Use safe borrow helpers for NostrDB lookups in networking and cache [untested]

[alltheseas]

Summary

• - 851e855 – Switch networking (findEvent, relay list) and EventCache to the
new borrow helpers to avoid holding raw NdbTxn values.

• ae9b648 – Add withNote and withProfile synchronous borrow helpers to
  Ndb.swift, copying to owned values before returning to callers.
• 34b1ea8 – Merge upstream-master into pr3336.
• 52115d0 – Fix profile crash.
• d651084 – Add AGENTS.md.
• 161fd96 – Add regression test for async waitFor.
• f628a34 – Return owned notes from Ndb async waiters.
• 37c4349 – Fix profile crash.

Checklist

Experimental Feature Checklist

Tip

This Pull Request is an experimental feature for Damus Labs, and follows a fast-track review process.
The overall requirements are lowered and the review process is not as strict as usual. However, the feature will only be available for Purple users who opt-in.

• I have read (or I am familiar with) the Contribution Guidelines.
• I have done some testing on the changes in this PR to ensure it is at least functional.
• I made sure that this new feature is only available when the user opts-in from the Damus Labs screen, and does not affect the rest of the app when turned off.
• My PR is either small, or I have split it into smaller logical commits that are easier to review.
• I have added the signoff line to all my commits. See Signing off your work.
• I have added an appropriate changelog entry to my commit in this PR. See Adding changelog entries.
  • Example changelog entry: Changelog-Added: Added experimental feature <X> to Damus Labs

Standard PR Checklist

• I have read (or I am familiar with) the Contribution Guidelines
• I have tested the changes in this PR
• I have profiled the changes to ensure there are no performance regressions, or I do not need to profile the changes.
  • Utilize Xcode profiler to measure performance impact of code changes. See https://developer.apple.com/videos/play/wwdc2025/306
  • If not needed, provide reason:
• I have opened or referred to an existing github issue related to this change.
• My PR is either small, or I have split it into smaller logical commits that are easier to review
• I have added the signoff line to all my commits. See Signing off your work
• I have added appropriate changelog entries for the changes in this PR. See Adding changelog entries
  • I do not need to add a changelog entry. Reason: [Please provide a reason]
• I have added appropriate Closes: or Fixes: tags in the commit messages wherever applicable, or made sure those are not needed. See Submitting patches

Test report

Please provide a test report for the changes in this PR. You can use the template below, but feel free to modify it as needed.

Device: [Please specify the device you used for testing]

iOS: [Please specify the iOS version you used for testing]

Damus: [Please specify the Damus version or commit hash you used for testing]

Setup: [Please provide a brief description of the setup you used for testing, if applicable]

Steps: [Please provide a list of steps you took to test the changes in this PR]

Results:

• PASS
• Partial PASS
  • Details: [Please provide details of the partial pass]

Other notes

[Please provide any other information that you think is relevant to this PR.]

[alltheseas]

Changes in this PR do NOT fix #3343

- damus/Core/Networking/NostrNetworkManager/ProfilesManager.swift:84-104 still uses streamIndefinitely to yield NdbNoteLenders and then hands raw NdbTxn<ProfileRecord?
    > objects back to UI consumers via ProfileStreamItem (damus/Core/Networking/NostrNetworkManager/ProfilesManager.swift:143-148). ProfileObserver continues to hold those
    transactions alive on the main actor (damus/Core/Nostr/ProfileObserver.swift:22-34). None of the lifetime/ownership guardrails that would prevent the observed mdb_page_get/
    mdb_txn_commit crashes were added, so the two crash stacks coming out of ProfilesManager.listenToProfileChanges and ProfileObserver.watchProfileChanges remain unchanged.
  - The preloading pipeline that is crashing in preload_event (Thread 12 with mdb_cursor_open in the 2025-11-25_23-11-17 log) is untouched. preload_event and the async Task
    spawned by preload_events still render note content in the background before plan.data is mutated (damus/Shared/Utilities/EventCache.swift:370-420 and damus/Shared/
    Utilities/EventCache.swift:444-460). The only change in this file is swapping .lookup_note(...).to_owned() for withOwnedNote at line 225, which is just a wrapper around the
    same copy and doesn’t impact the code paths that hit LMDB cursors.
  - The crashes where Ndb.close() traps during background termination (2025-11-26_10-13-28 and _15-12-26) also remain. Ndb.close is still a straight call to ndb_destroy with
    no coordination with outstanding query tasks or the TaskManager in SubscriptionManager (nostrdb/Ndb.swift:210-217). The branch doesn’t introduce any ordering between
    ContentView tearing down and long-running nostrdb work, so the assertion failure seen in the crash logs is still possible.
  - The new “owned” helpers and the change to waitFor don’t alter runtime behaviour in the app. withOwnedNote/withOwnedProfile simply move the existing .to_owned() call into Ndb
    and are only invoked in places that were already copying (EventCache lookup, findEvent, relay list manager). waitFor now returns NostrEvent?, but rg shows no production code
    ever calls this API (only the new unit test), so it cannot affect any of the recorded crashes.

  Given that none of the code paths implicated in the crash logs (ProfilesManager/ProfileObserver, preload_event/ContentRenderer, Kingfisher rendering, or the Ndb.close
  lifecycle) changed in this branch, none of the recorded crashes are addressed by this PR.

[alltheseas]

This PR is too big, and does not seem to fix outstanding crashes. I will instead tackle recent crashes one by one in separate PRs.