SDK-300 Add OnSuccess/OnFailure handlers to logoutUser by sumeruchat · Pull Request #1069 · Iterable/iterable-swift-sdk

sumeruchat · 2026-05-27T12:30:23Z

What

Adds a public IterableAPI.logoutUser(withOnSuccess:onFailure:) overload so SDK consumers can observe the logout sequence. The handlers signal that the local logout sequence completed and surface the triggered disableDeviceForCurrentUser result — they are an observability primitive, not a retry mechanism (the disable already auto-retries offline via SDK-297). Ticket: SDK-300.

Changes

Public API (IterableAPI.swift): New logoutUser(withOnSuccess:onFailure:) overload. Parameterless logoutUser() now delegates with nil/nil handlers, preserving Obj-C and source compatibility. Mirrors disableDeviceForCurrentUser's two-overload precedent.
Internal (InternalIterableAPI.swift): New logoutUser(withOnSuccess:onFailure:). Not-initialized → onFailure. autoPushRegistration on → handlers piped through the existing disableDeviceForCurrentUser(withOnSuccess:onFailure:) before identity cleanup (so the disable request goes out with the still-set userId/email), then local cleanup runs. autoPushRegistration off → local cleanup, then onSuccess(nil) synchronously.
logoutPreviousUser() untouched — internal user-switch paths keep their existing fire-and-forget contract.
Tests (AuthTests.swift): 5 new tests covering autoPush-on success, autoPush-on failure (verifies local cleanup STILL happens despite network error), autoPush-off synchronous success with no network, not-initialized failure, and the parameterless overload regression.

Impact

Breaking changes: None. The existing parameterless logoutUser() is unchanged; only adds a new overload.
Dependencies: None.
Performance: No-op — handlers are only invoked at the existing completion points.

Testing

How to test:

Run ./agent_test.sh AuthTests — all 42 tests pass (5 new).
Call IterableAPI.logoutUser(withOnSuccess:onFailure:) from a sample app with autoPushRegistration = true and a logged-in user; verify onSuccess fires after the disable network call returns.
With autoPushRegistration = false, verify onSuccess fires synchronously and no network request is made.
Without initializing the SDK (or without a logged-in user), verify onFailure fires with Iterable SDK is not initialized.
Confirm the existing parameterless logoutUser() still logs out without regression.

Adds a public `IterableAPI.logoutUser(withOnSuccess:onFailure:)` overload for observability around the logout sequence. logoutUser/logoutPreviousUser are local-only operations -- the only network side-effect is the disableDeviceForCurrentUser triggered when autoPushRegistration is on, and that already auto-retries offline via SDK-297. The handlers signal that the local sequence completed and surface the triggered disableDevice's result; they are not a retry mechanism. - Public: new `logoutUser(withOnSuccess:onFailure:)` on IterableAPI; the existing parameterless `logoutUser()` delegates to it with nil/nil handlers, preserving Obj-C and source compatibility. Mirrors the precedent set by disableDeviceForCurrentUser's two-overload shape. - Internal: new InternalIterableAPI.logoutUser(withOnSuccess:onFailure:). Not-initialized -> onFailure. autoPushRegistration on -> handlers piped through the existing disableDeviceForCurrentUser(withOnSuccess:onFailure:) *before* clearing identity (so the disable request goes out with the still-set userId/email), then local cleanup runs. autoPushRegistration off -> local cleanup, then onSuccess(nil) synchronously. - logoutPreviousUser() left untouched -- still used on internal user-switch paths with its existing fire-and-forget contract. - Tests in AuthTests cover: autoPush-on success, autoPush-on failure (local cleanup STILL happens despite network error), autoPush-off synchronous success with no network, not-initialized failure, and the parameterless overload regression. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

codecov · 2026-05-27T12:39:45Z

Codecov Report

❌ Patch coverage is 94.11765% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.31%. Comparing base (481f525) to head (dc2302a).

Files with missing lines	Patch %	Lines
swift-sdk/SDK/IterableAPI.swift	83.33%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1069      +/-   ##
==========================================
+ Coverage   71.27%   71.31%   +0.03%     
==========================================
  Files         112      112              
  Lines        9369     9389      +20     
==========================================
+ Hits         6678     6696      +18     
- Misses       2691     2693       +2

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

franco-zalamena-iterable · 2026-05-27T16:42:15Z

    }

    func logoutUser() {
-        logoutPreviousUser()


Before this PR, logoutUser() delegated to logoutPreviousUser() (line 251 on master), keeping a single source of truth for the eight-step logout cleanup sequence. After this PR, that body lives in two places:

logoutPreviousUser() (line 877), used by setEmail/setUserId user-switch paths.

logoutUser(withOnSuccess:onFailure:) (line 254), used by both public entry points.

@franco-zalamena-iterable Good catch and appreciate the attention to detail. This is fixed in c0193b7. logoutPreviousUser() now delegates to logoutUser(withOnSuccess:onFailure:) with nil handlers instead of carrying its own copy of the eight-step cleanup, so the sequence has a single source of truth again.

Behavior is unchanged for the setEmail/setUserId user-switch paths: a nil onFailure keeps the not-initialized guard a silent no-op, and a nil onSuccess makes the auto-push-off completion a no-op — exactly matching the old logoutPreviousUser() body.

Verified: build green, AuthTests 42/0, and UserMergeScenariosTests 21/0 (which exercises the user-switch paths through logoutPreviousUser).

Per Franco's review, collapse the duplicated logout cleanup sequence. logoutPreviousUser() now delegates to logoutUser(withOnSuccess:onFailure:) with nil handlers instead of carrying its own copy of the eight-step cleanup. The full sequence lives in exactly one place. Behavior is unchanged for the setEmail/setUserId user-switch paths: a nil onFailure keeps the not-initialized guard a silent no-op, and a nil onSuccess makes the auto-push-off completion a no-op. Verified: ./agent_build.sh green; AuthTests 42/0; UserMergeScenariosTests 21/0 (exercises the user-switch paths through logoutPreviousUser). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

franco-zalamena-iterable

Just one thing to address but non blocker anymore

Per Franco's review follow-up: now that logoutPreviousUser() delegates to logoutUser(withOnSuccess:onFailure:), the parameterless logoutUser() can keep calling logoutPreviousUser() as it did on master instead of calling the handler overload directly. Minimizes the diff and preserves the original delegation; behavior is unchanged (logoutUser() -> logoutPreviousUser() -> logoutUser(withOnSuccess:nil,onFailure:nil)). Build green; AuthTests 42/0. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Following a detailed handler-correctness review (no functional bug found), clarify the observability contract and close the one test gap: - Doc comment on the public logoutUser(withOnSuccess:onFailure:) spelling out that logout is local-only and the handlers reflect the triggered disableDevice result when autoPushRegistration is on. Notably, an onFailure (e.g. "no token present", or a failed disable request) does NOT mean the local logout failed. Also documents that handlers are in-process only (not persisted across app termination) and not guaranteed to be delivered on the main thread. - Test: autoPush ON with no registered token now asserts onFailure fires with "no token present", that local cleanup STILL completes, and that no disableDevice request is sent. Build green; AuthTests 43/0. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…success-failure-handlers * origin/master: SDK-493: Prepare for Release 6.7.2 (#1070)

joaodordio · 2026-05-29T15:59:45Z

Should this be updated with the same level of error as we have on logoutUser?

Good catch, addressed in 7708de1. Both disableDeviceForCurrentUser and disableDeviceForAllUsers completion overloads now fire onFailure("Iterable SDK is not initialized", nil) on the not-initialized guard, matching logoutUser. The parameterless overloads pass nil handlers, so they are unchanged. Added a test, AuthTests 44/0.

Per Joao's review, make the disableDeviceForCurrentUser and disableDeviceForAllUsers completion overloads call onFailure("Iterable SDK is not initialized", nil) on the not-initialized guard, matching logoutUser. The parameterless overloads pass nil handlers, so they are unchanged. AuthTests 44/0. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…success-failure-handlers * origin/master: SDK-469 Fix CodeQL string length conflation in deep link check (#1071)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…r-token-offline-retry * origin/master: SDK-300 Add OnSuccess/OnFailure handlers to logoutUser (#1069) SDK-469 Fix CodeQL string length conflation in deep link check (#1071) SDK-493: Prepare for Release 6.7.2 (#1070)

…ush-async-await * origin/master: SDK-108 Route registerToken through OfflineRequestProcessor for network-failure retry (#1068) SDK-300 Add OnSuccess/OnFailure handlers to logoutUser (#1069) SDK-469 Fix CodeQL string length conflation in deep link check (#1071)

sumeruchat self-assigned this May 27, 2026

sumeruchat requested review from franco-zalamena-iterable and joaodordio May 27, 2026 12:30

franco-zalamena-iterable requested changes May 27, 2026

View reviewed changes

franco-zalamena-iterable approved these changes May 29, 2026

View reviewed changes

Comment thread swift-sdk/Internal/InternalIterableAPI.swift

sumeruchat and others added 3 commits May 29, 2026 10:17

Merge remote-tracking branch 'origin/master' into fix/SDK-300-logout-…

ad74d48

…success-failure-handlers * origin/master: SDK-493: Prepare for Release 6.7.2 (#1070)

joaodordio approved these changes Jun 1, 2026

View reviewed changes

sumeruchat and others added 3 commits June 2, 2026 11:44

Merge remote-tracking branch 'origin/master' into fix/SDK-300-logout-…

f6e9356

…success-failure-handlers * origin/master: SDK-469 Fix CodeQL string length conflation in deep link check (#1071)

SDK-300 Drop em dash from logoutPreviousUser comment

dc2302a

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

sumeruchat merged commit 62a33d3 into master Jun 2, 2026
15 of 16 checks passed

sumeruchat deleted the fix/SDK-300-logout-success-failure-handlers branch June 2, 2026 13:35

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

SDK-300 Add OnSuccess/OnFailure handlers to logoutUser#1069

SDK-300 Add OnSuccess/OnFailure handlers to logoutUser#1069
sumeruchat merged 8 commits into
masterfrom
fix/SDK-300-logout-success-failure-handlers

sumeruchat commented May 27, 2026 •

edited by atlassian Bot

Loading

Uh oh!

codecov Bot commented May 27, 2026 •

edited

Loading

Uh oh!

franco-zalamena-iterable May 27, 2026

Uh oh!

sumeruchat May 29, 2026 •

edited

Loading

Uh oh!

franco-zalamena-iterable left a comment

Uh oh!

Uh oh!

joaodordio May 29, 2026

Uh oh!

sumeruchat Jun 2, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

sumeruchat commented May 27, 2026 • edited by atlassian Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Changes

Impact

Testing

Uh oh!

codecov Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

franco-zalamena-iterable May 27, 2026

Choose a reason for hiding this comment

Uh oh!

sumeruchat May 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

franco-zalamena-iterable left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

joaodordio May 29, 2026

Choose a reason for hiding this comment

Uh oh!

sumeruchat Jun 2, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

sumeruchat commented May 27, 2026 •

edited by atlassian Bot

Loading

codecov Bot commented May 27, 2026 •

edited

Loading

sumeruchat May 29, 2026 •

edited

Loading