refactor(llc, ui)!: route the read and delivery receipts through the generated client

[VelikovPetar]

Submit a pull request

Linear: FLU-923
Github Issue: -

CLA

• I have signed the Stream CLA (required).
• The code changes follow best practices
• Code changes are tested (add some information if not applicable)

Description of the pull request

Moves marking channels and threads read and unread, and the delivery receipts, onto the OpenAPI-generated client.

• StreamChatClient.markChannelRead, markThreadRead and markAllRead, and Channel.markRead and markThreadRead, return Result<MarkReadResponse> instead of throwing.
• markChannelUnread, markChannelUnreadByTimestamp and markThreadUnread, and Channel.markUnread, markUnreadByTimestamp and markThreadUnread, return Result<void>.
• markChannelsDelivered returns Result<MarkDeliveredResponse>. MarkChannelsDelivered, the callback ChannelDeliveryReporter takes, returns Future<Result<void>>; the reporter keeps the receipts on a failure, and still on a thrown error.
• The Channel guards return a failure instead of throwing: a missing read-events capability, and a markUnread on a locally counted channel whose message isn't loaded. An uninitialized channel still throws a StateError.
• On the local-unread path, Channel.markRead returns MarkReadResponse(duration: '0ms'), where it built an empty EmptyResponse.
• MessageDelivery moves to models/request/ as a freezed class: no toJson, value equality, copyWith.
• ChannelsRepository gains the four operations; channel_mapper.dart gains the envelope mappers and MessageDelivery.toRequest().
• The seven methods are removed from StreamChatApi.channel.
• MessageListUnreadController (internal) reads the Result to decide whether to retry a mark-read.

Generated	Public
DefaultApi.markRead (POST /api/v2/chat/channels/{type}/{id}/read)	StreamChatClient.markChannelRead, markThreadRead
DefaultApi.markUnread (POST /api/v2/chat/channels/{type}/{id}/unread)	StreamChatClient.markChannelUnread, markChannelUnreadByTimestamp, markThreadUnread
DefaultApi.markChannelsRead (POST /api/v2/chat/channels/read)	StreamChatClient.markAllRead
DefaultApi.markDelivered (POST /api/v2/chat/channels/delivered)	StreamChatClient.markChannelsDelivered
MarkReadResponse / MarkDeliveredResponse	same names (ours, freezed)
DurationResponse	void
DeliveredMessagePayload	MessageDelivery (kept)

Why

The calls take the API's envelopes rather than EmptyResponse, so a field the API adds later reaches callers without another break. The v2 routes are the same backend handlers as v1, and none is gated or in beta.

Notes for reviewers

• MarkReadResponse carries only duration. The API's event holds a thread whose parentMessage is a MessageResponse, which needs the message mappers; it's dropped with a TODO and can be added later without a break.
• The method names stay as in v10. Seven calls share four operations, so the spec's names can't tell most of them apart.
• MarkChannelsDelivered is retyped (a break beyond the usual Result one): with the old Future<void> signature a failed Result would have read as success and dropped receipts that were never sent.
• The generated requests send explicit nulls for the unused fields ({"message_id": null, "thread_id": "…"}); the server treats them as absent.

Testing

• stream_chat analyze clean, full suite green (1,912 tests); every package and the sample app analyze clean. generate_plan.py --check reports 0 problems.
• New client tests: client_mark_read_test.dart, client_mark_unread_test.dart, client_mark_all_read_test.dart and client_mark_channels_delivered_test.dart. The reporter tests cover a failed and a throwing callback.
• stream_chat_flutter's message_list_view tests pass; the rest of its suite fails only on CI goldens run on macOS, identically on the base branch.
• Live v1 vs v2 parity on the demo app, on throwaway channels deleted afterwards: all seven calls answer the same with the generated request bodies, and the real v2 responses (channel read, thread read, mark-all-read, delivered) decode.

Screenshots / Videos

Not applicable: no visible UI change.

🤖 Generated with Claude Code

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}