feat: add IBigSegmentStore interface + Redis and DynamoDB stores by beekld · Pull Request #536 · launchdarkly/cpp-sdks

beekld · 2026-05-19T22:38:11Z

Ticket: SDK-2363 · Follows #534

Summary

Adds the public IBigSegmentStore interface, the Membership / StoreMetadata value types, and concrete DynamoDB + Redis implementations. Schema strings match what the Relay Proxy writes.

auto redis_store = RedisBigSegmentStore::Create("redis://localhost:6379", "prefix");
auto dynamo_store = DynamoDBBigSegmentStore::Create("my-table", "prefix", options);

Design notes

synchronizedOn parsing. Stored as DynamoDB N / Redis string. Both stores reject malformed values (non-numeric strings, wrong DynamoDB attribute type) rather than silently returning 0, matching the existing dynamodb_source.cpp row validation.
No hashing here. The interface contract says callers pass the already-hashed base64 SHA-256 context key; the SDK will hash in the wrapper, not the store implementations.

Not in scope

BigSegmentsBuilder config plumbing.
BigSegmentStoreWrapper (LRU cache + staleness polling) and the hashing path.
Evaluator wiring / replacing the rules.cpp big-segments TODO.

Test plan

7 Membership unit tests pass.
9 RedisBigSegmentStore integration tests pass against Redis 7 (docker).
8 DynamoDBBigSegmentStore integration tests pass against DynamoDB Local.
Full server-sdk test suite (468 tests) still green.
Schema constants match what Relay writes (verified against the LocalStack/Redis fixture data).

Note

Medium Risk
New evaluation-adjacent persistence API with strict error handling on malformed store data; behavior is well-tested but not yet connected to flag evaluation, so misconfiguration would surface only after later wiring.

Overview
Introduces a Big Segments read path for the server SDK: public IBigSegmentStore plus Membership / StoreMetadata, with Redis and DynamoDB stores that match Relay’s key/schema layout (prefix-scoped, read-only).

GetMembership loads included/excluded segment refs per hashed context; GetMetadata reads the sync timestamp for staleness. DynamoDB uses consistent GetItem, validates String Set / Number attribute types, and errors on malformed rows instead of silent empty data; Redis uses set membership and string sync time with the same strict parsing. Membership::FromSegmentRefs applies spec rules (inclusion over exclusion). Integration tests cover empty stores, prefixes, and bad data; config, caching wrapper, hashing, and evaluator wiring are not in this PR.

^{Reviewed by Cursor Bugbot for commit b0ab11c. Bugbot is set up for automated code reviews on this repo. Configure here.}

… fixture

cursor

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0082f6e. Configure here.}

kinyoklion · 2026-05-27T21:40:41Z

+        return std::unique_ptr<RedisBigSegmentStore>(new RedisBigSegmentStore(
+            std::make_unique<sw::redis::Redis>(std::move(uri)),
+            std::move(prefix)));
+    } catch (sw::redis::Error const& e) {


Should we catch std::exception as well?

kinyoklion · 2026-05-27T21:42:19Z

+   public:
+    DynamoDBBigSegmentTests()
+        : table_name_("ld-dynamodb-big-segments-test"),
+          prefix_("testprefix"),


Should we have a test for an empty prefix?

kinyoklion · 2026-05-27T21:57:22Z

+            std::make_unique<sw::redis::Redis>(std::move(uri)),
+            std::move(prefix)));
+    } catch (sw::redis::Error const& e) {
+        return tl::make_unexpected(e.what());


Do we need to sanitize this for the redis user/password?

Or maybe make this and the dynamo version into something vague? Main concern is just a URI coming out which includes user/password or something else the customer puts into the URI and considers secret.

I wanna push back on this one. I think making it vague would lose the diagnostic value of the original error messages. I would hope that redis already wouldn't be putting sensitive info into error messages. But also, we aren't logging this here, either. I think the "correct" solution would be to have a sanitizer wherever we are logging, but that's a much bigger task.

kinyoklion · 2026-05-27T22:02:28Z

+IBigSegmentStore::GetMembershipResult RedisBigSegmentStore::GetMembership(
+    std::string const& context_hash) const {
+    std::string const include_key =
+        prefix_ + ":" + kIncludeKeyNamespace + ":" + context_hash;


nit: Would it be work it to cache prefix_ + ":" + kIncludeKeyNamespace and prefix_ + ":" + kExcludeKeyNamespace then just add on the context_hash?

kinyoklion · 2026-05-27T22:08:02Z

+    std::string prefix)
+    : redis_(std::move(redis)),
+      prefix_(std::move(prefix)),
+      sync_time_key_(prefix_ + ":" + kSyncTimeKey) {}


Should we check for an empty prefix?

.Net and JS do, but Java doesn't.

public RedisStoreBuilder<T> Prefix(string prefix) { _prefix = string.IsNullOrEmpty(prefix) ? Redis.DefaultPrefix : prefix; return this; }

Done

Yes, I think Java is probably wrong here

kinyoklion · 2026-05-27T22:12:34Z

+
+    auto const it = item.find(kBigSegmentsSyncTimeAttribute);
+    if (it == item.end()) {
+        return tl::make_unexpected(


Other implementations seem more lenient here. Though I wonder if there was an evolution where maybe this field didn't exist originally.

The spec seems to say we should be more lenient too. Done.

…ptions

…member

…ed, not error

beekld added 3 commits May 19, 2026 15:21

feat: add IBigSegmentStore interface + Redis and DynamoDB stores

fc17043

refactor: drop trailing colon from redis key namespaces; tighten test…

2d92bc7

… fixture

fix: inline Membership methods so dynamodb/redis dylibs can link

a3dfc38

beekld marked this pull request as ready for review May 20, 2026 18:50

beekld requested a review from a team as a code owner May 20, 2026 18:50

cursor Bot reviewed May 20, 2026

View reviewed changes

Comment thread libs/server-sdk-dynamodb-source/src/dynamodb_big_segment_store.cpp

fix: reject non-STRING_SET included/excluded in dynamodb big-segments

79af82e

cursor Bot reviewed May 27, 2026

View reviewed changes

Comment thread libs/server-sdk-dynamodb-source/src/dynamodb_big_segment_store.cpp

refactor: redis big-segments returns empty Membership instead of nullopt

0082f6e

cursor Bot reviewed May 27, 2026

View reviewed changes

Comment thread libs/server-sdk-redis-source/src/redis_big_segment_store.cpp

refactor: drop std::optional from IBigSegmentStore::GetMembershipResult

b99bf0c

kinyoklion requested changes May 27, 2026

View reviewed changes

beekld added 4 commits May 27, 2026 16:31

fix: redis big-segments handle empty prefix and catch unexpected exce…

44af937

…ptions

test: dynamodb big-segments empty-prefix smoke tests

44c8a54

refactor: cache redis big-segments key prefixes; drop unused prefix_ …

8453937

…member

fix: dynamodb big-segments treats absent synchronizedOn as never-sync…

b0ab11c

…ed, not error

beekld requested a review from kinyoklion May 27, 2026 23:53

Conversation

beekld commented May 19, 2026 • edited by cursor Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Design notes

Not in scope

Test plan

Uh oh!

Uh oh!

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

beekld commented May 19, 2026 •

edited by cursor Bot

Loading