| Code-Review | +1 |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
I am a bit worried about the UI (or any impact if the result of permission query is wrong). Perhaps it's a rare case and only happens once at singleton initialization.
For my curiosity, which is better between deferring the widget close and this approach?
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
I am a bit worried about the UI (or any impact if the result of permission query is wrong). Perhaps it's a rare case and only happens once at singleton initialization.
For my curiosity, which is better between deferring the widget close and this approach?
The UI impact is minimal — kNotDetermined just hides the "system settings" link in the PageInfo bubble. This only affects the very first bubble open before the posted task runs (next message loop iteration), which is unlikely in practice.
Deferring the widget close would mean changing core views infrastructure (Widget::CloseWithReason or the occlusion tracker), which is riskier. Fixing it at the source avoids the message-pumping COM call entirely, which is more targeted and prevents other potential re-entrancy from the same root cause.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |