<    March 2017    >
Su Mo Tu We Th Fr Sa  
          1  2  3  4  
 5  6  7  8  9 10 11  
12 13 14 15 16 17 18  
19 20 21 22 23 24 25  
26 27 28 29 30 31
01:11 fekepp joined
01:16 czart__ joined
02:48 s33se_ joined
03:36 blueness joined
04:06 asie joined
04:11 BitL0G1c joined
04:13 nasonfish joined
05:45 minimalism joined
07:05 vakartel joined
08:10 tty` joined
08:30 blueness joined
08:34 BitL0G1c joined
08:44 t0mmy joined
08:54 atmoz joined
09:09 t0mmy joined
09:36 stwa joined
09:45 <ncopa> something is broken with openjdk7
09:45 <ncopa> the build server hangs during build
09:50 royger joined
10:16 rrx joined
10:21 <ncopa> new version of icedtea seems to fix it
10:53 <pickfire> n
10:53 <pickfire> ncopa: Can we move s-nail to community? I think that is very stable.
10:55 <ncopa> where is s-nail?
10:55 <ncopa> it does not seem to be in testing
10:55 <ncopa> $ git log --grep s-nail
10:55 <ncopa> does not show anything
11:01 <ncopa> i had a copy in my "wip" dir though
11:12 t0mmy joined
11:12 vakartel joined
11:47 blueness joined
12:19 gromero joined
12:29 ferseiti joined
12:32 leitao joined
12:35 farosas joined
12:49 <ncopa> pickfire: i added s-nail package
12:49 <ncopa> im gonna push openssh 7.5 now
13:30 atmoz joined
13:38 dfs joined
14:21 t0mmy joined
15:09 blueness joined
15:13 fabled joined
15:15 <leitao> ncopa, would it be OK to skip check() for ppc64le during this 'bootstrap' phase?
15:17 <ncopa> possibly
15:17 <ncopa> are you hitting failing checks?
15:17 <leitao> for main/gd yes
15:18 <leitao> I usually fix make check tests, but, since we are in hurry for 3.6, I think we can do it later
15:18 <leitao> ncopa, creating a 'if' inside check() for ppc64le would be acceptable?
15:19 <ncopa> hm
15:19 <ncopa> <phone>
15:25 <* leitao> will grab something to eat. BRB
16:04 <ncopa> i think we dont want control that in the APKBUILD
16:04 <ncopa> we should probably skip the tests in /etc/abuild.conf
16:04 <ncopa> ...if that is possible
16:05 <ncopa> kaniini: is it possible to disable abuild check systemwide?
16:06 <pickfire> ncopa: Huh? I think I had sent one patch for s-nail
16:07 <kaniini> ncopa: options="!check" in abuild.conf may do it, also ABUILD_BOOTSTRAP=1
16:07 <pickfire> https://github.com/alpinelinux/aports/pull/1020
16:07 <pickfire> It was there but overwritten.
16:08 <pickfire> I hate github.
16:08 <ncopa> :)
16:08 <ncopa> its in testing now though
16:08 <pickfire> ncopa: No
16:08 <pickfire> Don't do it that way
16:08 <ncopa> no?
16:08 <pickfire> Ugly
16:08 <pickfire> I use s-nail to replace mail
16:09 <ncopa> do a new PR then
16:09 <duncaen> dont look at the history of the suid binary :P
16:09 <ncopa> can we get rid of the suid?
16:09 <duncaen> i think its not necessary in most setups
16:09 <ncopa> ok, maybe have it in separate package
16:10 <ncopa> so you'll have to work extra to actually get it
16:10 <pickfire> ncopa: http://ix.io/pe7
16:11 <duncaen> what do you patch?
16:11 <pickfire> ncopa: Ah, you missed the check.
16:12 <ncopa> duncaen: patch?
16:12 <pickfire> ncopa: Why not /usr/libexec?
16:13 <duncaen> pickfire: what do you patch in s-nail
16:13 <pickfire> duncaen: SID= NAIL=mail
16:13 <pickfire> As well as check()
16:14 <duncaen> ah, we have a patch in void for cross compiling, not sure if you need it
16:14 <ncopa> not yet i suppose
16:15 <ncopa> pickfire: how about symlink mail -> s-nail?
16:15 <ncopa> so you can use s-nail too if you want
16:15 <duncaen> we have NAIL=mail and symlinks
16:15 <ncopa> k
16:16 <pickfire> ncopa: That's a nice idea.
16:18 <ncopa> duncaen: how does that alternatives thing work?
16:18 <pickfire> What's the idea of changing excuberent tags to https://ctags.io/?
16:18 <pickfire> Universal ctags
16:18 <ncopa> pickfire: https://github.com/voidlinux/void-packages/blob/master/srcpkgs/s-nail/template
16:18 <pickfire> duncaen: Alternative as in void?
16:19 <pickfire> ncopa: I know that, I tried void before alpine.
16:19 <duncaen> we have xbps-alternatives, it just creates symlinks on install and if you use xbps-alternatives to switch it http://sprunge.us/HSJW
16:19 <pickfire> But alpine doesn't have that right?
16:19 <ncopa> we dont
16:20 <ncopa> thats why i ask
16:20 <ncopa> we have talked about implementing it
16:20 <duncaen> mail is the alternative group, second level the package, third level the linked file, source:target, if the source is relative its created in the target dir
16:20 <pickfire> Oh no, the reason why I don't use void is because of that xbps-*
16:20 <pickfire> Too long to type.
16:20 <duncaen> target:source
16:20 <pickfire> Their xbps-packages is too powerful.
16:21 <pickfire> Compared to alpine which is limited.
16:21 <duncaen> yea xbps-src is kind of big and hard to improve without being careful
16:21 <ncopa> duncaen: that looks interesting
16:21 <pickfire> ncopa: Ah, I long for the alternatives thing.
16:21 <ncopa> will have a closer look at that later
16:22 <kaniini> ncopa: did ABUILD_BOOTSTRAP work?
16:22 <pickfire> Because I can go python2 -> python3
16:22 <ncopa> kaniini: ah, just setting ABUILD_BOOTSTRAP should do it
16:23 <duncaen> one bad expirience was that we allowed to set bash as alternative to sh and the readline update broke the scripts that run at install time for users that used bash as sh :D
16:23 <pickfire> A
16:24 <pickfire> duncaen: IIRC there is duncaen in xbps as well.
16:24 <pickfire> Are you dev from there?
16:24 <ncopa> :D
16:24 <pickfire> Interesting
16:24 <duncaen> yes, im here to steal patches and share my fuckups :D
16:24 <pickfire> haha
16:25 <pickfire> duncaen: Lots of suckless suggested me to get void as well.
16:25 <pickfire> I like alpine but I don't like to stick everything to busybox.
16:25 <pickfire> Like I rather prefer ubase.
16:25 <pickfire> And iproute2
16:26 <duncaen> i like apk more after working with xbpss source and reading apk-tools a bit and i really like the stable branches
16:26 <ncopa> requires some work
16:26 <pickfire> duncaen: Sorry that I had switched to alpine from void.
16:26 <ncopa> maintenance work
16:27 <pickfire> ncopa: I hope alpine is in gsoc
16:27 <pickfire> So I can join.
16:27 atmoz joined
16:27 <ncopa> duncaen: we are working on ppc64le and s390x support now
16:28 <pickfire> ncopa: How do you do new support?
16:28 <ncopa> duncaen: iirc you beat us with the aarch64 support :)
16:28 <pickfire> I would like to see alpine running on various arm devices
16:28 <pickfire> Like odroid and pi.
16:29 <pickfire> I have quite a few models of those.
16:29 <duncaen> nice, i dont think we can add more architectures at the moment, qt5 is building currently since >11hours we dont have the resources :D
16:29 <ncopa> pickfire: we build crosscompiler, we build the base tools and then we set up a native build server
16:29 <ncopa> duncaen: i saw in #xbps
16:29 <pickfire> pi 1, pi 2, pi 3, c1, c2, u3
16:29 <ncopa> you need more build servers
16:29 <pickfire> ncopa: Yes
16:29 <pickfire> I would like to build on arm devices.
16:30 <duncaen> and we cross compile everything that is not x86 :s
16:30 <pickfire> ncopa: I think I can as well add travis-ci to my list of build servers.
16:30 <ncopa> duncaen: i know, i think we'd like to cross compile too
16:30 <ncopa> if not all, alot
16:31 <ncopa> pickfire: cant you use armhf packages?
16:31 <duncaen> why, isnt the arm64 box enough to build all arm packages?
16:31 <pickfire> Can, but not optimized like the one provided in alarm.
16:32 <pickfire> ncopa: I didn't know you are a spy in #xbps
16:32 <ncopa> pickfire: spy? im a contributor...
16:32 <pickfire> wow
16:32 <duncaen> would be a waste of time if we wouldnt communicate/share patches
16:32 <ncopa> https://github.com/voidlinux/xbps/pull/236
16:32 <ncopa> yeah
16:33 <pickfire> I am curious why don't we help each other?
16:33 <ncopa> lol
16:33 <ncopa> sorry pickfire but sometimes you just dont sound that smart :)
16:33 <ncopa> we do help eachother
16:33 <ncopa> why do you think i hang along in #xbps and duncan in here?
16:34 <pickfire> ncopa: Should I join there as well?
16:34 <ncopa> its up to you
16:34 <pickfire> Ah
16:35 <pickfire> ncopa: I wonder if we can do a -man and -doc separation because that's what I am really hoping for and as well no distro provides.
16:35 <ncopa> why?
16:35 <pickfire> Not just me, quite a few others as well.
16:35 <pickfire> -doc is total useless.
16:35 <pickfire> Fat=
16:36 <ncopa> not always
16:36 <ncopa> i think most packages have pretty small -doc?
16:36 <pickfire> I mean -doc generally contribute to the size of packages whereas -man wouldn't
16:36 <pickfire> And who would view an info page.
16:37 <duncaen> ah related to this, cross compiling and help2man sucks :D
16:37 <pickfire> duncaen: Do you guys do distributed compiling?
16:39 <duncaen> no we just have three build boxes one for glibc one for musl for all architectures and one for both arm64 libcs
16:40 <pickfire> ncopa: What about ours?
16:40 <duncaen> and updates like the current one, icu, libevent take ages rebuilding all the stuff with chromium/webkit bundled all browsers, libreoffice, qt
16:41 <pickfire> Yes
16:41 <duncaen> http://build.alpinelinux.org/
16:41 <ncopa> we have 1 big for x86_64, one big for x86, one aarch64 and one for armhf
16:41 <ncopa> the stable builders are containers
16:41 <pickfire> ncopa: Just 1?
16:41 <ncopa> yes one for each arch basically
16:42 <ncopa> i think we do the stable armhf branches on a wandboard
16:42 <duncaen> you dont use the aarch64 box to build arm packages?
16:42 <pickfire> duncaen: No
16:42 <ncopa> we have one x-gene that builds armhf
16:42 <ncopa> 8 core server
16:42 <ncopa> pretty good actually
16:42 <pickfire> ncopa: Why don't we use xbps build system?
16:42 <pickfire> I mean xbps build tracking system
16:43 <duncaen> and the thunderx?
16:43 <ncopa> and we havea thunder-x with 48 cores for aarch64
16:43 <ncopa> the reason is that the thunder-x cannot run 32bit arm
16:43 <duncaen> ah its just doing aarch64
16:43 <pickfire> Wow
16:43 <pickfire> 48 cores
16:43 <duncaen> oh didnt know this i thought this would work :s
16:44 <pickfire> ncopa: How large are alpine mirrors?
16:44 <ncopa> http://b2b.gigabyte.com/Density-Optimized/H270-T70-rev-100#ov
16:45 <ncopa> depends on which mirror
16:45 <pickfire> I like void linux sites, although bloated but served quite fast.
16:46 <duncaen> its github, thats why we dont have a https cert
16:46 <ncopa> duncaen: i think the worst thing that has been compiled when it comes to resource use
16:46 <ncopa> is llvm
16:46 <pickfire> Huh
16:46 <ncopa> the problem was link time
16:47 <ncopa> compiling with -j24 was no problem
16:47 <ncopa> but linking with -j24 was
16:47 <ncopa> i think LTO consumed extremely with memory
16:47 <duncaen> stop, i feel so bad for just having -j4 and building qt5 for >11hrs :D
16:48 <ncopa> :)
16:48 <ncopa> i think your problem is memory use
16:48 <ncopa> we had issues with armhf on wandbord
16:48 <ncopa> problem was not cpu but memory
16:48 <ncopa> and it ended up constantly swapping
16:48 <pickfire> Ah
16:49 <pickfire> ncopa: Who sponsors those boards?
16:49 <ncopa> gigabyte: https://alpinelinux.org/sponsors/
16:49 <ncopa> oh
16:49 <ncopa> we have a ppc64le and s390x from IBM too now
16:50 <pickfire> ncopa: wearetriple.com is quite nice too
16:50 <pickfire> Their mirror is nice.
16:51 <pickfire> © Copyright 2016 Alpine Linux Development Team all rights reserved | Privacy Policy
16:51 <pickfire> Lol
16:51 <pickfire> 2016
16:58 <mitchty> pickfire: http://www.contentious.com/2007/01/07/copyright-notice-is-the-year-really-necessary/
17:00 <mitchty> lawyers at my job have said don't change it once its there, the law doesn't behave the way you think
17:01 <pickfire> Ah
17:01 <mitchty> or leaving it out is better apparently, but talk to a real lawyer, even having just a copyright symbol and not the literal word "Copyright" can invalidate a notice
17:02 <mitchty> from what i remember/understand, changing the dates can cause issues with you basically saying the copyright only applies to things past that date, which means you could be stating copyright doesn't apply to earlier versions, but i'm not a lawyer and my eyes were more glazed over than a donut by that point
17:03 <pickfire> mitchty: What about © 2006-2016 suckless.org community | Impressum
17:03 <skarnet> I can confirm what you're saying
17:03 <skarnet> donuts *are* more interesting than copyright notices
17:03 <pickfire> Yeah
17:04 <pickfire> At least they are tasty
17:04 <pickfire> I have friends that study lawyar T_T
17:04 <pickfire> lawyer*
17:05 <mitchty> pickfire: i'm not a lawyer but same thing, you could be stating only stuff between 2006 and 2016 is under copyright, the rest not, again lawyer those questions, copyright is weird and not intuitive
17:05 <pickfire> Ah
17:06 <pickfire> So the one befor and after does not apply.
17:06 <mitchty> but i do know the copyright symbol isn't recognized as a valid statement of copyright, it has to be iirc Copyright or Copr or something else
17:06 <pickfire> Oh
17:06 <pickfire> But most used ©
17:06 <mitchty> the symbol not being legally recognized got drilled into me during that meeting :)
17:06 <pickfire> Rather "Copyright"
17:07 <mitchty> yeah, i dunno, its all greek to me, i just do Copyright now as its the most straightforward and "lawyer approved"
17:07 <pickfire> That's new to me as well.
17:07 <pickfire> I think I need to change my site as well.
17:07 <pickfire> Ah
17:08 <pickfire> Mine is CC-BY
17:08 <_ikke_> But don't you get copyright by default?
17:08 <mitchty> let me look up the doc
17:08 <pickfire> No returns
17:08 <pickfire> _ikke_: Yes
17:08 <pickfire> _ikke_: Others can get copyright by default as well.
17:11 <mitchty> looks like they updated it
17:11 <mitchty> The word Copyright, or the abbreviation Copr., or the symbol ©.
17:11 <mitchty> it was (C) i was thinking of rather sorry
17:12 dsabogal joined
17:12 <mitchty> ok, reading through this legal stuff is making me want to go to sleep again
17:14 <pickfire> :)
17:15 dsabogal joined
17:18 rdutra joined
17:19 <TemptorSent> ncopa, fabled : How many people would likely grab pitchforks and torches if I removed bootchart from the default init?
17:49 <kaniini> probably very few
17:49 <kaniini> tbh
17:51 blueness_ joined
17:51 <TemptorSent> Goodbye bootchart, you can come back if someone really needs you :)
17:52 <scadu> ncopa: I didn't now you're involved in Void. that's nice! some time ago I was thinking about it : >
17:53 <fabled> TemptorSent, yeah, go ahead. I was working earlier to fit in bootchart2 in instead. but it's still work-in-progress.
17:53 <TemptorSent> Anyone still reliant on 'root_size' out there?
17:54 <kaniini> BitL0G1c: community/modemmanager could really use some cleanup
17:54 <TemptorSent> fabled: I've got init basically torn apart and I'm starting to sort out what's actually required in all cases vs. what is only useful with given initfs features.
17:54 <kaniini> BitL0G1c: i fixed community/networkmanager, it was caused by corrupt dependency data being produced by community/modemmanager (and also a bug in pkgconf)
17:55 <TemptorSent> fabled: What kind of hooks does bootchart2 need? Also, is there any objection to spitting the init output to dmesg?
17:55 <kaniini> (and for that matter, also a bug in pkg-config)
17:56 <TemptorSent> fabled: I'd like to strip everything obsolete or unnecessary out, then add back just what's needed when we build.
17:57 <fabled> TemptorSent, bootchart2 just wants to be started as early as possibly; it infact normally recommends to override init= so it runs even before the initscript
17:57 <TemptorSent> fabled: Okay, is that somewhere I can take a look at it?
17:57 <fabled> TemptorSent, https://github.com/xrmx/bootchart
17:58 <TemptorSent> fabled: Also, do we use the keymaps properly currently? Should there be a cmdline option for selecting it?
18:00 <TemptorSent> fabled: It looks like bootchart2 is less invasive, this might be doable without a headache :)
18:00 <fabled> yes
18:01 <BitL0G1c> kaniini - ok thank you
18:01 <TemptorSent> fabled: Let me know if there are any other features I should consider while reworking this.
18:03 <TemptorSent> fabled: bootchartd isn't exactly slim, wow.
18:04 <fabled> well, it's semi-crappy code. it could be slimmer
18:04 <fabled> i was tempted to rewrite it, but i dunno.
18:04 <TemptorSent> fabled: What information do we actually need out of it?
18:04 <^7heo> ENOTIME?
18:05 <fabled> it makes log of all processes starting, stopping, their parent/child relations, cpu time used, wait states etc.
18:05 <fabled> it's just collecting info from kernel via netlink apis, and dumping it to files
18:05 <kaniini> scadu: i don't think he is, just that sometimes we share info
18:06 <TemptorSent> fabled: Got it. It's only marginally useful in the initrd then, right?
18:06 <fabled> TemptorSent, it's crucial. the point is to collect data what happens from power-on until you get boot prompt
18:07 <scadu> kaniini: fair enough. it's better then shitting on each other which happens sometimes.
18:07 <TemptorSent> fabled: Hmm, okay -- I'm guessing it's the timing information that you're looking for at that point? The rest is pretty much loggable...
18:07 <fabled> yes
18:08 <TemptorSent> fabled: I'm looking through to figure out what I need to do to support it cleanly.
18:08 <fabled> don't worry about it for now
18:08 <fabled> i think it's just few added files to image, and we can enable it via kernel init= cmdline option
18:08 <TemptorSent> fabled: Yeah, I'm just making sure I have a place to hook it.
18:09 <TemptorSent> fabled: It looks like it needs help with the handoff.
18:09 <TemptorSent> fabled: Not exactly well documented ;)
18:15 <gromero> for sure libssl.so comes from libressl, right?
18:16 <skarnet> yes, but why do you need to make that assumption?
18:21 <kaniini> fabled: what is your thoughts on haskell testing->community transition
18:23 <gromero> skarnet: looks like something is missing on ppct64le:
18:23 <gromero> alpine-ppc64le:~/git/aports/main/libressl$ apk info --contents libressl
18:23 <gromero> libressl-2.4.5-r0 contains:
18:23 <gromero> etc/ssl/x509v3.cnf
18:23 <gromero> etc/ssl/openssl.cnf
18:23 <gromero> etc/ssl/cert.pem
18:23 <gromero> usr/bin/openssl
18:23 <gromero> that's all
18:23 <gromero> mariabd build brakes...
18:23 <TemptorSent> gromero : Check the libs -- probably a missing dep
18:24 <skarnet> indeed, the shared libs seem to be missing :D
18:24 <gromero> TemptorSent: skarnet ok
18:24 <TemptorSent> gromero: apk add libressl-2.4-libssl
18:26 <duncaen> libressl-dev?
18:27 <skarnet> no, the shared libs are production, not development
18:27 <duncaen> but doesnt he try to build something?
18:28 <skarnet> if the .so's are missing there's a problem whether or not he's building
18:28 <kaniini> mariadb did not fail because of libressl
18:28 <kaniini> /home/buildozer/aports/main/mariadb/src/mariadb-10.1.22/storage/xtradb/include/ut0ut.h:89:30: fatal error: sys/platform/ppc.h: No such file or directory
18:28 <duncaen> yea they are in the subpackages -libssl -libcrypto, but -dev would pull them all in
18:29 <kaniini> he just need install libressl-dev if he is testbuilding
18:29 <kaniini> try: abuild installdeps
18:29 <skarnet> why the hell are there subpackages now
18:29 <gromero> alpine-ppc64le:~/packages/main/ppc64le$ apk info --contents libressl2.4-libssl
18:29 <gromero> libressl2.4-libssl-2.4.5-r0 contains:
18:29 <gromero> lib/libssl.so.39
18:29 <gromero> lib/libssl.so.39.0.0
18:29 <gromero> usr/lib/libssl.so.39
18:29 <gromero> usr/lib/libssl.so.39.0.0
18:29 <skarnet> you can't use libressl without both libssl and libcrypto anyway, why fragment the package
18:29 <kaniini> you can use libcrypto without libssl
18:30 <skarnet> yeah, that's really useful given that libssl is what, 10% of libcrypto's size?
18:32 <TemptorSent> Why do I feel like I'm implementing a sane version of dracut?
18:35 <skarnet> tbh that's probably what you're doing :P
18:35 <TemptorSent> bash for initrd -- because THAT'S sane, right?
18:36 <skarnet> it actually is if you don't care about size
18:36 <skarnet> (because portability isn't an issue)
18:36 <skarnet> obviously if you're concerned about resources, bash is a no-no
18:36 <TemptorSent> skarnet: Size, or speed, or number of deps, or just about anything else.
18:37 <TemptorSent> skarnet: I'm not even going to try for feature parity with dracut, but we might be able to at least support modular initram-init.
18:38 <TemptorSent> skarnet: Without 400 files!
18:38 <skarnet> keep initramfs simple. please.
18:38 <TemptorSent> skarnet: Yeah, that's the point -- init can be dead simple unless you NEED to do something different.
18:39 <duncaen> its not bash
18:39 <duncaen> they use bash just at build time, posix shell at boot time iirc
18:39 <TemptorSent> duncaen: Dracut? at least the bootchart module appears to want bash in the initfs.
18:40 <duncaen> maybe, never used it
18:41 <TemptorSent> Basically I'm cutting our initramfs-init down to just what's needed to establish the initramfs itself, then call whatever appropriate loader based on the kernel options.
18:42 <TemptorSent> duncaen: So the netboot code doesn't get tangled with the tmpfsroot code, nor the running from existing root.
18:43 <TemptorSent> Technically, we could just shove those in their own cpio archives and append them so we don't have anything floating around we don't need.
18:45 <duncaen> what is the problem iwth the current initramfs?
18:46 <TemptorSent> duncaen: Have you seen it? /usr/share/mkinitfs/initramfs-init
18:46 <TemptorSent> duncaen: Undebugable :)
18:47 <TemptorSent> duncaen: Also very difficult to add support for additional initfs features cleanly.
18:48 <TemptorSent> nfs/iscsi root support being a big one.
18:50 <TemptorSent> duncaen: Also support for things such as first-boot auto-install and full rescue mode is not currently doable.
18:50 <duncaen> ah ok
18:51 <duncaen> mkinitfs seemed sane compared to dracut
18:52 atmoz joined
18:54 <TemptorSent> duncaen: Yeah, mkinitfs has been cannabalized into my rework of mkimage (along with update-kernel). Perhapse it should be mkalpine :)
18:55 <TemptorSent> duncaen: The reason being is that mkinitfs had no idea how to figure out which packages it needed for a given feature, and the feautres were difficult to define with reasonable granularity.
18:55 <TemptorSent> duncaen: see https://github.com/TemptorSent/aports/tree/mkimage-refactor-scripts/scripts/mkimage
18:57 <TemptorSent> duncaen: The logic is nearing completion, with just a few more things to rearrange.
18:58 <TemptorSent> duncaen: Please take a look and let me know if there's anything major I've missed.
18:58 <duncaen> i guess local is portable enough by now
18:59 <duncaen> i skip quotes if i assign variables like foo=$bar
19:00 <TemptorSent> duncaen: I made sure there is no use of local outside of functions :) That's portable AFAIK
19:00 <duncaen> || ! warning: "ovl_create_link: ...
19:00 <duncaen> but local is not posix afaik
19:00 <TemptorSent> duncaen: I've run into gotchas with unquoted variable assignments in the past, but I can't recall which shell.
19:01 <duncaen> i think if its just one variable to another its ok to not quote them
19:02 <TemptorSent> duncaen: Nope, it can still get fubar if it's unquoted -- it will expand against the IFS
19:03 xming joined
19:03 xming joined
19:05 ahills joined
19:06 <gromero> kaniini: you are right ppc.h is a glibc dep and hence is missing. In mariadb-server, it's introduced here https://github.com/MariaDB/server/commit/64824a760d3ee4715d301dcdff541b66fac32992
19:07 <gromero> looking if musl provides the __ppc_get_timebase symbol...
19:12 <TemptorSent> Hmm, do we actually need anything other than bb-static in the initfs before we start loading APKs? IOOW, if we're booting to a fixed root, do we need anything else?
19:14 <skarnet> assume we don't. If it crashes when you test, you'll know. :P
19:15 TemptorSent joined
19:16 TemptorSent joined
19:22 tty joined
19:27 chris| joined
21:07 rrx joined
21:21 rokf joined
21:32 tmh1999 joined
21:54 <tmh1999> fcolista : can you update perl-test2-suite to either verion 67 or 70 ? Looks like current version 61 package got deleted in cpan.org and can only be retrieved via github
21:55 <tmh1999> fcolista : either 67 or 70 requires new perl deps, Term::Table Sub::Info
21:58 <kaniini> i'll look into it
21:58 <tmh1999> kaniini : thanks :)
22:19 gromero joined
22:44 emerson joined
23:02 stwa joined
23:16 czart_ joined