[MAJOR][feature]: Release/1.26.0 - #1908
Open
fidelianawar wants to merge 46 commits into
Open
Conversation
Restrict the refresh token credential (PRT) header so it is only added to the PkeyAuth challenge response when the submit URL host is a known/trusted AAD host, preventing the PRT from being leaked to untrusted hosts. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x] PR is independently mergeable (no hidden dependencies) - [x] Appropriate reviewers are assigned - [x] PR reviewed by code owner (required if Copilot-generated) - [x] SME or Senior IC assigned where required ## Proposed changes Two test-infrastructure updates so the IdentityCore-driven UI / lab tests keep working against current iOS sims and the LAB v2 service: **UI automation hardening** (`IdentityCore/tests/automation/ui_tests_lib/MSIDBaseUITest.m`) — addresses intermittent UI-test failures on iOS 18+ sims: - `aadEnterPassword:` / `tapPasswordSelectionButtonIfPresentInApp:` — scope the "Use your password" / "Other ways to sign in" lookups to `application.webViews` so iOS QuickType / Passwords AutoFill suggestions with the same label don't win the first-match query (which would otherwise open the empty system password picker on CI sims and loop the test). - New `dismissKeyboardIfVerifyEmailPagePresentInApp:` — on the MSA "Verify your email" interstitial the auto-focused email field raises the keyboard, which absorbs taps on the "Use your password" link even though the link reports as hittable. The helper taps the page header to defocus the email field and dismiss the keyboard so the link becomes truly tappable on the next polling tick. - `enterPassword:app:isMainApp:` — require `.isHittable` (not just `.exists`) on the secure text field so back-to-back `acquireToken` calls (e.g. `prompt=force` after a prior sign-in) don't type into a stale field left in the view hierarchy. **Lab API endpoint migration** (`IdentityCore/tests/automation/ui_tests_lib/lab_api/MSIDAutomation*APIRequest.m`) — point at the new ID4SLAB2 operations: `CreateTempUser`, `DeleteDevice`, `EnablePolicy`/`DisablePolicy`, and `Reset` are all suffixed `ID4SLAB2`. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [x] Engineering change - [x] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information Scope is test-only — no production source is touched. The four `MSIDAutomation*APIRequest` files only override `-requestOperationPath`, which is consumed exclusively by the lab automation client; `MSIDBaseUITest.m` lives under `IdentityCore/tests/automation/ui_tests_lib/` and is only linked by UI-test targets. No changelog entry is needed (test infrastructure only). --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [x] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## Proposed changes Merge main into dev for release 1.25.0 ## Type of change - [x] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information
…t-known-aad-hosts # Conflicts: # changelog.txt
Add MSIDPkeyAuthTag execution-flow tags for the added/skipped PRT-header decision in MSIDPKeyAuthHandler, recorded only when context.correlationId is available. Addresses review feedback on PR #1869. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…AAD hosts (#1869) Restrict the refresh token credential (PRT) header so it is only added to the PkeyAuth challenge response when the submit URL host is a known/trusted AAD host, preventing the PRT from being leaked to untrusted hosts. ## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x ] PR is independently mergeable (no hidden dependencies) - [x ] Appropriate reviewers are assigned - [x ] PR reviewed by code owner (required if Copilot-generated) - [x ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [x] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information
…1868) ## Problem `isNewMobileOnboardingFlow` is a `BOOL` (value type) on `MSIDInteractiveRequestParameters`. When the webview sets it to `YES` on encountering `msauth://enroll`, that mutation only affects the current params instance. Broker actions create **new** params objects from `ADBrokerRequest` — the flag is lost because BOOLs are copied, not shared. This is the same class of problem that `MSIDOnboardingBlobBuilder` solved for the onboarding blob — reference types are shared by pointer, value types are copied. ## Solution Introduce `MSIDMobileOnboardingState` — a minimal reference-type wrapper (same pattern as `MSIDOnboardingBlobBuilder`): ```objc @interface MSIDMobileOnboardingState : NSObject @Property (nonatomic) BOOL isNewMobileOnboardingFlow; @EnD ``` ### How it works 1. **`ADBrokerRequest`** creates one `MSIDMobileOnboardingState` per auth session, seeded from the IPC payload 2. **Every `MSIDInteractiveRequestParameters`** gets the same pointer: `params.mobileOnboardingState = request.mobileOnboardingState` 3. **Webview** sets `params.isNewMobileOnboardingFlow = YES` → computed setter writes to the shared state object 4. **All consumers** (serializers, follow-up actions) see the mutation immediately — no sync-back needed 5. **SSO extension hop** — the flag is serialized to JSON by `MSIDBrokerOperationInteractiveTokenRequest`, deserialized on the other side, and seeds a new `MSIDMobileOnboardingState` ### Why not just a BOOL? | Object | Persists across actions? | Webview can write? | |--------|------------------------|--------------------| | `ADBrokerRequest.isNewMobileOnboardingFlow` (BOOL) | ✅ | ❌ (webview has no access) | | `MSIDInteractiveRequestParameters.isNewMobileOnboardingFlow` (BOOL) | ❌ (recreated per action) | ✅ | | `MSIDMobileOnboardingState` (reference) | ✅ (shared pointer) | ✅ (via params) | No other BOOL on params has this requirement — all others (`forceUI`, `instanceAware`, `showHeadsUp`) are set once from IPC and never mutated mid-session. ## Changes - **New**: `MSIDMobileOnboardingState.h/.m` — shared mutable state holder - **Modified**: `MSIDInteractiveRequestParameters.h/.m` — added `mobileOnboardingState` property, `isNewMobileOnboardingFlow` is now a computed accessor forwarding to the state object Companion Broker4 PR: (pending) --------- Co-authored-by: Swasti Gupta <swagup@microsoft.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [x] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [X] All tests pass locally - [X] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## Proposed changes The proposed change updates MSIDFlightManager so assigning flightProvider completes immediately before the setter returns. Before the change, setFlightProvider: used dispatch_barrier_async . That means the assignment was placed onto the synchronization queue, but the caller did not wait for it to run. Code could set flightProvider and then immediately ask for a flight value, while the provider assignment was still waiting in the queue. In that case, MSIDFlightManager could still see flightProvider as nil and return the default value, usually NO . After the change, setFlightProvider: uses dispatch_barrier_sync . The assignment still goes through the same barrier queue, so the thread-safety model stays the same. The difference is that the caller now waits until the assignment is finished. Once flightProvider = provider returns, the provider is definitely stored and available to later reads. This fixes the CI flake because the failing tests inject a mock flight provider and immediately call code that reads from it. The old async setter allowed the read to happen too early. ## Type of change - [ ] Feature work - [X] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [X] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: Swasti Gupta <swagup@microsoft.com>
…native GetToken (POC) (#1872) ## Proposed changes Adds `MSIDBoundTokenProvider`, a Common Core seam that services a browser-native-message `GetToken` request **in-process** for a host such as OneAuth (embedded in Edge). On unmanaged iOS the platform SSO Extension is unavailable, so the host cannot silently invoke the broker through `ASAuthorizationSingleSignOnProvider`. Instead the host hands the typed `MSIDBrowserNativeMessageGetTokenRequest` to this provider, which owns the orchestration that would otherwise live behind the SSO Extension (silent BART SPA redemption, or an interactive broker flip). This PR is a **POC** that wires up the seam and proves the routing end-to-end; the real silent-redemption / interactive-broker-flip orchestration is layered on top of this provider in follow-up work. I have code changes in OneAuth that prove functionality. This provider must be created to support OneAuth when SSO EXT is not an option. This is for iOS only not Mac OS. In Edge native app when Browser sends BrowserNativeMessagingRequest OneAuth will check if SSO Ext is available. if SSO EXT is absent then OneAuth will use the path of MSIDBoundTokenProvider via Common Core to pass the request in. This class will later hold real validation logic and routing for the response to be obtained either through silent or interactive flow. A Similar class exist in Broker (thats how this works with SSO ext), a later PR will be created to see how the broker code can be broken down since this class is now required in Common core. I have discussed this with Sergei before he went OOF. ### What's included - `IdentityCore/src/oauth2/token/MSIDBoundTokenProvider.{h,m}` — the provider. Validates the request (nil → `MSIDErrorInvalidInternalParameter`; blank `clientId`/`redirectUri` → `MSIDErrorInvalidDeveloperParameter`) and returns a serialized browser-native-message response payload (`transport: in_proc_common_core`). - `IdentityCore/tests/MSIDBoundTokenProviderTests.m` — unit tests built from the real `MSIDBrowserNativeMessageGetTokenRequest` properties, covering the in-process success path and the missing-`clientId` validation path. ## Type of change - [x] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. Additive, self-contained POC seam with no existing call sites; consumed only by the parallel OneAuth POC. ## Additional information Paired with the OneAuth-side POC that routes `BrowserNativeMessagingGetTokenRequest` to this provider when the SSO Extension is unavailable. Branch: `maagubuzo/un_tb/poc/msidboundtokenprovider`, targeting `dev`. --------- Co-authored-by: Swasti Gupta <swagup@microsoft.com> Co-authored-by: Swasti Gupta <swastinitb@gmail.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com>
…job template (#1882) Convert pr-validation, msal_submodule_check and broker_submodule_check to use the central Pipeline YAMLs/shared/aces-macos-job.yml template for pool + Xcode + tool setup. Split broker steps into broker_build_steps.yml (repo-specific) so the hosted visionOS consumer keeps its own tool setup via broker_submodule_steps.yml. ## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…imulators (#1885) ## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) ## Proposed changes `ADBTokenBindingUITests` failures downstream (e.g. `testADRSDeviceRegistration_havingCAWithTokenBinding_andHavingSSOExtSecStorageDisabled_validateDeviceJoinIsECC`) were traced to `performAction:config:application:` tapping the automation host app's action buttons (e.g. "Acquire Token") without the touch ever actually reaching the button's real UIKit target-action handler. Confirmed via the simulator's unified log (`xcrun simctl ... log show`) — **not** just the XCTest driver's own captured log, which never surfaces output from the app under test: Between requesting a tap and XCTest synthesizing it, XCTest runs its own "make frontmost" dance (`Check for interrupting elements affecting <button>`, looking for a system alert/permission prompt from another app, followed by re-`Open`/`Activate`-ing our target app). That dance can happen *during* the tap call itself, so a coordinate resolved beforehand can go stale by the time the touch is actually delivered, and the touch is silently dropped. The button's target-action never fires, and the test hangs until a later, unrelated step times out ("Timed out waiting for the password field..." or "Wait for result pipeline"). Separately, XCTest's "scroll element to visible" step (run unconditionally as part of its tap synthesis) was observed corrupting the hit-point computation to `{-1, -1}` on iOS/iPadOS 26 simulators, even when the button never needed to scroll and had a perfectly valid, on-screen frame. **Fix:** - `performAction:config:application:` now taps action buttons via a new `tapActionButtonWhenHittable:application:` helper that explicitly activates the application and waits for it to settle in the foreground *before* resolving the button's tap coordinate, so any interruption dance happens before we snapshot the tap point rather than concurrently with the tap. - Taps via an explicit `XCUICoordinate` (bypassing `XCUIElement`'s own tap/scroll-to-visible synthesis path) to avoid the `{-1, -1}` hit-point bug. - Extended the existing `msidTap` cross-platform convention (`XCUIElement+CrossPlat`) to `XCUICoordinate` so the new coordinate tap follows the same tap-on-iOS/click-on-macOS pattern as the rest of the file, instead of duplicating the platform check inline. ## Type of change - [ ] Feature work - [x] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) Change is scoped entirely to the UI test automation library (`tests/automation`), not the shipped SDK — no runtime/customer-facing impact. ## Additional information Verified locally end-to-end against a real iPhone 17 / iOS 26.3.1 simulator with live lab credentials: all 4 previously-failing `ADBTokenBindingUITests` tests now pass consistently (previously always failed with "Timed out waiting for the password field..." or "Wait for result pipeline" timeouts caused by the dropped Acquire Token tap upstream). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
PR validation currently runs the automation health job even when a PR
only touches test files. This change narrows that behavior so test-only
changes continue through normal validation without invoking the
automation health check.
- **Change detection**
- Adds a lightweight `DetectAutomationHealthChanges` job to diff the PR
head against the target branch.
- Normalizes `System.PullRequest.TargetBranch` before fetching so the
comparison works whether Azure Pipelines provides a short branch name or
a full ref.
- **Automation health gating**
- Sets an output variable when any changed file falls outside
`IdentityCore/tests/`.
- Gates `AutomationHealth` on that output, so the job is skipped for
test-only PRs and still runs for any production/config/pipeline change.
- **Scope of skip behavior**
- Treats files under `IdentityCore/tests/` as test-only changes.
- Leaves the existing `Validate_Pull_Request` matrix unchanged.
```yaml
- job: AutomationHealth
dependsOn: DetectAutomationHealthChanges
condition: and(
succeeded(),
eq(dependencies.DetectAutomationHealthChanges.outputs['detectChanges.runAutomationHealth'], 'true')
)
```
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: Ameya Patil <amepatil@microsoft.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…osoft identity host list (#1883) ### Summary `cloud_instance_host_name` from the auth webview redirect URL was previously used to build the cloud authority without checking whether it was a recognized Microsoft identity host. This PR makes that handling more accurate: the value is only used when it is a known AAD public/sovereign cloud host, or a network environment already discovered via instance metadata. ### Changes - Add a recognized-host check in `setCloudAuthorityWithCloudHostName:` (`MSIDRequestParameters`) and `authorityWithUpdatedCloudHostInstanceName:error:` (`MSIDAADAuthority`). Unrecognized values are ignored and logged; the originally configured authority continues to be used. - Host matching is case-insensitive to align with cached `preferred_network` values (which may not be case-normalized). - Add unit tests for known public-cloud, known sovereign-cloud, unknown, nil, blank, and different-case host inputs. - Update the instance-aware interactive token request integration test to use a valid sovereign-cloud host. ### Impact - Sovereign-cloud and instance-aware flows are unaffected. - No public API changes. --------- Co-authored-by: kai <kasong@microsoft.com> Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
…n request in browser-native GetToken request (#1879) ## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## Proposed changes Simplifies how the Apple broker resolves the requested token type when parsing a browser-native `GetToken` request. The request now resolves `tokenType` as follows: 1. `tokenType` in `extraParameters` (extra query parameters) 2. `tokenType` at the top level of the request (fallback) `authenticationScheme` is no longer read at any point. The previously added `authenticationScheme` key constant has been removed. When `tokenType` resolves to `pop`, a valid `reqCnf` is still required (top-level or in `extraParameters`); otherwise the request is rejected. When no usable `tokenType` is present, the scheme defaults to Bearer, preserving existing behavior. ## Type of change - [x] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [x] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. - [ ] Medium – Errors could cause regression of 1 or more scenarios. - [x] Small – No issues are expected. ## Additional information - Related work item: Task 3596198 — Parse token_type from ESTS Response in Apple Brokers. - Updated unit tests in `MSIDBrowserNativeMessageGetTokenRequestTests` to cover `tokenType` in extra query parameters and at the top level, extraParameters-first precedence, PoP `reqCnf` validation in both positions, non-string `tokenType` handling, and regression tests confirming `authenticationScheme` is ignored. Full `MSIDBrowserNativeMessageGetTokenRequestTests` suite passes (39 tests, 0 failures). Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com>
…#1889) ## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Based on #1880 ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: ashok672 <ashok672@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…1891) ## Summary Gate the reqCnf presence validation for Pop token requests in `MSIDBrowserNativeMessageGetTokenRequest` behind a new ECS flight, `browser_core_disable_reqcnf_validation`. The flight is a **kill switch**: `MSIDFlightManager boolForKey:` returns `NO` when unset, so the validation is **enabled by default**. ECS can disable it to restore the prior fallback behavior (request initializes and falls back to the default Bearer scheme instead of failing when reqCnf is missing/blank on a Pop request). This matches the existing `browser_core_disable_pop` / `browser_core_disable_claims` kill-switch convention in the same file. The tokenType lookup-order (extraParameters first, then request) is unchanged. ## Changes - `MSIDConstants.h/.m`: add `MSID_FLIGHT_BROWSER_CORE_DISABLE_REQ_CNF_VALIDATION` (`browser_core_disable_reqcnf_validation`). - `MSIDBrowserNativeMessageGetTokenRequest.m`: gate the reqCnf nil/blank guard behind the flight. - `MSIDBrowserNativeMessageGetTokenRequestTests.m`: 3 new tests (flight on with nil/empty reqCnf -> init succeeds with default scheme; flight explicitly off -> still rejected) + `tearDown` resetting the flight provider. - `changelog.txt`: entry under TBD. ## Testing `IdentityCore Mac` scheme, `MSIDBrowserNativeMessageGetTokenRequestTests`: 42/42 passing. --------- Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
The PR validation job introduced in #1881 failed in Bash because `System.PullRequest.TargetBranch` was evaluated as command substitution instead of a pipeline variable. This broke target-branch resolution for the “test-files-only” change detector. - **Root cause** - Inline Bash used Azure DevOps variable syntax in a shell command context: - `$(System.PullRequest.TargetBranch)` → interpreted by Bash as `$(...)` command execution. - **Change** - Switched branch source to the Azure DevOps environment variable exposed to scripts: - `SYSTEM_PULLREQUEST_TARGETBRANCH` - Kept existing ref normalization behavior (`refs/*` guard) unchanged. - **Resulting behavior** - Target branch is now read correctly in PR runs. - Non-test change detection can compute diffs against the PR target branch as intended. ```bash # before target_branch_ref="$(System.PullRequest.TargetBranch)" # after target_branch_ref="${SYSTEM_PULLREQUEST_TARGETBRANCH:-}" if [[ "${target_branch_ref}" != refs/* ]]; then target_branch_ref="refs/heads/${target_branch_ref}" fi ``` Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
…put (#1895) ## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x] PR is independently mergeable (no hidden dependencies) - [x] Appropriate reviewers are assigned - [x] PR reviewed by code owner (required if Copilot-generated) - [x] SME or Senior IC assigned where required ## Proposed changes `+[NSString(MSIDExtensions) msidIsStringNilOrBlank:]` only special-cased `nil` and `NSNull`. For any other non-`NSString` argument it fell through to `!string.length`, sending `-length` to the object: ```objc if (!string || [string isKindOfClass:[NSNull class]] || !string.length) ``` When a mis-typed value reaches the check — e.g. a JSON/plist **boolean** decoded as `__NSCFBoolean` from an MDM SSO-extension configuration where a string was expected — this raises `-[__NSCFBoolean length]: unrecognized selector sent to instance` and crashes (SIGABRT). This was observed as a crash in the iOS `AuthenticatorSSOExtension` on the main thread during `-[ASSSOExtensionAuthenticationViewController beginAuthorizationWithRequest:]`, where SSO-extension configuration values are parsed. This change replaces the `NSNull`-only check with a positive `NSString` kind check, so any non-`NSString` (including `NSNull`, `NSNumber`/`__NSCFBoolean`, collections, etc.) is treated as blank and the method never messages a non-string: ```objc if (!string || ![string isKindOfClass:[NSString class]] || !string.length) ``` Adds a regression test covering `NSNumber`/`__NSCFBoolean`, `NSArray` and `NSDictionary` inputs. ## Type of change - [ ] Feature work - [x] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information Behavior for existing inputs is unchanged: `nil`, `NSNull`, empty and whitespace-only strings still return `YES`; non-blank strings still return `NO`. Only previously-crashing non-`NSString` inputs change from "throw" to `YES`. All `MSIDTestNSStringHelperMethods` tests pass locally (iOS, Xcode 26.5, iPhone 17 simulator), including the new regression test. Changelog updated under `TBD`. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…vider (#1894) ## PR Checklist (must be completed before review) - [x] All tests pass locally - [x] PR size is <= 500 LOC per PR Size Check policy - [x] PR is independently mergeable (no hidden dependencies) - [x] Appropriate reviewers are assigned - [x] PR reviewed by code owner (required if Copilot-generated) - [x] SME or Senior IC assigned where required ## Proposed changes `MSIDFlightManager`'s `boolForKey:` and `stringForKey:` tested `self.flightProvider` through the (unsynchronized) property getter **before** dispatching onto the synchronization queue, then re-read it inside the queue: ```objc if (self.flightProvider) { // unsynchronized read dispatch_sync(self.synchronizationQueue, ^{ result = [self.flightProvider ...]; // second read }); } ``` That first read races the barrier write in `setFlightProvider:`. When another thread swaps/releases the provider concurrently with a read, the read can end up messaging a freed provider (use-after-free). In the Microsoft Authenticator app this surfaced as an `objc_msgSend` crash inside `-[MSIDAccountCredentialCache checkFRTEnabled:error:]` at the flight-status read (`stringForKey:MSID_FLIGHT_CLIENT_SFRT_STATUS`), driven from a launch-time background thread that reads FRT/flight state while the main flow assigns a new `flightProvider` from a parsed broker device-info response. This change reads `_flightProvider` **once from inside the synchronization queue** into a strong local and messages that local, removing the unsynchronized access and keeping the provider alive for the duration of the call. It complements #1876 (which made the setter synchronous via `dispatch_barrier_sync`); the setter and readers are now fully serialized through the queue. Adds a concurrency regression test (`testConcurrentReads_whileProviderSwappedAndCleared_doNotCrash`) that interleaves provider swaps (including `nil`) with concurrent `boolForKey:`/`stringForKey:` reads. ## Type of change - [ ] Feature work - [x] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information All 15 `MSIDFlightManagerTests` pass locally (iOS, Xcode 26.5, iPhone 17 simulator), including the new regression test. Changelog updated under `TBD`. --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…ken (#1886) ## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Add a parameter for caller to pass nonce. This will allow getting nonce as a network request separately and pass it in here from its callback. This will also prevent deallocation of initial request object during async call. ## Type of change - [ ] Feature work - [x] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [x] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
…and complete customized XPC transport failures with errors (#1887) ## Summary Fixes ADO 3236668: https://dev.azure.com/IdentityDivision/_workitems/edit/3236668 `+[MSIDXpcSingleSignOnProvider canPerformRequest:]` has 5 distinct exit points that all collapse into a bare `NO`, making it impossible to root-cause the production spike of this client-side XPC gate rejecting requests without further instrumentation. This PR adds an additive, non-breaking `reason:` overload that reports which branch fired — at both the provider level and the controller level, since production callers (e.g. OneAuth's `MSAIBrokerClient`) call the controller's bare `canPerformRequest` class method directly and never see CommonCore's internal `MSID_LOG_WITH_CTX` output. ## Changes - **New enum `MSIDXpcCanPerformFailureReason`** (`MSIDXpcSingleSignOnProvider.h`), with one case per failure branch: - `None` (success) - `NoProviderInstalled` — no MacBrokerApp/CompanyPortal XPC component on device - `DeviceInfoRequestCreationFailed` — SSOExtension `getDeviceInfo` request object failed to construct - `DeviceInfoHandshakeError` — handshake completed with a hard error - `DeviceInfoHandshakeTimeout` — handshake did not complete within the 1s `dispatch_group_wait` window - `ValidateCacheProviderFailed` — no installed provider matches cached/available config - `UnsupportedOSVersion` — controller-level gate rejects the request on macOS < 13, before ever reaching the provider - **New `MSIDXpcCanPerformFailureReasonToString(reason)`** helper for non-PII logging. - **New `+[MSIDXpcSingleSignOnProvider canPerformRequest:reason:]`** overload that sets `*reason` at each return site. The original no-`reason` `canPerformRequest:` selector is kept as a thin wrapper (`reason:nil`) — fully source/behavior compatible with existing call sites and tests. - **Split timeout vs. hard-error**: both previously fell through to the same generic `validateCacheXpcProvider` failure; they're now tracked separately and reported as distinct reasons (timeout takes priority if both occurred). - **Updated both callers** — `MSIDXpcSilentTokenRequestController` and `MSIDXpcInteractiveTokenRequestController` `canPerformRequest` — to call the new overload and log the reason (name + int, no PII) via `MSID_LOG_WITH_CTX` when the result is `NO`. - **New `+[MSIDXpcSilentTokenRequestController canPerformRequest:]`** and **`+[MSIDXpcInteractiveTokenRequestController canPerformRequest:]`** overloads (reason out-param), mirroring the provider-level pattern, so external callers outside CommonCore (e.g. OneAuth) can retrieve and surface the failure reason to their own telemetry instead of only getting a bare `BOOL`. The existing no-arg `canPerformRequest` selectors remain thin wrappers (`canPerformRequest:nil`) — fully source/behavior compatible. - **Unit tests**: added coverage in `MSIDXpcSingleSignOnProviderTest.m` for all 5 provider-level failure branches and the success path. - **Changelog**: added/updated an entry under `TBD`. ## Testing Ran `./build.py --targets mac_library --no-xcpretty --no-clean`: - Build succeeded. - All 2496 tests passed (0 failures). ## Notes - No PII is logged — only the reason enum name/int and class/method name. - No behavior change for existing callers of any no-`reason` selector (provider or controller level). - Adoption in OneAuth (`MSAIBrokerClient.m`) to actually consume the new controller-level `reason:` overload and feed it into OneAuth's own telemetry (`MSAIBrokerEligibilityResponseFactory` / `appendExecutionFlow:`) is tracked separately, since that's a different repo/consumer. --------- Co-authored-by: kai <kasong@microsoft.com> Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
…load web session launches instead of after it completes so the banner fires while the user is away in Settings.
…ests for the moved scheduling method
…ule-only provider cannot crash
…ows (#1897) ## Summary Refactors the onboarding telemetry plumbing and adds new onboarding-blob steps for the new mobile onboarding flows. ### Telemetry - New passthrough onboarding-blob steps: `TokenRequestRetryStarted`, `TokenRequestRetryFailed`, `MDMEnrollmentFailed`, (`MSIDOnboardingBlobFieldKeys`). - `MSIDOnboardingBlobBuilder` additions to stamp the new steps. - `MSIDLocalInteractiveController.handleWebMDMEnrollmentCompletionResponse:` now stamps MDM status + TokenRequestRetry steps. > Note: The retry path stamps only `TokenRequestRetryStarted` and `TokenRequestRetryFailed`. There is no `TokenRequestRetrySucceeded` step — a successful retry is reflected by the terminal success signal of the overall flow. ### Webview - Navigation handling refactor (`MSIDOAuth2EmbeddedWebviewController`, `MSIDWebviewNavigationHandler/Delegate`, `MSIDWebviewNavigationDecisionResolver`, `MSIDAADOAuthEmbeddedWebviewController`). ### Tests - Added/updated unit + integration coverage for the new steps and navigation changes. --------- Co-authored-by: Swasti Gupta <swagup@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
…pt for notificaion after 75 seconds
…1905) ## Summary Forwards `x-app-name` / `x-app-ver` to Intune on the MDM enrollment webview request, **gated behind a Microsoft first-party (1P) app check** so third-party apps never leak their app identity. ## Changes - **`MSIDHelpers.isMicrosoftFirstPartyApp`** — returns YES when the running app's Apple Team ID (from `MSIDKeychainUtil`) is a known Microsoft team. - **`MSID_MICROSOFT_FIRST_PARTY_TEAM_IDS`** (in `MSIDConstants`) — the Microsoft-owned Apple Team IDs (`UBF8T346G9`, `SGGM6D27TK`, `9KBH5RKYEW`), mirroring the Broker-layer `microsoft1PAppsTeamIDs`. - **`NSBundle msidAppName`** — sibling to existing `msidAppVersion` (CFBundleDisplayName ?: CFBundleName). - **`MSIDWebviewNavigationDecisionResolver`** — attaches the two headers on the enroll request only when `isMicrosoftFirstPartyApp` is YES. ## Tests Added resolver unit tests exercising the real predicate via `teamId` swizzle: - `testEnrollURL_whenMicrosoftFirstPartyApp_shouldAttachAppNameAndVersionHeaders` - `testEnrollURL_whenNotMicrosoftFirstPartyApp_shouldNotAttachAppNameAndVersionHeaders` Full `MSIDWebviewNavigationDecisionResolverTests` suite passes on iOS simulator. --------- Co-authored-by: Swasti Gupta <swagup@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…e user is in Settings (#1901) ## Summary Relocates the MDM profile-installed local notification scheduling from the post-return `profile_download_complete` callback to the moment the profile-download `ASWebAuthenticationSession` is launched, so the reminder is armed *before* the user leaves for Settings. ## Changes - **`MSIDWebviewNavigationHandler.m`**: Adds `scheduleMDMProfileInstalledNotificationIfNeededForURL:`, invoked right before `transitionToSystemWebviewWithURL:`. It gates on the URL containing `downloadprofile`, reads the delay from the `MSID_FLIGHT_MDM_PROFILE_INSTALLED_NOTIFICATION_DELAY` flight (falling back to `MSIDMDMProfileInstalledNotificationDefaultDelay`), and forwards to the registered UX callback provider. Includes info/warning logs. - **`MSIDWebviewNavigationDecisionResolver.m`**: Removes the now-too-late scheduling in `decisionForProfileDownloadComplete:` (replaced with an explanatory comment) and the cancel call in `decisionForEnrollmentCompletionURL:`; drops the two imports that became unused as a result. - **`MSIDUXCallbackProtocol.h`**: Removes the now-unused `cancelMDMProfileInstalledNotification` declaration. ## Why On-device logs confirmed the in-app enrollment web callbacks (`profile_download_complete`, `in_app_enrollment_complete`) only fire *after* the user returns from Settings, i.e. when the app is foreground again and banners are suppressed. Scheduling at session launch lands the notification while the app is backgrounded (user still in Settings), so the banner is actually shown. The cancel path is no longer needed because the notification is armed at hand-off rather than speculatively. ## Related Paired with the Broker-side conformance-guard change (forwards `scheduleMDMProfileInstalledNotificationWithDelay:` to `MSIDUXCallbackProvider`) and the Authenticator host-app notification handler.
…1906) Support for Broker seeded-BRT silent device registration: - Add -isAADAuthority to base MSIDAuthority (returns NO) and override in MSIDAADAuthority (returns YES), so callers can gate AAD-only behavior without an isKindOfClass:[MSIDAADAuthority class] cast. - Add MSIDOnboardingBlobStepSeededBRTBootstrapStarted/Completed onboarding blob step constants (broker-only funnel, fanned out to mo_steps_list). Copilot-Session: e70061c1-fa52-46e5-a985-6ca86ad66d49 ## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: Swasti Gupta <swastigupta@rerr.local> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: Swasti Gupta <swagup@microsoft.com>
kaisong1990
reviewed
Jul 20, 2026
| MSIDNonceTokenRequest *nonceRequest = [[MSIDNonceTokenRequest alloc] initWithRequestParameters:nonceReqParams]; | ||
| __weak typeof(self) weakSelf = self; | ||
| [nonceRequest executeRequestWithCompletion:^(NSString * _Nullable resultNonce, NSError * _Nullable error) | ||
| if ([NSString msidIsStringNilOrBlank:self.nonce]) |
Contributor
There was a problem hiding this comment.
This changes executeRequestWithCompletion: from fetching its nonce internally to failing unless the caller pre-populates nonce. Existing callers that instantiate this request directly will start failing after the 1.26.0 update without a clear migration path. Can we either preserve the previous behavior or document this change in the header/changelog and point callers to MSIDDeviceTokenUtil?
kaisong1990
reviewed
Jul 20, 2026
| BOOL flowSucceeded = (endURL != nil && error == nil); | ||
| if (flowSucceeded && self.strongAuthSetupStarted) | ||
| { | ||
| // MDMEnrollmentFinished is stamped from the in_app_enrollment_complete redirect. |
Contributor
There was a problem hiding this comment.
This comment says MdmEnrollmentFinished is stamped from the enrollment-complete redirect, but the refactored code no longer calls addStep:MSIDOnboardingBlobStepMdmEnrollmentFinished anywhere. The old controller stamped it after MDM enrollment started, so successful MDM onboarding now loses its completion step. Can we restore that signal in the builder, or remove the step and update the contract if it is intentionally retired?
…rs (#1910) ## Summary Adds a `sourceApplication` property to `MSIDInteractiveTokenRequestParameters` (the invoking app's bundle identifier as reported by the OS). This is consumed by the broker (see the paired MSAL + Broker PRs) to gate first-party-only behavior on the mobile-onboarding MDM-enrollment request. The OS surfaces the caller bundle id only for apps signed under the broker's Apple Developer Team identifier, so a non-nil value effectively identifies a Microsoft 1P caller; third-party callers leave it nil. ## Changes - `MSIDInteractiveTokenRequestParameters.h`: new `@property (nonatomic, copy, nullable) NSString *sourceApplication;` (ARC-synthesized backing ivar). ## Testing - Consumed and validated in the Broker repo unit tests (`ADBrokerInteractiveControllerWithPRTTests`, `MSIDInteractiveTokenRequestParameters+ADBrokerTests`) — 29 tests pass. ## Related Part of a chained submodule change: - Common Core (this PR) → MSAL (submodule bump) → Broker (submodule bump + consumer logic). Co-authored-by: Swasti Gupta <swagup@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: agubuzomaximus <agubuzomaximus@gmail.com> Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com> Co-authored-by: Fidelia Nawar <fidelianawar@Fidelias-MacBook-Pro.local> Co-authored-by: Swasti Gupta <swastinitb@gmail.com>
## PR Checklist (must be completed before review) - [ ] All tests pass locally - [ ] PR size is <= 500 LOC per PR Size Check policy - [ ] PR is independently mergeable (no hidden dependencies) - [ ] Appropriate reviewers are assigned - [ ] PR reviewed by code owner (required if Copilot-generated) - [ ] SME or Senior IC assigned where required ## PR Title Format **Required Format:** `[Keyword1] [Keyword2]: Description` - **Keyword1:** `major`, `minor`, or `patch` (case-insensitive) - **Keyword2:** `feature`, `bugfix`, `engg`, or `tests` (case-insensitive) **Examples:** - `[MAJOR] [Feature]: new API` - `[minor] [bugfix]: fix crash` - `[PATCH][tests]:add coverage` ## Proposed changes Describe what this PR is trying to do. ## Type of change - [ ] Feature work - [ ] Bug fix - [ ] Documentation - [ ] Engineering change - [ ] Test - [ ] Logging/Telemetry ## Risk - [ ] High – Errors could cause MAJOR regression of many scenarios. (Example: new large features or high level infrastructure changes) - [ ] Medium – Errors could cause regression of 1 or more scenarios. (Example: somewhat complex bug fixes, small new features) - [ ] Small – No issues are expected. (Example: Very small bug fixes, string changes, or configuration settings changes) ## Additional information --------- Co-authored-by: agubuzomaximus <agubuzomaximus@gmail.com> Co-authored-by: Maximus Agubuzo <maagubuzo@microsoft.com> Co-authored-by: Fidelia Nawar <fidelianawar@Fidelias-MacBook-Pro.local>
15 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR Checklist (must be completed before review)
PR Title Format
Required Format:
[Keyword1] [Keyword2]: Descriptionmajor,minor, orpatch(case-insensitive)feature,bugfix,engg, ortests(case-insensitive)Examples:
[MAJOR] [Feature]: new API[minor] [bugfix]: fix crash[PATCH][tests]:add coverageProposed changes
Describe what this PR is trying to do.
Type of change
Risk
Additional information