Describe the bug
Two related issues with network retries:
1. Retries not cancelled on interrupt
When the user interrupts a query (e.g., pressing Ctrl+C or clicking "Stop"), any ongoing network retry backoff/sleep is not cancelled. The retry loop continues even after the user has explicitly stopped the operation.
2. Network retry information not displayed
When the system is retrying a provider request (e.g., waiting with exponential backoff), there is no UI feedback showing that a retry is in progress. The user sees no indication that the system is waiting and will automatically retry.
Expected behavior
- When a query is interrupted, any pending retry timers should be immediately cancelled
- The client should display a "Retrying in X seconds..." message so the user knows what's happening
- Need to design and implement the appropriate event shape for retry status updates
Where to look
crates/core/src/query.rs — retry logic around lines 1036-1059 and 1214-1240
crates/protocol/src/event.rs — event shapes for client communication
- Desktop and TUI apps — event consumption and display
Describe the bug
Two related issues with network retries:
1. Retries not cancelled on interrupt
When the user interrupts a query (e.g., pressing Ctrl+C or clicking "Stop"), any ongoing network retry backoff/sleep is not cancelled. The retry loop continues even after the user has explicitly stopped the operation.
2. Network retry information not displayed
When the system is retrying a provider request (e.g., waiting with exponential backoff), there is no UI feedback showing that a retry is in progress. The user sees no indication that the system is waiting and will automatically retry.
Expected behavior
Where to look
crates/core/src/query.rs— retry logic around lines 1036-1059 and 1214-1240crates/protocol/src/event.rs— event shapes for client communication