09:49 <jue> that was fast, I've reported the pam_lastlog2 issue yesterday, today we got a fix ;)

09:58 <jue> I've just added another issue because the pregenerated man-pages are not installed with meson, only with autotools

10:07 <beerman> Cool :) good stuff

14:18 <jaeger> I guess they're missing some unit testing upstream

14:21 <jaeger> Any thoughts on stenur's recent ML post about zstd support in pkgutils? Seems like a useful feature to add

14:38 <beerman> I feel very indifferent about it, surely wouldn't hurt, but also not very high up on my priority list

15:33 <beerman> ok. both lvm2 and cryptsetup will work if we omit LTO with util-linux

15:34 <beerman> not so nice to dance around, but if there is not anything else blocking the exchange to meson, I'd really like to switch it back

16:02 <jaeger> If it works properly, I have no objections

16:04 <beerman> Well, I have no use for cryptsetup anywhere, and I also don't use lvm2 on my private installs, so I can't test it out properly. Of course there is the testsuite with the code but I didn't run these (yet).

16:06 <beerman> But then again, unpaid hobby project, I am sure we have users that want to contribute running test suites, writing pull requests, etc., true to the CRUX spirit

16:06 <jaeger> If you like I can run it through an ISO build before the push

16:07 <beerman> I mean I feel confident enough thinking "what should break" but of course, as we saw with the autotools -> meson switch, this feeling can easily betray you :) I wouldn't have noticed it right away either if I did it

16:08 <beerman> a field test before going on another ride of the good ol' switcheroo might be approriate here :)

16:08 <beerman> just revert the last commit, disable lto and thats it

16:08 <beerman> of coures, rebuild all the libuuid.la and libmount.la etc references if you have them (again)

16:27 <jaeger> To clarify, revert the meson -> autotools commit, then disable lto, right? Not the pam_lastlog patch commit

16:29 <beerman> you can keep the patch, maybe it makes sense to pull the patch that is in util-linux-32 as well

16:29 <beerman> wait

16:32 <beerman> Pkgfile https://dpaste.com/9K5HHEHW9 syscalls.patch: https://dpaste.com/EMYC397GN footprint: https://dpaste.com/6WZ334Z7C

16:34 <beerman> that works for me with lvm2 and cryptsetup as is

16:37 <jaeger> ok, will test it on the ISO builder with a bootstrap

16:41 <beerman> cool :) lets see

16:43 <jaeger> Will take some time, not the fastest VM ever (it usually builds while I'm asleep so doesn't need much)

16:44 <beerman> no worries, we are not in a rush

16:47 <jue> I'd like to see man-pages in the util-linux package. If they don't fix the meson build it's easy for us to install them manually

16:57 <beerman> look, its the third individual that makes up the mysterious entity that makes all the decisions by themself

16:57 <beerman> yeah I missed the man pages. True, is there a ticket open for that already? I think I saw it

16:58 <jue> Yep, I did it

16:58 <beerman> cool

16:58 <jue> as I said already :)

16:59 <beerman> ah, sorry, I have my head around too many things at the same time :)

16:59 <jue> no worries :)

17:11 <farkuhar> probably better to install them manually anyway. The meson build with asciidoctor installed would create /usr/share/man/{man1/login.1,man8/sulogin.8}, which conflicted with shadow and sysvinit. Any fix they do upstream might still need patching on our end to avoid the conflict.

17:19 <jue> if so that's another bug in the upstream meson build: if we disable building the progs the man-pages shouldn't be installed

17:59 <beerman> ftr, i'd be fine with a pipeline generating these man pages for us, if we were to supply them ourselves and manually. but that kind of complexity I do not want to add into my personal workflow.

23:25 <jaeger> OK, the ISO bootstrap made it through stages 0 and 1 just fine, would have failed in stage 1 if it were going to