feat: add flag to redactEvent with edits

[krille-chan]

closes https://famedly.atlassian.net/browse/FM-732

[codecov]

Codecov Report

❌ Patch coverage is 12.50000% with 28 lines in your changes missing coverage. Please review.
✅ Project coverage is 59.21%. Comparing base (f9e6b68) to head (e261e44).

Files with missing lines	Patch %	Lines
...with_rel_types/msc_3912_redact_with_rel_types.dart	0.00%	14 Missing ⚠️
lib/src/room.dart	6.66%	14 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2389      +/-   ##
==========================================
- Coverage   59.29%   59.21%   -0.09%     
==========================================
  Files         161      162       +1     
  Lines       20300    20330      +30     
==========================================
+ Hits        12037    12038       +1     
- Misses       8263     8292      +29     

Files with missing lines	Coverage Δ
lib/src/event.dart	87.93% <100.00%> (+0.02%)	⬆️
...with_rel_types/msc_3912_redact_with_rel_types.dart	0.00% <0.00%> (ø)
lib/src/room.dart	75.29% <6.66%> (-0.98%)	⬇️

---

Continue to review full report in Codecov by Harness.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f9e6b68...e261e44. Read the comment docs.

[cursor]

PR Summary

Medium Risk
Opt-in redaction can issue many server requests and change timeline state when the MSC3912 fallback path runs; default behavior is unchanged.

Overview
Adds an opt-in redactAllEdits flag on Event.redactEvent / Room.redactEvent so redacting a message can also remove related m.replace edit events, not only the target event.

When redactAllEdits is true, the client calls /versions: if org.matrix.msc3912 is advertised, it uses a new redactEventWithRelTypes API helper (MSC3912 with_rel_types) in one request; otherwise it loads relating edits via getRelatingEventsWithRelType, redacts each edit (with MatrixException rate-limit retry), then redacts the original event. Default false preserves the previous single-redact behavior.

The MSC3912 extension is exported from matrix.dart for SDK consumers.

^{Reviewed by Cursor Bugbot for commit e261e44. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 2 potential issues.

Bugbot Autofix prepared fixes for both issues found in the latest run.

• ✅ Fixed: Edit fetch ignores pagination tokens
  • Manual edit redaction now paginates related m.replace events with from/nextBatch until all pages are processed.
• ✅ Fixed: Main redact lacks rate retry
  • The primary event redaction now uses the same retryAfterMs rate-limit retry path as edit redactions.

Or push these changes by commenting:

@cursor push ed4d881c74

Preview (ed4d881c74)

diff --git a/lib/src/room.dart b/lib/src/room.dart
--- a/lib/src/room.dart
+++ b/lib/src/room.dart
@@ -2491,24 +2491,47 @@
       );
     }
 
-    final edits = await client.getRelatingEventsWithRelType(
-      id,
-      eventId,
-      RelationshipTypes.edit,
-    );
-    for (final edit in edits.chunk) {
-      final txnid = client.generateUniqueTransactionId();
+    Future<String?> redactWithRateLimitRetry(
+      String redactionEventId,
+      String txid,
+    ) async {
       try {
-        await client.redactEvent(id, edit.eventId, txnid, reason: reason);
+        return await client.redactEvent(
+          id,
+          redactionEventId,
+          txid,
+          reason: reason,
+        );
       } on MatrixException catch (e) {
         final retryAfterMs = e.retryAfterMs;
         if (retryAfterMs == null) rethrow;
         await Future.delayed(Duration(milliseconds: retryAfterMs));
-        await client.redactEvent(id, edit.eventId, txnid, reason: reason);
+        return await client.redactEvent(
+          id,
+          redactionEventId,
+          txid,
+          reason: reason,
+        );
       }
     }
 
-    return await client.redactEvent(id, eventId, messageID, reason: reason);
+    String? nextBatch;
+    do {
+      final edits = await client.getRelatingEventsWithRelType(
+        id,
+        eventId,
+        RelationshipTypes.edit,
+        from: nextBatch,
+        limit: 50,
+      );
+      for (final edit in edits.chunk) {
+        final txnid = client.generateUniqueTransactionId();
+        await redactWithRateLimitRetry(edit.eventId, txnid);
+      }
+      nextBatch = edits.nextBatch;
+    } while (nextBatch != null);
+
+    return await redactWithRateLimitRetry(eventId, messageID);
   }
 
   /// This tells the server that the user is typing for the next N milliseconds

_{You can send follow-ups to the cloud agent here.}

^{Reviewed by Cursor Bugbot for commit e261e44. Configure here.}