You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Continuing the work started in #154873, @rebornix and I have identified a few common scenarios where we can smartly detect issues and direct the user to filing meaningful reports.
The text was updated successfully, but these errors were encountered:
For file open, we currently warm up markdown renderers and their rendered output before rendering the notebook, this makes markdown cells stable on first load. However, for output cells, we do not warmup and the renderers can take any amount of time causing some amount of flicker on load. In order to reduce flicker, we propose allowing some small amount of time for warming up the outputs say up to 250ms. In performant cases, this will mean no flicker for the user. In other case, it will be the similar to today. In that case, it would be nice to measure and display the renderer time and allow users to see and report it. No ideas yet have been proposed on how to show this.
For cell execution, there are several markers in time which we consider. The flow is as follows:
User clicks Run Cell
Cell is pending execution (no timer in UI)
Kernel starts executions (timer starts in UI)
Kernel finishes execution (timer still running in UI)
VS Code process kernel results (timer is showing elapsed time 3-5, timer in UI is updated to Kernel execution time 3-4)
Output renders execute and update outputs
Outputs are done rendering
We are concerned with 2 timespans:
VS Code Processing Time (span between 4 and 5)
Output Rendering Time (span between 5 and 7)
Our proposal is to detect when these times are above a threshold and indicate this to the user with a non-intrusive warning icon near the timer. Hovering this icon will reveal our stats and allow them to file an issue on the appropriate repo.
Continuing the work started in #154873, @rebornix and I have identified a few common scenarios where we can smartly detect issues and direct the user to filing meaningful reports.
The text was updated successfully, but these errors were encountered: