Skip to content

Testing quality e2e mutation offline loadtest - #349

Open
celina005 wants to merge 4 commits into
ethos-protocol:mainfrom
celina005:testing-quality-e2e-mutation-offline-loadtest
Open

Testing quality e2e mutation offline loadtest#349
celina005 wants to merge 4 commits into
ethos-protocol:mainfrom
celina005:testing-quality-e2e-mutation-offline-loadtest

Conversation

@celina005

@celina005 celina005 commented Aug 29, 2026

Copy link
Copy Markdown

Summary

Changes

  1. f416ba8 — Cross-platform beneficiary acceptance-flow E2E (Issue 1): new jobs in e2e-cross-platform.yml seed a shared vault+invitation fixture, then both iOS and Android open the invitation link via a cold app launch (the exact path Formalize the Beneficiary-Acceptance Token Requirement (or Deliberately Drop It) #109/Add Manage Beneficiary Screen to Android #195 broke) and verify acceptance server-side.
  2. 6231192 — Mutation-testing gap-fill for Stellar validation (Issue 2): added PIT/Mull configs, a shared/MUTATION_TESTING.md baseline doc, and 3 new fixtures/tests per platform closing gaps in the version-byte and mid-payload-checksum reject conditions.
  3. e5eeca4 — Offline→online check-in sync E2E (Issue 3): new Android/iOS jobs toggle real device network off/on around a real check-in queue drain, asserting final server state; includes a note for extending to deposit/withdraw once Add Manual "Retry Now" Control for Queued Offline Actions #240 ships.
  4. 361e907 — WebSocket reconnect-storm load test (Issue 4): a k6 script + methodology doc simulating a server-restart reconnect storm with a jitter on/off toggle to compare against Support Subscribing to Multiple Vault IDs Over a Single WebSocket Connection #253's backoff jitter, plus a results table to fill in.

Testing

Parity checklist

This project maintains a feature-parity table in PARITY.md that
tracks which features are implemented on iOS vs Android. Please answer the
questions below before requesting review.

  • This PR does not add, change, or remove any user-facing feature on either
    platform — no PARITY.md update needed.

    — OR —

  • This PR adds/changes/removes a user-facing feature. I have updated PARITY.md:

    • Updated the status symbol(s) for the affected row(s).
    • Added or updated "Notes" if the implementation is partial or has caveats.
    • Removed or updated any rows in the "Known gaps" table that this PR closes.

Related issues

Closes #288
Closes #289
Closes #290
Closes #291

@celina005
test(e2e): add cross-platform beneficiary acceptance flow to catch et…
f416ba8 
…hos-protocol#109/ethos-protocol#195

Add a full vault-creation -> invitation-link -> beneficiary-opens-link ->
acceptance e2e job to e2e-cross-platform.yml, run against a shared staging
fixture (one seeded vault/invitation reused by both platform jobs).
Both iOS and Android open the invitation deep link on a cold app launch
(simctl openurl / am start VIEW after a fresh install) so the token travels
through the same onCreate()/getIntent() cold-start path where ethos-protocol#109/ethos-protocol#195
forwarded the wrong beneficiary token, then independently verifies
server-side acceptance state rather than trusting client-side success.
@celina005
test(stellar): close mutation-testing gaps in address validator suites
6231192 
Add PIT (Android) and Mull (iOS) mutation-testing configs scoped to
StellarAddress, and document the process/baseline in
shared/MUTATION_TESTING.md. A manual walkthrough of the six reject
conditions in stellar-validation-spec.md against the existing fixtures
found two conditions with no dedicated test (version-byte mismatch, and a
checksum corruption outside the last character) plus a character-set
boundary case only tested mid-string. Add matching fixtures/tests to both
StellarAddressTest.kt and EthosProtocolTests.swift, generated directly from
the spec's algorithm, and record the new fixtures in the shared spec doc.
@celina005
test(e2e): add offline-to-online check-in sync scenario to e2e-cross-…
e5eeca4 
…platform.yml

PendingActionSyncWorker/CheckInSyncTask are covered at the unit level, but
unit tests mock away real WorkManager/BGProcessingTask scheduling. Add
Android and iOS e2e jobs that disable network on a real emulator/simulator,
queue a check-in offline, re-enable network, and assert the queue drains
via the real scheduler against the real staging API - then verify
server-side state directly rather than trusting the client's local queue
being empty. Deposit/withdraw queuing is left as a follow-up (tracked
behind ethos-protocol#240) with a note pointing at the same job to extend.
@celina005
test(load): add WebSocket reconnect-storm load test against staging
361e907 
Add a k6 script simulating N simultaneously-dropped clients reconnecting
via VaultEventSocket's exponential-backoff-with-full-jitter schedule
(ReconnectBackoff.delayForAttempt, ethos-protocol#253), with a JITTER_DISABLED flag to
reproduce pre-ethos-protocol#253 behavior for an ablation comparison against the same
staging endpoint. Document methodology, metrics to watch, and a findings
table to fill in once both runs are executed against staging.
@drips-wave

drips-wave Bot commented Aug 29, 2026

Copy link
Copy Markdown

@celina005 Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

1 participant

@celina005