Closed lukemarsden closed 5 years ago
Results from interactive testing:
yay this caching implementation is working really well now when i start downloading a big file, it stops using the result cache because dirtyData != 0 each time it moves the snapshot forwards when i stop the download, it notices that the fs stops changing and starts using the cache for every result so it’s not having to iterate over all the files (or do any snaps, mounts etc) when the filesystem is quiet and in all cases, it’s caching the file list from the latest dotmesh commit, so it’s only doing 50% of the work it was doing before
Should see slightly improved unzip performance as well, as there'll be less file I/O now hopefully.
Ready for review