyVGrid. Previously, the ZStack let the filled image push beyond the
aspectRatio constraint, causing inconsistent row heights.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
WHAT TO BACK UP section:
- Remove "New files only" / "Skip files already on NAS" toggle from Settings
- Remove BackupFilter.newFilesOnly field — duplicate protection is always active
(fileExists pre-check + hybrid filterIndex ensure NAS files are never re-uploaded)
NOTIFICATIONS section:
- Replace static decorative rows with real ToggleRow controls backed by
ConnectionStore (notifyOnStart=false, notifyOnComplete=true, notifyOnErrors=true)
- BackupEngine now gates notification calls behind these store flags
NETWORK section:
- WiFi row always shows ● On / ● Off indicator (was only shown when SSID known)
- Shows "Trusted" / "Unknown" when SSID is available, "On" / "Off" otherwise
Tests:
- makeAssets: add missing isScreenRecording parameter
- Rename test_backup_skipsExistingFiles_whenNewFilesOnly → _alwaysHardcoded,
remove filter.newFilesOnly assignment
- MockNASService: add downloadData, writeData, deleteFile to satisfy protocol
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
- NASTransferProtocol: add deleteFile(at:) to enable server-side deletion
- SFTPService: implement deleteFile using Citadel sftp.remove(at:)
- SMBService: implement deleteFile using SMBClient deleteFile(path:)
- SyncView FREE UP SPACE: two new cards —
• "Clean up screenshots" — shows count of phone screenshots already
backed up to NAS, opens review sheet for bulk PHPhotoLibrary deletion
• "NAS screenshot quarantine" — identifies NAS files that match phone
screenshots, offers to move them into a Screenshots/ subfolder via
download → writeData → deleteFile per file
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Phase 1 was only checking the Kisani manifest for sync status. Phone photos
uploaded by other tools had no manifest entry, showed as orange (phone-only)
even though they were already on the NAS.
Now Phase 1 falls back to a filename lookup against the full NAS directory
listing: manifest match → green (.synced), NAS directory match → green (.synced),
no match → orange (.phone). Phase 3 skips filenames already resolved in Phase 1.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Gallery NAS was only showing Kisani-manifest-tracked files (2 of 7432). Now
fetches the full NAS directory listing on load and adds Phase 3 items for files
not in the manifest (uploaded by other tools) — all 7432 files appear in the
gallery with thumbnails loaded lazily on scroll.
Settings network section: "No Wi-Fi" → "Wi-Fi" when connected but SSID is nil
(no location permission). Added cellular status row below Wi-Fi row.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
The root cause: buildManifestIndex was discarding manifest localIdentifiers and
switching to pure filename-only matching whenever dirCount > manifest.totalCount
(which is always true when the NAS has files from other tools). If filenames didn't
match exactly, Already Safe dropped to 0 on every stale-cache reconcile.
Fix: introduce ManifestIndex(manifest:nasFilenames:totalCount:) — a hybrid that keeps
Kisani's localIdentifiers as primary keys AND adds all NAS directory filenames for
filename-fallback. Used in:
- buildManifestIndex (full reconcile)
- fast path (when dirCount > cached.entries.count, previously gated at <= 1)
- BackupEngine step 3b (so "Back up again" only uploads genuinely missing files)
Also simplify countSafe / pendingIDs / step-4 filter to always try filename fallback
(removes the isFilenameOnly gate that blocked matches when the index had IDs).
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Use UIColor.secondarySystemGroupedBackground / tertiarySystemFill so
the segmented control adapts correctly to light and dark mode; darken
the shadow in dark mode so the selected pill is still visible.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
- Adds cellular row (antenna icon) below Wi-Fi in connection status section;
orange when active, muted when not in use
- Replaces "Retry Failed" header button in errors section with "Clear Errors"
(destructive, calls clearErrorHistory); retry action retained in error detail sheet
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Screen recordings are videos captured by iOS screen recording and are
almost never wanted in a photo backup. They are now excluded from all
counts and backup queues unless explicitly enabled.
Changes:
- BackupFilter.includeScreenRecordings (default: false) added
- PHAssetMediaSubtype.videoScreenRecording predicate applied in both
fetchAssets() and fetchResultForReconciliation() — excluded at the
PHFetchRequest level before enumeration
- LocalPhotoIndex.Record.isScreenRecording populated via makeRecord()
so the fast-path countSafe/count also respects the toggle
- LocalPhotoIndex.passes() checks includeScreenRecordings
- Settings "Screen Recordings" toggle added below Screenshots
This also resolves the transient count flash: screen recordings were
being included in the "need backup" tally on the initial fast-path
reconcile, then dropping when the filter was fully applied.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
BackupEngine was building pendingAssets from ManifestIndex(manifest:) which
only knows about files Kisani previously uploaded. When most NAS files were
uploaded by other tools, nearly all phone photos passed the pending filter
and "Back up again" would queue and attempt to re-upload everything.
Fix: step 3b builds a filterIndex from the cached NAS directory filenames
(or a fresh listDirectory scan when cache is unavailable) when the directory
has more files than the manifest. The filename-based filterIndex correctly
identifies only photos whose filenames are absent from the NAS, so "Back up
again" queues exactly the delta — not a full library re-upload.
The baseManifest/baseIndex are unchanged and still drive the manifest write.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
- buildManifestIndex now always scans the NAS directory on every full reconcile
(removed the `totalCount <= 1` guard) so NAS Archive count updates accurately
on every manual refresh, not just when the manifest is new/sparse
- The dir > manifest check still switches to filename-based matching when the
directory has more files than Kisani's manifest (uploaded by other tools)
- checkStatusRow shows NAS connection status (orange dot + "NAS Connected" /
red dot + "NAS Offline") below the timestamp line
- BackupFilter now conforms to Equatable; BackupView re-reconciles immediately
when the screenshot/video/RAW filter toggles change
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
- Move "X sec ago / refresh" row from inside statsStrip to directly below the
ring hero so the last-checked time is visible near the status it describes
- Remove "Checking status…" spinner from the action button — button shows
"Start backup" / "Back up again" at all times
- Remove "Checking…" spinner from the ring center (page 0) — ring always
shows the pending backup count or "All photos safe"
- Re-enable autoBackupEnabled default (true) now that already-safe counting
is accurate; auto backup will fire on LAN join when items need backup
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
When Kisani's manifest has 0–1 entries but the NAS directory contains thousands of
files uploaded by other tools, reconcile previously showed Already Safe = 1 because
countSafe only matched against the tiny manifest.
The validity check in buildManifestIndex now:
- Stores the filtered NAS filenames in NASManifestCache (directoryFilenames field)
- Returns a ManifestIndex(nasListing:) for filename-based matching against all NAS files
The fast-path reconcile now:
- Detects a sparse manifest (entries ≤ 1, dirCount > 1)
- Rebuilds ManifestIndex(nasFilenames:) from the cached filenames for O(1) filename matching
- Falls back to full NAS reconcile if cached filenames aren't available yet
Adds ManifestIndex.init(nasFilenames:) for fast-path reconstruction from cached strings.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Root cause of the alreadySafe=1 loop: after a backup where all assets are
collision-skipped (Object Name Collision), writeManifest() connects to the NAS
to write the updated manifest. If that connection fails or times out, the cache
is never updated with the 263 new entries, so the next refresh(force:true) reads
the old 1-entry cache and computes alreadySafe=1 → needBackup=263 → loop.
BackupEngine: after the upload loop, immediately merge all manifestEntries
(uploaded + collision-skipped + fileExists-skipped) into baseManifest and call
NASManifestCache.shared.update(). This happens before refreshAfterBackup so
the local truth is correct regardless of whether the NAS write later succeeds.
BackupStatusService.writeManifest: move NASManifestCache.shared.update() to
before transfer.writeData(). The local cache is now updated as soon as the
merged manifest is computed, decoupling local reconcile correctness from NAS
write success. The NAS write remains best-effort and non-blocking.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
Three changes that together eliminate the "Checking…" stall and excessive NAS
rescanning without requiring a database migration:
NASManifestCache: extend stale threshold from 5 min to 30 min. directoryCount
is now updated incrementally after every upload, and PHPhotoLibraryChangeObserver
keeps LocalPhotoIndex current — a full NAS roundtrip every 5 min was redundant
and forced the expensive reconcile path far more often than necessary.
BackupStatusService: add withTimeout(_:work:) using a racing ThrowingTaskGroup
so NAS operations (SMB connect, manifest download, directory listing) fail fast
instead of hanging indefinitely when the NAS is slow or unreachable. Timeouts:
15s for connect, 12s for manifest download, 12s for directory listings. On
timeout, BackupError.timeout is caught separately to preserve cached counts
instead of marking connectionState = .offline.
BackupStatusService: call LocalPhotoIndex.shared.loadOrBuild() at the start of
performRefresh before checking localIndexCount. Without this, AppDelegate's
concurrent loadOrBuild() task races with the first onActive() trigger; if it
loses, localIndexCount == 0 and the fast path is bypassed, forcing a full NAS
roundtrip on every cold launch.
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
SMBClient throws ErrorResponse with NTStatus.objectNameCollision (0xC0000035)
when a file already exists on the NAS. BackupError.uploadFailed wraps this,
hiding the real cause from the retry/failure logic.
Add isAlreadyExistsError() that unwraps BackupError.uploadFailed and checks
the underlying error for "object name collision", "already exist", or POSIX
EEXIST (code 17). When detected inside the retry catch block, the asset is
immediately marked as skipped and added to manifestEntries — no 3-retry
penalty (saves up to 6s per already-existing file), no false failure count.
The post-retry fileExists fallback is retained for edge cases where the file
lands on NAS but the error wasn't a collision (e.g. connection reset after
partial upload).
Also switch BackupQueueItem.UploadError to store the underlying transport
error's domain and description instead of the BackupError wrapper text, so
the SyncView failure list shows the real reason (e.g. "Access Denied",
"IO Timeout") rather than the generic "Failed to upload <file>".
Co-Authored-By: Kutesir <kutesir@provoc.ug>
Co-Authored-By: Sentry <sentry@provoc.ug>
