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
{{ message }}
This repository has been archived by the owner on Nov 20, 2020. It is now read-only.
A metric I used in the past to order cache entries for cleanup was:
timeSinceLastHit / (numberOfTimesHit + 1)
So entries which are hit twice as often will take twice as long to expire. (The +1 just avoids division by zero, if we don't record the creation of the cache entry as a "hit".)
Although for this project, a fixed timeout like "1 month old" for the disk-cache might be simpler, and avoid having to record hit statistics. "1 month unused" might be better, but we would need to record the last hit time.
The text was updated successfully, but these errors were encountered:
Just a placeholder to discuss this feature. :)
A metric I used in the past to order cache entries for cleanup was:
So entries which are hit twice as often will take twice as long to expire. (The +1 just avoids division by zero, if we don't record the creation of the cache entry as a "hit".)
Although for this project, a fixed timeout like "1 month old" for the disk-cache might be simpler, and avoid having to record hit statistics. "1 month unused" might be better, but we would need to record the last hit time.
The text was updated successfully, but these errors were encountered: