03:59 <dlcusa> emmett1, SiFuh: I have the MUSL setup completed with its default kernel and grub config. Initial issues are needing to get the root filesystem mounted rw and getting networking set up which setup never asks about). Compiled rc appears to be doing its job as expected, but I'm early in the testing.

07:25 <ppetrov^> hello future Crucial users

07:37 <SiFuh> Hello hardtab and softtab users

08:10 <ppetrov^> hm, i don't see stenur's ports at portdb anymore

08:22 <SiFuh> Hmm

08:26 <ppetrov^> did he anwer to your invitation?

08:28 <SiFuh> No

08:28 <SiFuh> ukky: https://dpaste.com/9RT6NZ5YU.txt

12:46 <ukky> SiFuh: those are gcc options, not make options. You could add them to CFLAGS.

12:49 <SiFuh> Oh I see

13:26 <farkuhar> Any response from stenur about working on Crucial with us?

13:28 <SiFuh> No

13:30 <farkuhar> Looking through stenur's ports, I'm reminded that he never had strong objections to linux-pam. In fact, some of his ports explicitly pass --with-pam to ./configure

13:33 <farkuhar> It's interesting to compare the way jue and stenur package their daemon ports. jue prefers to run pre-install to create the dedicated user and group, so that the Pkgfile build() can chown the files and directories. stenur seems to prefer waiting until post-install to reclaim the directories for the newly-created user/group.

13:36 <farkuhar> jue's approach wouldn't work so well in the pre-3.7 days, when "runscripts yes" was not the default in prt-get.conf. Without having run the pre-install script, you risk getting a footprint mismatch or a failed build.

13:48 <farkuhar> One advantage of stenur's approach is that you don't need fakeroot to build ports as an unprivileged user, provided you pass --uid 0 --gid 0 to the bsdtar command that archives $PKG. When contents under $PKG are expected to have owner/group different from root/root (as in jue's daemon ports), then fakeroot is the only way to build them as an unprivileged user.

14:30 <SiFuh> ukky: Got a small issue, maybe you'd know. if I use sys/reboot.h when I do halt -p it shutsdown the system but not completely off. If I use linux/reboot.h everything functions as it should.

14:31 <SiFuh> The other issue is linux/reboot.h isn't POSIX compliance like sys/reboot

14:32 <SiFuh> Oohh, there is another feature in sys/reboot.h hold on, I will try that.

14:34 <SiFuh> ukky: Nevermind, I switched from RB_HALT_SYSTEM to RB_POWER_OFF

15:10 <dlcusa> farkuhar, that sounds like a packaging standard regarding that difference might be a good idea.

17:15 <SiFuh> ukky: gcc -Wall -Werror -Wpedantic -ansi -pedantic -pedantic-errors -std=iso9899:1990 -D_POSIX_C_SOURCE=200809L -o rc rc.c This is fine.

17:15 <SiFuh> gcc -Wall -Werror -Wpedantic -ansi -pedantic -pedantic-errors -std=iso9899:1990 -D -D_POSIX_C_SOURCE=200809L -o rc rc.c <-- This with the extra -D that you had complains about missing headers that are there heh

17:20 <ukky> SiFuh: it was your interpretation of my '-D _POSIX_SOURCE'

19:56 <farkuhar> stenur never had strong objections to linux-pam, even going so far as to write a port for pam_xdg. Somehow I don't see stenur joining the crusade against eudev either.

20:00 <farkuhar> More plausible is that stenur would support the effort to reintroduce ./configure, in all those core ports that started relying on meson or cmake instead.

20:05 <farkuhar> stenur's recent post to the ML recalled the policy against listing core ports on the dependency line unless they're dynamically linked in. jue seemed resistant to stenur's suggestion of relaxing that policy. But opt/llvm actually does list some core ports as dependencies, in apparent violation of the aforementioned policy.

20:07 <SiFuh> farkuhar: Quite a few of olutmies have dependencies I saw to core ports

20:08 <farkuhar> https://lists.crux.nu/archives/list/crux@lists.crux.nu/message/SOWH5HEWLPDP7YIZYARIJVDNOJVIBNSB/ is the message I'm referring to, for those who aren't subscribed to the ML.

20:10 <ukky> If I would create a new distro, even core ports would be on dependency line.

20:15 <SiFuh> ukky: I think so too to be honest.

20:19 <farkuhar> SiFuh: ukky: So jue said "no" to stenur all those years ago, but quite a few of olutmies' ports even today have build-time dependencies in core that aren't dynamically linked in. Why the double-standard?

20:20 <SiFuh> farkuhar: Because there are no standards anymore.

20:24 <farkuhar> Or rather, there are multiple standards, and each maintainer is free to decide which precedent to follow. Consider the different ways of creating a dedicated user for a daemon port (opt/chrony versus stenur/lighttpd, as a concrete example).

20:32 <ukky> Ideally, user/group should only be created by pkgadd. Ideally, pkgmk should have a way to specify user/group other than root in .footprint.