On Tue, Oct 07, 2025 at 08:26:08AM -0400, Michael Stone wrote:
I assume this is yet another case where something was de facto taken over by systemd and then arbitrarily shitcanned when the systemd folks lost interest.
This was an intentional change; upstream systemd views this mechanism as inherently insecure (they're not wrong) and defaulted permissions to a more secure baseline to force migrations to existing, established, mature alternatives (like flock).
I mention all that to say: This was not a lack of interest, nor was it arbitrarily; if anything this was due to a lot of attention being paid to fix a long-standing design weakness encoded in the FHS.
The ctte decision here is basically "Yes, this is right, but we need to coordinate a sensible transition". TBH this change would be interesting even without systemd forcing it.
Paul (without any hats on) -- ⢀⣴⠾⠻⢶⣦⠀ Paul Tagliamonte <paultag> ⣾⠁⢠⠒⠀⣿⡁ https://people.debian.org/~paultag | https://pault.ag/ ⢿⡄⠘⠷⠚⠋ Debian, the universal operating system. ⠈⠳⣄⠀⠀ 4096R / FEF2 EB20 16E6 A856 B98C E820 2DCD 6B5D E858 ADF3
Attachment:
signature.asc
Description: PGP signature