Commit Graph

6 Commits

Author SHA1 Message Date
klopez4212
b1d60f98e2 Bring synced Solo chats and projects to mobile (#8337)
## Summary
Bring desktop Solo chats and project organization to mobile, with synced
collapsible sections, shared unread highlights and tab badges, agent
selection, shared search, native iOS creation sheets, and inline typing
bubbles that transition into replies.

Resolve Solo placement for every chat entry route, retain encrypted
organization snapshots for offline reading, and let users resume or
discard unfinished chat and project setup without deleting existing
messages.

### Related issue
Related: #6565; expandable agent transcripts remain separate work.

### Testing
- Fresh Solo read state seeds each active chat from its captured initial
last-message timestamp before exposing row/section/tab unread state.
Baselines survive intervening channel updates and reset across
identity/community boundaries while surviving read-state readiness
changes. Regressions cover multiple chats, events before read-state
readiness and during/after seeding, and identity changes; removing
either the seed or the display gate fails them.
- Chat completion reconciles a retained Solo-save journal even when
placement is already confirmed. False/throwing cleanup regressions
ensure the outer intent remains until reconciliation and only one
placement is published. Typing terminal-close loops remain bounded; the
retry budget resets after an uninterrupted eight-second health interval
or transport reconnect.
- Typing retry budgets reset after sustained recovery; invalid Solo
updates retain cached rows with a retry warning that survives failed
refreshes. Project names use the editors’ 120-grapheme limit, including
emoji and combining characters. Regression coverage exercises repeated
recovery, invalid live/refresh heads, and character-limit boundaries.
- Rebased onto current main, preserving the newer Search controls,
Huddle timeline filtering, and message-media layout.
- Failed project journal writes reload storage before claiming recovery
and reopen the entered name for retry. Typing subscription rejections
are caught; terminal closures clear stale indicators and retry with a
bounded backoff, canceled on disposal. Regression tests cover both
false/throwing storage writes and initial/late subscription rejection.
- Explicit ordinary destinations skip optional Solo lookup waits.
Generic destinations perform one bounded lookup in the destination
(search no longer duplicates it); an unavailable classification falls
back to ordinary sending per the product owner's explicit choice.
Positively identified Solo destinations use Solo providers.
- Terminal Solo subscription closure preserves cached rows and exposes
live-update Retry. Page coverage verifies warning visibility, usable
rows during recovery, and a subsequent terminal close. Removing the
warning or the immediate ordinary-route guard fails its regression.
- Generic channel entries share an ordinary default when Solo
classification is unavailable, per the requested product behavior. Known
Solo placement takes precedence; callers can explicitly require
unresolved placement to remain blocked. Bounded initial lookup preserves
known Solo deep-link routing.
- Dirty newest snapshots and offline-save warnings survive transport
reconnects; Retry persists that same event even if the relay returns an
older head. Community changes replace the service, and agent-directory
reconnect refresh remains intact.
- Pending-save discard reloads failed removals. Repeated reconnect
testing confirms typing subscriptions are released on provider rebuild
and disposal; removing the existing cleanup fails the regression.
- Ordinary Home/create routes explicitly preserve private-channel
sending during Solo failures; identified Solo sends use the Solo sender
and explicit unresolved routes remain fail-closed. Solo unread replies
beyond the initial 200-event page are loaded through the oldest-unread
boundary.
- Chat/project/agent cleanup reloads failed removals, retains
same-process recovery, and protects replaced records. False/throw
regressions verify retries do not duplicate creation or membership.
Route, unread, and cleanup guard-removal mutations each fail their
regression.
- Ordinary channels, DMs, and forums remain sendable when optional Solo
sync fails; explicit unresolved placement stays blocked. Controlled
cache races retain the newest verified snapshot, keep usable chats
visible, and retry offline persistence without stale writes. Both
protection-removal regressions fail as expected.
- Confirmed journal cleanup checks false/throwing removal and
exact-record identity. Interrupted agent choices are visible
immediately, identify the saved agent, and cover Resume/Later/Discard
with accessible name/key-fallback controls.
- Notification deep links preserve the contiguous history cursor;
same-second messages infer the newest agent consistently. Failed
offline-cache writes surface Retry, with false/throwing storage failures
and offline restart covered.
- Composite history cursors cover 350 same-second events plus older
history; removing the ID cursor makes the regression fail. Agent changes
now journal before granting access, with restart/failure and
Resume/Later/Discard widget coverage. Signed malformed profiles remain
isolated.
- Native iOS avatar tests passed on the simulator: visible-row loading,
three-request concurrency, streamed byte limits, response validation,
downsampling/pixel limits, cancellation, and visible avatar updates.
- Solo and Home share observed-event/read-marker logic. Widget coverage
verifies unread rows, collapsed-section indicators, the tab badge before
first Solo visit, message/thread reads, forced unread, and filtering
archived/nonmember/unplaced chats.
- Chat intent is saved before ownership and organization preflight;
tests cover interrupted ownership lookup and a disconnect while the
native creation form is open.
- Mobile analysis, formatting, file-size checks, and full mobile tests
passed at `1d4c41354`.
- Project creation persists its encrypted name/UUID before connectivity
checks or relay reads. Restart tests cover Resume/Later/Discard, scope
isolation, stale-intent rejection, and lost acknowledgements. Removing
the pre-network save makes all four production-flow regressions fail.
- Regression coverage includes notification/Activity routing, inline
replies, unknown-placement send protection, restart recovery/discard,
and encrypted offline cache isolation/refresh.
- Earlier feature build exercised on a physical iPhone with existing
account/community data; review follow-ups have not been reinstalled on
the device. UIKit agent-row accessibility checks passed on an iOS
simulator.
- Earlier cross-platform `just ci` stopped on the existing `buzz-acp`
test `signals_cancel_startup_turn_and_pending_stdin_without_hanging`
(process exit-code assertion); no Rust production changes are included
in this PR.

### Remaining review limitation
Agent discovery is bounded to 10 pages / 1,000 profiles with automatic
retries disabled. Owner-scoped discovery remains open because the
current relay ignores `#auth` filters; supporting that query requires a
relay change.

### Snapshots
Sample data, shared Flutter/Android surface. Native iOS uses platform
sheets and controls.

| Solo: projects and chats | New chat | Choose an agent |
| --- | --- | --- |
| <img
src="https://raw.githubusercontent.com/block/buzz/d9b29aa079a3b9a88e8ed89b70df8d6af06fa41b/pr-8337--01-solo.png"
width="240" alt="Solo with collapsible projects and chats" /> | <img
src="https://raw.githubusercontent.com/block/buzz/d9b29aa079a3b9a88e8ed89b70df8d6af06fa41b/pr-8337--02-new-chat.png"
width="240" alt="New chat form" /> | <img
src="https://raw.githubusercontent.com/block/buzz/d9b29aa079a3b9a88e8ed89b70df8d6af06fa41b/pr-8337--03-agent-picker.png"
width="240" alt="Agent selection list" /> |

---------

Signed-off-by: kenny lopez <klopez4212@gmail.com>
Signed-off-by: Logan Johnson <loganj@squareup.com>
Co-authored-by: Terp <5856f6900e8e0a4e0ea1f4f8d2ee33e4753d2119bc122f94ecafc660b46c0511@buzz.block.builderlab.xyz>
Co-authored-by: Logan Johnson <loganj@squareup.com>
Co-authored-by: Larry <627498bd4bd1f281a16431e3c6cce3b5c25b6692798c78672298aefbf2f8f8b5@buzz.block.builderlab.xyz>
2026-10-15 17:07:23 -04:00
Logan Johnson
9322289364 feat(mobile): follow the portable mention rules, including DMs (#8102)
🤖
## Summary
The mobile app's @mention chooser now follows the same rules as the
desktop app (buzz-app). The same typed text finds the same people in the
same order on both apps, and sending works the same way.

- **Same matching and order.** Mobile uses the shared mention rules and
runs their shared test cases. A multi-word search such as `@Mary J`
still searches the community directory, so it can find "Mary Jane" even
when she is not in the channel.
- **DMs can mention people outside the DM.** The chooser marks them "not
in DM". Nobody can be added to a DM, so the message sends without a
prompt. Outside people go out as references, so they are not notified.
This matches buzz-app #510.
- **Channels still ask.** In a stream or forum, mentioning someone
outside the channel asks "Invite" or "Do nothing" before the message
sends. A channel with no type tag counts as a stream.
- **Sessions do not search the directory.** Only session members appear.
- **A failed directory search stays visible.** The chooser shows the
error and a Retry button. Before, it closed.
- **Search results stay in their community.** If you switch communities
while a search runs, its people and its error do not show in the new
community. The search starts again there.
- **A failed send check keeps the draft.** Before a message with a
mention sends, the app checks who is in the channel. If that check
fails, the message stays in the compose bar and a notice offers Retry.
Before, nothing was sent and nothing was shown.
- **Invite stays in its community.** If you switch communities while the
"Invite" question is open, Invite adds nobody.

### Screenshots
The real compose bar widget, rendered by Flutter's test renderer with
the app's fonts and theme. The data is sample data. These are not device
captures.

| DM: outside person is "not in DM" | Stream: outside person is "not in
channel" |
|---|---|
| <img width="320" alt="DM chooser"
src="https://github.com/user-attachments/assets/0e884518-2386-4cd1-9f19-16661aaa194e"
/> | <img width="320" alt="Stream chooser"
src="https://github.com/user-attachments/assets/5395e4e1-032e-459b-98c6-0126c060c4d9"
/> |

| Stream: send asks first | Session: members only | Search failed: Retry
|
|---|---|---|
| <img width="260" alt="Stream ask on send"
src="https://github.com/user-attachments/assets/2b988257-3c6a-445d-acfa-fca47b9752fe"
/> | <img width="260" alt="Session chooser"
src="https://github.com/user-attachments/assets/e193551f-f5bf-41d9-aa29-bc8665c643f9"
/> | <img width="260" alt="Search failed with Retry"
src="https://github.com/user-attachments/assets/e8c4b9bb-067e-4fe9-9fdd-4e5691bc66be"
/> |

### Details
- `mobile/lib/shared/mentions/mention_rules.dart` holds the rules: query
syntax, match tiers, sort order, Space selection of an exact unique
name, stable rows while results arrive, and `p` and `mention` tags. A
`p` tag notifies the person. A `mention` tag is a reference and does not
notify.
- `mobile/test/shared/mentions/mention-rules.fixtures.json` is an exact
copy of the buzz-app fixture file. `mention_rules_test.dart` runs every
case.
- Known difference: desktop gives the directory to a channel with an
unknown explicit type (not stream, forum, DM or session). Mobile does
not. No such type exists today.

### Related issue
None found. Desktop counterpart: block/buzz-app#510. #7513 also edits
`send_message_provider.dart` and may conflict.

### Testing
`just mobile-install mobile-check mobile-test` passes: analyze clean,
2998 tests pass on `49c44653` (before rebase). The pre-push hook ran
`just mobile-check && just mobile-test` again on the rebased head
`de1ca080` and passed.

Each review fix was mutation-checked: removing the fix makes its new
test fail.

---------

Signed-off-by: Larry <627498bd4bd1f281a16431e3c6cce3b5c25b6692798c78672298aefbf2f8f8b5@buzz.block.builderlab.xyz>
Co-authored-by: Larry <627498bd4bd1f281a16431e3c6cce3b5c25b6692798c78672298aefbf2f8f8b5@buzz.block.builderlab.xyz>
2026-10-05 18:54:49 -04:00
Logan Johnson
d074576871 fix(mobile): bind same-name mentions to exact selected identities (#7385)
<!-- Draft PR body for #7385 · fix(mobile): bind same-name mentions to
exact selected identities · https://github.com/block/buzz/pull/7385 -->

🤖

## Summary

If two people in a channel share a display name, mobile couldn't tell
them apart in mentions: picking the second could overwrite the first's
selection, a rendered mention linked whichever same-name person matched
first, and a later rename or a shorter name could re-bind the text to
the wrong recipient. This PR binds every mention to the exact selected
identity:

- Each same-name selection keeps its own recipient instead of
overwriting by name; conflicting picks get a qualified label like `Name
(key…)`.
- The longest matching label wins, so a shorter or interior name can
never claim part of a longer one and steal its identity.
- Rendering resolves recipients by the signed identity key from the
event's tags — never from message text alone — so qualified labels stay
correct regardless of tag order and survive later renames; an untagged
ambiguous label blocks shorter mentions instead of silently re-binding.

Ports the landed Desktop exact-recipient behavior (see
`docs/mention-editor.md`).

### Related issue

- Fixes: N/A. No mobile issue; Desktop's landed exact-recipient fixes
are the reference this ports.
- Independent base (`main`). #7387 (child) persists these exact
selections in saved drafts.
- Landing note: branches in this series overlap in the composer — when
rebasing, keep exact/durable mention bindings, the explicit
invite/reference-only choice, the account/visit/revision fences, and
authorization before membership preparation and publication; don't
resolve conflicts by taking either side wholesale.
- Draft — not requesting merge yet; the security advisory run for this
range timed out without results (no verdict).

### Testing

- Regressions cover same-name collisions, prefix/overlap, removal, tag
order, and renames.
- At `acd841354a692243f1cb4c04059baad42f1d2abb`: `just mobile-check` and
full `just mobile-test` pass (2,082 tests). The earlier full `just ci`
receipt linked below is reused only for unchanged non-mobile
code/tooling, not claimed as a rerun at this head.
- Previous-head evidence: `just mobile-check`, the full mobile test
suite, and full local `just ci` all pass — receipts in the [exact-head
evidence
comment](https://github.com/block/buzz/pull/7385#issuecomment-5574635836).
- Verification is widget-test level; no native device or simulator run
is claimed.

To see it: mention two teammates with the same display name — both stay
distinct, the second shows a qualified label, and the rendered message
keeps both correct even after either renames.

### Screenshots

Flutter production-widget test renders — not native-device screenshots
or acceptance captures.

| Scenario | Before | After |
|---|---|---|
| Qualified mention chip for same-name recipients, at 200% text scale
(deliberate stress fixture) | ![Before: the short chip is followed by
the distinguishing key spilling out as raw
text](https://github.com/user-attachments/assets/f1e40acc-24f1-47f7-b2f7-e3ba2348c169)
| ![After: the full qualified label wraps inside the bounded chip,
keeping the @ and robot glyph
visible](https://github.com/user-attachments/assets/b949e44f-e3f6-4c72-98ff-68139b2e1ab3)
|

<details>
<summary>Capture provenance</summary>

Rendered by the Flutter widget engine in a `flutter test` run
(production widgets, production theme; no device or simulator). Before:
this PR's declared base `3c7f288c60d67df78577b237e27c3dfc8831aaa1`.
After: its head `39afd73b0adfde14164f4b10dbd089cb498312b6`.

</details>

Signed-off-by: Logan Johnson <loganj@squareup.com>
2026-09-11 11:19:17 -04:00
Logan Johnson
82656ffea0 fix(mobile): standardize public-key identity display on npub (#7493)
🤖

## Summary

In the mobile app, anyone who hasn't set a display name shows up as a
raw 64-character hex key (e.g. `3a5d4f9c…`) — unreadable, and
unrecognizable as the same identity across screens. Profile and Settings
also let you copy that raw hex. Nostr public keys have a standard
readable form — `npub1…`, the same encoding other Nostr apps and our
desktop app already display. This PR makes every mobile identity surface
render npub instead:

- **Unnamed people everywhere** — message and thread authors, reactions,
typing indicators, member lists, channel details, DM headers and tiles,
inbox, search, forum cards, Pulse notes and reply context, mention
suggestions, and invite rows — now show a compact npub label: first 8 +
last 4 characters of the full npub joined by an ellipsis
(`npub1abcd…wxyz`), the same truncation desktop uses. Previously these
showed truncated raw hex.
- **DM fallback avatars and blank names** — 1:1 DM tiles and headers key
their fallback avatar to the same non-self counterpart the label names,
including self-first participant order; a self-DM keeps its
hex-key-derived initial. Blank or whitespace-only display names fall
back to the compact npub instead of rendering empty, while nonblank
authored names render verbatim (padding included).
- **Profile sheet → "Copy public key"** now copies the full canonical
npub — never raw hex. When the identity string isn't a valid public key,
the copy tile is disabled, so a malformed key never reaches the
clipboard.
- **Settings → Identity (pubkey)** displays and copies the full npub; an
invalid identity reads "Identity unavailable" with copy disabled.
- **Invalid identities never leak truncated raw hex** into the UI
anywhere — they render a neutral "Unknown identity" label.
- **Unchanged on purpose:** display names and verified handles (NIP-05 —
the `name@domain` badge) still render as before. Unnamed avatars keep
distinct per-key initials, derived from the underlying hex key rather
than the npub — otherwise every unnamed key would render the same "N"
initial. Event IDs are not public keys, so they keep their hex
truncation (in Pulse's "Replying to", the parent author shows npub while
an event-id fallback still shows hex). The nevent share link, private
keys, and internal hex storage are untouched. Inputs that accept a key
(invite/member entry) accept both hex and npub and keep working in hex
internally.

### Related issue

N/A. Searched open issues/PRs for npub identity display on mobile —
closest related: none found. Desktop's parallel npub standardization
lives in the stacked desktop PRs (#7488 foundation, #7489 controls,
#7495 display surfaces); this is the independent mobile slice (based
directly on `main`, not on those branches).

### Testing

At exact head `5a620e420a1fd57d9d8011ac26434eed32fcf765` (base: `main`
`44316ff72`; 40 files, +1,345/−154):

- Full mobile suite: 2,098 tests passing (`cd mobile && flutter test`);
`flutter analyze` clean; `dart format --set-exit-if-changed .` clean —
the same checks CI runs.
- Widget/unit coverage at production seams: compact labels and hex-keyed
avatar initials for DM headers/tiles, member rows, mention suggestions,
and Pulse reply context; DM fallback avatars keyed to the labeled
counterpart (self-first order and self-DMs); blank/whitespace
display-name npub fallback with nonblank authored labels verbatim,
including the Activity inbox sender and profile-sheet heading (each with
its own empty/whitespace production-seam regression); full-npub copy and
disabled-copy semantics in profile and settings; invalid-key
suppression; and hex↔npub input round-trips.

Verified via unit and widget tests — no device/simulator validation is
claimed.

### Task provenance

Buzz channel: `1f0e4a3d-7e01-4efe-bb16-843b357f85c9`

Task:
buzz://message?channel=1f0e4a3d-7e01-4efe-bb16-843b357f85c9&id=86b34eb4bd84a1472419e9af22636c011c0fe273e3c196f967d7a36996e149b6

---------

Signed-off-by: Logan Johnson <loganj@squareup.com>
Co-authored-by: Larry <627498bd4bd1f281a16431e3c6cce3b5c25b6692798c78672298aefbf2f8f8b5@buzz.block.builderlab.xyz>
2026-09-09 14:08:24 -04:00
klopez4212
822c5ab231 Hide Huddles in mobile agent DMs (#6676)
## Summary
- hide the mobile Huddle action in one-to-one agent DMs
- preserve Huddles for human DMs and group DMs

## Testing
- just mobile-check
- flutter test (1,662 tests)

---------

Signed-off-by: kenny lopez <klopez4212@gmail.com>
Signed-off-by: Princess Donut <b238ea756dee4d98afa5883fc7f1de61eeabe65bf700e3a5a5a80db5e42e2c2b@buzz.block.builderlab.xyz>
Co-authored-by: Princess Donut <b238ea756dee4d98afa5883fc7f1de61eeabe65bf700e3a5a5a80db5e42e2c2b@buzz.block.builderlab.xyz>
2026-08-25 08:18:22 +01:00
klopez4212
06582ee6f0 Render mobile agent mention chips (#3702)
## Summary

- Render selected agent mentions as visible bot chips in the mobile
composer.
- Recognize agent profiles consistently when rendering message-body
mentions.
<img width="630" height="1368" alt="Screenshot 2026-07-30 at 07 54 16"
src="https://github.com/user-attachments/assets/035b46bf-ee78-4ee5-82fc-84591415ed7c"
/>

## Validation

- `flutter test test/features/channels/compose_bar_test.dart
test/features/channels/message_content_test.dart`
- `flutter analyze`

---------

Signed-off-by: kenny lopez <klopez4212@gmail.com>
2026-07-30 10:02:48 -07:00