If there’s a hidden danger in these modes, better to leave the switch requiring octal tunes! -- Brian Sniffen > On Feb 8, 2018, at 8:40 PM, Daniel Kahn Gillmor <dkg@fifthhorseman.net> wrote: > >> On Tue 2018-02-06 14:43:56 -0500, Daniel Kahn Gillmor wrote: >> There are legitimate cases (public archives) where a user might >> actually want their archive to be readable to the world. >> >> "notmuch insert" historically used mode 0600 (unreadable by group or >> other), but that choice doesn't appear to have been specifically >> justified (perhaps an abundance of caution?). >> >> This patch also adjusts the default mode used for --create-folder, to >> be mode 0755 before the application of the umask. >> >> If the user wants "notmuch insert" to create files or folders that are >> not readable by group or other, they can set their umask more >> restrictively. > > I'm now having second thoughts about this. > > postfix's local delivery agent has apparently been delivering with mode > 0600 for nearly 20 years: > > https://github.com/vdukhovni/postfix/blame/master/postfix/src/local/maildir.c#L188 > > And dovecot's lda defaults to 0600 on delivery: > > https://sources.debian.org/src/dovecot/1:2.2.33.2-1/src/lib-storage/mail-storage.c/?hl=2591#L2591 > > So maybe there's something i don't know about why a delivery agent would > want to have this restrictive mask? > > Perhaps a better way to fix this is with a new option to notmuch insert. > > on IRC, bremner suggests something flexible like --mode=0600 > > I'm more inclined to keep it simpler and more usable (most people don't > know octal, let alone unix permissions bits) and just have a boolean > --world-readable which defaults to false (and switches between modes > 0600 and 0644 for files, and 0700 and 0755 for directories). > > Any thoughts? > > --dkg > _______________________________________________ > notmuch mailing list > notmuch@notmuchmail.org > https://notmuchmail.org/mailman/listinfo/notmuch _______________________________________________ notmuch mailing list notmuch@notmuchmail.org https://notmuchmail.org/mailman/listinfo/notmuch