feat!: Support getExperimentFlag, trackEvent and trackExposure methods - #214

Important

Review skipped

This PR was authored by a bot without an assigned CodeRabbit review seat. Ask an organization administrator to assign a seat to the bot. See the seat-assignment guide. After the seat is assigned, request a review again.

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: 97ac6ca3-111d-46df-a7b4-17f8d3acf48c

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

• 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.

Merging this PR will not alter performance

✅ 15 untouched benchmarks

_{Comparing openhands/sdk-spec-v0.7.1 (7086f72) with main (745945c)}

The spec run failed because v0.7.1 adds the events capability to this SDK's scenarios (core or events or local-evaluation), and the steps file had no Create the SDK client with events enabled step. The SDK had no event tracking on main, so this re-applies the earlier event work.

Changes

• Re-applied 8e598d8 (feat: Track events and experiment exposures) as 54834fa. It adds EnableEvents, GetExperimentFlag, TrackEvent, TrackExposureEvent and FlushEvents, with unit tests in EventsTests.cs. It also wires the spec's events steps in sdk-spec/steps.cs.
• Re-applied 3b73d6a (fix: Wait for in-flight event batches on flush and drop unserialisable ones) as 40dfb66.
• Resolved two conflicts with main:
  • SdkVersion.cs: kept the 9.1.0 release version and made it internal, because events report it as sdk_version.
  • steps.cs: kept both the local-evaluation client step and the events client step.
• The events features have not changed since v0.6.0, the version the earlier work targeted, so the code needed no other changes.
• v0.7.1's own change adds empty-string flag values (""). They already pass through unchanged in remote and local evaluation, so the SDK needed no change for them.

Verified locally: build and dotnet format --verify-no-changes pass, Flagsmith.Client.Test passes (72 passed, 1 skipped), and the spec at v0.7.1 passes all 9 scenarios with core or events or local-evaluation.

The last red run on 40dfb66 was not caused by the code. In Test (Flagsmith.Client.Test, 10.0), dotnet restore could not download Newtonsoft.Json from nuget.org ("Connection reset by peer"), so no tests ran. Test (Flagsmith.Client.Test, 9.0) passed all its tests (72 passed, 1 skipped), and the matrix then cancelled it because the other job had failed. I could not re-run the failed jobs: the token returns "Resource not accessible by integration".

Changes

• 7086f72 adds unit tests for v0.7.1's one change: a flag whose value is an empty string is returned as "", in remote and local evaluation. The SDK already behaved this way, and the spec run confirms it, but no unit test covered it. The push also runs CI again.

Verified locally: dotnet format --verify-no-changes passes, and Flagsmith.Client.Test passes (74 passed, 1 skipped).