- Move LargeSoftCache from amethyst/model to commons/jvmAndroid/model/cache
so both Android and Desktop share the same WeakReference cache implementation
- Change ICacheProvider return types from Any? to proper types (User?, Note?)
since these model classes already live in commons
- Remove unnecessary casts in ThreadAssembler, ChatNewMessageState, SearchBarState
- Improve LargeSoftCache.cleanUp() to use single-pass iterator (zero allocation)
- Update Android imports in LocalCache, LargeSoftCacheAddressExt, and test
Per Vitor's feedback on PR #1905: BoundedLargeCache evicts arbitrarily.
LargeSoftCache uses WeakReferences so GC respects the reference graph.
Both screens now show cached data instantly on navigation instead of
showing loading states while re-fetching from relays.
- BookmarksScreen: reads cached BookmarkListEvent from addressableNotes,
seeds public bookmark events from notes cache
- ReadsScreen: seeds long-form notes (kind 30023) from notes cache
Relay subscriptions still run to fetch fresh data in parallel.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Cached value now stays visible until relay data exceeds it, preventing
the count jumping from cached→0→ticking back up on each navigation.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Profile screen now reads initial values from cache so navigating back
shows data instantly instead of re-fetching everything from scratch.
- Seed displayName/about/picture from User.metadataOrNull() on compose
- Add followerCounts/followingCounts maps to DesktopLocalCache
- Write counts on each update, read on profile open
- Follower count still ticks up live (good UX) but starts from cached value
- 4 new tests for profile count caching + metadata cache reads
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause of "0 relays, 0 notes, 0 follows": all screens used
connectedRelays as the relay set for subscriptions, but NostrClient
uses relay-on-demand — relays only connect when openReqSubscription
is called with their URL. This created a deadlock: screens waited for
connected relays, but relays only connect when screens subscribe.
Fix: use relayStatuses.keys (registered/available relays) which are
populated by addDefaultRelays() at startup. openReqSubscription
triggers connection on-demand.
Affected screens: FeedScreen, UserProfileScreen, ThreadScreen,
NotificationsScreen, BookmarksScreen, ReadsScreen, NewDmDialog.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add RelayHealthIndicator component — shows elapsed time since last
relay event. Hidden when <30s (healthy), shows "45s ago" / "3m ago"
when stale. Placed above BunkerHeartbeatIndicator in NavigationRail.
Reads from Coordinator's subscriptionHealth StateFlow.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Rewrite FeedScreen from inline EventCollectionState + 8 rememberSubscription()
calls to cache-centric DesktopFeedViewModel pattern:
- ViewModel keyed on feedMode with DisposableEffect cleanup (destroy())
- FeedState pattern matching (Loading/Loaded/Empty/Error)
- FeedNoteCard takes Note instead of Event, reads counts from Note.flowSet
- DisposableEffect for note.clearFlow() cleanup in LazyColumn
- DisposableEffect for Coordinator interaction subscription lifecycle
- LaunchedEffect for rate-limited metadata loading via Coordinator
- Extract FeedHeader into separate composable
- Add destroy() to DesktopFeedViewModel (ViewModel.clear() is internal in KMP)
- Update UserProfileScreen to look up Note from cache for FeedNoteCard
Removes ~350 lines of inline state management (EventCollectionState,
zapsByEvent, reactionIdsByEvent, replyIdsByEvent, repostIdsByEvent,
followedUsers, bookmarkList, 8 rememberSubscription blocks).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add try-catch in consumeEvent() so one bad event doesn't kill the
pipeline. Add requestInteractions()/releaseInteractions() with Job
tracking via ConcurrentHashMap for screen-scoped subscription lifecycle.
Add SubscriptionHealth tracking (lastEventReceivedAt, eoseReceived)
exposed as StateFlow for UI consumption.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
toNoteDisplayData() now reads User.toBestDisplayName() from cache
instead of always showing npub. Falls back to npub if no metadata.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Resolve LruCache vs LargeCache: use LargeCache (lock-free ConcurrentSkipListMap)
with BoundedLargeCache wrapper for size enforcement on put()
- Confirm LargeCache available on desktop via quartz jvmAndroid source set
- Detail FeedNoteCard rewrite: Note model field mapping, 5 subscription removals
- Fix filter examples to use filterIntoSet (LargeCache API)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add quartz protocol package for NIP-86 JSON-RPC relay management with
NIP-98 HTTP authorization, and a management UI accessible from the
relay information screen when the relay advertises NIP-86 support.
Quartz (nip86RelayManagement/):
- Nip86Method: all 21 method constants from the spec
- Nip86Request: serializable JSON-RPC request with factory methods
- Nip86Response: response model with typed result parsing helpers
- Nip86Client: builds NIP-98 auth headers, serializes requests,
parses responses for each result type (pubkeys, events, kinds, IPs)
Amethyst:
- Nip86Retriever: OkHttp-based HTTP executor for NIP-86 calls
- RelayManagementViewModel: state management for all NIP-86 operations
- RelayManagementScreen: tabbed UI (Pubkeys, Events, Kinds, IPs,
Settings) with add/remove dialogs, moderation queue, and relay
settings fields
- Route.RelayManagement added to navigation
- "Manage Relay" button shown in RelayInformationScreen top bar when
relay's supported_nips includes "86"
https://claude.ai/code/session_01QckCAm1T8pJqqmURBiNtaQ
Add support for NIP-77 (Negentropy Syncing) protocol messages using
the negentropy-kmp library for efficient set reconciliation between
client and relay.
New files:
- NegOpenCmd, NegMsgCmd, NegCloseCmd: Client-to-relay commands
- NegMsgMessage, NegErrMessage: Relay-to-client messages
- NegentropySession: Client-side reconciliation orchestrator
- NegentropyServerSession: Server-side reconciliation handler
- NegentropyManager: High-level sync manager with listener callbacks
Updated serializers:
- Jackson MessageSerializer/Deserializer for NEG-MSG, NEG-ERR
- Jackson CommandSerializer/Deserializer for NEG-OPEN, NEG-MSG, NEG-CLOSE
- Kotlin MessageKSerializer/CommandKSerializer for all NIP-77 types
Compatible with strfry relay's negentropy implementation via the
negentropy-kmp library which implements the same Protocol V1 spec.
https://claude.ai/code/session_01Dc6W1G1jURAAR9kzyVvk6g
Add a new screen allowing users to send NIP-62 relay deletion requests.
Users can select a specific relay or ALL RELAYS, pick a date (data
created before that date will be requested for deletion), and provide
an optional reason. Includes confirmation dialog with clear warnings
about the irreversible nature of vanish requests.
https://claude.ai/code/session_019Xrprdfq6pVN8beYrYUSr4
Derive the label from the existing searchTerm — if it ends with .bit,
show a chain-emoji label above each user result. No new StateFlow or
ViewModel state management needed.
Remove namecoinResolvedUser StateFlow and the chain-emoji label
from search results. .bit resolution still works through the
existing NIP-05 path — just no special UI decoration.
Address @vitorpamplona's review comments on the last commit:
1. Remove NamecoinResolutionState sealed class and status banners
— the existing NIP-05 resolver already handles .bit, no need for
separate Namecoin-specific UI state (spinner, chain emoji, error)
2. Remove separate toNip05IdOrNull() resolution path
— widen the existing `if (term.contains('@'))` gate in both
SearchBarViewModel and UserSuggestionState to also accept bare
.bit domains (synthesize _@domain.bit → existing NIP-05 flow)
3. Remove d/ and id/ identifier support from search
— per feedback, only NIP-05 style resolution (.bit) in search
4. Remove special Namecoin result rendering in ImportFollowList
— just show UserLine for all results, no separation needed
5. Keep a lightweight 'Namecoin' label on resolved users in search
— namecoinResolvedUser StateFlow tracks the .bit-resolved user,
shown as a small label above the UserCompose row
Net: -405 lines, +50 lines across 6 files.
Widen the existing NIP-05 resolution gate in UserSuggestionState and
SearchBarViewModel to also accept bare .bit domains and d//id/ Namecoin
identifiers. A synthesised Nip05Id is passed to the existing
nip05Client.get() path — which already routes .bit to the
NamecoinNameResolver — so no separate resolution logic is needed.
Add NamecoinResolutionState (Idle/Resolving/Resolved/Error) as a
StateFlow so the UI can show progress. ImportFollowListSelectUserScreen
and SearchScreen display an animated status banner (spinner while
resolving, chain emoji on success, warning on error) and label
Namecoin-resolved profiles in the results list.
The field no longer means 'trust all certificates' — since the Samsung
One UI 7 fix, it means 'use the pinned trust store (hardcoded +
TOFU-pinned certs + system CAs) instead of system-only trust.'
The old name was actively misleading and could cause a future
contributor to interpret the flag literally, potentially reintroducing
the trust-all pattern that Samsung Knox and GrapheneOS reject.
Renamed across all 4 files:
- ElectrumxServer.kt: field definition + updated KDoc
- ElectrumXClient.kt: createSocket() usage + comments
- NamecoinSettings.kt: parseServerString() usage + comments
- NamecoinSettingsTest.kt: test assertions
Add support for NIP-B0 web bookmarks with a complete protocol
implementation in Quartz and a full UI in Amethyst for adding,
editing, browsing, and deleting web bookmarks.
Quartz:
- WebBookmarkEvent (kind 39701) as an addressable event
- Tag builder/parser extensions for title, published_at, hashtags
- Registered in EventFactory
Amethyst:
- WebBookmarksScreen with FAB to add, and per-card edit/delete/open
- WebBookmarkEditDialog for add/edit with URL, title, description, tags
- WebBookmarkFeedFilter querying addressable notes by kind + pubkey
- Account.sendWebBookmark() and Account.deleteWebBookmark()
- LocalCache.consume() for WebBookmarkEvent
- Drawer navigation entry with Language icon
- Route.WebBookmarks and navigation registration
https://claude.ai/code/session_01UzfLJttwuJzovtb8HX5F9n
Address compatibility and security issues identified by reviewing
GrapheneOS source (hardened Conscrypt, strict TLS enforcement):
1. .onion: prefer pinned factory, fall back to trust-all
GrapheneOS patches Conscrypt with stricter TLS enforcement that may
reject no-op X509TrustManagers even for proxied sockets. The .onion
server's cert is already in PINNED_ELECTRUMX_CERTS (same operator as
electrumx.testls.space), so we now use cachedPinnedSslFactory() as
the primary path. onionSslFactory() (trust-all) is kept as a fallback
on SSLHandshakeException only — this handles cert rotation or unknown
.onion servers gracefully.
2. TOFU: require explicit user confirmation before pinning
Previously, Test Connection auto-pinned every cert on success. An
attacker performing MITM during first test would get their cert
permanently trusted. Now each new cert triggers an AlertDialog showing
the full SHA-256 fingerprint. Users must explicitly accept ('Trust')
or reject each cert — aligning with GrapheneOS's philosophy of
explicit trust decisions.
3. Clarify trustAllCerts semantics
The field name is misleading post-refactor: it now means 'use pinned
trust store' not 'trust all certificates'. Added TODO to rename to
usePinnedTrustStore and updated inline comments to prevent future
misinterpretation.
Note: hostname verification on raw SSLSocket is a pre-existing gap
(not introduced by the Samsung fix PR) — SSLSocket.createSocket() uses
the host parameter for SNI only, not hostname verification. A follow-up
should add endpointIdentificationAlgorithm='HTTPS' or fingerprint-based
verification for pinned certs.
* 'main' of https://github.com/vitorpamplona/amethyst:
reuse hex methods.
spotless
delete voice files on failuer
added progress and test plan to doc
fix: delete abandoned compressed video when larger than original fix: eagerly delete intermediate temp files in upload pipeline fix: delete temp file from MediaCompressorFileUtils after image compression refactor: simplify temp file cleanup logic - Remove TOCTOU anti-pattern (file.exists() before file.delete()) - Consolidate double deleteTempUri calls into single conditional - Remove restating comments
analysis for temporary files
fix: clean up temp files after upload completes
perf: optimize ChaCha20 and XChaCha20-Poly1305 for speed
style: apply spotless formatting to ChaCha20/Poly1305 sources
feat: replace libsodium with pure Kotlin ChaCha20/Poly1305 implementation
nrobi144