On 2025-12-04 19:02:43 +0100, intrigeri wrote: > One potential UX improvement, derived from a suggestion from Vincent, > is this: when we know we're going to do a full cache build, i.e. > when the cache sub-directory for the current kernel×policy feature set > does not exist, we could say so in the logs. This could be a useful > feature request upstream.
I've just reported https://gitlab.com/apparmor/apparmor/-/issues/575 > I would classify this as "low priority, good are patches welcome" > given IIRC it's the first time, since I'm working on AppArmor in > Debian, that someone complains about this. Yes, searching for "Job apparmor.service/start running" on the web currently gives only 4 matches: * 2 for this Debian bug report; * 1 that corresponds to a different issue; * 1 that corresponds to a probable cache rebuild, but this is just a boot log (without any discussion about it). But I could find another complaint, which does not quote the above message: https://www.reddit.com/r/linuxquestions/comments/1jphu5r/apparmorservice_dragging_boot_times/ "'systemd-analyze blame' reveals that apparmor.service is taking 21 seconds to start" (and no information about its reproducibility). -- Vincent Lefèvre <[email protected]> - Web: <https://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / Pascaline project (LIP, ENS-Lyon)
