Commit Graph
751 Commits
Author SHA1 Message Date
Kgothatso NgakoandClaude Opus 5 a9532ff115 fix(frost): resume a session whose own turn was cut short
Every step a session takes is driven by a message arriving, and a device
killed part way through its own pass is waiting on one that has already
been delivered. The signer set is recorded, so it will never be processed
again; the partial signature it would have produced was never made. The
coordinator asking again cannot reach past it either -- what a resend
asks a member for is the message this device believes it already sent --
so the session is a dead end on both sides, and in a room with several
open proposals it is the one that was mid-pass that stops.

So one sweep that no message drives: on opening a room, every session in
it that is not settled and whose owner has answered is handed to
`advance`, which does what it always does. It publishes nothing new --
each step asks whether its output exists rather than whether a stage says
it ran -- and only answered sessions are passed in, so the approval gate
is not approached, let alone reached past. A refusal counts as an answer
for the same reason: a device killed between deciding and publishing says
it on the next open rather than never.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-08 14:33:13 -05:00
Kgothatso NgakoandClaude Opus 5 1e59fd2334 fix(frost): a knock past the nonces is for the signers, not the quiet
A resend had one idea of who it was for at every stage -- the share
holders it had heard nothing at all from -- and that is a question only
the first stage can answer. A session reaches
COLLECTING_PARTIAL_SIGNATURES because enough members offered a nonce, so
every member who could be holding it up has been heard from: the set came
back holding the members who were never chosen and owe nothing, while the
chosen signer whose partial signature went missing was named nowhere.
Once the unchosen had answered or refused the set was empty, and the
control gave up before it reached the branch written for the stall it was
meant to repair.

So the set asks what the stage can answer. Of the signers the coordinator
chose, who has yet to send a partial signature -- which is the half of a
resend only they can carry out, since a partial signature is read by
nobody but the coordinator and no other device can repeat it. COMPLETE
keeps the old reading, and correctly: the signature is the coordinator's
own and nobody answers it.

The screen now asks the manager who is owed rather than deriving it from
the rows beside the button, the way it already asks about approvals: the
same reading decides whether a resend publishes anything, and two copies
of a rule that changes with the stage is a screen offering a knock the
manager would refuse. The transcript line and the notice say what the
members named are short of, since past the nonces they have answered once
already.

The test that looked like it covered this passed because it wrote the
stage into the row with no nonce on file anywhere -- a state the protocol
cannot produce, and exactly the state that hid the bug. It now reaches
the stage by offering a real nonce.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-08 14:32:09 -05:00
Kgothatso NgakoandClaude Opus 5 e4657df08e test(frost): cover the transport every ordinary session actually uses
Nine tests for gaps found by auditing the suite against the branches rather than
by counting tests.

The big one: every existing resend test ran on a NIP-17 room, which is only a
group's *first* signature. Every ordinary session runs in the Marmot admin room,
and that path had never executed -- no coverage at all for the store it writes
to, the tags it omits, or the id guard on that store. Given the bug fixed in
daf8365a was transport-specific, this was the gap most likely to hide the next
one. Three tests now cover it: a resend is one group event naming nobody with
the silent tag still on it, a nudged member republishes the same bytes as an
inner event, and the id guard holds on that store too.

Also covered, each of them a branch that had never run:

 - resending while collecting partial signatures, which repeats the signer set.
   One of three arms of the stage table, and the arm a stalled signer needs: they
   have the proposal and their nonce is in, and the signer set that says what to
   sign against never arrived.
 - the same stage table sending no signer set before one is chosen.
 - republishing a partial signature -- the third kind `republishOwn` repeats, and
   the one whose loss stalls a session that has every other piece.
 - refusing a resend whose stored events will not parse.
 - ignoring a silent tag on anything but a proposal, so the nudge cannot be
   triggered by any message at all.
 - the singular transcript wording.

Two notes on what is still not covered. `broadcast` returning false needs four
identical publishes in four consecutive seconds behind a five-minute cooldown;
forcing it means injecting a clock, and a racy test of it would be worse than
the guard. And neither the screen section nor the view model has a test, though
both are established patterns in this project.

Found the same `maxBy`-on-a-shared-second trap again while writing these, which
is now a helper per transport rather than a thing to remember.

1288 tests, no failures. m3Audit: all budgets met.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:33:10 -05:00
Kgothatso NgakoandClaude Opus 5 daf8365a6e fix(frost): a resend names the whole room, because the p-tags are its identity
The resend narrowed its NIP-17 p-tags to the members it was for, on the
reasoning that a wrap is sealed per recipient so repeating one for the whole
room wastes wraps. Both halves of that were wrong.

A receiver does not read the p-tags as a recipient list. It hashes them together
with the sender to work out which room the payload belongs to --
`GiftWrapPayload.aggregatedParticipantsPublicKey`, via
`ChatRoom.deriveChatRoomId`, which is what `GiftWrapSeal.decryptGiftWrapPayload`
files an inbound payload under. Leave a member out and the id changes, so the
resend is filed under a room that does not exist; `getOrCreateChatRoom` then
mints that room rather than rejecting it. A repaired session would have become a
phantom two-person conversation in the reader's list, and the session would have
stayed stalled.

There was no saving to trade against it either: `sealGiftWrapPayload` wraps for
every participant of the room it looks up, not for the p-tags it was handed, so
a narrowed payload cost exactly as many wraps as a full one.

Who a resend is for travels in `frost_silent`, which is not part of that hash,
so nothing is lost by naming everybody.

Marmot was never affected -- a group event names nobody and is routed by its
h-tag.

The two tests that asserted the narrowing now assert the opposite, which is the
property worth having: the p-tag set does not change between a publish and its
repeat. The p-tags were the right thing to assert all along; they just mean
something other than what they look like.

Found while answering whether this is backwards compatible with older receivers.
It is -- no new kind, an unknown tag that old builds ignore, and every
duplicate-folding gate in `record` and `acceptProposal` predates this work. This
bug was never a compatibility bug; it broke old and new receivers identically.

1279 tests, no failures. m3Audit: all budgets met.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:22:52 -05:00
Kgothatso NgakoandClaude Opus 5 2edd05f3ca fix(frost): a broadcast that queued nothing records nothing
`broadcast` now says whether it queued anything, and `publishOwn` records the
local copy only when it did. The copy is what stops `advance` from publishing
that message again, so writing one for a message that never went out is a
message the group waits on forever -- the exact opposite of the trade the
queue-then-record order exists to make. The comment on `publishOwn` had that
trade written down in one direction only.

Reachable only through the id-collision guard added with the resend, which
refuses after four tries rather than overwriting a sent row. Four consecutive
seconds of the same message is not something a person with a button can do, but
"unreachable" and "safe when reached" are different claims.

Also brings the docs up to what was built: the resend note is in the index, the
skipped-keys note points at it as the recovery for the failure it describes --
a new message is encrypted at a generation the receiver has not passed, which is
the whole mechanism -- and the plan records the two things landing it turned up,
plus `ask_again` being reused from the catalogue rather than duplicated.

1279 tests, no failures. m3Audit: all budgets met.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:09:57 -05:00
Kgothatso NgakoandClaude Opus 5 e0f20b2b63 test(frost): the case the resend exists for, end to end
The headline test is the one that goes through `acceptProposal` from the outside:
Bob's device has no session at all -- his copy of the first proposal decrypted to
nothing, which is terminal and which the sender cannot see -- and the resent
proposal opens it.

Two databases rather than a hand-written row, because an event's id is rebuilt
from its own fields under the author the *receiver* resolves, never the
proposer's word for it. A fabricated event cannot pass that check, which is the
point of the check, so the proposal has to be one a coordinator really made.

What the rest of them pin, in rough order of how much it would cost to get
wrong:

 - a named member republishes the *same bytes*, not a fresh nonce. Two nonces
   for one message against two aggregates is how a secret share is extracted.
 - a named member whose owner has not decided publishes nothing, and is still
   asked. A message from the coordinator must never put a device's share behind
   events its owner has not agreed to.
 - a nudge from somebody who is not the coordinator does nothing.
 - a resent proposal delivered twice mints no new nonce seeds.
 - a republished refusal is counted once. A resend makes duplicates routine, and
   counting one twice abandons a 2-of-3 session on a single no.
 - two resends in one second get two ids.

Plus the state the screen reads: who counts as quiet (not the refusers, not this
device), who is offered the resend (the coordinator), and that the cooldown
leaves the offer standing while quieting the button.

1279 tests, no failures. m3Audit: all budgets met.

Plan: docs/frost-session-resend.md, Phase 8.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:06:04 -05:00
Kgothatso NgakoandClaude Opus 5 aee19737f6 feat(frost): the knock, under the list of doors it is for
The resend gets its affordance on the signing screen, below the member list and
above the approval block. The list is already the screen's answer to whose door
to knock on; this is the knock.

Coordinator only, and the state says why rather than leaving it to the button:
`acceptProposal` takes a session's coordinator from the author of the proposal
that opened it, so a resend from anybody else would have a member who was
missing the session believe the resender coordinates -- and then drop the real
coordinator's signer set.

`silent` is derived on the UI state rather than collected. Everything it is made
of is already there -- the share holders, who offered a nonce, who signed, who
refused -- and a second flow would be a second thing that can disagree with the
rows the screen draws. A member who refused is not silent: they answered.

The cooldown disables the button rather than hiding the section, because a
control that vanishes after a tap reads as one that failed, and what happened is
that it worked. `resend` hands back whether anything went out, so the snackbar
says which: asked again, or everyone has since answered and there was nobody
left to ask.

The transcript line gets its own icon. A resend must not read as a second
proposal, which is the one thing a session can never have -- two proposals under
one session id would be asking for two signatures out of one secret nonce.

`ask_again` already existed in the catalogue for re-querying a list's relays and
means exactly this, so it is reused rather than duplicated.

m3Audit: all budgets met. jvmTest: 1260 tests, no failures.

Plan: docs/frost-session-resend.md, Phases 5-7.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 15:01:09 -05:00
Kgothatso NgakoandClaude Opus 5 01e022fdcc feat(frost): a proposer can resend a signing session
A session can sit in COLLECTING_NONCES forever waiting on a member who would
have agreed, because the message asking them never arrived. MLS moved its
ratchet past the generation it came in at, which is terminal at the receiver and
invisible to the sender (docs/mls-skipped-keys.md); or a disappearing-message
group expired it off the relay; or the member joined the room after the proposal
was made, holding a share from an earlier ceremony. None of them heals on its
own, and nothing asked the sender to resend.

`FrostSigningManager.resend` repeats what the coordinator wrote -- the proposal,
and whichever of the signer set or the signature the stage has reached. A resent
message is a *new* message: new created_at, new id, encrypted at a generation
the receiver has not passed, carrying a fresh expiry. That is why it works where
the first copy did not.

Two and never three. Publishing in bursts is what the skipped-keys bug feeds on,
and a recovery for it that re-creates it is no recovery, so the stage table is a
minimum rather than "send everything".

The resent proposal names the members the coordinator has heard nothing from, and
a named device republishes what *it* already sent -- the stored bytes, under its
own key. That is the only half that can repair a nonce lost on the way *to* the
coordinator: `advance` republishes a message it has no local copy of, so a device
that published and recorded its nonce would never send it again. It publishes
nothing new, so a nudge cannot talk a device into putting its share behind events
its owner has not approved.

Coordinator-only, both halves, and that is correctness rather than etiquette:
`acceptProposal` takes coordinatorPublicKey from the proposal's author, so a
non-coordinator resending to a member with no session row would have that member
drop the real coordinator's signer set.

Also fixes a bug `broadcast` had all along. An outbound row's primary key is a
hash over its created_at, stored to the second, so a message re-published inside
that second hashes to the id of the copy it repeats -- and the upsert put an
already-sent row back on the queue with marmotGroupEventId cleared, double-sending
it and losing its mapping to the group event it became. It now steps past a taken
id rather than writing over it, and refuses after four tries.

Schema 24 -> 25 for FrostSigningSession.resentAt, which is the cooldown and the
second line of defence on that collision. Phase 4's column, pulled forward
because Phase 2 needs it.

Plan: docs/frost-session-resend.md, Phases 2-4.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 14:56:15 -05:00
Kgothatso NgakoandClaude Opus 5 7bc00af304 feat(frost): a tag naming the members a resend is for
A proposer can resend what the proposer wrote and nothing else. A nonce or a
refusal relayed by somebody else is filed under the relayer -- `record` keys a
signer's message on the inner event's own pubkey -- and for a refusal that is a
way to manufacture one, which in an n-of-n group abandons a session nobody
refused.

So a resent proposal names the quiet members instead, and their own devices
republish what they already said, under the keys that said it. This is the tag
that names them: `frost_silent`, all-or-nothing like `frost_signers`, because a
list naming the wrong people is worse than no list at all.

Additive on the wire. A build that does not know the tag ignores it and accepts
the proposal it rides on, which is the half of a resend that fixes the common
case -- a member who never saw the proposal because MLS dropped the generation
it arrived at (docs/mls-skipped-keys.md) or a disappearing-message group expired
it off the relay.

Plan: docs/frost-session-resend.md, Phase 1.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-10-06 14:46:18 -05:00
Kgothatso b9140c6a27 fix(ios): build ktor clients on Darwin so wss/https work on a device
CIO has no TLS implementation on kotlin/native, so every secure
connection died at the handshake with "TLS sessions are not supported
on Native platform". Pick the engine per platform -- CIO on android
and the desktop jvm, Darwin on ios -- and pin both the relay socket
client and coil's image fetcher to it.
2026-09-15 09:42:38 +02:00
Kgothatso aa9312f0ee build(ios): unbreak the ios link tasks
linkDebugTestIosArm64: Kotlin/Native rejects backticked test names
containing a comma ("Name contains illegal characters: ','"), so the
commas come out of the commonTest names.

linkReleaseFrameworkIosArm64: Kotlin 2.4.10's release backend crashes
in StaticInitializersOptimization on this module's call graph, so
RemoveRedundantCallsToStaticInitializersPhase is disabled for the
release native links. The pass is an optimization only.
2026-09-15 08:58:59 +02:00
Kgothatso 987df2c8f6 feat(ios): declare publish cancellation and the Room constructor to swift
@Throws(CancellationException) on the publish path so swift sees a
cancellation as one, @ConstructedBy so Room can build the database on
ios, an unused import out of Platform.ios.kt, and the development
TEAM_ID for signing.
2026-09-15 08:58:53 +02:00
Kgothatso NgakoandClaude Opus 5 79ff9ee1fc fix(chat): the detail pane's view models are the selected room's, not the first one's
In an expanded window the transcript beside the room list showed the
first room opened and stayed there: pick another room and the list
moved on while the pane did not. HomeScreen wraps the pane in
key(chatRoomId), which rebuilds the composition when the selection
changes, but viewModel() does not read composition keys. It scopes to
the nearest ViewModelStoreOwner, and in the pane that is the home
destination's one back stack entry for every room, so the second room's
viewModel(factory = ...) got the first room's ChatRoomMessagingViewModel
and ChatMessageListViewModel back with its factory ignored. A phone never
showed it because a pushed room is its own entry with its own store.

The pane now provides LocalViewModelStoreOwner with
rememberViewModelStoreOwner() under the room key: a store per room,
cleared when the selection moves on, and parked in the home entry's
store so that it survives a rotation. Keying the two viewModel() calls
on the room would have unstuck the pane too, at the price of keeping
every room ever opened -- four DB collectors each -- in the home entry's
store for as long as the home screen lives.

ChatPaneSelectionJvmTest opens two rooms and comes back to the first,
told apart by what was said in them since a name is on the row as well
as the bar; its third case checks the room left behind stops collecting,
which is the case the keyed alternative fails. The conformance doc's
account of the hand-built pair records the trap.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 20:10:13 +02:00
Kgothatso NgakoandClaude Opus 5 30685924db feat(curated): the date field type, and an unknown type is kept aside rather than dropped
CuratedFieldType had no `date`, so a schema declaring one lost the field
at parse: entries were read without it, the accept flow copied a
suggestion into a canonical entry without it, validation never asked for
it, and the entry published cleanly -- and was refused by every client
that did know the type, for the required field it lacked. Fourteen of
bitcoin.necropolis.report's entries went out that way without a `died`.

`Date` is the NIP's tenth type: `YYYY`, `YYYY-MM` or `YYYY-MM-DD`, at
whatever precision the record allows, with `min` and `max` bounding the
year. `CuratedDate.parse` is the check -- month 1-12, day on the calendar
for that month and year, leap days included -- and the value is kept as
written. `checkValue` returns early the way the numeric branch does, so
`max` is a year and never a length cap. The value is not numeric, so
`isNumeric` stays false; `isBounded` is what the editor keys its min/max
inputs off instead, labelled as years for a date.

The root failure was wider than one type: any type this build does not
know was dropped the same way. A `field` tag of an unknown type is now
kept aside as CuratedUnsupportedField, the way CuratedFieldConfig.others
keeps unknown config keys. Reading is unchanged -- an entry is checked
against the fields a reader understands, which is the NIP's tolerance --
but the accept flow refuses to propose under a required one, naming the
type, and the schema editor carries it through an edit and writes it back
as it came, so re-signing a list here cannot change its rules unseen.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 19:30:02 +02:00
Kgothatso NgakoandClaude Opus 5 984db0a180 Merge branch 'curated' into claude/quartz-1-15-2-compat-c8267f
Both sides had taken schema version 21: curated for the nullable
FrostSigningSession.signDeclinedAt (one refusal takes a member out of the n),
this branch for the Marmot current profile's three columns, with 22 and 23
behind it for the routing table and the publish obligations. Curated keeps 21.
This branch's three migrations move up one -- 21->22, 22->23, 23->24 -- and
the database is at 24; the comment on 22 says they moved and why.

The exported schemas for 22 and 23 could not be hand-edited, since each is a
snapshot of the whole entity set with an identity hash over it, and could
not be produced from the merged tree, which only exports its own version. So
each was regenerated the way the originals were: the tree of the commit that
introduced it (Phase 3 for 22, Phase 4 for 23), with curated's entity change
and its 21.json dropped in, the version bumped, and KSP run. The result is
what a device on curated's 21 migrates through: 22 adds the three columns
and nothing else, 23 adds MarmotRouting, 24 adds MarmotPublishObligation and
MarmotOutboundGate together with MarmotCommitResult.obligationId, and every
one of them carries signDeclinedAt. 24.json is the merged tree's own export.

Everything else merged itself: ChatMessage's kinds and line types, the
transcript's icons, the strings, the nav host. Full jvmTest 1220 green,
Android debug compile, M3 audit.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 08:09:54 +02:00
Kgothatso NgakoandClaude Opus 5 9b404e2035 fix(frost): the signing screen lists the key's share holders, not the room
The members list on "Sign with the group's key" was the room's roster, and a
room is not the set that can sign with its key. A member invited after the
ceremony holds no share: the proposal they receive is dropped before it
becomes a session on their device, and no nonce or refusal of theirs would
count. The screen drew them a row anyway, in a list whose whole purpose is to
say whose door to knock on.

The list is now the ceremony's share holders -- the members who put a host
key into it, read off DkgParticipantMessage through the new
FrostSigningRepository.observeShareHolders. That is the same set
FrostSigningManager places signer ids over, so what the screen lists cannot
disagree with whose messages the session counts, and a ceremony cannot
complete without every participant's host key, so for a completed one it is
the whole participant set. Names still come off the room, in the room's
order; a share holder the room no longer holds keeps a placeholder row, since
their share did not leave with them.

The view model follows the session to its ceremony with flatMapLatest, and
the list is empty until the ceremony has been read rather than falling back
to the room -- a fallback would flash everybody and then correct itself.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 07:53:21 +02:00
Kgothatso NgakoandClaude Opus 5 a7f3917e65 fix(frost): one refusal takes a member out of the n; it takes n - t + 1 to abandon
"Don't sign" ended the session for everybody: decline() failed it locally and
broadcast a FAILURE, which every other device failed on arrival. In a group
whose whole reason for running a ceremony was that no one member is needed,
one member's no was everybody's.

A refusal is now its own kind, DECLINE (30330), and one FrostSignerMessage per
member, like a nonce. Every device counts the refusals it has received from
ceremony participants against the same arithmetic -- FrostSigningSession.
declinesThatAbandon(), which is n - t + 1 -- and fails the session only on the
refusal that leaves fewer than t who could still sign. Nobody broadcasts the
abandonment, because nobody has to: each device reaches it the way each writes
its own transcript, and a FAILURE sent to say so would blame one member for the
group's arithmetic. In a 2-of-3 group the first refusal changes nothing the
other two can do; the second ends it. FAILURE keeps meaning abandon-and-blame,
for the faults no quorum can work around, and a build predating the new kind
still sends its refusals that way and is honoured as it always was.

decline() takes approve()'s shape: the decision goes on the row, as the new
nullable signDeclinedAt (Room v21, an AutoMigration), and advance() publishes
it whenever the row says so and no copy is on file -- so a device killed
between deciding and sending says it on the next message rather than never. A
member who refused is not asked again, but the session stays open around them
and closes when the group's signature arrives. Refusals from a member holding
no share of the key are turned away at the door.

fail() now leaves a COMPLETE session alone. A member who refused before their
device had seen the signature has refused something that no longer needed
them, and a late FAILURE used to flip a signed session to failed.

The transcript gets a TYPE_FROST_DECLINED line for a refusal the session
survives, with the count -- "declined to sign. It takes 2 of 3 members, and 2
could still sign." -- and the tipping refusal is written as the FAILED line in
that member's name, saying both that they declined and that it was the one
that ended it. A reader's own refusal settles their request line, and only
theirs; nothing earns a tick. The signing screen marks who declined and says
how many could still sign, and the proposal list says "You declined." where it
said "You agreed."

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 07:37:29 +02:00
Kgothatso NgakoandClaude Opus 5 de5b2bd4a6 docs(marmot): the v2 plan is built -- where the build chose differently, and where the rollout stands (v2 Phase 9)
docs/marmot-v2.md, Phase 9. The phase has no code in it; it is the rollout,
and the rollout is a flip and a retirement that wait on the people this
app's users invite. What the phase owes the document is the repo's
convention for a plan that has been built: the status paragraph, the table
of where the implementation chose differently, the thing found on the way
that belongs to nobody's phase, and the commit the account was written
against. All four are here now, and the phases themselves are kept as
written, because they are the reasoning and the code reads better against
the argument it came from than against a summary of itself.

The table has twenty-two rows, and the first is the one to read before any
of the code: the plan said the library's engine replaces the app's pipeline,
and the build made a room's profile pick its pipeline instead -- legacy
rooms keep the app's path, comparator and cache included, and retire with
the profile; current-profile rooms go through `MarmotEngine`. The library's
processor has no carve-out for the bare gift wrap a legacy direct message is
carried in, so under it every one of those would have stopped decrypting,
and the legacy pipeline is frozen anyway. Most of the other rows follow from
that one: one store rather than four, no message-id column, the rotation
manager constructed and unused, the engine hanging off the database rather
than an identity's scope. The rest record the smaller turns -- the envelope's
kind, the chronicle cap measured escaped, the KeyPackage not last-resort, the
admin refused before a leave is attempted, captions from the participants
diff rather than a snapshot column, no rename, the two non-chat kinds only,
the four Phase 8 scenarios spread over four tests -- and the three
follow-ups that closed lines the phases' own commits had missed: the
outbound hold while a pass is open, the rejected commit discarded, and the
1210/1009 bubbles.

The thing found on the way is quartz 1.15.2's: its convergence rebuilds a
candidate branch by replaying its commits against the retained parent, and
a member cannot replay a commit it authored, so a fork this device is half
of is scored with its own branch stale and two committers swap branches
rather than agree. Every device that only observed still agrees, which is
the case the app depends on and the first fork test holds; the second pins
the swap and says what to do the day it fails. The paragraph says the
practical exposure is small and why, and that the Phase 5 follow-up on
withdrawing a losing branch's rows waits on the same fix.

Phase 9's own section gains "Where it stands": everything ships built and
unflipped -- `DEFAULT_FOR_NEW_ROOMS` is legacy, both slots are published, no
room is on the current profile unless a caller asked by name -- so the first
of the three one-way doors is behind us and the other two wait on the flip.
The README's row and reading-order paragraph now say the plan has been
built and point at the table's first and last rows.

One code change, a comment: the plan's Phase 5 asked that the two library
gaps the rewrite does not close -- the retained-epoch decrypt without sender
verification, and the skipped keys that do not survive `saveState` -- be
said where the new code is, rather than let the rewrite look like it closed
them. `MarmotEngine`'s class doc now says both.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 00:16:10 +02:00
Kgothatso NgakoandClaude Opus 5 2bbe1fc9bc feat(marmot): a commit every relay refuses is discarded, and the room is free again
docs/marmot-v2.md, Phase 6: "a terminal rejection resolves FAILED, discarding
the pending state with the live group already correct." The phase's commit
built the CONFIRMED half and left every other outcome pending, on the
grounds that the broadcast queue reported a timeout and a relay's refusal
the same way -- both were "failed" -- and a timeout must stay pending, since
its retry re-sends the same bytes and may yet be accepted. The cost of that
was a room whose commit a relay refused outright: its obligation never
resolved, `canPrepareLocalCommit` stayed false, and every later invite,
removal or disband was refused with "try again in a moment" for as long as
the room existed. The refusal was re-sent on every launch and refused again.

The two are distinguishable, and now are. `RelayPool` already surfaces an
`["OK", id, false, reason]` as a `NostrPublishException` carrying the relay's
reason, apart from socket errors and timeouts; `BroadcastStatus.of(error)`
turns that into a status. A reason a relay will keep giving -- `blocked:`,
`invalid:`, `pow:`, `restricted:` -- is the new "rejected", which the requeue
on launch leaves alone: the same bytes to the same relay get the same answer.
A `rate-limited:` or `error:` refusal is "failed" like a timeout, because
those change. And a `duplicate:` refusal is "published": the relay is saying
it has the event, which is what published means, and it is the answer a
byte-identical retry gets from a relay that took the first copy. The four
statuses that were string literals across the sync pump, the navigation gate
and the DAO are named on the same object.

`DatabaseNostrRepository.broadcastProcessed` gains the relay's reason and a
"rejected" branch. One relay's refusal decides nothing -- another may still
accept, and a timeout on another is retried -- so the obligation is
discarded only when every request row for the event is "rejected":
`MarmotEngine.discardRejected` resolves it FAILED, the gate drops the clone,
the live group having never advanced, and the room can prepare a commit
again. Nothing is told to convergence, since a commit no relay carries is not
a branch any member will see. The room says what happened, in the relay's
words, because the screen that asked closed when the commit was made: an
invite gets the existing `TYPE_MEMBER_INVITE_FAILED` line, and a removal or
a disband step gets a new `TYPE_CHANGE_REJECTED` line, since the caption
written when it was asked for would otherwise stand uncorrected.

`PublishBeforeApplyJvmTest` gains the plan's third case: with a two-relay
scope, one refusal leaves the obligation, a timeout on the other leaves it,
the other's refusal discards it -- epoch unchanged, the peer not a member,
the line naming the reason -- and the same peer can be invited again. 1187
tests green, Android debug compile, M3 audit.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 00:12:52 +02:00
Kgothatso NgakoandClaude Opus 5 0b3780c604 feat(marmot): nothing goes out against a state that may lose
docs/marmot-v2.md, Phase 5, the last line of "Welcome and outbound", which the
phase's commit left out: before encrypting or committing, read
`convergenceStatus(groupId).allowsOutboundWork`. While a convergence pass is
open the room's canonical state is provisional -- two members committed
against the same epoch and the pass has not picked between them -- and a
message keyed under the branch that loses is one no peer will ever read,
while a commit against it makes a three-way fork out of a two-way one. The
library defines the flag and nothing in it reads it; the app is where the
outbound work is, so the app reads it.

`MarmotEngine.awaitSettled(routingId)` is the read. It is a short wait rather
than a refusal because a pass settles at the quiescence window -- one quiet
second, five at the outside -- so a caller that waits gets its work done in
order, whereas one that threw would hand its row back to a queue that
re-emits only when its table changes, which a pass settling does not do. The
wait ticks `settleDueConvergence` itself rather than relying on `settleLater`
having been armed, and gives up after three ticks with
`MarmotConvergencePendingException`, whose message the add-member screen
shows like the other two refusals; the bound is short because most of the
callers hold a Room transaction while they wait.

Five callers: the message send in `encryptAndSendMarmotInnerEvent` (the
current-profile branch; a legacy room has no convergence to wait for), the
invite past the founding Add, `publishGatedCommit` (so removal and both
disband steps), and `leave`, each after the publish gate's own check and
before the manager stages anything.

The observer fork test now settles the app's side through `awaitSettled`
instead of resolving it directly -- asserting outbound work is refused before
and allowed after -- and finishes with what settling is for: a message the
app sends once settled is an `ApplicationMessage` on Dana's side. 1186 tests
green, Android debug compile, M3 audit.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 00:04:39 +02:00
Kgothatso NgakoandClaude Opus 5 437fd3e18b fix(marmot): a received kind:1210 or kind:1009 is not a bubble
docs/marmot-v2.md, Phase 7, the line the phase's commit missed: "a received
kind:1210 or kind:1009 gets no bubble". Both are kinds the current profile
defines as not-chat. A 1210 is a group system row -- member added, member
left -- and the profile is explicit that one arriving over the wire is one
member's assertion rather than a fact; this device derives its own from the
commit it applied (`TYPE_MEMBER_ADDED` and its siblings, written by
`processGroupMembershipChanges`), so the received one adds nothing it can
trust. A 1009 is an in-place edit, which the app does not apply. Until now
either fell off the end of `applyInnerEvent`'s `when` into `unsupported()`,
which is the raw-JSON bubble, so a White Noise member leaving a room would
have put their client's system row into the transcript as a block of JSON
under the caption this device wrote for the same fact.

Both kinds now return null: no line. The bubble stays for every other
unknown kind, which is narrower than the plan's "neither does any other
unknown kind". Two reasons. The `TextNoteEvent` arm reaches `unsupported()`
deliberately -- a kind:1 the room did not sign is a member's own note
travelling through the room and has always rendered as this -- and the
bubble is the only trace an unrecognised payload leaves; retiring it would
make the next new kind a silent drop, which is the failure mode the chronicle
note complains about from the other side. The Phase 9 table records the
narrowing.

`MarmotSystemRowJvmTest` holds both: the two kinds make no row, and kind
31337 is still the bubble.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:59:18 +02:00
Kgothatso NgakoandClaude Opus 5 daac3962fa test(marmot): prove convergence where the app relies on it, and pin where quartz cannot yet
docs/marmot-v2.md, Phase 8. The plan's last code phase is not code: it is the
test that shows the app's forks settle the way Amethyst's do. Phase 7 wired
`MarmotInboundProcessor`'s convergence into the app -- commits admitted as
forks open a pass, `MarmotEngine.settleLater` resolves it at the cutoff,
`resolveConvergence` is the reindex's way in -- but nothing yet held the
property that is the whole point of a pass: two clients that freeze the same
batch pick the same branch. This commit holds it, and in holding it found the
one place quartz 1.15.2 does not deliver it.

What the first test proves
--------------------------

Alice's group, with Bob promoted to admin, Dana as a plain member, and the
app joined last. Every commit Alice makes on the way there is fanned out to
whoever is already in, so all four sit at one epoch before anything forks.
Then Alice and Bob each add a different member against that epoch: two
commits, one parent. The two observers -- the app, fed through the real
ingest path (`storeInApp`, so `NostrDao.indexCurrentProfileGroupEvent`, the
engine, the row), and Dana, driven through the processor directly -- are
handed the two commits in opposite orders. Each applies the first as a plain
linear commit and admits the second as a fork, which opens a pass on both.
`resolveConvergence` is called on each without waiting for the cutoff -- the
timers are scheduling, not semantics -- and the assertions are the property
itself: both settled, both one epoch past the base, same members, same group
context byte for byte, exactly one of the two additions surviving, and the
app's `ChatRoom.mlsGroupState` row carrying what its engine holds.

That is the case the app depends on. A device that only observes other
members' commits is every device most of the time; the app's own commits go
out through the publish gate one at a time, and the plan's Phase 7 argument
was that a device which never applies before publishing rarely authors one
side of a fork at all.

What the second test pins
-------------------------

The case it *can* author: the app, promoted to admin, adds peer A through
`DatabaseChatRepository.inviteMember` and confirms the publish; the other
admin adds peer B against the same epoch; each is handed the other's commit;
each resolves its pass. The expectation was agreement. The result is a swap.

On each side, the commit that side authored is scored STALE and the other's
ACCEPTED, and both sides rewind -- onto each other's branch. The cause is in
the library's candidate scoring, and it is structural rather than a tie-break:
a pass rebuilds every branch by replaying its commits against the retained
parent state, and a member cannot replay a commit it authored. The committer
holds no ciphertext for its own copath in the UpdatePath, so
`MlsGroup.restore(parent).processFramedCommit(ownCommit)` fails with
"UpdatePath at common ancestor carries no ciphertext for us", the replay
fails, and the branch is scored as if it could not be reached. The state that
branch leads to is already in hand -- it was retained when the commit was
recorded -- and needs no replay; using it is the library's fix, not this
app's. Confirmed with a direct probe against `MlsGroup` before the test was
written, so the pin is on the library and not on the app's wiring.

The test asserts the defect exactly as observed -- own commit STALE on both
sides, both rewound, the two sides' contexts differing, peer B on the app's
side and peer A on the other's -- and one thing that must hold regardless:
the app's row matches its engine. Its last assertion carries the note that
the day the two sides agree, the library has been fixed and the test should
become the agreement it was written to prove. Until then, an app device that
authors one side of a fork against another admin's commit will take the
other admin's branch, and the other admin will take the app's; the group is
still one group on every device that only observed, which is the case the
first test holds.

The fixture
-----------

`MarmotStandIn` gained `commitLocally`, which is what a stand-in's commit had
been missing: `recordLocalCommit` with the pre-commit snapshot, the way
Amethyst tells its own processor about its own commits. Without it, a
competitor commit authored against the same parent is an orphan rather than a
fork, and no pass opens. `inviteApp` now commits through it, and
`lastCommitEvent` turns the newest commit into the kind:445 the other
members are handed, which is what lets the first test fan Alice's commits out
before the fork.

Verified: full jvmTest 1184 tests green, Android debug compile, M3 audit.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:55:43 +02:00
Kgothatso NgakoandClaude Opus 5 c4a46e892f feat(marmot): leaving, removal, disbanding, and captions derived from the state (v2 Phase 7)
The lifecycle of a current-profile room, through the publish gate. docs/marmot-v2.md,
Phase 7.

Leaving used to be a local flag and a "Left group" chat line: the group never
learned of it, and the leaf stayed in the tree holding the group's keys. A
current-profile room is now left the way v2 leaves: a standalone SelfRemove
proposal on the wire as a kind:445, the LEAVING gate raised first and durably,
the local state dropped, the room closed on this device with its own line. It
is not a commit, so it waits on no relay -- there is no state of ours it would
make canonical. An admin's device that receives the proposal commits it on
receipt, as the library expects, so a leaver is out of the tree without anyone
pressing anything; the commit goes through the gate like every other. The
library refuses a SelfRemove from an admin outright -- an admin self-demotes
first -- so an admin's leave is refused here before anything is raised or
dropped, in words a person can act on: hand over admin, or, for the only
admin, disband. Legacy rooms leave the way they always did.

An admin removes a member with stageRemoveMember through the gate; the library's
admin-depletion guard refuses a commit that would leave no admin. An admin
disbands with stageEnableDisbanding and stageDisband -- two commits when the
group predates the lifecycle component, one otherwise -- under the DISBANDING
gate, which is durable so a crash between raising it and publishing does not
lose the request; the second step is taken by continueDisband when the first
is accepted. Both are on the room detail screen, for a device that administers
a current-profile room: a remove action on every member row but its own, and a
disband button beside leave, with an explainer that says nobody rejoins. A
disbanded room says so and hides every write entrance.

The captions are derived, never received. v2 spells them as kind-1210 system
rows and is explicit that a row arriving over the wire is one member's
assertion, not a fact; here they are written by processGroupMembershipChanges
from the difference between the room's members before a commit and after it --
"X joined the group", "X left the group", "The group was disbanded", "You were
removed from the group" -- on every device that applies the commit, including
the one that made it, which is why an admin's own removal no longer writes a
line of its own: every device says the same thing at the same point. A commit
that removed our own leaf, or disbanded the group, closes the room here as well.
A commit this device made is applied at the relay's OK, where the room now
follows the state it gained. MarmotCommitResult.NO_INVITEE marks a commit that
invites nobody, so the OK sends no Welcome for it.

Tests: GroupLifecycleJvmTest, with quartz's own managers standing in for the
other member, covers an admin committing a member's leave on receipt and the
room saying they left; a removal applied on both sides once the commit lands; a
disband ending the group for everyone, closing the room and refusing what
follows; an admin's leave refused with its reason; and a member's leave sending
the proposal the other side stages, dropping the state and closing the room.
The stand-in is now one fixture, MarmotStandIn, shared with the interop test.
Full JVM suite: 1,182 tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:43:11 +02:00
Kgothatso NgakoandClaude Opus 5 f617560cf9 feat(marmot): publish before apply (v2 Phase 6)
A commit this device makes into an established current-profile room is no
longer applied and then broadcast. It is prepared on a clone by the manager,
recorded as a durable obligation before anything is broadcast, sent, and
installed only when a relay in its scope accepts it. docs/marmot-v2.md, Phase 6.

The library's gate explains why apply-and-roll-back is not the same thing:
between applying a local commit and undoing it on failure there is a window in
which this client's canonical state is an epoch no peer has, and a crash inside
it makes the fork permanent. Recording the obligation before the broadcast
closes the mirror-image window -- a commit peers accepted that this device has
no memory of preparing, and would mint a replacement for at the next start.

Two new tables (schema 22 -> 23), MarmotPublishObligation and MarmotOutboundGate,
back RoomPublishObligationStore; MarmotEngine holds a MarmotPublishGate over it
and restores it after the groups. MarmotCommitResult -- the row that already
held a commit's Welcome until the relay's OK -- carries the obligation id, and
DatabaseNostrRepository.broadcastProcessed resolves it on "published" through
MarmotEngine.confirmPublished, which installs the pending state, tells
convergence of the commit we authored (without which a peer's competing commit
would look like an orphan rather than a fork we are half of), and only then
lets the Welcome go out: the MIP-02 ordering enforced by sequence rather than by
hope. A second OK for the same event from another relay finds no obligation
and changes nothing.

The founding Add keeps its exception. A group that is only its creator has no
peer a failure to publish could fork, so it installs at once -- the library's
satisfyEmptyObligation draws the same line, "creation and its immediately
following founding Add, never a later commit". While an obligation is pending
the room cannot prepare another commit: a second invite is refused with
MarmotCommitPendingException, shown as itself on the confirmation screen, rather
than forked; the membership note's batching closed the race inside one invite,
and this closes it between invites. The broadcast queue's requeue re-sends the
same signed event, which is the byte-identical retry the rule demands, and a
"failed" from the pump -- a timeout or a relay's refusal, which the pump does
not tell apart -- leaves the obligation pending rather than discarding a state
some relay may have taken. Legacy rooms are unchanged.

Tests: PublishBeforeApplyJvmTest pins the founding Add installing at once, a
later commit prepared and queued but not applied, the second Welcome held back,
a concurrent invite refused, the OK installing the state and queueing the
Welcome, a second OK changing nothing, and an obligation restored by an engine
rebuilt over the same database and installed by a late OK. Full JVM suite:
1,177 tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:27:40 +02:00
Kgothatso NgakoandClaude Opus 5 8e1037dff1 feat(marmot): the library's engine over Room, for current-profile rooms (v2 Phase 5)
A current-profile room is now read and written by quartz's own managers:
MlsGroupManager holds its live group and persists through a store over the
app's tables, MarmotInboundProcessor reads its kind:445s -- settling same-epoch
commits through the convergence engine and recognising redelivery by the
Marmot message id -- and MarmotOutboundProcessor builds the ones this device
sends. Everything in docs/marmot-v2.md's "v2 says" column that is not a wire
shape lives in those three, and the app no longer carries its own copy of
their logic for these rooms.

Where the build chose differently from the plan: **a room's profile picks its
pipeline**. The plan had one engine for every room. Two things stood against
that once the code was open. The library's processor has no carve-out for the
bare gift wrap a legacy room carries a direct message in -- it would refuse
every legacy DM as an author that is not the sender -- and the manager's
installState declines a state whose group context matches the live one, so a
legacy path that mutates the manager's instance and then asks it to persist
would persist nothing, while its own retained-epoch list would drift from the
manager's. So legacy rooms keep MarmotInboundManager over MlsGroupCache,
unchanged, and retire with it; current-profile rooms go through MarmotEngine.
A room is on one profile for life, so the two never hold the same group, which
is what keeps the manager's live instance the single writer of a
current-profile room's state. For the same reason the library's KeyPackage
rotation manager is constructed but not fed: Welcomes are matched to this
device's bundles by the app's own profile-aware table, by reference, and the
processor reads the rotation manager only on its own Welcome path. The
vendored CommitOrdering therefore stays, for legacy rooms.

RoomMlsGroupStateStore resolves the library's key -- the routing address --
through MarmotRouting to the room, keeps the state in ChatRoom.mlsGroupState
as before, and keeps the retained epochs as MarmotRetainedEpochSecret rows,
which until now were written by the outbound path and read back by nothing:
every epoch the inbound path retained was a map in an object, gone at the next
launch. The engine is owned by the database (CurareDatabase.marmotEngine),
because the DAOs that need it are made by Room from that instance and can be
handed nothing else; it restores every group once, seeds convergence with each,
and loads a room made or joined later from its row on first ask.

Inbound, NostrDao.indexMarmotGroupEvent branches on ChatRoom.marmotProfile()
-- read off the stored state's group context without rebuilding the group --
and the result handling every kind:445 shares (membership changes, the chat
line, the signing and ceremony dispatch) is one function for both pipelines.
The processor names the group by its address; the result is re-addressed to
the room once. A pass a commit opened is driven to its cutoff by a ticking
coroutine, since a quiet group has nothing else to drive it. Outbound, an
application message in a current-profile room is built by the outbound
processor over the engine's group, and an invite or batched add is staged on a
clone by the manager and installed at once -- retained, persisted -- with the
callers' own state writes skipped for these rooms; installing on a relay's OK
rather than at once is what Phase 6 changes. A reindex closes the room's open
convergence pass before replaying.

Tests: CurrentProfileInteropJvmTest puts two implementations in one group --
the app's engine over a CurareDatabase and a bare library stand-in playing
Amethyst -- and has the stand-in invite the app through a gift-wrapped Welcome,
messages cross in both directions (the app's read by the stand-in as a
canonical payload), a commit from the other side move the room's membership
and epoch, and a second engine over the same database read a message from the
epoch before that commit -- the case the old pipeline could not pass.
RoomMlsGroupStateStoreJvmTest pins the store's contract. Full JVM suite: 1,175
tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:20:02 +02:00
Kgothatso NgakoandClaude Opus 5 9b5e4c95fc feat(marmot): current-profile rooms, and a routing address apart from the room id (v2 Phase 4)
After this the app can create rooms on the current profile -- behind
MarmotProfile.DEFAULT_FOR_NEW_ROOMS, which stays legacy until the rollout flips
it -- and every path keyed on the `h` tag reads the room through its address.
docs/marmot-v2.md, Phase 4.

The room's id is a key. A room built around a threshold key is keyed on the id
derived from that key -- the one it signs as and every chronicle is verified
against -- and every member can compute it unaided. v2's routing component
says of `nostr_group_id` the opposite: random, never derived from any public
key, because a derivable address lets a relay link a group's traffic to the
identity behind it, which here is the key that authors the group's published
entries. Both are right about their own value, so they are two values now. A
new MarmotRouting table (schema 21 -> 22) maps an address to its room; a
current-profile room this app creates mints a random one, and every other
room -- legacy, or made by another client -- has a row whose address is its
own id, or no row at all and is addressed by its own id, which every reader
falls back to. The inbound ingest, the reindex, the negentropy answer for a
`#h` filter, the chat list's sync request and the live subscriptions all
resolve or emit addresses; LocalChatRoom carries its routings as a relation so
the subscription planner needs no query; the outbound commit, application and
Welcome templates are stamped with the room's address rather than its id.

A joiner keys a room on the group's `nostr_group_id`, which for a
current-profile room would be the address, and the creator, the joiner's own
DKG session and every chronicle know the room by its derived id. The
components carry no field for it, so it rides in the group profile's
description as one line -- `Room id: <hex>` -- the way the derivation path
already does, and MlsGroup.roomId() reads it back first. It is a claim by the
creator, exactly as the legacy `nostr_group_id` was, and is trusted no further.

MarmotGroupFactory is now the one place a group is minted. The three sites --
a direct message, the convenient room, and the derived room MarmotGroupCreation
builds -- say what the room is, and the factory decides the profile: a legacy
group is the 0xF2EE extension as before; a current-profile group is quartz's
own CurrentProfileGroupFactory with the admin policy, routing, profile and the
creator's proof-bearing leaf, and no 0xF2EE at all, because RFC 9420 requires
every member to support every GroupContext extension and an MDK leaf does not
advertise it. Creating a direct-message room reads the identity from the chat
repository, which now takes it, so the creator's leaf can carry its proof.

A peer's KeyPackage is picked by the room's profile rather than by age: a
current-profile room takes the newest current-profile package or none, since
the legacy package is deliberately the newest a peer publishes. A leaf on the
wrong profile is refused a step before the library refuses it, with a reason
the member can act on -- their package is from an older profile -- shown on
the confirmation screen and in the room's line rather than as "try again".

Tests: MarmotGroupFactoryJvmTest pins both profiles' shapes, the address
minted apart from the room id, the room id riding in the description, the
read-only creator refused on the current profile, and which leaves fit which
groups; MarmotGroupCreationJvmTest adds a current-profile room keeping its
derived id with a routing row, a legacy room unchanged, and a kind:445 under
the address filed in the room; MarmotRoomIdentityTest the description line;
LiveSubscriptionPlanTest the addresses watched. Full JVM suite: 1,166 pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 23:03:52 +02:00
Kgothatso NgakoandClaude Opus 5 99dba546d8 feat(marmot): identity proofs and the current-profile KeyPackage (v2 Phase 3)
After this a peer on Amethyst 1.15.2 or White Noise can add this app's users to
their current-profile groups, and the app joins. docs/marmot-v2.md, Phase 3.

A current-profile leaf carries an app data dictionary with three components: the
component ids it supports, an explicitly empty safe_aad list, and the 104-byte
account identity proof -- the account's Nostr key signing a kind-450 template that
is never published and names this leaf's own Ed25519 signature key. MLS never
checked that the account in a credential agreed to the signature key beside it;
the proof is what makes a leaf unforgeable, and a current-profile group refuses a
leaf without one at add time. Only the account signer can make it, and it has to
exist before the leaf is built, so publishMarmotKeyPackageBundle now takes the
KeyPair rather than the public key it used to. The current-profile package comes
from quartz's own CurrentProfileGroupFactory -- the same leaf shape Amethyst
publishes, including the agent-text-stream component the reference client
requires of every leaf it invites -- and its kind:30443 from
KeyPackageEvent.buildCurrentProfileFrom: framed content, the 0x0006 extension,
the 0x0008 and 0x000a proposals, an app_components tag, no encoding and no relays
tag. KeyPackageUtils.isValid checks all of it, and the tests run that validator
because it is the one a peer runs on ours.

The legacy package is still published, in a second slot, for as long as builds
that decode kind:30443 content as a bare KeyPackage are inviting people (Phase 9
retires it). It is stamped one second after the current-profile package, every
time both are published: an inviter picks a peer's newest package by created_at,
and an old build then fails on a framed one, so the legacy package has to stay
the newest. Each profile now has a slot whose d-tag is minted once and reused on
every rotation, so a replacement supersedes the previous event on relays instead
of accumulating beside it -- the "TODO: Persist this" on the random d-tag.

Schema 20 -> 21, an AutoMigration: MarmotKeyPackageBundle.profile and .dTag, and
MarmotKeyPackage.profile filled at ingest from KeyPackageEvent.isCurrentProfile().
Rows from before read back null, which every reader takes as legacy. The notary
keeps one publishable bundle per profile rather than any bundle at all, because
a current-profile peer consumes only the current-profile package and the legacy
one still being active would otherwise mean nothing was republished for them.
The key-package screen takes the identity, not just its public key, and refuses
to publish for one that cannot sign rather than reporting "published".

KeyPackages now also go to the account's NIP-65 write set, unioned with the
default DM relays -- transports/nostr.md has a conformant peer look there and
nowhere else, "there is no dedicated KeyPackage relay list" -- and the member
and peer sync requests ask for kind 10002 alongside the legacy 10051 hint.

Tests: CurrentProfileKeyPackageJvmTest pins one package per profile with the
legacy one a second later, both passing the peer's validator, the proof binding
the leaf to the account, a rotation reusing its slot, and the join: a group made
by quartz's own factory standing in for Amethyst adds the app's package, the
Welcome arrives as a NIP-59 gift wrap through the sync funnel, and the app lands
in a room keyed by the routing id from the 0x8004 component with the inviter as
admin. The npub round-trip test now waits for both bundles. Full JVM suite: 1,153
tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 22:46:22 +02:00
Kgothatso NgakoandClaude Opus 5 3cdce6a908 feat(marmot): send the canonical app payload, and an envelope for direct messages (v2 Phase 2)
What goes into MLS is now the Marmot app event the spec defines: six members and
no `sig`. The rumor serialised as an Event carried `"sig":""`, which a conformant
decoder refuses outright -- a payload with a signature would be a standalone relay
event, republishable as a signed statement by its author. The id does not move,
because NIP-01 never hashed the signature, so every row filed under an inner
event's id stays valid; and the build already in the field reads the new shape,
because Event.fromJsonOrNull defaults a missing signature. This switch alone needs
no gating and is made everywhere at once. The `encoding` tag comes off the
application kind:445 while here -- transports/nostr.md forbids it on every shape.
docs/marmot-v2.md, Phase 2.

A direct message could not make that switch as it was. Its payload is a NIP-59
wrap signed by a throwaway key, which a conformant peer refuses twice over: for
the sig member and for an author that is not the MLS sender. In a current-profile
room the wrap now travels inside an envelope -- an unsigned app event of kind
30330 authored by the sender, one `p` tag for the recipient, the wrap's JSON as
content. The envelope passes the author check like any payload, so the carve-out
in mip03Rejection that admitted bare wraps is confined to legacy rooms by a new
`legacyRoom` flag: in a current-profile room a bare kind:1059 is refused as every
other conformant client refuses it. Every property the throwaway key was chosen
for survives -- the envelope is unsigned and so a claim rather than a proof, the
wrap inside is still signed by nobody, and the seal-to-leaf binding in classify
still decides who wrote the words. The line for a message this device cannot read
now hangs off the row the payload was filed under (the envelope's id in a
current-profile room), which ChatMessage's foreign key requires.

The envelope is gated, and the gate is the room's profile: an old build renders
it as a raw-JSON bubble, so a legacy room keeps sending the bare wrap, and a
current-profile room -- which admits only leaves carrying an identity proof,
which only a build past Phase 3 can make -- never holds an old build. That is the
flag for every wire change after this one.

A conformant decoder bounds the whole app payload at 64 KiB, and a chronicle page
rides inside one as JSON-escaped content. The new bounds test caught what the
arithmetic missed: dense JSON payloads escape to half again their size, so a 48
KiB page became a 70 KiB app event. The packer now measures a page by
ChronicleEvent.escapedByteSize -- the bytes it will occupy inside the envelope --
against MAX_PAGE_BYTES, lowered from 64 to 48 KiB, and MAX_PAGE_EVENTS drops from
128 to 96 so the count cap still binds under it, as the comment on the two caps
insists. The receiver's raw-bytes check is unchanged and now holds by
construction.

Tests: MarmotMip03CarveOutTest becomes the test that the exemption is confined to
legacy rooms and that an envelope passes the ordinary check; MarmotDirectMessageTest
round-trips the envelope through the strict decoder and opens the wrap inside;
MarmotPayloadBoundsTest asserts the largest chronicle pages and an envelope are
conformant app events, including the dense-content page measured the way the
packer measures it. The ingest test's tamper case flips a decoded byte rather
than the base64 tail, which could change nothing but padding. Full JVM suite:
1,148 tests pass. The direct-message and chronicle notes record the changes.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 22:32:30 +02:00
Kgothatso NgakoandClaude Opus 5 a68cad06a4 feat(marmot): read both Marmot profiles (v2 Phase 1)
After this the app can be a member of a current-profile group and can invite a
current-profile peer -- an Amethyst 1.15.2 user -- into one of its own legacy
rooms. Nothing it sends changes. docs/marmot-v2.md, Phase 1.

Seventeen sites read `currentMarmotData()`, which is the legacy 0xF2EE
extension and is null on a current-profile group: ten for the room id, and the
Welcome path threw MarmotMissingNostrGroupDataExtension on that null, so a v2
Welcome was not a room the app could not read but an exception it could not get
past. The others read the admins -- SubgroupManager's "only an admin can make a
subgroup" and processGroupMembershipChanges' adminAt stamp both saw no admins at
all on a v2 group -- the relays, the name and the retention. MlsGroupProfile.kt
writes the fallback once per field (current profile first, 0xF2EE second) and
the sites read roomId(), roomName(), roomDescription(), roomRelays(),
retentionSeconds() and the library's currentAdminIdentities(). The retention
one is the bug the library's own outbound processor had: read only 0xF2EE and
disappearing messages simply did not disappear in v2 groups.

A current-profile kind:30443 frames its KeyPackage in an MLSMessage, and the
invite decoded whatever the ingest stored as a bare KeyPackage, taking the
frame's version and wire-format words for the package's own. The ingest now
validates with KeyPackageUtils.isCryptographicallyValid -- the check a
conformant peer runs on ours, covering tags, signature, lifetime and on the
current profile the identity proof; nothing checked any of it before -- and
stores the package bare, so MlsGroup.proposeAdd and every other consumer of the
row read the same bytes whichever way it was published.

The inbound author check decodes the canonical payload strictly first
(MarmotAppEvent: six members, no sig, an id that rehashes) and falls back to
Event.fromJsonOrNull only for the legacy rumor shape every build before Phase 2
sends. The fallback would accept the canonical shape too, but by defaulting the
signature rather than checking anything.

A Welcome is matched to this device's bundle by the KeyPackageRef the Welcome
itself names, in RFC 9420 order, before the `e` tag is consulted: the tag is
what the inviter thinks it used, and MDK stamps the id of a cached copy, so a
peer who rotated while an inviter kept inviting from cache was unjoinable by
the tag alone. WelcomeBundleMatch is pure and scans the handful of bundle rows
rather than adding a column. The exception the tag-only path threw when the tag
was missing has no caller left and goes.

Tests: MarmotKeyPackageIngestJvmTest runs both published shapes through the
sync funnel, invites a current-profile package into a legacy room and sees the
member in the tree, and refuses a tampered package; WelcomeBundleMatchJvmTest
picks the addressed bundle in either candidate order, none when none is ours,
and survives an unreadable row; MarmotAppPayloadTest pins that the legacy rumor
and its canonical form share an id, that the strict decoder refuses a sig
member and a forged id, and that the fallback reads both. The KeyPackage test
fixture's lifetime drops from 90 days to the app's own 30 because 1.15.2 caps
it at 84 and the ingest now checks. Full JVM suite: 1,140 tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 22:15:13 +02:00
Kgothatso NgakoandClaude Opus 5 96a0c81341 build(marmot): quartz 1.15.2, with the two things it broke (v2 Phase 0)
The bump to 1.15.2 -- the release that carries the Marmot current profile --
breaks the compile in exactly two places, and the rest of the ~130 quartz
symbols the app imports are untouched.

CommitOrdering is gone from the library. Its inbound processor now settles
same-epoch commits through a convergence engine and calls the old rule
superseded "and deliberately": created_at and the event id are both transport
metadata the sender chooses and MLS never authenticates, so a member could win
every race by backdating. The app's pipeline does not go through that
processor -- it drives MlsGroup itself -- so the rule is vendored unchanged into
managers/CommitOrdering.kt, with a header saying why the library dropped it,
until Phase 5 of docs/marmot-v2.md replaces the pipeline. NostrDao and
MarmotInboundManager only change their import.

GroupEventResult gained AppMessageOnCandidateBranch and RefusedByLifecycle, so
the exhaustive when in ChatMessage.fromGroupEventResult needs both. Neither is a
delivery -- the first is a payload that decrypted only on a losing fork, which
the library is explicit must not be rendered, the second is input a disbanded
or unrecoverable group refuses -- so both get no chat line, the same as a
duplicate. The app never constructs either today.

Verified rather than assumed: the two sources jars were diffed file by file
against the app's imports; all 1,128 JVM tests pass on 1.15.2; the rewritten
Curve25519 field arithmetic matches the RFC 8032 and RFC 7748 vectors bit for
bit (the app persists raw MLS private keys, so this mattered); the app's own
KeyPackages sit inside the new 84-day lifetime ceiling at 30 days; no new
EventFactory kind collides with the app's.

One-way door: MlsGroupState's blob goes from version 2 to version 4 and 1.14.0
has `require(version in 1..2)`, so a room re-persisted under this build cannot
be opened by the previous one again.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 22:04:41 +02:00
Kgothatso NgakoandClaude Opus 5 8e502aa857 feat(groups): a new curated schema starts from one of curare.to's templates
"Add schema" opened the editor on the two fields every list has and left the
member to build the rest. curare.to's /new page does it the other way round:
a "Start a list" picker of four shapes -- links & text, discussion, catalogue,
custom -- and an editor seeded from the one chosen. The button now pushes
SelectCuratedSchemaTemplateScreen, that picker on a phone, and a card pushes
the editor with the template's key, the same word the site takes as
?template=. Back from the editor is the picker and another shape; when a new
list's proposal goes out the signing session replaces both.

CuratedSchemaTemplate is lib/protocol/templates.ts field for field, with
CuratedDeriveRule for the [rule:<key>] mark the site records in the
identifier field's hint -- its own convention, not the NIP's, and the thing
that lets that site's form derive an entry's d for a list the group starts
here. The hint is written byte for byte as the site writes it; the app still
composes no suggestions and applies no rule. Null or an unknown key seeds
Custom, which is the bare two fields the editor opened on before, so a
screen handed no template behaves as it did.

The picker's one departure from the site: its cards list a template's
fields by label, the way the group's schema cards do, rather than counting
them -- what an entry asks for is what the choice is made on.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 20:25:37 +02:00
Kgothatso NgakoandClaude Opus 5 272b631633 feat(ui): a member on the room's screen opens the profile preview
The member rows on the chat room detail were the last "Profile detail" stub
in the app: each pushed ImplementationPendingRoute. They now push
ProfilePreviewRoute by the member's public key, which the nav host already
serves, so a member arrives with the cached profile at once and the relays
refreshing it, and the bar's action opens the chat that exists with them or
starts one.

The preview rather than the profile detail, because the detail is keyed by a
kind 0's event id and a member whose profile has not arrived yet -- the
"LOADING..." row MemberProfileSync exists for -- has none. The public key is
the one thing every row has. A test pins both members opening the same
screen, the named one and the one still named by its key.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 14:44:57 +02:00
Kgothatso NgakoandClaude Opus 5 80735df86a feat(ui): the preview's action is the app's extended FAB
curated gained ExtendedFab while the preview was being built: every screen's
bottom-bar action, with an enabled that borrows the disabled colours, marks the
node disabled and drops the press. The preview had chosen a Button for exactly
the reason that widget removes -- a FAB had no disabled state -- so after the
merge it was the one screen whose primary action was shaped somewhere else.
Now it is that FAB in a bottom bar, the indicator in its icon slot while the
key package is still owed, and the reason line stays in the body above it.
Hidden, bar and all, for a read-only identity.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 13:43:20 +02:00
Kgothatso Ngako f4b0347e90 Merge branch 'curated' into claude/chat-creation-profile-preview-ae07be 2026-09-13 13:39:54 +02:00
Kgothatso NgakoandClaude Opus 5 ca6c16beb0 refactor(ui): one extended FAB for every screen's primary action, with a gap in it
Twenty screens end their bottom bar in an `ExtendedFloatingActionButton` --
twenty-one buttons, one screen having two -- and eighteen of those were built
from the content-lambda overload with an `Icon` directly followed by a `Text`.
That overload puts nothing between the two, so the icon touched the label on
every one of them: visible in any render, and unnoticed because every FAB in
the app looked the same. Thirteen of the eighteen also carried the same
twenty-line block for looking disabled, because M3 gives a FAB no `enabled`.

**`ExtendedFab`, once, in widgets/buttons.** A label, an icon slot, `enabled`
and `onClick`. It draws `space150` between icon and label, which is the 12dp
M3 names `ExtendedFabEndIconPadding`; it borrows the disabled colours every
other button in the app uses, marks the node disabled so a screen reader does
not announce a button it is happy to press, and drops the press -- so the
thirteen call sites lose their `if (!canX) return@…` guards along with the
block and the `buttonColors` each of them declared. An icon slot rather than
an image, because two screens put a progress indicator where the icon was
while the action they started is out, and they still do.

**Not M3's `text`/`icon` overload, which would have spaced them for free.**
Its source wraps the label in `clearAndSetSemantics {}`: the label leaves the
merged semantics tree, a screen reader is told only what the icon says, and a
test cannot find the button by what it says -- five screen tests do. The
three FABs that were on that overload move to the widget too, and two of them,
Home's "New chat" and "Start the key ceremony", had a decorative icon beside
the cleared label and so no accessible name at all until now.
`ReadOnlyEntrancesJvmTest` had gone to the unmerged tree to find "New chat"
for this reason; the habit is harmless and stays. Every icon's description is
as it was, since several tests reach their FAB by it.

`ExtendedFabJvmTest` pins the decision: found by its label in the merged tree,
the icon before the label with a gap between, and disabled both announced and
refusing the press. The two docs that recorded the old state -- the
conformance account's "left alone" and the npub plan's unmerged-tree note --
each say what superseded them.

Verified with :composeApp:m3Audit (all budgets met, 0 dp literals in spacing
positions), :composeApp:compileDebugKotlinAndroid and :composeApp:jvmTest
(1088 tests, 3 new).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 13:31:47 +02:00
Kgothatso NgakoandClaude Opus 5 d2a4298d25 feat(groups): broadcast every accepted entry from one button, each to where its own would go
The entries screen had one "Broadcast" under each card, and each opened
`BroadcastGroupSignedEventScreen` for that one entry: a group that has accepted
twenty entries into its lists sent them with twenty trips through the relay
picker. This is the other button. "Broadcast all", an extended FAB in a bottom
bar, sends every entry on the screen, newest first, one after another, and says
under each card what its relays answered. Behind no gate, like the button under
each card: sending what the group has already signed takes no share of the key.

**Each entry goes where its own broadcast screen would have started.** The list
that screen seeds -- `defaultRelaysFor`: the group's General write relays or the
app's publish set, plus the list's own relays, minus what the group has blocked
-- is what a member who reads nothing and presses the button sends to, and this
button is that member with twenty entries. The group's relay lists are read
into the loaded state for it; the list's relays come off the entry, which
carries its schema, so the rule is handed that one schema rather than all of
the group's. Nothing is shown before the press, which is the difference from
the one-event screen and the reason that screen stays: a refusal is answered
there, where the relay can be changed and the send tried again.

**One entry at a time, on purpose.** The rows fill in top to bottom, so the
member can see where the batch is; and a public relay sent twenty events in
one burst answers "rate-limited" to most of them, which would make the button
worse than the presses it saves. Each entry's send has the same ceiling as a
single broadcast and the batch has no other, since a long list is meant to
take a while. It goes to the pool directly rather than through the queue, for
the reason `BroadcastGroupSignedEventViewModel` gives, and it is only as
durable as the press: leaving the screen ends it where it is.

**The send is shared rather than copied.** What a relay's answer becomes -- an
OK, a rejection in the relay's words, a failure, or no answer when the whole
send runs out of time -- was decided inside `broadcast()` on the one-event view
model. It is now `sendToRelays` on that model's companion, beside
`defaultRelaysFor` and `wireEventOf`, and `broadcast()` keeps only the state
around it; the thirteen tests on it pass unchanged. The words for an answer
moved the same way, out of the relay row and into `RelayOutcome.label()`, which
both screens now use.

**What the member is told.** A line above the cards -- "Sending entry 3 of
12…", then "Accepted by every relay for 11 of 12 entries." -- persistent rather
than a snackbar, because it is the answer the member came for and a snackbar
would be gone before the last card had been read. Under each card, "Accepted
by n of m relays." in the one-event screen's words, and one line per relay
that did not say OK, named with its reason, because "blocked: not on the allow
list" is something a member can act on and a count of two out of three is not.
An entry with nowhere to go -- everything the seed would name blocked -- is
marked so rather than sent to nobody. The FAB looks disabled while a batch is
out, and a second press is refused.

`GroupCuratedEntriesViewModelJvmTest` pins where each entry goes, that the next
waits for the last to settle, what the answers become, and the refusals; the
screen test drives the button end to end and checks that each answer sits
between its card and that card's own broadcast. The shared fixture entry now
carries the `a` root a real entry has -- without it the seed rule could not
find the list's relays, which the first run of these tests found.

Verified with :composeApp:m3Audit (all budgets met),
:composeApp:compileDebugKotlinAndroid and :composeApp:jvmTest (1085 tests,
7 new).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 13:31:32 +02:00
Kgothatso NgakoandClaude Opus 5 d649098504 feat(chat): the preview's button knows whether it can be pressed
Phase 4 of docs/npub-profile-preview.md. Phases 1-3 moved the twenty-second
dead end one screen later: a person found on an indexer, a press, and
InitiatingNewChat waiting for a key package that someone who has never opened
Curare will never have published. The preview now reads the answer before the
press. A room that already exists opens; a key package that is here enables
"Start new chat"; one that is not, once the relays have had their time,
disables it with the reason under it and a try-again that asks the DM relays
alone.

The existing-room check was a loop with a TODO inside initiateNewChat; it is
ChatRepository.findDirectMessageRoom now -- a room whose members are exactly
the two keys, or the one key when they are the same -- read by both screens and
pinned by a test on an in-memory database. The old size-one branch would have
opened any room the peer sat in alone, whoever they were.

Both halves are observed together, as the chat screen does: a key package
arriving after the profile touches a different table, and a collect on the
profile alone would never see it. A room is read once up front, because a key
package is meant to be used once and the relays need not hold a fresh one for
someone you already talk to.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 13:19:51 +02:00
Kgothatso NgakoandClaude Opus 5 923b59ec32 feat(ui): the new-chat dialog opens the person, and no longer starts the chat
Phase 3 of docs/npub-profile-preview.md, and the user-visible change: "Direct
message via npub" now shows who this is before anything is created. The
dialog's button reads "Find profile" and pushes ProfilePreviewRoute; the chat
is started from there.

The field keeps what it parsed rather than whether it parsed. Its check and its
decode used to be two different functions, which is how "nostr:npub1…" passed
the one and failed the other, closing the dialog with nothing to show for it;
CredentialParser.npubOrNull is the sign-in decode exposed for an address field
-- an nsec is null there, because a field asking whom to write to must never
take a secret for an answer -- and the enabled button and the route it pushes
read the same key. The nip05 branch still lands on the placeholder, and the
test says so. "Start chat" goes with its only reader.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 12:59:26 +02:00
Kgothatso NgakoandClaude Opus 5 8973c0058e feat(ui): the profile preview screen, reachable by nothing yet
Phase 2 of docs/npub-profile-preview.md. ProfilePreviewRoute is keyed by public
key, because that is all the entrance has, and ProfilePreviewScreen draws the
four states over a constant bar: the npub above the indicator while the relays
are asked, the person -- avatar, name, nip05, the whole of the about, the npub
that was pasted -- and under them "Start new chat", which hands over the route
the dialog pushes today and creates nothing here. Not found is the sign-in
screen's sentence with a try-again; a read-only identity is shown the person and
nothing to do, and ReadOnlyEntrancesJvmTest gains the row.

Registered in the host with the room replacing the preview, as it replaces the
profile detail. No button navigates to it until Phase 3.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 12:55:19 +02:00
Kgothatso NgakoandClaude Opus 5 37c6f7eabb feat(chat): the lookup behind a profile preview, and the sync it shares with the chat screen
Phase 1 of docs/npub-profile-preview.md. No screen yet: a view model that finds
the person behind a pasted npub and says when it has stopped looking, over the
row the device already holds -- cached first -- and two asks of the relays.

The three-kind filter the chat screen queued privately is now
DirectMessagePeerSync.requests, and the kind 0 alone goes to the indexers as
profileRequests: a key pasted from elsewhere has lived elsewhere, and a
three-kind filter sent to an indexer comes back with the kind 0 and nothing
else. "A placeholder is not a profile" was spelled out twice as a comparison
against GENESIS_AT; it is Profile.isResolved() now, and both sites read it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 12:50:15 +02:00
Kgothatso NgakoandClaude Opus 5 88577bb5c3 test(identity): several profiles, both ways, and a boot
Phase 8 of docs/multiple-profiles.md: the seams of Phases 1 through 7 in a
row, through the real view models on an in-memory database with a real key
store.

A is made from a phrase through the create view model and the seed writer,
the way Landing makes the first profile: both files written, and listed
once as a profile with a wallet attached, under the wallet's id, with the
credential's key. A is opened with a node behind it -- a PhoenixBusiness
that was never started, which is enough to make it one the switch has to
stop. B's npub is signed in from inside through the sign-in view model; the
switch stops A's node, B is open and read-only, the machine routes it from
its placeholder, and nothing was ever started for a public key. A kind 1
queued for A while B is open waits three seconds unsigned; back to A, and
it goes out, with nothing more stopped since B had no node. C is created
from inside through the bare-key writer: a key, not a second wallet, routed,
with no node, and the switch to it stops A's node again. C and B are
forgotten through ForgetIdentity with the view model's writers, the default
is cleared and the selector reset: A is listed alone, its credential is the
only entry left in the file, B's and C's account rows went with them, and A
itself cannot be forgotten because a wallet is attached. Then a fresh view
model over the same directory, which is what a cold boot is: the repair
finds nothing to do, A is listed from its credential rather than derived,
and StartupChoice opens it without asking.

A navigation state is matched to a profile by whichever step it is at --
UnsignedProfile, UnqueuedProfileSynchronization, UnannouncedProfile through
the event the broadcast request names, ProfileLoaded -- because with the
notary running, A's account moves from unsigned to announced-pending while
the test looks, and what the test asserts is whose account it is, not which
step it reached.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 01:57:05 +02:00
Kgothatso NgakoandClaude Opus 5 9bb340042d feat(ui): leaving one profile among several
Phase 7 of docs/multiple-profiles.md. Small, because the forget sequence was
built right.

Sign out of a read-only identity and "forget this key" for a bare key both
run ForgetIdentity and then the tail -- re-list, clear the default since
Phase 3, show the selector with what remains or Landing if nothing does --
and that is the right behaviour with several profiles as well; nothing in
the sequence changes. A profile with a wallet attached answers
WalletAttached from Phase 1 and is offered neither exit: its sign-out row
stays the pending route, for the reason both sign-in plans gave, and it is
now also the only way such a profile can leave, since the repair would
write its credential back. What changes is that the user is no longer stuck
behind it -- "switch profile" is the row above.

The not-found screen learns about the others. It offers "try again" and,
for a read-only identity, "use a different key", which forgets it; with
another profile on the device it now offers "switch profile" as well, for
every kind, pushing the switcher -- which has a back button, so the screen
is still there if the user changes their mind. The identity that was not
found stays listed; forgetting it remains a separate decision. The nav host
passes the callback only when the listing holds more than one.

Tests: ReadOnlyEntrancesJvmTest -- with another profile on the device the
not-found state offers the switch beside the read-only exit and it calls
through once; a signing identity that was not found gets it above the
set-up form; with one profile there is nothing to switch to.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 01:44:44 +02:00
Kgothatso NgakoandClaude Opus 5 759199f2af feat(sync): a fetch queue per identity, and the inbox reopened on activation
Phase 6 of docs/multiple-profiles.md. The one data problem, closed from both
ends.

The two queues that fetch -- SynchronizeNostrEventRequest and
NegentropySynchronizeRequest -- gain a nullable ownerPublicKey, the identity
that asked, because what a fetch brings back is opened with the fetcher's
key. The pumps ask for their own: the head of the queue among this
identity's requests and the ones nobody owns, so that a request another
identity queued waits for that identity. Rows from before the column read
back null, meaning "the device's", which is what they were, and any identity
may drain them. The broadcast queue deliberately gets no owner: a signed
event is anyone's to carry, and holding A's outgoing message until A is
opened again would be a delivery failure the user would never be told about.
Room version 20, an AutoMigration for two nullable columns, with the schema
export committed beside its predecessors.

The stamp is the repository's, not the call site's. Eighteen view models and
two DAO paths queue requests, and every one of them does so as the active
identity -- a screen cannot queue anything as anyone else. DatabaseNostrRepository
takes the identity flow at construction, which meant moving the view model
above the repositories in the nav host, and stamps every request it queues
where the caller stamped nothing; a negentropy request carries its owner
into the REQ it becomes. The DAO's own requests -- the placeholder-profile
syncs it plants while indexing, and the participant syncs of a Marmot join
-- stay unowned, on purpose and against the plan's sketch: they fetch public
kinds that need no key to open, and any profile that is open may as well
fetch them.

The other end: wraps that were fetched under the wrong key -- before this,
or by a read-only identity whose nsec was pasted later, the case the npub
plan's DAO guard left with the words "the key that opens this one may be
signed in later" and no code behind them. storeNostrEvent never indexes an
event it already holds, so a wrap that arrived under the wrong key stayed
closed for good: the live subscription re-received it, the DAO saw a known
id, and returned. NostrDao.reopenInbox finds every wrap addressed to the
active key that no seal names and runs each through indexNostrEvent again
under the right key, one transaction per wrap so that one that cannot be
opened rolls back its own changes and the next is still tried; one pass,
since wraps do not depend on one another. It runs from the sync pumps'
collector once per activation, for an identity that can sign, before the
live inbox and the broadcast pump are launched -- so the sweep and the live
subscription are not opening the same wrap at once; both are idempotent by
event id.

Found on the way: GiftWrapSeal.giftWrapMessageId, a foreign key to the wrap
a seal came out of, existed and was never written, so nothing in the
database could say which wraps had been opened. decryptGiftWrapSeal now sets
it. A seal from before reads back null, is swept once, and comes back with
the link -- the reindex is idempotent, and persistInboundChatMessage already
files a message it has once.

Tests: ReadOnlyGiftWrapDaoJvmTest -- a wrap for us stored under another
profile's key is stored and not opened, a re-delivery under our key changes
nothing, the sweep opens it once and finds nothing the second time; the
other profile's sweep and a read-only pair open nothing; a wrap stored
read-only is opened once the key is here. OwnedRequestQueueJvmTest -- A's
request is offered to A and not to B, nobody's to both, oldest first; the
repository stamps what it queues with the identity that is open, nothing
when none is, and keeps an owner a caller named.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 01:41:14 +02:00
Kgothatso NgakoandClaude Opus 5 2f42023081 feat(identity): adding a profile from inside, as a key and not a second wallet
Phase 5 of docs/multiple-profiles.md. The two entrances, and the two things
that were only safe with one profile.

Under the list, on the switcher and on the startup selector alike, two rows:
"sign in with a key", to the sign-in screen, and "create a new profile", to
the create screen. Both push a screen that has a back button. Landing is not
one of them -- it has no back button and clears the stack on its way out,
because it is the first-run screen, and pushing it from inside would make it
a pushed screen on some days and a root on others. A device with one profile
can now be given a second, which it could not: Landing showed only when the
device held nothing.

A profile created from inside is a bare key. CreateProfileScreen generated
twelve words behind its form -- a seed, a node, a wallet, and at the NIP-06
path the key that becomes the profile -- which is right for the first
profile on a device, the one the wallet will belong to, and wrong for the
second: a user who wants another profile has not asked for another
Lightning node, another set of channels, or a second phrase with funds
behind it. The screen now takes a NewProfileWriter and does not know which
it was given: seedProfileWriter from Landing, bareKeyProfileWriter from the
switcher, thirty-two random bytes written the way a pasted nsec is.
CreateProfileRoute carries the choice. The view model writes the secret
first, then the six bootstrap events for the key it derives, then hands the
id to the tail -- so the events are only ever queued for a key the device
holds, and they are in the database before the tail activates the identity,
an ordering the sign-in screen documents as load-bearing and the create flow
used to get away with by luck. The tail gains the popUpTo(0) the sign-in
tail has. A second wallet is not offered from inside; a pasted recovery
phrase still brings its own.

"End this" goes. It wiped the database -- every profile's rooms, MLS state,
key packages and queues -- from the screen whose purpose is to add a
profile; it was only ever a development exit, and with several profiles on
a device it was a way to lose the others. wipeDatabase leaves the view model
and stays on the repository for the tests that use it.

A key already on the device is refused with the id it is listed under.
WriteNostrCredentialResult.AlreadyExists and WritingSeedState.Error.
SeedAlreadyExists carry it -- the wallet's id where a seed derives the key
-- and the sign-in screen's error state gains one action, "switch to it",
which is the sign-in tail with that id. From Landing "already on this
device" was the whole message; from inside it is a sentence with an obvious
next step, and a duplicate paste becomes the fastest switch in the app.

Found on the way, and fixed in the jvm actuals: DataStoreManager caches one
GlobalPrefs per process behind an unsynchronised check-then-set, and
DataStore refuses a second instance over the same file. Two first calls at
once -- the listing on IO and a sign-in's write, or the startup screen on
Main -- could both see the empty cache and both build one, and the loser
threw "multiple DataStores active for the same file" at first use. The test
for this phase hit it one run in three. JvmGlobalPrefs is now the one way
the jvm target reaches the prefs, under a lock; Android goes through the
Application's single instance and never raced; the library's own callers
come at node start, when the cache is long populated.

Tests: AddProfileFromInsideJvmTest, through the real view models on an
in-memory database -- A open, B's nsec committed and switched to, and back,
both accounts intact with their own kind 0; A's key pasted again refused
with A's id; a profile made through the bare-key writer lists as a bare key
with the kind 0 first of its six bootstrap rows, none signed yet, and no
node ran. SignInToProfileViewModelJvmTest and IdentityWriterJvmTest assert
the ids the three refusals carry. ProfilesScreenJvmTest: the two rows call
their callbacks and switch nothing, under the caption that the open profile
stays. CreateProfileScreenJvmTest: the state that carried "end this" is
composed and the words are not there.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 01:24:14 +02:00
Kgothatso NgakoandClaude Opus 5 25464595f4 feat(ui): the switcher, showing the nostr profile and what the device holds for it
Phase 4 of docs/multiple-profiles.md. One pushed screen, and the startup
selector brought up to match it.

ProfilesRoute, reached from the profile tab's row -- "Switch profile", where
it said "Change account" and went to the pending screen -- and drawn by
ProfilesScreen: a top app bar with a back button, and WalletsSelector as its
whole content with the open profile above the divider and the others below.
Tapping another row is switchToIdentity and nothing else; the observer
restarts the app as that identity through its lock gate, so the screen is
never around to see the result. Tapping the open row does nothing. Three
states, said so: loading while the list is read, which in practice is never
seen; an unreadable key file, through ErrorState with a retry that re-lists;
and the list. Empty cannot happen while something is signed in and the screen
does not pretend it can. One column at every width. The row is offered to
every kind: a read-only identity can leave for another profile the same way
a signing one can.

ProfilesViewModel joins the flows the app already holds -- the listing state,
the identities, the active identity, the wallet metadata -- to the profile
rows for the listed keys, through NostrRepository.observeProfilesOf, one map
that moves when any of them does. Nothing in it writes.

What a row shows changes. WalletsSelector showed "Default name" over a random
emoji for every row, because nothing in this app writes the wallet metadata's
name, and two profiles side by side both called Default name and told apart
by a bech32 string is not a switcher. The row now takes the nostr profile --
humanReadableNameOrPubkey over ProfileAvatar, the widgets the profile tab
draws itself with -- with the npub on the second line, and falls back to the
npub over the metadata's emoji for a key the device has no kind 0 for yet:
one just created, or a read-only one that was never found. And it says what
the device holds for the profile before the tap: "Wallet" for a seed
attached, beside the "Read only" that was already there, and nothing for the
plain case of a bare key. It is the difference between a profile that can be
signed out of and one whose sign-out is a wallet question, and between two
profiles with the same name. The startup screen draws the same widget and
gets the same map through rememberProfilesOf, so both lists look the same.
The selector's dead globalPrefs parameter goes with the rewrite.

Tests: ProfilesScreenJvmTest, on the unmerged tree, with a wallet-attached A
open, a bare B and a read-only C -- A by its nostr name with its npub and
"Wallet" under it, B and C by their npub with only C labelled, no "Default
name" anywhere, the open one drawn above the others; tapping A calls
nothing and tapping B calls the switch once with B's id; an unreadable key
file is an error whose retry is counted; the bar has a back button.
ReadOnlyEntrancesJvmTest asserts "Switch profile" for both kinds.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 00:50:37 +02:00
Kgothatso NgakoandClaude Opus 5 00fc996995 feat(identity): which profile opens, on launch and after a switch
Phase 3 of docs/multiple-profiles.md. The startup screen learns to agree with
the switch.

The table that picks what to open is lifted out of the screen's remember into
StartupChoice.resolve, where it can be read and tested, because its order was
the whole decision and it was wrong in the one way a switch would hit:
"!startWalletImmediately -> null" sat above "desiredWalletId != null", and
nothing ever set the flag back to true. So after the first visit to the
selector, every sign-in on that device landed on the selector instead of the
profile it had just signed in. Two rows move: what a switch or a sign-in
named goes above the user's earlier "show me the list", since it is the more
specific instruction; and the single-identity row drops below it, since when
both are set they name the same thing.

The default is written. GlobalPrefs.getDefaultWallet has been read by startup
since Phoenix and saved by nothing in this app, so a cold boot with two
profiles was the selector every time with no memory of which one was open.
setActiveIdentity now saves the id it activates -- every activation goes
through it, startup for all three kinds and the two tails through startup --
so the default is always the profile most recently open. The two forget tails
clear it, since the default is the one memory of an identity that would
otherwise outlive it. Both writes are wrapped: a default that could not be
recorded is a selector on the next boot, not a crash now. The screen reads
the saved value as a default only when it names a wallet, which is the one
thing left to decide at the call site.

The startup screen's literals go to the catalogue, and "wallet" goes with
them except in the one place a wallet is what is starting: "Starting wallet"
is shown from StartupViewState.StartingBusiness, which only the branch with
a wallet attached reaches, and stays. The rest become "Preparing profiles",
"Opening profile", "Decrypting", "Loading preferences", "Unlock to continue"
and "Could not load the profiles on this device"; the selector's title is
"Choose a profile", and select_a_wallet goes.

Tests: StartupChoiceTest in commonTest, one case per row and the two
orderings this phase is for -- a desired id opens even after the user once
asked for the list, and one identity with the list asked for still shows it.
IdentitySwitchJvmTest gains the cold boot: activate A then B, and a fresh
view model over the same directory resolves B without asking; forget the
default, and it resolves nothing. Found on the way and recorded in the test:
DataStoreManager caches the global preferences once per process, bound to
whichever directory was current when the first test in the JVM asked -- the
same trap the nsec plan recorded for user preferences -- so the test reads
the default through the view model's own instance rather than one of its
own.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 00:43:21 +02:00
Kgothatso NgakoandClaude Opus 5 9b3a19278f feat(identity): a switch that tears down what it should
Phase 2 of docs/multiple-profiles.md. The transition, made correct, with
nothing yet calling it from a screen: the view model, one manager, and three
one-line actuals.

switchToWallet becomes switchToIdentity, and does what a switch is: remember
which one, set startWalletImmediately back to true -- a switch is the user
saying which one, and the flag was only ever the user saying "show me the
list"; nothing set it back before because nothing could switch -- clear the
active identity, and stop the node of the profile being left, if it had one.
The identity is cleared before the node is stopped, so that every collector
that could reach the business is cancelled before the stop runs. The test is
business != null, not the kind: whether the profile being left has a node
behind it is the fact, and the kind is how it currently comes to be true.
stopPlatformBusiness is an expect beside updateBusinessActiveInUI, for the
reason that one is: BusinessManager is a per-platform object. Its three
actuals call stopBusiness, which existed on every platform and nothing
called. The manager is injected into the view model as a function so that
the branch can be pinned without a node.

RelaysSocketManager.observeActiveUserId becomes a child of collectLatest --
the shape the pumps and the notary already had. It used to keep one job per
pubkey on its own scope and cancel only the job for the pubkey being
started, which meant a switch left the previous identity's observer running:
two observers feeding updateRelayPools, and the one pool following whichever
relay list emitted last. The map goes; a null identity closes nothing, since
the pool is shared by pumps a null identity has already cancelled, and the
next identity's list replaces it through changeRelays as it always did. The
scope is injectable and relayUrls is exposed, both for the test.

The sign-in and create tails keep their own navigate beside the observer's,
now with a comment saying why: the navigation state is a StateFlow, an
update to a state equal to the current one emits nothing, and the explicit
navigation is what guarantees the stack moves even on the day the state
does not.

Tests: IdentitySwitchJvmTest, through the real SovereignWalletViewModel with
a real notary and navigation machine on an in-memory database -- A open and
a kind 1 queued for A is signed; switch to B, the identity clears, the
machine goes to startup, B is activated the way startup does and routed from
its account; a kind 1 queued for A now waits three seconds unsigned, and
goes out when A is back. Leaving a profile with a PhoenixBusiness behind it
-- constructible without a node, since everything in it is lazy -- stops
that node once; leaving nothing, or a bare key, stops nothing.
RelaysSocketManagerSwitchJvmTest over a fake relay repository and fake
sockets: after the switch the pool holds B's relays, a re-emission of A's
list changes nothing -- the line that fails against the old observer -- and
an identity whose list is still empty leaves the pool as it was, which is
the behaviour updateRelayPools has always had for an empty list.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 00:27:06 +02:00
Kgothatso NgakoandClaude Opus 5 3008137e3f feat(identity): a seed's key is a credential, and the seed is the wallet attached to it
Phase 1 of docs/multiple-profiles.md. No library change: the file format, the
encrypted writer and the manager all exist, and the app already wrote the
credentials file from the seed writer -- to delete a public entry a seed
superseded. This changes what it writes there, and what the listing believes.

Until now a seed's nostr key was never in nostr-credentials.dat. It was
derived from the words at listing, to know which npub to show, and from the
running node at activation, to know which key to sign with -- so a seed-backed
profile existed only as a derivation, and the app had to start a Lightning node
to find out who it was. Both sign-in plans made "one key, one file" an
invariant, and it was the wrong one: it said a wallet is a profile. Now the
credentials file is the list of profiles and a seed is a wallet attached to the
entry its key derives.

writeMnemonic writes two things, credential first: a Secret for the derived key
under its x-only pubkey, replacing a public entry where there is one, and then
the seed. Credential first because a crash between the two leaves a bare-key
profile the phrase completes, which is a valid thing to hold and says the model
out loud -- the profile exists, then a wallet is attached to it. The one
refusal it drops is the phrase of a key held as a bare secret, SeedAlreadyExists
under the nsec plan: the device did not have that wallet, so this is the
profile acquiring the wallet that derives it. The entry stays a secret for the
same key, the id becomes the wallet's, and the bare key's preference files go
with the old id -- the profile's preferences are the wallet's now, fresh, which
is right since there is a new secret to back up. The same seed twice is still
refused, by wallet id, and is the only way a profile with a wallet attached is
offered its phrase again.

SeedCredentials.reconcile is the repair for every seed already on a device,
beside migrateFromNostrKeys in listIdentities and shaped like it: a named,
idempotent write, one file write for however many seeds are missing, nothing at
all on a device with none or one already repaired. A failed write is a result,
not a throw, and carries the map that was read: the listing goes on with it and
merge derives the key of a seed that has no credential, so a seed this could
not repair is still listed. A failed write must never hide a wallet. The plan
had the repair running before the credentials file was read; it runs after,
and hands the listing what it returns, so the file is decrypted once -- the
doc now says so.

StoredIdentity.merge inverts: the credentials are the list, and each seed is
attached to the Secret its key derives -- listed once, as Mnemonic under the
wallet's id, carrying the credential's key -- where before the seeds were the
list and a secret for a seed's key was listed twice under two ids. Mnemonic
gains privateKey. A Public for a seed's key is skipped with a log line, since
the next repair upgrades it; a seed with no credential is listed by derivation,
with a log line. IdentityKind.Mnemonic's doc changes to what the kind now
means: the name records the attachment, not the source.

setActiveWallet takes the StoredIdentity.Mnemonic and builds the identity from
the credential's key, with the node's derivation as a cross-check -- a check(),
because a node disagreeing with the credentials file is the one corruption
worth refusing to run under, and it cannot fail for a file the repair wrote.
The node still starts for a profile with a wallet attached: not for the key any
more, but for what startNewBusiness does besides -- metadata, preferences,
last-used build, and on Android the channel watcher a restored Phoenix phrase
may need. Making it lazy is now one branch and is named as its own decision.

forgetNostrCredential refuses a second thing: a key a seed derives. The seed
would derive it again and the next repair would write it back, so a forget
that succeeded would undo itself. NotACredential becomes WalletAttached in the
writer's result and ForgetIdentity's outcome, since that is now the only reason
a signing profile cannot be forgotten -- a key not in the file at all can, since
the repair, only be a seed's key the repair could not write.

The two sign-in docs' tables each gain a row pointing here for the invariant
this supersedes.

Tests: SeedCredentialsJvmTest against a device from before -- seed.dat written
directly, no credentials -- writes exactly what is missing in one write; a
device already repaired is not written to again, checked by the file's bytes
since a rewrite would carry a fresh iv; a public entry for a seed's key is
upgraded; bare keys are untouched; and a write that fails, with the key store
locked, is reported with the map that was read and the wallet is still listed.
IdentityWriterJvmTest drives the callback-shaped seed writer through a
CompletableDeferred on a real Main dispatcher, since it reports after a real
one-second delay: a phrase writes both files, its nsec and npub are then
duplicates and its forget is WalletAttached; the same phrase twice is refused;
the phrase of a bare key attaches, under the wallet's id, with the bare key's
preferences gone; the phrase of a key held read-only attaches and the entry
becomes a secret. StoredIdentityJvmTest lists a secret-plus-seed once with the
credential's key, a seed without a credential by derivation, and a public entry
for a seed's key as the wallet only. Not under test: the activation's
cross-check, because a PhoenixBusiness cannot be built without a node.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-13 00:20:22 +02:00
Kgothatso Ngako 86cb876b77 Merge branch 'claude/npub-readonly-signin-ec4398' into curated 2026-09-12 22:24:40 +02:00
Kgothatso NgakoandClaude Opus 5 315e93315a test(identity): the preview end to end, with nothing signed, and the upgrade
Phase 7 of docs/npub-sign-in.md.

NpubPreviewRoundTripJvmTest is the nsec round trip with the repository
made real: an in-memory database under the app's own DAOs, because the
assertion this exists for is about what is not in it. An npub is signed in
the way the sign-in screen does it, listed the way startup does, activated
as a read-only identity, and handed to NavigationViewModel: it lands on
UnqueuedProfileSynchronization. The kind 0 the relays would answer with is
indexed through the read-only key pair: ProfileLoaded. And with the real
NotaryViewModel watching the same rows for longer than its key package
delay, nothing was signed -- the only unsigned row for the pubkey is the
placeholder, still at genesis; nothing is queued for a signature; no key
package bundle; no broadcast request; no node.

The contrast that makes those assertions worth having: the same harness
as a signing identity does sign -- the notary makes a key package bundle
within its delay. Without it, "nothing was signed" could be true of a
harness in which nothing can be signed.

Then the upgrade: the nsec of the key held read-only signs in over it
through the same view model, under the same id, leaving one credential
that is now a secret, one listed identity, and one account -- the second
sign-in planted nothing.

The full jvm suites pass: 988 in the app, 159 in the library.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-12 20:51:16 +02:00