Marquel | morning. | 19:54 |
---|---|---|
Arsen | ello! | 20:12 |
Arsen | 8PM here | 20:12 |
Marquel | Arsen: here too. ;) | 20:13 |
Arsen | quite a sleep schedule | 20:13 |
Marquel | given that last night was down to three hours, this night ends pre-dawn and there's labour ahead -yes. | 20:14 |
bb|hcb | Morning, Marquel :) | 20:52 |
bb|hcb | Ariadne, Arsen, blueness: let's coordinate how the pending PRs are reviewed - I prefer not to push them without a review... But in case noone volunteers, I will have to :) | 22:04 |
Arsen | uh, I have limited time generally these days, what'd you like me to do? | 22:05 |
Ariadne | i will look at them tonight | 22:05 |
bb|hcb | Basically I need an ACK/NAK for the PRs; I have already handled the ones from other people | 22:06 |
Ariadne | in future feel free to ping me with a URL to a PR | 22:06 |
bb|hcb | https://github.com/eudev-project/eudev/pulls | 22:06 |
bb|hcb | Arsen: I understand :) no worries, take your time | 22:07 |
Ariadne | syncs look ok | 22:08 |
Ariadne | bitwise looks ok | 22:09 |
Ariadne | the only one i’m skeptical on is the selinux one | 22:10 |
bb|hcb | Can you also mark them as reviewd on GH? | 22:10 |
bb|hcb | selinux_context_t is char * | 22:11 |
Arsen | iirc security_context_t was a thrown out alias | 22:11 |
bb|hcb | it got deprecated as a type, but the API is not deprecated | 22:11 |
Arsen | yes, that's what I mean | 22:12 |
Arsen | Ariadne: https://github.com/SELinuxProject/selinux/blob/d155b410d4bbc90d28f361b966f0429598da8188/libselinux/include/selinux/selinux.h#L16-L21 | 22:12 |
Arsen | lgtm | 22:12 |
Ariadne | ok then :) | 22:13 |
Arsen | ty :D | 22:13 |
Ariadne | in future please include references to such things :) | 22:14 |
bb|hcb | Sure, thanks for the review :) | 22:14 |
Ariadne | alpine doesn’t use selinux so i do not know anything about it | 22:14 |
bb|hcb | I am involved (willingly or not) with quite the variety of different types of systems, so lots of unwanted knowledge comes my way... | 22:15 |
Arsen | selinux is only slightly annoying, it's fiiiiiiine | 22:16 |
bb|hcb | I believe that with the PR from slicer69 (about quotes handling in udev rules) we are really much closer to the upstream behaviour now | 22:17 |
bb|hcb | My plan is to check the diff on the rest of the files and file some more PRs, in case there are things to merge | 22:17 |
bb|hcb | And maybe it will be a time for a release afterwards? | 22:18 |
Arsen | that'd be quite nice | 22:19 |
bb|hcb | Ugh, I forgot one step before that - review and merge the distro specific patches. gnu_srs: can you propose which of the Devuan patches are good for upstream merge and which ones are not? | 22:20 |
bb|hcb | Ariadne: about alpine patches, is this the place to look: https://git.alpinelinux.org/aports/tree/main/eudev?h=master - are these changes suitable for upstream? | 22:24 |
Ariadne | load-fbcon probably is | 22:46 |
Ariadne | default-rules is designed to make /dev/kvm 0660 | 22:46 |
bb|hcb | I can't say; from the point of view of my machines both are OK, but we need to consider other (unknown to us) use cases | 22:49 |
Arsen | what's kvm by default otherwise? | 23:03 |
Arsen | 0660 on my machine too, I don't remember doing any modifications | 23:04 |
bb|hcb | mine is 0666, root:kvm | 23:04 |
Arsen | gentoo doesn't apply patches to add KERNEL=="kvm", GROUP="kvm", MODE="0660" | 23:04 |
Arsen | presumably, that comes from systemd | 23:04 |
Arsen | hm, I don't see it in that tree | 23:05 |
Arsen | oh, I'm looking at my stripped tree | 23:05 |
bb|hcb | and there isn't any kvm related rule | 23:06 |
Arsen | ./rules.d/50-udev-default.rules.in | 23:07 |
Arsen | 93:KERNEL=="kvm", GROUP="kvm", MODE="{{DEV_KVM_MODE}}", OPTIONS+="static_node=kvm" | 23:07 |
Arsen | it's there | 23:07 |
Arsen | we can probably add that | 23:07 |
Arsen | meson_options.txt | 23:07 |
Arsen | 283:option('dev-kvm-mode', type : 'string', value : '0666', | 23:07 |
Arsen | ^ the default on systemd | 23:07 |
Arsen | realistically, it's probably fine to keep it 0660, it can be overridden | 23:08 |
bb|hcb | I agree that a more restricted approah upstream is better; that can be tuned by any distro; we only need to mention in release changelog so the change wouldn't go unnoticed | 23:09 |
Arsen | sure | 23:09 |
bb|hcb | gnu_srs: What is your opinion on the fbcon patch? (https://git.alpinelinux.org/aports/tree/main/eudev/load-fbcon.patch) | 23:19 |
Arsen | Ariadne: do you guys build udev with libkmod? | 23:21 |
bb|hcb | # we need disable kmod til busybox depmod generates binary formats: | 23:23 |
bb|hcb | --enable-kmod \ | 23:24 |
bb|hcb | looks like yes, but there is a comment, that implies that was not always the case | 23:24 |
Arsen | in that case, I'm not sure if that kmod is necessary | 23:24 |
Arsen | I'm sure ariadne can elaborate | 23:24 |
bb|hcb | just checked, devuan builds eudev with the upstream default for libkmod-dev (default is on) | 23:27 |
bb|hcb | i have also skimmed through the patches - there are 11, most are with debian/ubuntu origin; the only devuan specific one is reversing the default of network interface names | 23:38 |
bb|hcb | IMO, most of those are good for inclusion; maybe leaving interface default to be a distro decision is the best way | 23:39 |
bb|hcb | i see no point in changing the upstream default | 23:40 |
bb|hcb | let's wait for gnu_srs to say what approach is best for those patches | 23:40 |
Arsen | link? | 23:41 |
bb|hcb | w8 | 23:43 |
bb|hcb | https://git.devuan.org/devuan/eudev/src/branch/suites/unstable | 23:44 |
bb|hcb | patches are in debian/patches, only the ones in debian/patches/series are applied | 23:45 |
Arsen | add_hwdb.7.patch seems applicable so far, maybe | 23:47 |
Arsen | udev-hwdb-Support-shipping-pre-compiled-database-in-.patch check if this was happening upstream somehow | 23:48 |
Arsen | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=657948 seems like it should be addressed on the installed side, honestly | 23:49 |
Arsen | I dont think Load-sg-module-from-80-drivers.rules.patch ought to be merged | 23:49 |
Arsen | Updates-to-default-device-permissions.patch is likely distro specific | 23:49 |
Arsen | https://git.devuan.org/devuan/eudev/src/branch/suites/unstable/debian/patches/Set-default-polling-interval-on-removable-devices-as.patch should be checked with upstream too, this sounds like it was likely raised to sysd | 23:51 |
bb|hcb | these are really old ones, coming from debian packaging of udev for wheezy (kernel 3.1.x) | 23:52 |
Arsen | likely addressed upstream then | 23:52 |
bb|hcb | and eudev in devuan is 3.2.9, which is pretty close to master branch | 23:52 |
bb|hcb | not master, but 3.2.10 | 23:53 |
Arsen | Always-install-non-arch-specific-files-in-lib-udev.patch hm, what files are these? | 23:53 |
Arsen | reverse-net-ifnames-logic.patch :^) i already talked about this | 23:53 |
Arsen | man_udev.xml.patch is fine too, it's just a typo | 23:53 |
bb|hcb | we didn't touch anything related, so close to master too | 23:53 |
bb|hcb | that is debian terminology for arch independent stuff; in this case rules live in /usr/lib/udev | 23:54 |
bb|hcb | /lib/udev, sorry | 23:55 |
Arsen | ah I see | 23:55 |
Arsen | /lib rather than /lib64 | 23:55 |
Arsen | that's also down to the distro frankly | 23:55 |
bb|hcb | distro specific | 23:56 |
bb|hcb | lib64 --> lib on debian and derivates | 23:56 |
Arsen | yeah | 23:56 |
bb|hcb | and its use is discouraged... now i see that it is not a link and contains ld-linux*.so only | 23:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!