Summary
Preserve and defer the idea of replacing the iPhone root NavigationSplitView with a path-driven NavigationStack.
The original local experiment targeted a SwiftUI compact-mode race where rapid mailbox navigation or foreground return could leave stale/blank content, incomplete transitions, or mismatched SwiftUI/UIKit lifecycle state. Current main still uses NavigationSplitView, but it has since accumulated additional navigation contracts and targeted mitigations, so the historical implementation is reference material, not a merge candidate.
This issue is intentionally deferred. Revisit it only if compact-mode navigation remains user-visible or split-view-specific workarounds continue accumulating.
Preserved branch and artifacts
Everything needed to revisit the experiment now lives in this repository:
The two documents are byte-identical preservation copies. The patch was verified in a disposable worktree to reconstruct the original two modified Swift files exactly at the experiment's source base.
Why it cannot be applied directly
Since the experiment, navigation gained behavior that the old patch does not model, including:
- real programmatic message and Account pushes under ADR-IOS-054;
- account-scoped notification-tap resolution, sentinel IDs, and unresolved-tap handling;
- richer locally-authored and server-draft open authority;
- additional settings, template, sheet, and cover routes;
- lifecycle guards for collapsed-split-view quirks.
The original plan also says “not started” even though an implementation patch exists, and its build, regression, and physical-device checklists remain incomplete. The historical claim that the shell replacement fixes the framework bug is therefore unproven.
Revisit requirements
When this work is resumed:
Current action
None. Keep this issue open with the deferred label; there is no PR, merge, or release attached to this preservation branch.
Summary
Preserve and defer the idea of replacing the iPhone root
NavigationSplitViewwith a path-drivenNavigationStack.The original local experiment targeted a SwiftUI compact-mode race where rapid mailbox navigation or foreground return could leave stale/blank content, incomplete transitions, or mismatched SwiftUI/UIKit lifecycle state. Current
mainstill usesNavigationSplitView, but it has since accumulated additional navigation contracts and targeted mitigations, so the historical implementation is reference material, not a merge candidate.This issue is intentionally deferred. Revisit it only if compact-mode navigation remains user-visible or split-view-specific workarounds continue accumulating.
Preserved branch and artifacts
Everything needed to revisit the experiment now lives in this repository:
b03cffca4The two documents are byte-identical preservation copies. The patch was verified in a disposable worktree to reconstruct the original two modified Swift files exactly at the experiment's source base.
Why it cannot be applied directly
Since the experiment, navigation gained behavior that the old patch does not model, including:
The original plan also says “not started” even though an implementation patch exists, and its build, regression, and physical-device checklists remain incomplete. The historical claim that the shell replacement fixes the framework bug is therefore unproven.
Revisit requirements
When this work is resumed:
main; do not cherry-pick or blindly apply the preserved patch.Stringpath values.Current action
None. Keep this issue open with the
deferredlabel; there is no PR, merge, or release attached to this preservation branch.