Add a private provider-neutral SQLite WAL cache with transactional migrations, retained raw payloads and tombstones, per-window opaque checkpoints, and reserved outbox/conflict tables. Verify migrations, permissions, identity invariants, concurrent access, atomic pull-page rollback, and future-schema rejection with deterministic tests. No OAuth, networking, secrets, or Graph synchronization is included yet.
74 lines
3.3 KiB
Markdown
74 lines
3.3 KiB
Markdown
# Roadmap
|
|
|
|
## Foundation — current milestone
|
|
|
|
- C++20 domain model and deterministic six-week month layout
|
|
- Responsive, theme-native ANSI month view with keyboard navigation
|
|
- Appointment focus, dense-day traversal, and full-frame detail reader
|
|
- Validated add/edit/delete forms with confirmation and failure rollback
|
|
- Guarded atomic `.ics` persistence, advisory locking, and round-trip tests
|
|
- Bounded session undo/redo, exact external-change detection, and backup recovery
|
|
- Windowed recurrence expansion, compatible RDATE, EXDATE, IANA TZID, and
|
|
multi-day segment cues
|
|
- Occurrence-aware appointment search with exact result-to-grid navigation
|
|
- Session calendar toggles shared by month rendering, focus, and search
|
|
- Read-only 42-day agenda with bounded recurrence and exact return navigation
|
|
- Configurable week start with Monday-default CLI and consistent grid queries
|
|
- Explicit atomic import/export with validation, collision refusal, and backups
|
|
- Meson build, Nix development shell, desktop entry, and Hyprland launcher
|
|
- Seeded sample data when explicitly requested, never silent data mutation
|
|
|
|
## 0.1 — complete local calendar
|
|
|
|
- Advanced recurrence rule editing and detached RECURRENCE-ID overrides
|
|
- Embedded VTIMEZONE support
|
|
- Screen-reader-friendly linear view and full Unicode display-width handling
|
|
|
|
## 0.2 — durable cache and provider boundary — in progress
|
|
|
|
Completed cache foundation:
|
|
|
|
- Provider-neutral SQLite WAL cache in a private local file
|
|
- Versioned, all-or-nothing schema migrations
|
|
- Raw provider payload retention, inactive calendars, and event tombstones
|
|
- Per-calendar/window opaque cursors committed atomically with pulled pages
|
|
- Normalized event instances for immutable snapshots
|
|
- Reserved transactional outbox and conflict records for later remote writes
|
|
|
|
Next provider-boundary slices:
|
|
|
|
- Injectable HTTP, PKCE authentication, and deterministic fake transports
|
|
- Secret Service credentials; OAuth secrets are never stored in SQLite
|
|
- Generic `SyncProvider` contract and observable sync status
|
|
|
|
## 0.3 — CalDAV
|
|
|
|
- Discovery and incremental sync for Nextcloud and standards-compliant servers
|
|
- ETag-aware writes, offline mutation queue, conflict UI
|
|
- TLS and failure-path integration tests
|
|
|
|
## 0.4 — hosted providers
|
|
|
|
- Google Calendar adapter with incremental synchronization
|
|
- Microsoft 365 read-only discovery and Graph delta synchronization, followed by
|
|
ETag-aware remote writes after the conflict boundary is proven
|
|
- Account/calendar management UI and per-calendar ANSI identity
|
|
|
|
The current cache work is only the prerequisite durable boundary. No network,
|
|
OAuth, or Graph synchronization is implemented in that phase. After the cache
|
|
passes its verification gates, the immediate sequence is injectable HTTP plus
|
|
PKCE and fake transports, then read-only Microsoft Graph delta synchronization.
|
|
|
|
## 1.0 — distribution quality
|
|
|
|
- Stable configuration/data format and documented recovery procedures
|
|
- Nomarchy package, Hyprland defaults, man page, shell completions
|
|
- Performance budgets exercised against large and adversarial calendars
|
|
- Fuzzed iCalendar parser, accessibility review, translations
|
|
|
|
## Quality gates
|
|
|
|
Every milestone requires clean warning-enabled builds, unit and integration
|
|
tests, sanitizer runs, terminal-state recovery after signals/errors, no network
|
|
access without explicit account setup, and no loss of unknown calendar data.
|