On Sat, Jul 10, 2021 at 11:37:33PM +0200, Thorsten Glaser wrote: > OK, thanks for the maintainer approval. Accordingly, I have > just uploaded the attached debdiff. I chose to direct the > output to syslog instead of the bitbucket so it is not lost > if someone indeed needs it so it isn’t a regression.
> - invoke-rc.d apache2 reload > + invoke-rc.d apache2 reload 2>&1 | logger -t apache2.logrotate I for one believe the old behaviour was superior for the common case of "success" -- no news is good news, no need to bother anyone just because logrotate does its job, especially that the restart is already logged elsewhere. On the other hand, logging failures is a damn good idea. But then: • three weeks before the release is no time for such meddling • it should be coded in sysv-rc/runit/etc instead of every daemon (the author of the patch that regressed this was probably used to a certain inferior rc system that fails to notify an interactive user whether a reload succeeded or was a no-op due to in-chroot policy/etc, which is a major no-no) Thus, thank you TG for the for-Bullseye fix! Meow. -- ⢀⣴⠾⠻⢶⣦⠀ ⣾⠁⢠⠒⠀⣿⡁ If you ponder doing what Jesus did, remember than flipping tables ⢿⡄⠘⠷⠚⠋⠀ and chasing people with a whip is a prime choice. ⠈⠳⣄⠀⠀⠀⠀
