Arch Linux User Repository

Hi, thx for providing this package!

Could you change in https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=brscan4#n26

./mk-udev-rules opt/brother/scanner/brscan4/{Brsane4.ini,models4/*.ini} > etc/udev/rules.d/40-$pkgname.rules

to:

gawk -f mk-udev-rules opt/brother/scanner/brscan4/{Brsane4.ini,models4/*.ini} > etc/udev/rules.d/40-$pkgname.rules

Otherwise it breaks when "mk-udev-rules" is cached on mount point with "noexec" flag.

updated to 0.4.5_1 @ron2138 I changed the upstream URL to something more unspecified because the one you provided did not have the latest updates. Thanks for reporting upstreams - please keep us informed if you get an answer.

1. The upstream URL recorded above can not be found

. I believe the this is the current URL.

2. Replying ettavolt comment on 2018-06-13 05:13 and Harey comment on 2018-06-12 09:20: On 2018-06-19 I sent a message through Email Support. Got #180619-000271 reference number for follow up. A reply by email was they will provide this to the software engineers.

Man (https://www.freedesktop.org/software/systemd/man/udev.html) says that /etc/udev/rules.d/ is for local administrator. Local = per-machine. Package isn't machine-dependent (though it may be useful only for some that have Brother's scanner connected). Having rules in /usr/lib would allow a simple disabling via linking /etc/udev/rules.d/40-brscan4.rules to /dev/null. And this way won't interfere with package updates (while deleting file does).

The output of 'pacman -Qo /usr/lib/udev/rules.d/' makes me feel it is better to leave this directory alone ;). As far as you don't have a pressing argument I will stay with /etc/udev/rules.d/. One could interpret 'vendor' as 'Archlinux' as well...

Currently, /etc/udev/rules.d/40-brscan4.rules are installed by brscan4. I think it would be better if that file will be installed under /usr/lib/udev/rules.d/. The reason is that, as far as I can tell, /usr/lib/udev/rules.d/ are for vendor files, while /etc/udev/rules.d are for local files.

So you should file a bug for sane in Archlinux's bugtracker and see what the devs think about the situation...

Works, thank you. But for a bit different reason. ☺

So, what's happening with my DCP-1510 (as I understand it now):

40-brscan: MODE 664, GROUP lp, libsane_matched=yes

49-sane: MODE 664, GROUP scanner

50-udev-defaults: GROUP="lp" (because ENV{ID_USB_INTERFACES}==":0701??:")

70-uaccess: TAG+uaccess (because libsane_matched=yes)

Systemd somehow adds current-user ACL for uaccess-TAGged devices (this package owns udev-defaults and uaccess). Since brscan4 has priority of 40 (higher than 70 of uaccess) scanner works for current user without modification of his group. I think Arch's sane has to generate ACL-based udev rules (instead of what's generated now).

Here you can remove MODE and GROUP actions. The combination of sane.rules and udev-default.rules will do exactly the same (checked for DCP-1510).

ettavolt: OK, this seems quite logical to me. I pushed a new version of the package with the change you proposed. Let's see what happens ;)

Sorry, forgot to enable notifications. Basically udev is applying files from /etc and /usr in lexical order (without directory paths). Sane's rules are chowning scanner ports to 'scanner' group if ENV{libsane_matched}=="yes" (see the bottom of /usr/lib/udev/rules.d/49-sane.rules). "brscan4" is lexically after "49-sane". That's why ENV{libsane_matched}="yes" in "brscan4" will be applied after sane's rules and thus a Brother's scanner port won't be chowned to 'scanner'.

BTW, I wonder why is there MODE=xxx GROUP=xxx in the sane's rules file? This changes are reapplied by ENV{libsane_matched}=="yes" condition…