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
In LiGet we have following approach for caching, which can be ported to BaGet:
Every minute we read events from catalog reader of nuget.org.
For any package ID, when an event happens for it, we invalidate cache entries for it.
This allows to cache queries for all versions of packages. BaGet can know all versions of each package, without having to do lookup to nuget.org each time. This allows to serve some queries directly by baget, rather than making extra trip to nuget.org.
In worst case, BaGet does not know about packages which were pushed/unlisted up to 1 minute ago.
This should be done after profiling #9 , we should know if there is actually significant benefit from this.
The text was updated successfully, but these errors were encountered:
In LiGet we have following approach for caching, which can be ported to BaGet:
This allows to cache queries for all versions of packages. BaGet can know all versions of each package, without having to do lookup to nuget.org each time. This allows to serve some queries directly by baget, rather than making extra trip to nuget.org.
In worst case, BaGet does not know about packages which were pushed/unlisted up to 1 minute ago.
This should be done after profiling #9 , we should know if there is actually significant benefit from this.
The text was updated successfully, but these errors were encountered: