Hey folks, reading groups functions for openbsd i could realize that in order to have supplementary groups retrieved, the routine pass through the entire database group. For each of the group entry, it scans if the given user matches in the list of the group member. I believe this approach does not scale (O(n*m)). Am i wrong?
Since this is a many-read-few-write database, what about an aditional register for each of the users with supplementary group? In this case a single database probe could be enough. Where am i wrong? Thanks in advance.
