better UX for cached evaluation failures #223

Closed
opened 2024-04-04 09:10:36 +00:00 by yu-re-ka · 3 comments
Member

As a user I might run a command and see the error
error: cached failure of attribute '<...>'

I might not remember why this failed the last time.

For new users, this situation can be very frustrating, because they do not know how to get the information why this command fails.

Describe the solution you'd like

Best would be to save the error to the eval cache, and show it together with a hint that this is from a cached failure.

Describe alternatives you've considered

I'm currently just adding --impure to force re-evaluation. Adding an accompanying hint to the cached failure message would also be an improvement.

Additional context

N/A

## Is your feature request related to a problem? Please describe. As a user I might run a command and see the error `error: cached failure of attribute '<...>'` I might not remember why this failed the last time. For new users, this situation can be very frustrating, because they do not know how to get the information why this command fails. ## Describe the solution you'd like Best would be to save the error to the eval cache, and show it together with a hint that this is from a cached failure. ## Describe alternatives you've considered I'm currently just adding `--impure` to force re-evaluation. Adding an accompanying hint to the cached failure message would also be an improvement. ## Additional context N/A
Owner

Btw you can force re-evaluation with --no-eval-cache or --option eval-cache false. However. I would like to kill this with fire.

Is there any reason to cache failures in first place honestly??

Btw you can force re-evaluation with `--no-eval-cache` or `--option eval-cache false`. However. I would like to kill this with fire. Is there any reason to cache failures in first place honestly??
qyriad added the
ux
label 2024-04-04 13:16:01 +00:00
Owner

Is there any reason to cache failures in first place honestly??

none that can be reasonably argued for; in a perfectly deterministic evaluator (which we do not have!) it would make sense though. we should just remove the failure cache entirely, the eval success cache is already sketchy and the eval failure cache even more so

> Is there any reason to cache failures in first place honestly?? none that can be reasonably argued for; in a perfectly deterministic evaluator (which we do not have!) it would make sense though. we should just remove the failure cache entirely, the eval *success* cache is already sketchy and the eval *failure* cache even more so
Owner
https://gerrit.lix.systems/c/lix/+/771
Sign in to join this conversation.
No milestone
No project
No assignees
3 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: lix-project/lix#223
No description provided.