Closed jrmarino closed 1 month ago
I don't know about this now. Especially when rvn comes, scanning will be really fast plus isn't this distributed across workers?
This is no longer necessary. Rather than call rvn (or ravensw before it) to extract the metadata, ravenadm has imported rvn-format code and can open the files directly. This was an exponential increase in speed, especially since it's done in parallel.
with nearly 7000 packages, it's taking longer and longer to scan them all. come up with a scheme to do this without opening the packages. sqlite?
basically - as each package is created, update the register. The assumption here is nobody alters the packages -- that you can rely that the information in the register matches the information in the package.
This should speed up scanning tremendously.