fix(tui): use the is_running property for app guards that never fired (#894)

[aniketshukla1]

Summary

textual's App.is_mounted is a method taking a widget, not a property — so a bare self.is_mounted is an always-truthy bound method, and every if not self.is_mounted: return guard in the TUI was dead. Worst case, the dead guard let query_one run against a missing screen, raise, get caught by the surrounding except, and re-schedule via call_after_refresh — looping instead of returning.

Closes #894.

Changes

Switch the 13 app-level guards in strix/interface/tui/app.py to App.is_running (the bool property these guards actually want). The one correct widget.is_mounted (a Widget property) is left unchanged.

This resolves the mypy "Function 'is_mounted' could always be true in boolean context" and pyright reportUnnecessaryComparison findings from the issue. Confirmed by introspection that App.is_running is a property and App.is_mounted is a function, so not self.is_running is now a real check.

[greptile-apps]

Greptile Summary

Corrects TUI lifecycle guards by checking whether the Textual app is running instead of treating the App.is_mounted method as a boolean.

• Updates 13 app-level guards in strix/interface/tui/app.py.
• Leaves the widget-level widget.is_mounted property check unchanged.

Confidence Score: 5/5

The PR appears safe to merge, with the corrected guards preventing TUI operations after the app stops running.

The changes consistently replace always-truthy app method references with the intended boolean lifecycle property, and no reachable blocking or non-blocking defect remains.

Important Files Changed

Filename	Overview
strix/interface/tui/app.py	Replaces ineffective app-level is_mounted method checks with boolean is_running lifecycle checks; no actionable regression was identified.

_{Reviews (1): Last reviewed commit: "fix(tui): use the is_running property fo..." | Re-trigger Greptile}