feat: update Mattermost rendering guidelines for standups and enhance daily logs with recent validation details

This commit is contained in:
2026-05-13 07:44:30 -06:00
parent bc467c8a8c
commit 286adf2d6c
11 changed files with 83 additions and 22 deletions

View File

@@ -3,13 +3,13 @@ type: daily
project: fidelity
date: 2026-05-12
status: active
focus: [pdiap-12284, pdiap-15836]
work-items: [PDIAP-12284, PDIAP-15836]
blockers: [xflowviewmaker-xflowsdk-dependency-mismatch]
focus: [pdiap-12284, pdiap-15836, rest-validation]
work-items: [PDIAP-12284, PDIAP-15836, PDIAP-15838]
blockers: []
tags:
- daily
- fidelity
updated: 2026-05-12
updated: 2026-05-13
---
# 2026-05-12
@@ -31,18 +31,21 @@ updated: 2026-05-12
- Files reported changed: `FlowConfig.swift`, `FlowViewBuilder.swift`, `FlowViewMaker.swift`, and `XFlowManager.swift`.
- Validation reported: SwiftUI remains the default host mode; explicit UIKit host path preserves dismissal-completion behavior; delegate callbacks/session cleanup still happen after confirmed dismissal; existing `FlowConfig` standard init call sites remain compatible; edited files had no file-level diagnostics.
- Build validation is not green yet. XFlowViewMaker's module build failed because its podspec currently resolves `XFlowSDK 2.8.48`, which does not expose the new host-mode type/API. XFlowSDK SwiftPM validation in that environment is also blocked by registry configuration: no registry configured for the `fidelity-src` scope.
- After manual dependency/build handling, the Fid4 compile path is now green for the latest `FlowViewBuilder` shape. The source-level fixes moved setup side effects out of the `@ViewBuilder`, kept `@ViewBuilder` confined to host-view construction, and preserved internal type safety without reintroducing `AnyView`.
- `PDIAP-12284` was moved to In Progress. `PDIAP-15836` may become a subtask of `PDIAP-12284`; Jeff planned to ask Quy how to handle the Jira structure and story points.
- David explained that "host-mode resolution" is shorthand for deciding whether XFlow builds through the SwiftUI host path or the temporary UIKit/`UIHostingController` path. XFlowViewMaker makes that decision before calling XFlowSDK; SwiftUI is the default and UIKit is selected only when the flag explicitly enables it.
- For REST validation, David confirmed Fid4 can be pointed at production through a plist value, which is the preferred non-pushed path. Testing with the production plist still showed GraphQL being used in the tested case. Raj appears to have a production account for follow-up production validation.
---
## Next Steps
- Keep the implementation changes if the local diff matches the reported boundaries, then perform a minimal dependency-alignment follow-up so XFlowViewMaker validates against the updated XFlowSDK branch/API instead of pod version `2.8.48`.
- Separate code-correctness validation from environment/dependency blockers when reporting status.
- After dependency alignment, rerun the targeted symbol/error/compile pass to confirm no stale enum names, host defaults, or dismissal sequencing regressions remain.
- Move to runtime validation now that compile is green: first run a representative Fid4/XFlow flow with the default configuration and confirm logs show the SwiftUI host path is selected.
- Then run a controlled UIKit-host experiment by forcing/simulating `iOS-XflowUIKitHostEnabled == true`, and confirm logs show the UIKit host path while preserving dismissal-completion behavior.
- Keep the evidence log-based and separate default SwiftUI behavior from the explicit UIKit-flag path.
---
## Blockers
- XFlowViewMaker compile validation is currently blocked by dependency mismatch against `XFlowSDK 2.8.48`, not by a confirmed code-level error in the host-mode changes.
- XFlowSDK SwiftPM validation is blocked in the current environment by missing `fidelity-src` registry configuration.
- No current source-level compile blocker after manual dependency/build handling. Runtime host-path validation is still pending.