2017-08-14 06:44:29 <ncopa> #7681
2017-08-14 06:45:19 <ncopa> #7681
2017-08-14 06:45:46 <ncopa> the problem was that it didnt handle https
2017-08-14 06:53:08 <kaniini> ncopa: lua-aports dependency solving seems broken
2017-08-14 08:32:41 <ncopa> kaniini: i heard
2017-08-14 08:42:37 <clandmeter> ncopa can you fix the builders?
2017-08-14 08:43:01 <ncopa> ok
2017-08-14 08:50:30 <ncopa> should be ok now
2017-08-14 12:33:40 <danieli> ncopa: hvorfor bugger den?
2017-08-14 12:58:51 <PureTryOut[m]> kaniini: could you verify https://github.com/alpinelinux/aports/pull/2097 one more time for me?
2017-08-14 12:58:58 <PureTryOut[m]> *check
2017-08-14 13:10:57 <danieli> looks correct
2017-08-14 13:29:24 <andypost> ncopa, looks tests are disabled now so I reverted tests you commented out https://github.com/alpinelinux/aports/pull/2079
2017-08-14 18:32:01 <NK_> Does anybody know about untrusted signatures when running abuild-apk update?
2017-08-14 18:35:11 <NK_> I've been trying to use the mkimage script with the rpi profile to make an ISO.  When the script runs this abuild-apk update, the apkindex in the repo http://dl-cdn.alpinelinux.org/alpine/edge/main/armhf always comes up as having an untrusted signature which stops the mkimage script
2017-08-14 18:37:00 <NK_> For reference I've been using the alpine wiki page https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage
2017-08-14 19:24:40 <ncopa> stateless: bug and patch in fortify-headers: http://bugs.alpinelinux.org/issues/7681
2017-08-14 21:47:10 <kaniini> question
2017-08-14 21:47:28 <kaniini> say i am packaging something that is both python2 and python3 but runs as a daemon and i want to bundle an init script for it
2017-08-14 21:47:33 <kaniini> how should i handle that
2017-08-14 21:48:52 <_ikke_> If it's a daemon, don't you just choose which one you run?
2017-08-14 21:49:13 <_ikke_> you as a packager I mean
2017-08-14 21:49:18 <Shiz> does it include a library part
2017-08-14 21:49:42 <Shiz> if yes, package py-foo and py3-foo, with the daemon part as foo package using py3
2017-08-14 21:49:44 <Shiz> otherwise, just use py3
2017-08-14 21:49:46 <Shiz> that'd be my choice
2017-08-14 21:54:47 <mosez> kaniini: i have packaged cli tools with py3 only.
2017-08-14 21:55:04 <mosez> that's anyway the only version jirutka suggests :)
2017-08-14 21:55:05 <_ikke_> yeah, if it's not a library, then it should not matter
2017-08-14 21:55:23 <_ikke_> it's just a command you run, you should not care whether it's py2 or py3
2017-08-14 21:56:04 <mosez> btw, my prs are waiting for another review :P
2017-08-14 21:57:08 <kaniini> it's celery
2017-08-14 21:57:13 <kaniini> so it's a library and daemon
2017-08-14 21:57:15 <mosez> #1962, #1963, #1964, #1965, #1966, #1967, #1968, #2115 on github 8)
2017-08-14 21:57:20 <kaniini> that's the problem
2017-08-14 21:57:35 <mosez> oh... it's linking to bugs and not to github :(
2017-08-14 21:57:39 <kaniini> you might want to use celery with either py2 or py3, and the version you choose needs to
2017-08-14 21:57:45 <kaniini> match the version your app is using
2017-08-14 21:57:59 <kaniini> %1962
2017-08-14 21:58:06 <kaniini> hm
2017-08-14 21:58:21 <mosez> submitted github pull requests... idk if there is also an alias
2017-08-14 21:58:49 <_ikke_> did algitbot just came alive?
2017-08-14 21:59:47 <mosez> _ikke_: because of all the messages? :)
2017-08-14 22:04:33 <_ikke_> mosez: yea
2017-08-14 22:05:31 <_ikke_> mosez: But I just saw you linking to them
2017-08-14 22:23:15 <mosez> yes, didn't knew that algibot links it to bugs.alpinelinux.org
2017-08-14 22:23:45 <_ikke_> mosez: The reason was I heard messages earlier about it not responding
2017-08-14 22:23:53 <_ikke_> that's why it suddently looked 'alive'
2017-08-14 22:24:13 <mosez> ah, ok
2017-08-14 23:31:00 <foxkit> does db-doc really need to be 92 MB of C#, Java, and Rust docs that we don't even build language bindings for...?
2017-08-14 23:40:27 <Shiz> probably not :)
2017-08-15 19:22:54 <opendata> https://github.com/alpinelinux/aports/pull/2143
2017-08-15 19:32:44 <PureTryOut[m]> ooh nice
2017-08-15 20:06:47 <PureTryOut[m]> is Natanael Copa in this channel?
2017-08-15 20:10:10 <_ikke_> yes
2017-08-15 20:10:21 <_ikke_> but he probably is not online right now
2017-08-15 20:12:19 <PureTryOut[m]> what's his IRC name? so I can ping him later
2017-08-15 20:12:54 <_ikke_> ncopa :)
2017-08-15 20:13:00 <PureTryOut[m]> thanks!
2017-08-15 20:23:42 <danieli> it's 10.23 pm in our time zone
2017-08-15 20:23:47 <danieli> _ikke_: norsk du og?
2017-08-15 20:24:22 <PureTryOut[m]> Jeg snakker ikke Norsk 😞
2017-08-15 20:24:27 <danieli> hahahah
2017-08-15 20:25:10 <PureTryOut[m]> I actually know that sentence, no Google Translate 😉 "smøre", "takk". that's about all I know
2017-08-15 20:25:16 <mitchty> is og norwegian for to be?
2017-08-15 20:25:35 <PureTryOut[m]> "as well" I think
2017-08-15 20:26:15 <mitchty> ah, google gave me it as "and", that makes more sense then
2017-08-15 20:26:30 <danieli> mitchty: "og" is "and"
2017-08-15 20:26:43 <danieli> binding word
2017-08-15 20:26:57 <TBB> so ogre would translate as andre
2017-08-15 20:27:05 <danieli> wat
2017-08-15 20:27:11 <PureTryOut[m]> lol
2017-08-15 20:27:13 <danieli> 'ogre' would probably be troll, not sure
2017-08-15 20:27:18 <mitchty> ah, got it thanks danieli, the german I know helps... a tiny bit :)
2017-08-15 20:27:20 <danieli> anyway, this is #alpine-devel
2017-08-15 20:27:54 <ncopa> PureTryOut[m]: hi
2017-08-15 20:27:54 <danieli> mitchty: ich kann auch deutsch sprechen!
2017-08-15 20:28:33 <PureTryOut[m]> ncopa: ooh hi. I was wondering, since you maintain the poppler package. could you either enable qt5 support or make a separate (sub)package with support for it?
2017-08-15 20:28:43 <PureTryOut[m]> also, an update has been released for it on the 1st of august
2017-08-15 20:28:58 <ncopa> PureTryOut[m]: sounds like a good idea, but not now and not tomorrow
2017-08-15 20:29:08 <ncopa> likely not this week
2017-08-15 20:29:16 <ncopa> can you file an issue on bugs.a.o?
2017-08-15 20:29:25 <PureTryOut[m]> sure!
2017-08-15 20:29:29 <ncopa> and if i dont do it this week, please ping me next
2017-08-15 20:29:35 <mitchty> danieli: auch gut, aber Ich spreche kein Norwegisch
2017-08-15 20:29:37 <PureTryOut[m]> I will, thanks!
2017-08-15 20:29:49 <danieli> mitchty: i gathered as much
2017-08-15 20:30:51 <danieli> ncopa: got any resources for quickly getting into Alpine development / maintaining? I'm not too familiar with the ecosystem and paradigms
2017-08-15 20:31:41 <ncopa> we build packages in the aports git repo, where we have shell build scripts, APKBUILDs
2017-08-15 20:31:53 <danieli> gotcha - it isn't harder than that?
2017-08-15 20:31:54 <ncopa> which is parsed with 'abuild'
2017-08-15 20:32:15 <ncopa> abuild -r
2017-08-15 20:32:30 <ncopa> you need create an abuild key
2017-08-15 20:32:43 <danieli> we can do this in PM, or otherwise just link me if you have some resources online
2017-08-15 20:32:53 <ncopa> abuild-keygen -a -i
2017-08-15 20:33:02 <ncopa> there should be lots of stuff in the wiki
2017-08-15 20:33:06 <ncopa> wiki.alpinelinux.org
2017-08-15 20:33:09 <danieli> gotcha, will check it out
2017-08-15 20:56:24 <_ikke_> danieli: nei
2017-08-15 20:56:49 <danieli> _ikke_: ok
2017-08-16 07:59:41 <PureTryOut[m]> kaniini: could you do one last check of https://github.com/alpinelinux/aports/pull/2097?
2017-08-16 14:06:19 <fcolista> fabled, ncopa: can some of you delete namecoin-0.13.1rc1  package ?
2017-08-16 14:07:48 <clandmeter> fcolista, what is wrong with namecoin tarball?
2017-08-16 14:08:04 <fcolista> upstream changed something without relasing a new tarball
2017-08-16 14:08:13 <fcolista> so checksum does not change
2017-08-16 14:08:17 <fcolista> *change
2017-08-16 14:08:19 <fcolista> and fails
2017-08-16 14:08:37 <clandmeter> any reference to this change?
2017-08-16 14:08:43 <fcolista> but builders does not download it
2017-08-16 14:08:48 <fcolista> no reference
2017-08-16 14:11:25 <clandmeter> the tag is from on Aug 29, 2016
2017-08-16 14:11:43 <fcolista> yes I saw that
2017-08-16 14:12:04 <clandmeter> so the previous build was from before that time?
2017-08-16 14:12:37 <fcolista> right
2017-08-16 14:12:42 <clandmeter> its weird
2017-08-16 14:12:52 <clandmeter> cause it had a bump  after that tag
2017-08-16 14:12:59 <clandmeter> and it build correctly
2017-08-16 14:13:11 <clandmeter> which means the github tarball has changed
2017-08-16 14:15:46 <fcolista> another option is that the build has been triggered by an APKBUILD update
2017-08-16 14:16:20 <fcolista> github somehow changed the package from aug 2016
2017-08-16 14:16:34 <fcolista> but we nebver realized because the package was never rebuilt
2017-08-16 14:17:20 <fcolista> well
2017-08-16 14:17:31 <fcolista> last was :
2017-08-16 14:17:32 <fcolista> commit b8152a3e29d0146426f217975e8b047943a250fc
2017-08-16 14:17:32 <fcolista> Author: Jakub Jirutka <jakub@jirutka.cz>
2017-08-16 14:17:32 <fcolista> Date:   Wed Aug 16 00:43:14 2017 +0200
2017-08-16 14:17:32 <fcolista>     community/namecoin: rebuild against protobuf-3.3.2
2017-08-16 14:23:19 <clandmeter> would be interesting if we could compare sources
2017-08-16 15:09:52 <vkrishn> clandmeter, we may wanna look at some of the tools here https://reproducible-builds.org/tools/
2017-08-16 15:11:48 <vkrishn> but for same pkg different version a simple git compare should also work
2017-08-16 19:47:52 <hiro> who here is  Head_on_a_Stick  ?
2017-08-16 20:26:19 <PureTryOut[m]> how does the `-dev` subpackage determine what files belong to it?
2017-08-16 20:28:21 <PureTryOut[m]> does it just copy all `.h` and `.cmake` files?
2017-08-16 20:28:48 <Shiz> https://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n1586
2017-08-16 20:28:56 <Shiz> you can override it, but this is the default impl :)
2017-08-16 20:29:38 <PureTryOut[m]> oooh thanks, that's useful
2017-08-16 20:38:52 <nidan_> Does anyone know if xen-4.8.1-r4 is ok? There is a script named /etc/xen/scripts/xen-backend.agent mentioned in /etc/xen/README that seems to be missing.
2017-08-16 20:39:29 <nidan_> Or maybe it's a bug in the xen-package itself?
2017-08-17 06:47:46 <fcolista> hi
2017-08-17 06:59:49 <abbierwolf> hi
2017-08-17 07:00:07 <abbierwolf> Did you receive my email on the kimchi and wok aports?
2017-08-17 07:02:07 <fcolista> abbierwolf, nope
2017-08-17 07:03:17 <abbierwolf> fcolista, I have used your email adress in the maintainer line
2017-08-17 07:03:40 <fcolista> correct
2017-08-17 07:03:47 <fcolista> still, I've not receive it
2017-08-17 07:04:53 <abbierwolf> fcolista, I have some updates on the wok and kimchi aports containing upgrades to the latest versions and some patches to let the server startup and run
2017-08-17 07:05:06 <fcolista> good
2017-08-17 07:05:21 <fcolista> can you create a PR to GH for that?
2017-08-17 07:06:00 <abbierwolf> fcolista, On the aports project or on a private repo?
2017-08-17 07:06:18 <fcolista> aports project it's fine
2017-08-17 07:06:27 <fcolista> Thanks!!!
2017-08-17 07:06:43 <abbierwolf> fcolista, ok i will create a pull request soon
2017-08-17 07:06:59 <fcolista> Thanks a lot
2017-08-17 07:07:21 <abbierwolf> fcolista, I already added a new aport novnc to the aports so this can be used as a new dependency
2017-08-17 07:18:20 <fcolista> abbierwolf, cool
2017-08-17 07:18:22 <fcolista> thanks
2017-08-17 07:19:23 <hiro> did you guys get anywhere with zfs in installer scripts?
2017-08-17 07:19:57 <hiro> or would i need to install to a normal disk, then just copy the content into a second zfs disk
2017-08-17 07:20:18 <hiro> also, do you guys use grub for booting directly from zfs?
2017-08-17 07:20:29 <hiro> or is this not recommended?
2017-08-17 08:17:45 <fcolista> abbierwolf, you around?
2017-08-17 08:39:53 <abbierwolf> fcolista, At this moment yes
2017-08-17 09:12:03 <abbierwolf> fcolista, If its about the commit message format. I saw I have to add the name of the aport in front. I will do it next time.
2017-08-17 09:25:58 <abbierwolf> fcolista, When will the new version be available in the testing repository? Is there a nightly build running?
2017-08-17 10:40:58 <fcolista> abbierwolf, it's available in edge
2017-08-17 10:41:04 <fcolista> which is the rolling release
2017-08-17 10:41:25 <fcolista> about the commit message: yes, it was wrong and I overlooked it
2017-08-17 10:41:28 <fcolista> my mistake
2017-08-17 10:41:42 <abbierwolf> fcolista, which repository can I use best to get it as quick as possible?
2017-08-17 10:42:25 <fcolista> dl-cdn.alpinelinux.org
2017-08-17 10:42:39 <abbierwolf> fcolista, ok thank you
2017-08-17 10:42:51 <fcolista> yw
2017-08-17 10:44:19 <abbierwolf> fcolista, I see the new packages in the pkgs.alpinelinux.org when will they be available in the repository?
2017-08-17 11:48:10 <fcolista> abbierwolf, which repository are you using?
2017-08-17 11:48:16 <fcolista> and which arch?
2017-08-17 12:53:42 <abbierwolf> fcolista, I am using dl-cdn.alpinelinux.org and x86_64 arch
2017-08-17 12:53:59 <fcolista> and you don't have the package updated?
2017-08-17 12:59:07 <fcolista> abbierwolf, you can't see them because are not yet built
2017-08-17 12:59:14 <fcolista> http://build.alpinelinux.org/
2017-08-17 12:59:34 <fcolista> the builder is stucekd...
2017-08-17 12:59:37 <fcolista> *stucked
2017-08-17 12:59:37 <fcolista> uff
2017-08-17 13:00:31 <fcolista> jirutka, edge-builder is stucked
2017-08-17 13:00:41 <fcolista> commit cc1147f83743b36e30d189b33aa3fca7ffa8b78b
2017-08-17 13:00:41 <fcolista> Author: Jakub Jirutka <jakub@jirutka.cz>
2017-08-17 13:00:41 <fcolista> Date:   Wed Aug 16 00:43:13 2017 +0200
2017-08-17 13:00:41 <fcolista>     testing/criu: rebuild against protobuf-3.3.2
2017-08-17 13:01:43 <abbierwolf> thanks I see
2017-08-17 17:20:41 <leo-unglaub> alright boys, lets do some demage to my new AMD RYZEN CPU .... lets build all x64 packages and see how long it takes
2017-08-17 17:30:40 <scadu> Woop woop
2017-08-17 17:31:18 <Shiz> nice
2017-08-17 17:34:10 <leo-unglaub> i wanted to safe up for a AMD threadripper, but 1000€ is way to much for me :)
2017-08-17 17:34:28 <leo-unglaub> the thing right now messing up my benchmark is the download time
2017-08-17 17:34:39 <leo-unglaub> is there a way to prefetch all downloads and then start building?
2017-08-17 17:39:42 <Shiz> you can do a separate abuild fetch for everything
2017-08-17 17:42:17 <leo-unglaub> i did that, but for some reason that i dont get some parts still get refetched
2017-08-17 18:38:57 <candy`> morning
2017-08-17 19:59:38 <kaniini> hi, does anybody require any experimental build runs for anything
2017-08-17 19:59:47 <kaniini> because i am tearing down my build cluster tonight
2017-08-17 19:59:53 <kaniini> as i need to move it to my house
2017-08-17 20:00:01 <Shiz> nt me]
2017-08-17 20:00:03 <Shiz> not me*
2017-08-17 20:00:48 <_ikke_> kaniini: what are you using for your build cluster?
2017-08-17 20:01:47 <kaniini> ikke: 2x dual intel e5-2670v3, 128GB RAM
2017-08-17 20:02:32 <kaniini> ikke: 3x quad intel e7-4820v4, 512GB RAM
2017-08-17 20:02:50 <_ikke_> Wow, that's not nothing
2017-08-17 20:02:53 <foxkit> dibs on the DS10Ls
2017-08-17 20:02:56 <foxkit> >.>
2017-08-17 20:03:11 <kaniini> u can fetch them from my garage
2017-08-17 20:03:32 <danieli> I'll be right there, I'm on the next plane
2017-08-17 20:03:42 <foxkit> is your apartment's living room exposed to the elements enough to call it a garage, or is that your new place
2017-08-17 20:04:59 <kaniini> new place
2017-08-17 20:05:02 <kaniini> i'm just shoving it all
2017-08-17 20:05:06 <kaniini> in the garage for now
2017-08-17 20:05:55 <kaniini> the security system is Internet of Shit enabled
2017-08-17 20:06:32 <kaniini> so just run hacksecuritysystem.py as downloaded from 0day.today
2017-08-17 20:06:34 <kaniini> andhelp yourself
2017-08-17 20:09:03 <foxkit> lol
2017-08-17 20:48:07 <foxkit> exfat-utils is unmaintained: how do you 'save' a package like that?  do I just grab it and update it and mark myself as maintainer?
2017-08-17 20:48:55 <_ikke_> foxkit: correct
2017-08-17 20:51:35 <ncopa> git mv to testing
2017-08-17 22:12:19 <ncopa> stateless: ping
2017-08-17 22:12:43 <ncopa> stateless: have you seen  https://bugs.alpinelinux.org/issues/7681 ?
2017-08-17 23:13:34 <foxkit> what is the going opinion on running `make check` on packages that output messages like "No tests yet"
2017-08-17 23:14:08 <foxkit> personally I think it is better than doing !check because when they do add tests, there will be no more work necessary
2017-08-17 23:14:08 <foxkit> but I was just wondering community consensus
2017-08-17 23:14:10 <Shiz> a counter-argument would be that it gives a false impression of tests being ran
2017-08-17 23:14:19 <Shiz> and thus being less subject to any future scrutiny over packages that run testsuites
2017-08-17 23:14:41 <Shiz> (even for things like statistical purposes)
2017-08-17 23:15:04 <Shiz> (i'm not set on any opinion myself, just presenting a thought that arose in me)
2017-08-18 00:42:55 <foxkit> http://lists.busybox.net/pipermail/busybox/2016-February/083907.html
2017-08-18 00:43:09 <foxkit> funny how first result in ddg for 'busybox useradd' is thread by alpinists :)
2017-08-18 04:36:07 <scadu> > personally I think it is better than doing !check because when they do add tests, there will be no more work necessary
2017-08-18 04:36:08 <scadu> I'd not expect it will be no more work. Half of those tests might fail and then the fun begin
2017-08-18 06:01:44 <fcolista> morning all
2017-08-18 10:00:44 <danieli> morning
2017-08-18 10:00:48 <danieli> lol
2017-08-18 10:12:01 <opendata> how can i use { } in an APKBUILD  (for downloading files)
2017-08-18 10:22:33 <_ikke_> opendata: What do you mean?
2017-08-18 10:22:51 <_ikke_> What does { } have to do with downloading files/
2017-08-18 10:22:53 <_ikke_> ?
2017-08-18 10:29:42 <opendata> _ikke_ so instead of  repeating the same url 5 times i defined a list of the files i need then put url{$_files}
2017-08-18 10:30:06 <opendata> _files is the variables name
2017-08-18 10:30:47 <_ikke_> why are you downloading files yourself btw in an APKBUILD?
2017-08-18 10:30:58 <_ikke_> Shouldn't you list those in sources?
2017-08-18 10:31:09 <opendata> that is sources
2017-08-18 10:32:28 <_ikke_> the shell is ash
2017-08-18 10:32:40 <_ikke_> so you need to use something ash supports
2017-08-18 10:33:13 <opendata> what does ash support?
2017-08-18 10:33:40 <_ikke_> opendata: can you show an example of what you are trying to achieve?
2017-08-18 10:34:46 <opendata> https://ptpb.pw/rGDZ
2017-08-18 10:36:00 <_ikke_> I don't think ash does support those kind of expansions
2017-08-18 10:38:19 <opendata> ikke anything i could do instead
2017-08-18 10:38:59 <_ikke_> Turn it around, define the URL as a variable
2017-08-18 16:10:29 <ncopa> morning
2017-08-18 16:11:19 <_ikke_> ncopa: hey
2017-08-18 16:11:56 <ncopa> opendata: you can do a loop
2017-08-18 16:12:42 <ncopa> for _i in $_files; do source="$source https://.../$i"; done
2017-08-18 16:13:09 <ncopa> or as _ikke_ suggested, use url as variable
2017-08-18 16:26:26 <opendata> ncopa, good suggestion
2017-08-18 16:27:37 <ncopa> opendata: we do something similar in main/bash for the patches
2017-08-18 17:35:01 <ncopa> i am fighting rethinkdb
2017-08-18 17:35:24 <ncopa> we need paxmark mksnapshot when they build v8
2017-08-18 17:35:34 <ncopa> but they fetch the sources themselves
2017-08-18 17:36:11 <ncopa> and there seems to be no mksnapshot target so i cannot run make in multiple steps
2017-08-18 17:36:41 <ncopa> our current hack no longer works (make || paxmark $out/mksnapshot; make)
2017-08-18 17:36:53 <ncopa> second make there will rebuild the mksnapshot apparently
2017-08-18 17:37:05 <ncopa> this is really not fun
2017-08-18 21:58:21 <ollieparanoid[m]> hi guys! has there been a discussion already about having an "flag package outdated" feature on pkgs.alpinelinux.org?
2017-08-18 22:06:59 <ncopa> ollieparanoid[m]: i think we already have one
2017-08-18 22:07:53 <ollieparanoid[m]> ncopa: oh, I found it now >_< thanks!
2017-08-18 23:28:25 <ncopa> does anyone have a clue what might have deleted all the distfiles/v3.3/* files around june - july 2017?
2017-08-19 10:36:32 <vkrishn> ncopa, is it just v3.3 or others too ? (distfiles missing) ^^^
2017-08-19 10:36:56 <vkrishn> they also seem kinda short list in browser ;)
2017-08-19 10:37:53 <vkrishn> this is not our 1st minor server disaster, is it!
2017-08-19 10:40:05 <vkrishn> btw, happy weekend morning!
2017-08-19 10:41:15 <vkrishn> I think I have distfiles for v3.5 & v3.6 lying around if needed, will check later
2017-08-20 07:22:33 <candy`> morning everybody
2017-08-20 08:15:16 <candy`> someone know if we can still boot alpine through pxe ?
2017-08-20 08:17:09 <candy`> after taking a lookg a the wiki page (https://wiki.alpinelinux.org/wiki/PXE_boot) and searching into the latest release of alpine, it appear (maybe I'm wrong) that since the commit ba27888b4576ceab7413ab9104d0aeda50990832 we are not able to use nfs because initramfs-init has been refactor in favor of nlplug-findfs which doesn't allow nfs but only zfs, mdadm, lvm
2017-08-20 18:46:46 <PureTryOut[m]> could #2097 please be merged? https://github.com/alpinelinux/aports/pull/2097
2017-08-20 18:48:07 <PureTryOut[m]> huh, didn't know that was a thing. I'll make sure not to use #<number> anymore
2017-08-21 11:48:34 <leo-unglaub> hey :)
2017-08-21 11:50:24 <leo-unglaub> i am still confused why my alpine intall always waits during boot for the press of a button
2017-08-21 11:50:47 <leo-unglaub> the strange thing is that this only happens when openbsd is the host for the virtual machine
2017-08-21 11:50:52 <leo-unglaub> l
2017-08-21 12:13:48 <zaolin> hey does alpine provides intel microcode updates ?
2017-08-21 12:20:01 <^7heo> zaolin: intel provides them.
2017-08-21 12:20:26 <zaolin> Sure but does the distro offers the microcode as package
2017-08-21 12:20:33 <^7heo> not AFAIK no.
2017-08-21 12:21:07 <zaolin> okay so it's missing atm
2017-08-21 12:21:34 <^7heo> AFAIK yes.
2017-08-21 12:21:45 <^7heo> you might want to ask Shiz or fabled or ncopa about that tho.
2017-08-21 12:21:58 <^7heo> they might know better than I.
2017-08-21 12:22:29 <Shiz> zaolin: we don't
2017-08-21 12:22:44 <Shiz> leo-unglaub: any idea where in the boot process this happens?
2017-08-21 12:23:01 <zaolin> just asking because we are running servers with alpine.
2017-08-21 12:23:03 <^7heo> tbh it would be nice to have a package for that; but since it's offered by intel already...
2017-08-21 12:23:05 <zaolin> Two options
2017-08-21 12:23:09 <^7heo> oh remote...
2017-08-21 12:23:12 <zaolin> initramfs with the kernel module
2017-08-21 12:23:13 <^7heo> yeah I see the point now.
2017-08-21 12:23:28 <zaolin> or microcode_ctl tool/service
2017-08-21 12:23:40 <^7heo> nah you can also use an OOB console
2017-08-21 12:23:43 <^7heo> and do it via the bios
2017-08-21 12:24:29 <zaolin> yeah but you have down time
2017-08-21 12:24:43 <zaolin> The bios/uefi is not really the solution for microcode updates
2017-08-21 12:24:53 <^7heo> with any reboot you have downtime.
2017-08-21 12:25:03 <zaolin> not with microcode-ctl
2017-08-21 12:25:19 <^7heo> well, you need to reboot for the microcode to be updated, don't you?
2017-08-21 12:25:36 <^7heo> (honest question - I only assume so, I don't know)
2017-08-21 12:25:46 <zaolin> no
2017-08-21 12:25:55 <^7heo> So the microcode updates live
2017-08-21 12:25:57 <zaolin> microcode is peristent
2017-08-21 12:26:00 <^7heo> with running software
2017-08-21 12:26:06 <zaolin> it vanish after reboot
2017-08-21 12:26:12 <^7heo> ah
2017-08-21 12:26:28 <^7heo> yeah okay so you update the memory but not the live image?
2017-08-21 12:26:30 <TBB> aports has in its non-free subsection a package called intel-ucode
2017-08-21 12:26:41 <TBB> the amd one is probably in linux-firmware
2017-08-21 12:26:44 <^7heo> (trying to understand the best I can from two contradicting statements)
2017-08-21 12:26:52 <^7heo> TBB: yeah AFAIK it is.
2017-08-21 12:27:25 <zaolin> Microcode can only be loaded during runtime. It vanishes after changing power states
2017-08-21 12:27:35 <^7heo> yeah but loaded from?
2017-08-21 12:27:38 <^7heo> that's the thing.
2017-08-21 12:27:42 <zaolin> OS
2017-08-21 12:27:44 <zaolin> for example
2017-08-21 12:27:45 <^7heo> really?
2017-08-21 12:27:50 <zaolin> there is a kernel module
2017-08-21 12:27:56 <zaolin> yes
2017-08-21 12:28:01 <^7heo> there's no other source for the microcode than the hard drive?
2017-08-21 12:28:04 <^7heo> (or SSD)
2017-08-21 12:28:14 <zaolin> Or you update the BIOS and the bios does the microcode update
2017-08-21 12:28:36 <^7heo> I would have expected it to be written to an EPROM
2017-08-21 12:28:43 <zaolin> I am talking about intel cpus not SSD
2017-08-21 12:29:01 <^7heo> and I'm talking about hard drives or SSDs
2017-08-21 12:29:14 <^7heo> since it's basically the same, from a CPU POV in this context.
2017-08-21 12:30:03 <zaolin> Yeah but that works different. They have non-persistent updates
2017-08-21 12:30:10 <^7heo> "they"?
2017-08-21 12:30:18 <^7heo> Look, I don't think we're getting anywhere ;)
2017-08-21 12:30:25 <^7heo> I'll just read what I find online on the subject
2017-08-21 12:30:35 <^7heo> thanks for your time and your attempts at explaining what you know ;)
2017-08-21 12:31:50 <^7heo> "Initially, CPU instruction sets were hardwired." - And now it's not. Well that makes sense to have a microcode then.
2017-08-21 12:31:58 <^7heo> I didn't know it changed.
2017-08-21 12:32:08 <TBB> so yeah, if you want the OS to make the update, you place intel-ucode.img in your /boot and list it as your first initrd in the bootloader configs. that should be it
2017-08-21 12:32:28 <^7heo> does the package do that?
2017-08-21 12:32:45 <leo-unglaub> @Shiz: its in the bootloader
2017-08-21 12:32:59 <leo-unglaub> for some reason it waits until the selected os gets confirmed
2017-08-21 12:33:26 <leo-unglaub> its the default config, i guess that for some reason the bootloader does not run the countdown and instead waits for a keystroke
2017-08-21 12:33:34 <^7heo> leo-unglaub: I would expect it to be an OpenBSD feature.
2017-08-21 12:33:53 <leo-unglaub> but how? the bootloader is alpine
2017-08-21 12:33:54 <^7heo> leo-unglaub: I would expect the countdown to happen normally; and the automatic boot to fail.
2017-08-21 12:34:06 <^7heo> leo-unglaub: executing on an OpenBSD machine, right?
2017-08-21 12:34:17 <^7heo> leo-unglaub: so ultimately running OpenBSD code, or?
2017-08-21 12:34:19 <leo-unglaub> yes, in vmd on ann openbsd system
2017-08-21 12:34:36 <^7heo> yeah so the alpine code is executed on OpenBSD software and kernel
2017-08-21 12:34:51 <^7heo> so I'd really check what happens on OpenBSD at the same time
2017-08-21 12:35:02 <^7heo> because that really sounds like a security feature from OpenBSD to me.
2017-08-21 12:35:11 <^7heo> (ie. no automatic execution)
2017-08-21 12:35:30 <leo-unglaub> it runs on openbsd software, but at this point the alpine bootloader does not know that. it thinks its bootet in a seabios
2017-08-21 12:35:33 <Shiz> zaolin: microcode is non-free so we can't package it
2017-08-21 12:35:42 <Shiz> as TBB noted, there's a package in the nonfree section t5hough
2017-08-21 12:35:45 <^7heo> the OpenBSD software is very aware of itself tho.
2017-08-21 12:35:51 <^7heo> and that's what I'm talking about.
2017-08-21 12:36:07 <zaolin> Shiz: Okay but Alpine runs on non-free hardware. Makes no sense to me too
2017-08-21 12:36:10 <^7heo> (and since it's running the alpine machine, it's also aware of the software of that machine)
2017-08-21 12:36:27 <Shiz> we can't control the hardware :)
2017-08-21 12:36:35 <^7heo> Shiz: can't the non-free package download the relevant microcode and update it?
2017-08-21 12:36:46 <^7heo> Shiz: I mean, automatically.
2017-08-21 12:37:00 <^7heo> (without containing the microcode, ofc)
2017-08-21 12:37:50 <Shiz> that's what the apkbuild does already
2017-08-21 12:38:19 <Shiz> well not automatically, we have to manually update it on new releases because the urls are unpredictable
2017-08-21 12:38:43 <Shiz> leo-unglaub: anyway, so it's a syslinux thing
2017-08-21 12:39:23 <^7heo> Shiz: I meant automatically from a user POV, not from our POV.
2017-08-21 12:39:31 <Shiz> abuild -r is automatic
2017-08-21 12:39:33 <^7heo> (and you seemed to answer it does)
2017-08-21 12:39:38 <^7heo> ok
2017-08-21 12:39:43 <Shiz> :P
2017-08-21 12:39:51 <^7heo> automatic from our POV would be bad anyway
2017-08-21 12:39:57 <^7heo> since it wouldn't lead to reproducable builds.
2017-08-21 12:40:10 <^7heo> reproducible? whatever. you got me.
2017-08-21 12:40:38 <Shiz> leo-unglaub: maybe vmd emits some kind of default keystroke
2017-08-21 12:40:43 <Shiz> which would interrupt the syslinux timeout?
2017-08-21 12:40:49 <Shiz> just throwing out ideas here
2017-08-21 12:40:55 <Shiz> accidentally or otherwise
2017-08-21 12:41:22 <^7heo> Shiz: what do you think of the OpenBSD system having some security feature implemented preventing the automatic loading of code at boot?
2017-08-21 12:41:24 <leo-unglaub> i had the same idea 2 weeks ago ... and it would be the most logical thing ... but i did not find anything the the openbsd vmd code that would do that
2017-08-21 12:41:35 <^7heo> (the VM host I mean)
2017-08-21 12:41:36 <Shiz> nonsense
2017-08-21 12:41:43 <^7heo> why?
2017-08-21 12:41:44 <leo-unglaub> nah, that would make no sence at all
2017-08-21 12:41:52 <Shiz> code is already loaded
2017-08-21 12:41:55 <Shiz> namely, the syslinux code
2017-08-21 12:42:01 <^7heo> ok so the jump would be prevented?
2017-08-21 12:42:19 <Shiz> it wouldn't even get to syslinux if it had such a 'feature'
2017-08-21 12:42:22 <leo-unglaub> that is not possible ... the hypervisor cannot just manipulate software running inside during runtime
2017-08-21 12:42:24 <^7heo> Shiz: ok
2017-08-21 12:42:34 <^7heo> leo-unglaub: well it technically can
2017-08-21 12:42:41 <^7heo> leo-unglaub: but yeah it would be rather bad.
2017-08-21 13:02:24 <zaolin> aports inte-ucode does everything I need
2017-08-21 13:05:50 <TBB> zaolin: excellent, you gave it a go? that's good to know
2017-08-21 13:31:09 <zaolin> TBB: yep
2017-08-21 13:51:54 <Shiz> jirutka: ping
2017-08-21 13:54:06 <clandmeter> github: ping
2017-08-21 13:54:18 <Shiz> https://0x0.st/nD0.patch
2017-08-21 13:54:36 <Shiz> https://0x0.st/nDG.patch
2017-08-21 13:54:42 <Shiz> jirutka: two patches for alpine-chroot-install ^
2017-08-21 13:54:50 <Shiz> (if you apply them, could you cut a new release?)
2017-08-21 14:47:31 <zaoqi> apk search racket;apk search scheme    =   echo      !
2017-08-21 14:56:36 <Shiz> ?
2017-08-21 14:56:56 <^7heo> yeah, same here: '?'
2017-08-21 14:58:24 <zaoqi> I can't find racket, scheme packages.
2017-08-21 14:59:01 <^7heo> $ apk search scheme
2017-08-21 14:59:01 <^7heo> tinyscheme-1.41-r0
2017-08-21 14:59:08 <^7heo> I dunno, I get something
2017-08-21 15:05:14 <zaoqi> I can't find mit-scheme package.
2017-08-21 15:06:44 <jcmdln> it appears that only tinyscheme is packaged at the moment, most likely you will need to build your desired interpreter/compiler chain yourself.
2017-08-21 19:46:41 <stateless> ncopa: was off the grid for a week, i will check the bug report tomorrow :)
2017-08-21 19:46:59 <stateless> i take it you are carrying a patch at the moment?
2017-08-21 20:10:05 <ncopa> stateless: not yet
2017-08-21 20:10:18 <ncopa> but i'll do so unless you say its a completely wrong thing to do
2017-08-21 21:47:43 <stateless> ncopa: ok i will let you know tomorrow :)
2017-08-21 21:50:50 <ncopa> stateless: thank you!
2017-08-22 01:00:52 <foxkit> a note for all: Qt 5.9 is LTS release, so it can probably safely be moved to main and slowly we can deprecate old broken insecure unsupported qt4
2017-08-22 01:01:12 <foxkit> and please don't bump to 5.10 in main
2017-08-22 01:25:18 <scamper> The procps package installs /bin/pgrep, but that comes after /usr/bin/pgrep (from busybox) in PATH.
2017-08-22 01:25:30 <scamper> Where do I report issues like this?
2017-08-22 02:02:53 <Xe> scamper: probably here
2017-08-22 02:08:47 <scamper> Xe, is that an authoritive answer? :) Other sensible place I considered was the devel mailing list.
2017-08-22 02:09:22 <Xe> scamper: what branch of alpine are you on?
2017-08-22 02:09:33 <scamper> Xe, edge.
2017-08-22 02:09:39 <scamper> as installed by lxd.
2017-08-22 02:10:04 <Xe> what do you get when you `$ env | grep PATH`?
2017-08-22 02:11:00 <scamper> PATH=/root/bin:/root/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/root/go/bin:/root/src/fzf/bin
2017-08-22 02:11:00 <scamper> GOPATH=/root/go
2017-08-22 02:11:00 <scamper> FPATH=<lots>
2017-08-22 02:11:28 <scamper> Tested without my rcs, too.
2017-08-22 02:11:37 <Xe> i have a terrible idea if you're open to it
2017-08-22 02:11:37 <scamper> And /usr/bin is still before /bin
2017-08-22 02:11:42 <Xe> as a shim
2017-08-22 02:11:52 <scamper> Sure.
2017-08-22 02:11:57 <Xe> prepend /bin to your $PATH
2017-08-22 02:12:38 <scamper> Hmm. I think that's worse than just rm /usr/bin/pgrep
2017-08-22 02:12:57 <Xe> no, rm-ing that will only make it come back when you upgrade busybox
2017-08-22 02:13:17 <scamper> Maybe it's not so bad, then. On my main distros /usr/bin is symlinked to /bin.
2017-08-22 02:13:29 <scamper> So it won't hurt my dotfiles too much.
2017-08-22 02:13:32 <Xe> if you can't trust /bin, what can you trust?
2017-08-22 02:13:54 <Xe> ya know?
2017-08-22 02:13:57 <scamper> I'm sure I can trust it, but I might want it overridden
2017-08-22 02:14:32 <Xe> yeah, then later dotfiles can prepend or append depending if they want to override $PATH or not
2017-08-22 02:19:12 <scamper> Looks like I can edit /etc/profile without alpine-baselayout crushing it.
2017-08-22 02:21:04 <scamper> Unrelated: I built apk-files on my arch machine and copied it into alpine, but it won't execute. They're both x86_64 and apk-files is in go, so it should be a static build. Any ideas?
2017-08-22 02:21:33 <Xe> $ ldd that-binary
2017-08-22 02:21:54 <Xe> pro tip: CGO_ENABLED=0 will build static binaries every time in go
2017-08-22 02:28:28 <scamper> "CGO_ENABLED=0 go build github.com/jessfraz/apk-file" doesn't generate a static binary and go get gives "go install net: open /usr/lib/go/pkg/linux_amd64/net.a: permission denied"
2017-08-22 02:33:00 <scamper> Looks like it's a known bug with distro installed go: https://github.com/golang/go/issues/18981https://github.com/golang/go/issues/18981
2017-08-22 02:37:40 <scamper> Workaround `go build -o <output file> github.com/jessfraz/apk-file`
2017-08-22 02:54:25 <foxkit> scamper: that is a bug
2017-08-22 02:54:39 <foxkit> scamper: the other utilities are supposed to override busybox, I reported something similar and it was resolved
2017-08-22 02:55:18 <foxkit> http://bugs.alpinelinux.org/issues/7412
2017-08-22 03:02:45 <scamper> foxkit, thanks: https://bugs.alpinelinux.org/issues/7737
2017-08-22 04:01:44 <foxkit> things that piss off the fox: kmod developers disabled the ability to static link
2017-08-22 04:02:18 <foxkit> I don't really like to static link everything, but core system things should be, and they took that away
2017-08-22 04:02:21 <foxkit> and it makes me mad
2017-08-22 04:26:33 <foxkit> "Welcome to AdAClie Linux"
2017-08-22 04:26:38 <foxkit> well, something doesn't seem to like utf-8 :s
2017-08-22 06:22:35 <fcolista> hi guys
2017-08-22 07:12:52 <zaoqi_> kindle-x : https://paste.gnome.org/pk1lcazx2 https://paste.gnome.org/pktyraukg /usr/bin/xfce4-session: line 1: syntax error: unexpected word (expecting ")")
2017-08-22 07:15:35 <PureTryOut[m]> so, since I got the majority of Plasma Mobile working, I'd like to upstream it to the Alpine repos. now, in total there are 80 required KDE packages. should I submit them one by one or all in one go? and if one-by-one, do I have to wait with packages that depend on others from KDE till there dependencies are merged first?
2017-08-22 07:15:55 <PureTryOut[m]> *their dependencies
2017-08-22 07:24:01 <zaoqi>  /usr/bin/xfce4-session: line 1: syntax error: unexpected word (expecting ")") kindle-x : https://paste.gnome.org/pk1lcazx2 result: https://paste.gnome.org/pktyraukg
2017-08-22 10:43:03 <stateless> ncopa: i applied the patch and created a new release as well :)
2017-08-22 16:01:10 <ncopa> stateless: thanks!
2017-08-22 20:02:11 <ollieparanoid[m]> ncopa, kaniini: any thing I can do to help getting this PR merged? (let consolekit2 replace consolekit) https://github.com/alpinelinux/aports/pull/2097
2017-08-22 20:22:58 <nmeum> ollieparanoid[m]: could you elaborate why this is necessary/useful?
2017-08-22 20:25:33 <ollieparanoid[m]> nmeum: consolekit is deprecated, consolekit2 is a maintained fork
2017-08-22 20:26:18 <ollieparanoid[m]> and consolekit2 is needed for kde stuff, and could also be used by kaniini's gnome aports (he said he would look into using it at some point)
2017-08-22 20:26:32 <ollieparanoid[m]> (yeah, we're porting kde stuff ... I can give you more context if you like)
2017-08-22 20:32:45 <kaniini> i rather just replace consolekit with consolekit2 outright
2017-08-22 20:33:19 <PureTryOut[m]> well, consolekit2 is actually maintained where consolekit is not, and it's a drop-in replacement. so that'd be alright with me
2017-08-22 20:33:29 <kaniini> sorry have been dealing with my ISP, they decided to let their infrastructure to rot in my new neighbourhood
2017-08-22 20:33:35 <kaniini> switching ISPs
2017-08-22 20:33:39 <kaniini> tomorrow
2017-08-22 20:34:07 <kaniini> will be able to give static ipv4 to my build/porting machines at that time too
2017-08-22 20:35:16 <ollieparanoid[m]> oh, I understand. that sucks.
2017-08-22 20:37:20 <PureTryOut[m]> kaniini: so, want me to make a PR that replaces consolekit with consolekit2 then?
2017-08-22 20:37:31 <kaniini> yes
2017-08-22 20:37:38 <kaniini> that seems like a good plan
2017-08-22 20:37:38 <PureTryOut[m]> ok I'll do that then
2017-08-22 20:38:13 <PureTryOut[m]> I'll close that other PR then
2017-08-22 20:38:15 <kaniini> i think that resolves what nmeum was asking too
2017-08-22 20:38:27 <kaniini> from what i gathered, he was wondering what the point of having both was
2017-08-22 20:38:40 <nmeum> yes, indeed
2017-08-22 20:39:06 <PureTryOut[m]> it's a good question, and I'm not sure either. I just didn't straight up want to replace a core package for my first submitted package 😛
2017-08-22 20:39:53 <PureTryOut[m]> so, remove consolekit, move consolekit2 from testing to main, keep the name consolekit2, and make it provide consolekit?
2017-08-22 20:40:12 <nmeum> while at it we might also want move consolekit and packages depending on it from main/ to community/
2017-08-22 20:40:49 <PureTryOut[m]> 
ACTION uploaded an image: Screenshot_2017-08-22_22-40-42.png (13KB) <https://matrix.org/_matrix/media/v1/download/matrix.org/bpLGoVQyMdGolxUWezbWwYZF>

2017-08-22 20:40:53 <PureTryOut[m]> those? ^
2017-08-22 20:41:38 <foxkit> personally I would see if they work on consolekit2
2017-08-22 20:42:16 <PureTryOut[m]> well, I guess it's time to install a DE on my old laptop then. I guess Mate will do
2017-08-22 20:43:10 <foxkit> PureTryOut[m]: I probably do not want to ask you this question, but for the "80 KDE packages" PR, do you have them supporting i18n?
2017-08-22 20:43:18 <PureTryOut[m]> yes
2017-08-22 20:43:28 <foxkit> so there is -lang subpackages and such?  :D
2017-08-22 20:43:34 <PureTryOut[m]> on most of them yes
2017-08-22 20:43:39 <foxkit> oh thank you so much, this is such a relief
2017-08-22 20:43:55 <PureTryOut[m]> https://github.com/postmarketOS/pmbootstrap/tree/postmarketos-ui-plasma-mobile/aports/kde
2017-08-22 20:44:08 <PureTryOut[m]> I try to use -doc and -lang everywhere
2017-08-22 20:44:21 <PureTryOut[m]> well and -dev where needed
2017-08-22 20:44:27 <PureTryOut[m]> and even -libs in one place I think
2017-08-22 20:46:54 <foxkit> PureTryOut[m]: hm, this looks okay.  why not 5.8 LTS though?
2017-08-22 20:47:15 <foxkit> that has more stability, better for a distro that supports releases for years
2017-08-22 20:47:18 <nmeum> PureTryOut[m]: maybe just replace consolekit with consolekit2 as a first step. we can move consolekit2 and packages needing it from main to community later on
2017-08-22 20:47:24 <PureTryOut[m]> foxkit: because we're packaging Plasma Mobile, not just Plasma Desktop. since it doesn't even have a proper release yet, we want to be as bleeding edge as possible there
2017-08-22 20:47:36 <PureTryOut[m]> nmeum: sure
2017-08-22 20:47:39 <foxkit> PureTryOut[m]: ah ok.  makes sense then.
2017-08-22 20:48:01 <foxkit> PureTryOut[m]: I might take some of this work and 'backport' to 5.8 LTS for Adélie, if you wouldn't mind, because we're shipping desktop and I'd like it to be stable :P
2017-08-22 20:48:06 <PureTryOut[m]> I got Plasma Mobile running already, with some kinks here and there
2017-08-22 20:48:16 <PureTryOut[m]> ooh sure no problem
2017-08-22 20:48:25 <PureTryOut[m]> (Adélie is upcoming release?)
2017-08-22 20:48:49 <foxkit> 1.0-alpha3 is supposed to come out next week, it won't have X yet though.. sadly.. because there's too many other things I'm trying to fix
2017-08-22 20:48:59 <PureTryOut[m]> we can just put 5.10 and higher in testing till a next stable release is done by the KDE folks
2017-08-22 20:49:43 <foxkit> if it wasn't clear, Adélie is a fork/whatever of Alpine (I guess like postmarketOS) focused on desktop stuffs
2017-08-22 20:49:53 <foxkit> and yeah, shipping a desktop distro without X is silly
2017-08-22 20:49:59 <PureTryOut[m]> ooh, didn't know that
2017-08-22 20:49:59 <foxkit> lol
2017-08-22 20:50:09 <PureTryOut[m]> yeah like postmarketOS then, pmOS being focussed on mobile
2017-08-22 20:50:28 <PureTryOut[m]> do note that KDE 5.8 might not have this (I think required) patch https://phabricator.kde.org/R108:6ebc3d65f032ef79f2ce7db0fa86016aadbbf18f
2017-08-22 20:50:36 <PureTryOut[m]> actually
2017-08-22 20:50:43 <PureTryOut[m]> never mind, that's only required if you want Wayland
2017-08-22 20:50:54 <PureTryOut[m]> (which we're going for, but you might not)
2017-08-22 20:51:37 <foxkit> PureTryOut[m]: https://adelielinux.org/
2017-08-22 20:52:02 <foxkit> woah, that got finished?  amazing!
2017-08-22 20:52:24 <foxkit> we were not planning on shipping Wayland for 1.0, because it's still so experimental, but that's looking really good for 2.0 :)
2017-08-22 20:52:25 <PureTryOut[m]> interestingh
2017-08-22 20:52:32 <PureTryOut[m]> do you have a GUI for apk as well?
2017-08-22 20:52:43 <foxkit> not yet, that's something we're working on
2017-08-22 20:53:02 <nmeum> kaniini: also started writing a gui for apk https://github.com/kaniini/apk-gtk
2017-08-22 20:53:02 <PureTryOut[m]> please make it's interface responsive, so we can use it on postmarketOS as well 😄
2017-08-22 20:53:11 <foxkit> really need package categories beyond main/community/testing though
2017-08-22 20:53:25 <foxkit> like "games" and "server stuff" and "productivity" etc..
2017-08-22 20:53:43 <PureTryOut[m]> yeah, same for pmOS probably
2017-08-22 20:53:50 <PureTryOut[m]> nmeum: ooh nice, thanks for linking
2017-08-22 20:53:56 <kaniini> foxkit: spec it out
2017-08-22 20:54:00 <foxkit> PureTryOut[m]: all the GUIs I have written in past three years are written in Qt5 and tested on pentium 3/900MHz system for speed.  hopefully that is close enough
2017-08-22 20:54:03 <kaniini> PureTryOut: it's mostly for system upgrades though
2017-08-22 20:54:16 <foxkit> kaniini: yeah I'm still thinking on how best to do it.
2017-08-22 20:54:23 <PureTryOut[m]> foxkit: definitely! Qt5 is quite good for mobile
2017-08-22 20:54:40 <PureTryOut[m]> kaniini: still better than opening a command line on your phone to update lol
2017-08-22 20:55:05 <kaniini> for OTAs, apk 3 will be much better
2017-08-22 20:55:24 <kaniini> because it will be possible to apply the update as a single atomic image
2017-08-22 20:55:34 <foxkit> PureTryOut[m]: my last qt5 project was game boy emulator, full native game boy speed on G3/600 MHz with link cable and all the rest ^^  pass more of 'blarggs tests' (game boy emulator test suite) than most others too ;)
2017-08-22 20:55:45 <PureTryOut[m]> kaniini: nice!
2017-08-22 20:55:47 <foxkit> kaniini: yeah I am looking forward to the thing where it lets you 'change' a package for another
2017-08-22 20:55:55 <foxkit> kaniini: like add and delete as one transactional unit
2017-08-22 20:55:56 <PureTryOut[m]> foxkit: awesome stuff 😃
2017-08-22 20:56:00 <foxkit> kaniini: will be necessary to change /bin/sh and such
2017-08-22 20:56:22 <foxkit> I believe that feature is slated for APK 3
2017-08-22 20:56:26 <kaniini> foxkit: swap operation is tricky, still thinking on the semantics for it
2017-08-22 20:56:37 <foxkit> kaniini: yeah I can imagine.
2017-08-22 20:57:38 <PureTryOut[m]> so, since I'm still quite new to this, how do I actually test a package without having it in a repo somewhere? does `apk add consolekit2` somehow figure out that it needs the APKBUILD on my local machine, or do I need to add a local repo somewhere?
2017-08-22 20:57:56 <_ikke_> PureTryOut[m]: you need to add a local repo
2017-08-22 20:58:28 <_ikke_> which is quite easy
2017-08-22 20:58:59 <PureTryOut[m]> do you have a link to a wiki page?
2017-08-22 20:59:35 <PureTryOut[m]> I find all kinds of stuff except the information I want lol
2017-08-22 21:00:46 <przemoc> it's off-topic, I guess, but Qt has some OTA solution based on OSTree.  does anyone checked Qt OTA or OSTree as how useful, safe and efficient they are out-of-the-box?  just curious.
2017-08-22 21:01:20 <_ikke_> PureTryOut[m]: I'm not aware of a tutorial
2017-08-22 21:01:53 <przemoc> (I asked here, because foxkit is not on -offtopic channel)
2017-08-22 21:04:51 <PureTryOut[m]> aah, I see `abuild` builds to `~/packages`, which is already a repository on itself it seems. I guess I just need to add it to `/etc/apk/repositories`?
2017-08-22 21:05:57 <skarnet> that's how I do it. Not sure if it's the kasher way.
2017-08-22 21:06:11 <skarnet> kosher*
2017-08-22 21:06:16 <PureTryOut[m]> sounds fine too me
2017-08-22 21:06:45 <foxkit> that's what I do too
2017-08-22 21:06:47 <okeuday_bak> Hi, could someone review/merge https://github.com/alpinelinux/aports/pull/2140 ?
2017-08-22 21:06:54 <foxkit> well mine is in /srv/staging/packages but :)
2017-08-22 21:08:07 <foxkit> przemoc: OSTree / QtOTA is as safe as you want it to be.  it can do GPG signing, TLS with PKP, custom hooks, and so on.  I can't really comment on efficiency because I've never done more with it than just playing around with silly things.
2017-08-22 21:08:17 <PureTryOut[m]> `WARNING: Ignoring /home/****/packages/x86_64/APKINDEX.tar.gz: No such file or directory`
2017-08-22 21:08:18 <PureTryOut[m]> huh
2017-08-22 21:08:32 <PureTryOut[m]> ooh nvm
2017-08-22 21:08:41 <skarnet> if it's the first time you abuild a package, ignore the message
2017-08-22 21:08:48 <PureTryOut[m]> yeah ok just was being awfullyl stupid
2017-08-22 21:08:55 <PureTryOut[m]> neh I forgot to specify `main`
2017-08-22 21:09:26 <przemoc> foxkit: ok, thanks
2017-08-22 21:26:09 <PureTryOut[m]> well, installing Mate is one hell of a job, but it seems to work fine on consolekit2 without any changes
2017-08-22 21:27:30 <PureTryOut[m]> actually wait no, Mate seems to have installed consolekit itself...
2017-08-22 21:27:58 <PureTryOut[m]> and I can't remove it due to dependency stuff... the Mate guide said to manually tell apk to install consolekit, so I thought using consolekit2 there would do it
2017-08-22 21:32:24 <PureTryOut[m]> I guess `replaces=` is needed
2017-08-22 21:34:18 <PureTryOut[m]> ugh, even `--force` doesn't uninstall the package
2017-08-22 21:45:37 <PureTryOut[m]> euhm I might just be understanding the output of `apk info` wrong. does consolekit show as installed *because* consolekit2 provides consolekit?
2017-08-22 21:51:21 <PureTryOut[m]> I guess it is... since both packages provide the same files, they literally can't be installed together. and since consolekit2 shows as installed too, and it being the only package that I saw in the list of stuff to be installed
2017-08-22 22:08:17 <PureTryOut[m]> https://github.com/alpinelinux/aports/pull/2166
2017-08-22 22:08:38 <PureTryOut[m]> PR created. Let the reviewing and testing begin!
2017-08-22 22:12:34 <PureTryOut[m]> well that's an interesting Travis failure. as the package it breaks on isn't even related to consolekit in anyway. no dependency anywhere
2017-08-22 22:13:11 <PureTryOut[m]> never mind I'm blind, it does have a dep
2017-08-22 22:13:52 <PureTryOut[m]> still not sure why it fails though
2017-08-22 22:14:15 <PureTryOut[m]> the `cgmanager` dependency seems to be masked?
2017-08-22 22:40:38 <nmeum> PureTryOut[m]: yep
2017-08-22 22:41:14 <PureTryOut[m]> well, is that a problem for my PR then? what can I do about it?
2017-08-22 22:42:22 <nmeum> PureTryOut[m]: cgmanager is a package from community/ packages in main/ can't depend on packages in community/
2017-08-22 22:42:45 <nmeum> that's the reason why it is masked
2017-08-22 22:42:46 <PureTryOut[m]> ooh, obviouisly
2017-08-22 22:42:51 <PureTryOut[m]> *obviously
2017-08-22 22:43:37 <nmeum> so you either need to move cgmanager from community/ to main/ or consolekit from main/ to community
2017-08-22 22:43:54 <nmeum> the latter would probably be preferable though I don't know what packages depend on consolekit since these would have to be moved as well then
2017-08-22 22:44:37 <PureTryOut[m]> maybe I can get rid of the cgmanager dep entirely
2017-08-22 22:44:46 <nmeum> that would of cause also be an option
2017-08-22 22:44:50 <PureTryOut[m]> as according to the consolekit2 readme it's "recommended on Linux, for process tracking"
2017-08-22 22:45:02 <PureTryOut[m]> so, not necessary, just recommended
2017-08-22 23:17:34 <Cred> Hi, when using an older Alpine release but package from edge, should the edge package declare minimum `depends=` in the APKBUILD?
2017-08-22 23:36:26 <Shiz> depends if the dependencies are not automatically picked up by the .so scanner, which they usually are
2017-08-22 23:56:29 <Cred> hm. So in this particular case I had old versions of the dependencies causing failures. Only after one by one walking through upgrades did things start to work.
2017-08-22 23:58:19 <Cred> I guess what I want to understand is, if I use old Alpine releases and pull in edge packages should it be me or the package who does dependency checks like that
2017-08-23 15:56:17 <ncopa> looks like redhat is scared of alpine: http://crunchtools.com/comparison-linux-container-images/
2017-08-23 15:58:42 <iamthemcmaster> as a former debian/ubuntu user, i can say that there have definitely been a few instances of things that they have a package for that alpine doesn't, but it's a MILLION times easier to contribute a new package to alpine than it is to contribute one to either debian or ubuntu
2017-08-23 16:04:00 <Shiz> >life cycle: unknown
2017-08-23 16:04:09 <Shiz> then they haven't really bothered checking the wiki for our release process
2017-08-23 16:04:52 <Shiz> >automated testing: none found
2017-08-23 16:04:55 <Shiz> not really paying attention
2017-08-23 16:08:12 <iamthemcmaster> the fact that they talk about their 10 year life cycle tells me they don't get it, anyways... i don't need to use an OS that supports a 10 year old version of python inside of a container, thanks
2017-08-23 16:37:35 <TBB> yeh, that article looks like the author didn't really properly research his stuff
2017-08-23 16:38:57 <TBB> admittedly the glibc incompatibilities are an issue, but nothing that can't be gotten over. the citation about "gets to 200 mb like everyone else" is false too; I do packaging and imaging of both CentOS and Alpine extensively and the size difference is clear
2017-08-23 16:41:32 <TBB> iamthemcmaster: they get it allright, it's a valid point in many, many organizations
2017-08-23 16:42:41 <TBB> we could've chosen CentOS as our product base instead of Alpine but we didn't, and that has meant a lot of extra work for us. but for our project it doesn't matter; for many other projects it probably does
2017-08-23 16:47:47 <ncopa> iamthemcmaster: im glad to hear that you find it simple to contribute with new package. that was one of the goals
2017-08-23 16:49:34 <ncopa> yes, he has some valid points but he mixes in too much false/misleading/outdated info to back his claims
2017-08-23 18:31:29 <vkrishn> <brew>
2017-08-23 18:31:30 <vkrishn> some mixed thoughts thats brewing for a week,
2017-08-23 18:31:30 <vkrishn> 1. have DragonEgg gcc plugin build in alpine
2017-08-23 18:31:30 <vkrishn> 2. this would allow developers to re-compile some apps and cross benchmark with regular compile
2017-08-23 18:31:30 <vkrishn> 3. I think AL if plans for LLVM the gcc fork should continue.
2017-08-23 18:31:30 <vkrishn> </brew>
2017-08-23 18:31:34 <vkrishn> apk fetch from parell above (pun on paragraph+barrel)
2017-08-24 03:29:28 <foxkit> why does abuild's APKBUILD explicitly depend on libressl?  does abuild use libressl in some way? o.O
2017-08-24 06:49:16 <nmeum> foxkit: abuild-keygen needs it for generating an RSA keypair
2017-08-24 06:49:24 <nmeum> abuild-sign needs it as well of cause
2017-08-24 09:50:23 <mosez> jirutka: maybe you can take another look at my pull requests? https://github.com/pulls?q=is%3Aopen+is%3Apr+author%3Atboerger+user%3Aalpinelinux
2017-08-24 18:44:28 <foxkit> is there a standard for prerelease version strings?
2017-08-24 18:44:33 <foxkit> like 2.8.0-pre1
2017-08-24 18:45:36 <skarnet> yes
2017-08-24 18:45:56 <skarnet> it's encoded somewhere in the abuild C code, I don't remember where exactly
2017-08-24 18:46:03 <skarnet> let me have a look
2017-08-24 18:46:15 <foxkit> I thought abuild was in shell o.O
2017-08-24 18:46:23 <skarnet> um, apk, sorry
2017-08-24 18:46:59 <skarnet> there it is: https://git.alpinelinux.org/cgit/apk-tools/tree/src/version.c#n75
2017-08-24 18:47:50 <skarnet> and IIRC it comes from the Gentoo rules for version strings
2017-08-24 19:01:44 <ncopa> yes its very similar to gentoo rules for version strings
2017-08-24 19:35:15 <foxkit> thank you both :)
2017-08-24 20:38:02 <foxkit> ...
2017-08-24 20:38:11 <foxkit> I did not realise our fuzzy version matching stuff was merged in, very nice!
2017-08-24 21:52:55 <ncopa> it seems that coreos has a fork of grub that can do TPM and kernel verification
2017-08-24 21:53:04 <ncopa> i wonder if we should use that instead of gnu grub
2017-08-24 21:53:13 <ncopa> we need grub for aarch64 and ppc64le
2017-08-24 21:53:22 <ncopa> so we need to support it
2017-08-24 21:53:38 <ncopa> i wonder if we should use it default on all arches
2017-08-24 21:53:57 <ncopa> its used for xen too
2017-08-24 21:54:15 <ncopa> im no fan of grub. its a big ugly monster
2017-08-24 21:57:40 <danieli> should probably look at coreos' grub fork if the licensing is all right
2017-08-24 21:59:23 <TBB> that sounds awesome
2017-08-24 21:59:38 <TBB> (and I realise my response does not :)
2017-08-24 22:08:55 <foxkit> ncopa: grub2 is getting slimmer these days.  they stopped doing so much BS with it and made it really modular
2017-08-24 22:09:20 <foxkit> ncopa: we ship it as default in Adélie on ppc32, it uses just around 700 KB and the config syntax is readable again
2017-08-24 22:09:38 <ncopa> i looked at /etc/grub.d/*
2017-08-24 22:09:43 <ncopa> and, well
2017-08-24 22:09:50 <ncopa> at least it works
2017-08-24 22:09:54 <ncopa> and managed to boot alpine
2017-08-24 22:09:55 <danieli> it works™
2017-08-24 22:10:32 <ncopa> my reasoning is that we need support it anyway, so we could probably as well use it on all platforms (except s390x)
2017-08-24 22:28:23 <Shiz> i would very much prefer to keep syslinux.
2017-08-24 22:29:22 <foxkit> syslinux's EFI support is non-existant
2017-08-24 22:29:26 <foxkit> is really the other problem
2017-08-24 22:29:37 <foxkit> I know it's "supposed" to have EFILINUX, but I have yet to see that boot a real computer
2017-08-24 22:30:01 <foxkit> it doesn't work on:  apple imac, apple macbook pro, dell xps, intel s5000 server board, intel s5400 server board, supermicro X10 workstation board
2017-08-24 22:30:22 <Shiz> well, i don't want efi boot
2017-08-24 22:30:24 <Shiz> :)
2017-08-24 22:30:24 <foxkit> grub2 works on all of them, though I only use grub2 on the dell
2017-08-24 22:30:32 <foxkit> the others are kernel efi stub
2017-08-24 22:30:36 <TBB> gummiboot does handle efi and kernel verification is done by the efi firmware
2017-08-24 22:30:40 <skarnet> the definition of a real computer according to foxkit :)
2017-08-24 22:31:03 <foxkit> skarnet: I meant that EFILINUX works in virtual machine like VirtualBox and KVM
2017-08-24 22:31:09 <foxkit> but not on a physical machine
2017-08-24 22:31:15 <ncopa> gummiboot is sort of dead upstream
2017-08-24 22:31:16 <foxkit> sorry for confusing words
2017-08-24 22:31:23 <skarnet> I'm just pulling your leg
2017-08-24 22:31:25 <foxkit> XD
2017-08-24 22:31:46 <ncopa> i suppose we need to support syslinux anyways
2017-08-24 22:31:47 <TBB> ncopa: naturally as it's probably melted into systemd (which it is, it's called systemd-boot now)
2017-08-24 22:31:54 <foxkit> no, real computer is SGI Onyx or Talos 2 :)
2017-08-24 22:31:55 <ncopa> i know
2017-08-24 22:32:04 <Shiz> let me reword what i said
2017-08-24 22:32:08 <ncopa> Shiz: can you have a look at this? http://lists.alpinelinux.org/alpine-aports/4829.html
2017-08-24 22:32:09 <Shiz> i don'tw ant forced efi boot
2017-08-24 22:32:09 <skarnet> real computer is AMD Elan SC520
2017-08-24 22:32:11 <Shiz> :p
2017-08-24 22:32:17 <Shiz> h yeah
2017-08-24 22:32:30 <Shiz> this is weird because i cant repro the issue
2017-08-24 22:32:32 <foxkit> skarnet: mmmm, Elan
2017-08-24 22:32:32 <ncopa> just give me an ACK or NACK
2017-08-24 22:32:32 <TBB> but yeah, syslinux is handy and it works almost everywhere
2017-08-24 22:32:46 <foxkit> let me be clear that I do not advocate for removal of syslinux
2017-08-24 22:32:48 <ncopa> syslinux only works on x86 + x86_64
2017-08-24 22:32:59 <Shiz> and im not sure what its doing in the patch, so i have to take a closer look
2017-08-24 22:33:12 <ncopa> thanks
2017-08-24 22:33:26 <foxkit> and in fact we recommend syslinux on x86_32 in all cases https://wiki.adelielinux.org/wiki/HOWTO:Manually_install_Ad%C3%A9lie_Linux#Install_the_boot_loader
2017-08-24 22:33:39 <foxkit> but grub2 is better for EFI, and necessary on PowerPC
2017-08-24 22:33:42 <Shiz> and i don't like the manual time adjustment
2017-08-24 22:34:11 <foxkit> the only other bootloader for PowerPC would be yaboot, which does not work with musl at all, and doesn't support ext4 or btrfs or LUKS or anything modern like that
2017-08-24 22:34:22 <foxkit> it only barely supports ext3, ext3 was new when yaboot was last maintained :/
2017-08-24 22:35:33 <foxkit> skarnet will hate me for having a low bar, but grub2 is the least-bad option, and I don't think anyone here really wants to get down and dirty with writing a better boot loader (but I would appreciate the effort!)
2017-08-24 22:35:55 <skarnet> I'm afraid you're right and all bootloaders suck
2017-08-24 22:36:58 <foxkit> as someone who /has/ written bootloader code before: imagine the worst APIs possible (this is your firmware), now imagine impossible constraints (512 bytes for stage1 on x86, 64 KB for the entire stage2 on ppc, no memory map on ARM); boot loaders are about 10x worse than you are imagining right now
2017-08-24 22:38:30 <foxkit> so we can keep syslinux for the simpler/lightweight case of just booting a BIOS based x86_{32,64}, and uboot for some ARM SoC
2017-08-24 22:38:46 <foxkit> but grub2 is really only option for EFI x86_64, PPC, and aarch64 :/
2017-08-24 22:38:59 <foxkit> that's my opinion anyway
2017-08-24 22:41:28 <danieli> grub2 is the only suitable option indeed, except developing something ourselves, which would be... well, mission impossible
2017-08-24 22:45:41 <ncopa> foxkit: i think you are right
2017-08-24 22:49:26 <TBB> I've tried twice to look at grub2 configuration and both times got so disgusted that I just went back to gummiboot. but at least grub is usable on a wide range of platforms; I have no idea if gummiboot is even portable outside of x86*
2017-08-24 23:04:04 <ncopa> i think there have been some arm related commits but i dont think it works
2017-08-24 23:04:15 <ncopa> i dont think it works with 32bit x86 either tbh
2017-08-24 23:04:48 <danieli> systemd.. grrr
2017-08-24 23:05:05 <ncopa> i currently use gummiboot on my laptop
2017-08-24 23:05:14 <danieli> systemd ate gummiboot
2017-08-24 23:05:15 <ncopa> its kind of simple to set up
2017-08-24 23:05:23 <ncopa> yes, i know it did
2017-08-24 23:05:30 <danieli> black hole
2017-08-24 23:05:34 <ncopa> so gummiboot is basically dead up stream
2017-08-24 23:05:38 <danieli> yep
2017-08-24 23:05:43 <danieli> sounds about right
2017-08-24 23:06:19 <danieli> shame really
2017-08-24 23:06:28 <ncopa> *shrug*
2017-08-24 23:06:36 <ncopa> well, yes, it was sort of nice
2017-08-24 23:07:15 <danieli> it was an alternative boot manager is one thing
2017-08-24 23:07:19 <danieli> now it's just.. systemd-something
2017-08-24 23:25:11 <danieli> ncopa: are you familar with coreboot's grub fork? does it work out of the box with alpine?
2017-08-24 23:30:47 <ncopa> danieli: no, im not
2017-08-24 23:30:56 <ncopa> i just tested that gnu grub works
2017-08-24 23:31:01 <danieli> oh I see
2017-08-24 23:31:09 <danieli> I see.. I'll set up a VM, clone it, build it, and test
2017-08-24 23:31:18 <ncopa> the only thing  i noticed was that it says "GNU/Linux" in the menu
2017-08-24 23:31:24 <danieli> yep...
2017-08-24 23:31:25 <ncopa> which is sort of annoying :)
2017-08-24 23:31:33 <danieli> I suppose we'll have to fork it hahaha
2017-08-24 23:31:53 <ncopa> coreboot?
2017-08-24 23:32:00 <ncopa> do you mean coreos?
2017-08-24 23:32:06 <danieli> I meant coreos
2017-08-24 23:32:11 <danieli> jesus christ, how did I make that mistake
2017-08-24 23:32:17 <danieli> #toolibre
2017-08-24 23:32:25 <danieli> no wait, that's libreboot, also a thing
2017-08-24 23:34:23 <danieli> hmm.. the alpine website isn't too good on phones
2017-08-24 23:35:48 <foxkit> danieli: that's nothing, the first bug I filed on alpine's redmine was from an iPhone 3GS because my desktop was crashing when I started X
2017-08-24 23:35:56 <foxkit> that was... a less than stellar experience
2017-08-24 23:36:08 <danieli> foxkit: jeez..
2017-08-24 23:37:17 <foxkit> alpinelinux.org seems ok on my pixel
2017-08-24 23:37:20 <foxkit> what's going on with it?
2017-08-24 23:37:42 <danieli> the responsive design doesn't seem to work too well with the hamburger menu button and the text on the image
2017-08-24 23:38:18 <danieli> try emulating a device with chrome, f12 then ctrl+shift+m
2017-08-24 23:40:42 <foxkit> I don't have Chrome installed on any device (due to instability, bloat, and the fact a chrome 0day cost me my skype credentials)
2017-08-24 23:40:59 <danieli> point taken. I run Chromium because I push to it
2017-08-24 23:41:00 <foxkit> but it works fine on Firefox/Android 7, Safari/iOS 10, and Opera/iOS 9
2017-08-24 23:42:19 <danieli> uploading
2017-08-24 23:42:36 <danieli> one sec
2017-08-24 23:43:20 <danieli> foxkit: from my s7 edge https://i.imgur.com/SYWVauE.jpg
2017-08-24 23:44:07 <danieli> the hamburger menu doesn't work properly, the logo is a bit large in comparison to the menu button
2017-08-24 23:44:26 <danieli> the text over the mountain picture wraps and looks bad on smaller phones
2017-08-24 23:47:49 <foxkit> http://i.imgur.com/0Hr03wX.png the scaling is junky on firefox
2017-08-24 23:48:03 <foxkit> http://i.imgur.com/qkmyehg.png http://i.imgur.com/oQONrW2.png
2017-08-24 23:48:17 <foxkit> but the menu does work on both devices (sorry, didn't screencap the open menu from firefox/android)
2017-08-24 23:48:30 <danieli> it does work, but doesn't look very good
2017-08-24 23:50:04 <foxkit> you said the menu doesn't work properly
2017-08-24 23:50:32 <danieli> "doesn't seem to work too well" design wise
2017-08-24 23:50:41 <danieli> my bad, i was more ambigious than i intended to
2017-08-25 04:50:36 <danieli> building coreos' grub to run some testing..
2017-08-25 06:08:03 <fcolista> hi all
2017-08-25 06:12:06 <danieli> hi you
2017-08-25 07:03:00 <clandmeter> danieli,  you can submit patches for our www. its on our github.
2017-08-25 07:07:55 <danieli> clandmeter: ah great, I did a quick google search but didn't spot it - do you know the name of the repo?
2017-08-25 07:08:20 <clandmeter> im sure you can find it :)
2017-08-25 07:08:26 <danieli> I'll look again :p
2017-08-25 07:09:31 <danieli> found it - strange repo name
2017-08-25 07:11:16 <clandmeter> it is?
2017-08-25 07:11:25 <danieli> s/strange/unusual/
2017-08-25 07:11:54 <clandmeter> its a static website generator
2017-08-25 07:12:09 <danieli> guessed so, currently grepping through the code
2017-08-25 07:13:45 <clandmeter> ncopa prefers to use/create his own tools (eat your own dog food model™)
2017-08-25 07:14:18 <danieli> dogfooding is nice, i like it a lot.. but i also do rely on third party libraries and tools, although i could rewrite most of them given enough time
2017-08-25 07:28:22 <fcolista> readline in edge is broken :-/
2017-08-25 07:28:33 <fcolista> it seems due to 23606ad3dba2ece01a669e6123c3426049561f7a
2017-08-25 07:29:19 <fcolista> nah
2017-08-25 07:29:23 <fcolista> i'm an idiot
2017-08-25 07:29:24 <fcolista> nm
2017-08-25 07:29:40 <danieli> i know the feel
2017-08-25 07:32:15 <fcolista> umh
2017-08-25 07:32:17 <fcolista> actually not
2017-08-25 07:32:33 <fcolista> main/readline built fine
2017-08-25 07:32:41 <fcolista> but has errors:
2017-08-25 07:32:45 <fcolista> https://dpaste.de/dqrz
2017-08-25 07:33:19 <fcolista> simply the build does not stop
2017-08-25 07:33:32 <fcolista> so it went unnoticed
2017-08-25 07:33:40 <fabled> parallel build issue?
2017-08-25 07:33:53 <fcolista>  make -w -j8
2017-08-25 07:34:00 <fcolista> smells like that
2017-08-25 07:34:14 <danieli> sounds and looks like it too
2017-08-25 07:34:28 <fcolista> running with -j1
2017-08-25 07:34:44 <fcolista> no
2017-08-25 07:34:49 <fcolista> still same errors
2017-08-25 11:33:52 <vkrishn> what happend to 'lilo' become a "vintage boot" !
2017-08-25 11:34:00 <vkrishn> and http://sourceforge.net/projects/elilo/ seems to support efi
2017-08-25 11:39:17 <vkrishn> https://www.centos.org/docs/5/html/5.2/Virtualization/chap-Virtualization-Configuring_ELILO.html
2017-08-25 11:40:58 <vkrishn> while deciding to add grub on cdrom (ahem!!), can /boot/grub files can be sorted to only include whats needed ?
2017-08-25 13:17:53 <^7heo> I got an issue with what I think is the Alpine kernel
2017-08-25 13:18:15 <^7heo> I have no sound when I replug my usb headset, but I have sound at boot.
2017-08-25 13:20:20 <^7heo> https://git.alpinelinux.org/cgit/aports/tree/main/linux-hardened/config-virthardened.x86_64#n2761
2017-08-25 13:20:30 <^7heo> I think it might help to set that to yes.
2017-08-25 13:20:53 <^7heo> And also the next line
2017-08-25 13:20:59 <^7heo> ncopa: ^
2017-08-25 13:21:02 <^7heo> Shiz: ^
2017-08-25 13:21:15 <^7heo> (RFC)
2017-08-25 13:24:43 <andypost> can I get this accepted https://github.com/alpinelinux/aports/pull/2079
2017-08-25 13:26:46 <fabled> andypost, looks mostly good. what are the changes to disabled tests?
2017-08-25 13:27:24 <andypost> fabled, last time they failed but ncopa disabled tests
2017-08-25 13:27:43 <fabled> are they fixed now?
2017-08-25 13:29:09 <andypost> fabled, they should be fixed with previous release
2017-08-25 13:29:35 <andypost> fabled, they exactly what 7.1.6 fixes
2017-08-25 13:30:42 <andypost> fabled, but travis using outdated kernel ... see https://github.com/alpinelinux/aports/pull/1851
2017-08-25 13:32:06 <andypost> fabled, and locale is not configured in travis, that's why all locale tests fails
2017-08-25 13:32:20 <andypost> and time related tests as well
2017-08-25 13:33:51 <fabled> i'm running the tests now to see what happens in my lxc
2017-08-25 13:45:37 <fabled> andypost, there's a tons of tests, no wonder it fails on travis-ci due to running out of time
2017-08-25 13:48:35 <andypost> fabled, yes that's a second trouble with travis
2017-08-25 13:49:17 <andypost> fabled, any idea how to move forward with that?
2017-08-25 13:49:46 <fabled> eventually, i hope we get to use our own CI that also test-builds others arches
2017-08-25 13:49:47 <andypost> fabled, php does not accept fixes to upstream if they have no tests
2017-08-25 13:50:05 <andypost> so test suite will grow
2017-08-25 13:51:39 <fabled> FAILED TEST SUMMARY
2017-08-25 13:51:39 <fabled> ---------------------------------------------------------------------
2017-08-25 13:51:39 <fabled> IntlDateFormatter: several forms of the calendar arg [ext/intl/tests/dateformat_create_cal_arg_variant4.phpt]
2017-08-25 13:51:39 <fabled> IntlDateFormatter: setCalendar()/getCalendar()/getCalendarObject() [ext/intl/tests/dateformat_get_set_calendar_variant4.phpt]
2017-08-25 13:51:41 <fabled> numfmt_get_locale() [ext/intl/tests/formatter_get_locale_variant2.phpt]
2017-08-25 13:51:43 <fabled> ..
2017-08-25 13:52:52 <fabled> maybe keep those three tests disabled at least still
2017-08-25 13:53:18 <fabled> andypost, ^
2017-08-25 13:54:02 <andypost> fabled, good idea, will update PR asap
2017-08-25 13:54:30 <andypost> at least bugs should be tested
2017-08-25 16:03:08 <ncopa> hi
2017-08-25 17:19:59 <foxkit> for #7737 can busybox be instructed to install applet symlinks to a different directory?  typically pgrep and pkill are in /bin instead of /usr/bin because they are "core" utilities that may be used before /usr is mounted
2017-08-25 17:20:31 <foxkit> I know the other option would be to move procps pgrep and pkill to /usr/bin (and then successfully override busybox) but that feels like wrong solution :/
2017-08-25 17:20:41 <ncopa> foxkit: i have normally patched busybox for that
2017-08-25 19:38:25 <danieli> hi
2017-08-25 19:59:16 <ncopa> hi
2017-08-25 20:08:10 <okeuday_bak> Hi, could someone review/merge https://github.com/alpinelinux/aports/pull/2140 ?
2017-08-25 20:33:34 <iamthemcmaster> ncopa: is this the right place to ask for PR reviews? if so I have one to request too :)
2017-08-25 20:33:45 <ncopa> okeuday_bak: done
2017-08-25 20:34:01 <okeuday_bak> ncopa: thanks!
2017-08-25 20:34:01 <ncopa> iamthemcmaster: i suppose it is...
2017-08-25 20:34:11 <ncopa> iamthemcmaster: url?
2017-08-25 20:34:14 <iamthemcmaster> https://github.com/alpinelinux/aports/pull/2155
2017-08-25 20:35:36 <iamthemcmaster> i wasn't sure if there was any more formal process for moving something from testing to community, if so i'm happy to do whatever
2017-08-25 20:37:00 <ncopa> we dont have any more formal process than that
2017-08-25 20:38:31 <ncopa> iamthemcmaster: you should not need depends="zlib"
2017-08-25 20:38:49 <ncopa> abuild should autodetect the libz.so dependency
2017-08-25 20:39:05 <ncopa> i dont think you need zlib-dev in makedepends_dev either
2017-08-25 20:39:34 <ncopa> $ tpaste < pkg/libhtp-dev/usr/lib/pkgconfig/htp.pc
2017-08-25 20:39:34 <ncopa> http://tpaste.us/BeV9
2017-08-25 20:39:48 <ncopa> there are no references to zlib in the .pc
2017-08-25 20:40:03 <ncopa> so if zlib is really needed, it is a bug and the .pc should be fixed
2017-08-25 20:40:17 <iamthemcmaster> hmm i'll mess with it and see
2017-08-25 20:40:23 <danieli> tpaste - another sprunge fork?
2017-08-25 20:40:31 <danieli> ooh lua
2017-08-25 20:40:45 <ncopa> yup
2017-08-25 20:40:52 <danieli> i forked sprunge and dockerized it, python + flask + redis
2017-08-25 20:41:10 <ncopa> clandmeter wrote tpaste.us in lua
2017-08-25 20:41:11 <danieli> chrome 60 broke the data: url in <a>
2017-08-25 20:41:23 <ncopa> tpaste.us is fast
2017-08-25 20:41:29 <danieli> redis is lightning fast
2017-08-25 20:41:41 <danieli> i called my sprunge fork 'resprunge' - very creative
2017-08-25 20:41:46 <ncopa> :)
2017-08-25 20:43:04 <ncopa> iamthemcmaster: i think you are right about the zlib-dev dep in makedepends_dev
2017-08-25 20:43:22 <ncopa> $ grep zlib  pkg/libhtp-dev/usr/include/htp/* | tpaste
2017-08-25 20:43:22 <danieli> https://p.duniel.no/AhSH?sh
2017-08-25 20:43:22 <ncopa> http://tpaste.us/NRLk
2017-08-25 20:43:50 <danieli> ncopa: building coreos/grub choked at the hardening flags
2017-08-25 20:43:52 <danieli> :(
2017-08-25 20:44:22 <ncopa> did you use the current grub APKBUILD?
2017-08-25 20:44:55 <danieli> I modified it very slightly, but yes - will retry changing some environmental stuff, see if it helps
2017-08-25 20:45:17 <ncopa> iamthemcmaster: you dont need the `|| return 1` anymore
2017-08-25 20:45:19 <danieli> without the pakbuild, i came across some issues in terms of flex and others
2017-08-25 20:46:48 <foxkit> I think I finally found a limitation of abuild that is above my head to fix
2017-08-25 20:46:50 <foxkit> conditional patching
2017-08-25 20:47:10 <ncopa> ok?
2017-08-25 20:47:11 <foxkit> it whines about sha512sums, and if I try to conditionally modify sha512sums it just dies
2017-08-25 20:47:32 <iamthemcmaster> ncopa: ok got it. so i should leave it in makedepends and depends_dev, but drop it from depends?
2017-08-25 20:48:08 <iamthemcmaster> it looks like the build process didn't find it as a depend for libhtp-dev w/o having it in there, but did for libhtp
2017-08-25 20:48:18 <ncopa> iamthemcmaster: and you need add a check() function
2017-08-25 20:48:32 <ncopa> $ git diff | tpaste
2017-08-25 20:48:32 <ncopa> http://tpaste.us/xWdJ
2017-08-25 20:48:51 <ncopa> iamthemcmaster: are you ok with the above change? ^^^
2017-08-25 20:49:09 <iamthemcmaster> yeah that looks good to me
2017-08-25 20:49:25 <ncopa> we should probably have a comment about the zlib-dev dep
2017-08-25 20:50:44 <iamthemcmaster> do i need to incorporate all of that and update the PR?
2017-08-25 20:51:00 <ncopa> ok with this? http://tpaste.us/yadl
2017-08-25 20:51:13 <ncopa> i think i can --amend it here this time
2017-08-25 20:51:31 <ncopa> or git rebase -i fixup it
2017-08-25 20:51:34 <iamthemcmaster> yup that looks good. i can add a check() function to the suricata package too
2017-08-25 20:54:39 <ncopa> have you tested that the init.d works?
2017-08-25 20:55:17 <iamthemcmaster> not extensively but it works on my firewall at home, at least
2017-08-25 20:55:35 <ncopa> thats enough i think
2017-08-25 20:57:10 <ncopa> the init.d could probably be somewhat better but i think its good enough
2017-08-25 20:57:20 <ncopa> there are some whitespace damage
2017-08-25 20:57:42 <ncopa> space before tab
2017-08-25 20:58:40 <ncopa> iamthemcmaster: are you ok with this? http://tpaste.us/YnNz
2017-08-25 20:59:14 <ncopa> the init.d change is only removal of space before tab
2017-08-25 20:59:48 <ncopa> one more question, can it use normal lua instead of luajit?
2017-08-25 21:00:02 <clandmeter> i forgot about tpaste, it actually is running on packet.net but its still using the nginx proxy on nl infra :)
2017-08-25 21:00:22 <iamthemcmaster> ncopa: hmm. not sure. i can certainly test that out and see
2017-08-25 21:00:28 <ncopa> luajit will likely need paxmarking on -hardened kernel
2017-08-25 21:00:56 <ncopa> iamthemcmaster: if you are ok i'll push the above diff with your commits
2017-08-25 21:01:28 <ncopa> clandmeter: thats why it takes 1/10 sec instead of the normal 1/100sec :)
2017-08-25 21:02:00 <clandmeter> huh? it should be slower now.
2017-08-25 21:02:17 <ncopa> thats what i'm saying
2017-08-25 21:02:23 <ncopa> its 10 times slower
2017-08-25 21:02:26 <clandmeter> oh yes
2017-08-25 21:02:27 <clandmeter> lol
2017-08-25 21:02:41 <ncopa> and im just kidding. i havent tested it
2017-08-25 21:02:47 <iamthemcmaster> ncopa: yeah that looks good to me. only thing i'm confused by is dropping the ../ in those couple places in package()
2017-08-25 21:03:19 <ncopa> iamthemcmaster: we normally install "$srcdir"/$pkgname.initd ...
2017-08-25 21:03:45 <ncopa> there should be a  symlink there
2017-08-25 21:04:02 <ncopa> this is so we in future can set srcdir=/var/something
2017-08-25 21:04:12 <iamthemcmaster> oh i see
2017-08-25 21:04:16 <iamthemcmaster> then yes, looks good to me
2017-08-25 21:07:30 <ncopa> iamthemcmaster: i pushed it
2017-08-25 21:07:32 <ncopa> thanks!
2017-08-25 21:07:55 <iamthemcmaster> awesome! thank you!
2017-08-25 21:08:08 <iamthemcmaster> do community packages get built with numbered releases? or are they only installable from edge?
2017-08-25 21:10:08 <ncopa> they are built with stable releases
2017-08-25 21:10:28 <ncopa> and have support for the last stable relase cycle
2017-08-25 21:10:41 <ncopa> so it will be in v3.7 now and supported til v3.8 is out
2017-08-25 21:14:28 <iamthemcmaster> awesome
2017-08-25 21:14:33 <iamthemcmaster> thanks for your help.
2017-08-25 21:15:41 <ncopa> thank you too!
2017-08-25 21:38:57 <foxkit> is there a reason fortify-headers only works with optimisations enabled?  seems silly limitation, and is making it really hard to debug a failure
2017-08-25 21:39:31 <foxkit> everything is inlined and optimised out so all I know is "the function, which calls strcmp, memcpy, and strstr, is around where the failure is" which doesn't help at all :/
2017-08-25 21:40:51 <foxkit> temporary hack seems to be "-O0 -D__OPTIMIZE__=1" in CFLAGS
2017-08-25 22:30:57 <ncopa> stateless: ^^^
2017-08-25 22:31:42 <ncopa> reminds me, there was a new release
2017-08-25 23:10:41 <kaniini> yes, there is actually a reason
2017-08-25 23:11:04 <kaniini> some behaviours that fortify depends on require optimizations to be enabled in earlier GCCs
2017-08-25 23:13:54 <foxkit> ahh, that makes sense
2017-08-26 19:36:07 <foxkit> /usr/lib/gcc/mips-foxkit-linux-musl/6.4.0/../../../../mips-foxkit-linux-musl/bin/ld: .gnu.hash is incompatible with the MIPS ABI
2017-08-26 19:36:43 <foxkit> why does Alpine force gnu-style hash on?  just curious, before I disable it on mips
2017-08-26 21:30:54 <foxkit> would "the entire tree can be crossed" be an eventual goal?
2017-08-26 21:32:42 <skarnet> I don't think it's a desirable goal
2017-08-26 21:33:11 <skarnet> it's difficult to cross properly, and for instance autotools does it wrong
2017-08-26 21:33:50 <skarnet> so the entire tree is probably made of things that would silently break when crossed
2017-08-26 21:35:14 <skarnet> it's safer to cross-bootstrap a small subset needed to natively build the rest of the tree
2017-08-26 21:35:55 <foxkit> I'm not sure my 400 MHz MIPS32 SoC is up to the task.
2017-08-26 21:36:23 <skarnet> would you prefer having incorrect binaries?
2017-08-26 21:36:39 <foxkit> the CPU could probably be left to its own devices for three months, but the 512 MB RAM with no manner to upgrade (kernel ABI limitation limits process memory to 512 MB anyway) is a significant blocker
2017-08-26 21:36:47 <skarnet> qemu
2017-08-26 21:37:01 <foxkit> I know for a fact qemu is /not/ up to the task
2017-08-26 21:37:12 <foxkit> if you recall, we tried to build s6-dns on qemu
2017-08-26 21:37:15 <skarnet> I know of your experiences with qemu
2017-08-26 21:37:17 <foxkit> it did not run on a real MIPS
2017-08-26 21:37:40 <foxkit> GXemul might work
2017-08-26 21:38:50 <skarnet> didn't s6-dns run on the real MIPS? I thought it just exposed a bug, and when the bug was fixed, the qemu-built one worked, am I wrong?
2017-08-26 21:42:08 <foxkit> skarnet: it exposed a bug /in qemu/ that meant the guest still saw little endian despite emulating big endian.
2017-08-26 21:42:20 <foxkit> it worked on my mipsel box, but not mips itself
2017-08-26 21:42:47 <skarnet> that's bad indeed XD
2017-08-26 21:46:45 <skarnet> anyway if I had decision making powers, I'd work on fixing emulators and build stuff natively, using emulators if real machines don't have enough power. Seems less error-prone than cross-building. And I'm pretty sure landley would agree for instance.
2017-08-26 21:49:18 <foxkit> gxemul docs say that the malta 4Kc, only 32-bit MIPS supported, can only boot NetBSD and RISCOS, not Linux
2017-08-26 21:49:20 <foxkit> :/
2017-08-26 21:50:23 <skarnet> can't you boot Linux from NetBSD? :)
2017-08-26 22:50:23 <foxkit> full cross gcc requires linux-headers when targeting mips
2017-08-26 22:59:12 <foxkit> no idea how to make that work, so that stops mips development for this weekend
2017-08-27 05:58:58 <foxkit> whoever did the bootstrap.sh stuff, thank you so much; you have made a much better, faster, more reliable system than gentoo's crossdev
2017-08-27 05:59:04 <foxkit> I've brought up three full arches today, and started two more
2017-08-27 05:59:14 <foxkit> it took multiple days per arch on gentoo
2017-08-27 06:16:46 <kaniini> foxkit: octeon
2017-08-27 06:17:10 <kaniini> foxkit: is the way to get fast mips
2017-08-27 06:17:45 <foxkit> what about mips32?  afaik, the kernel does not do 64->32 chroots
2017-08-27 06:17:51 <foxkit> and all octeons that I am aware of are mips, not mips32
2017-08-27 17:51:52 <mitchty> foxkit: out of curiosity could you post your notes, i've been trying to build an i386 cross compiler from x86_64 and have been running into nothing but trouble with it
2017-08-27 17:55:14 <mitchty> i presume i'm doing something dumb, as i'm to the point I need to patch binutils
2017-08-27 18:01:58 <foxkit> mitchty, "scripts/bootstrap.sh x86"?
2017-08-27 18:02:14 <foxkit> x86 was the easiest one for me...
2017-08-27 18:02:34 <foxkit> mips and ppc required binutils patches, and arm doesn't even work :)
2017-08-27 18:04:28 <mitchty> foxkit: bah, it always dies building binutils due to headers it can't find
2017-08-27 18:05:17 <mitchty> ok i'll keep poking at it with blunt rusty spoons
2017-08-27 18:11:33 <foxkit> headers?
2017-08-27 18:17:02 <mitchty> yep, i get this when i try bootstrapping https://gist.github.com/mitchty/091859f357e7f5005cd4bcf5c88af739, so when I set V=1 on the binutils APKBUILD, the problem appears to be with libiberty when building gold
2017-08-27 18:19:14 <mitchty> anyway, haven't debugged it much further but wasn't sure how many people use bootstrap.sh to know if i should start breaking out the winter nights to debug
2017-08-27 18:27:45 <foxkit> I don't see an error
2017-08-27 18:29:01 <mitchty> hold on, i'm running it again with SHELL="/bin/bash -vx", my memory might be off
2017-08-27 18:29:50 <mitchty> here's the errro
2017-08-27 18:29:54 <mitchty> err error rather https://gist.github.com/mitchty/fc1265e0c08b40b4dba2250b78e79d3e
2017-08-27 18:30:04 <mitchty> libiberty doesn't seem to be finding its includes
2017-08-27 18:32:02 <mitchty> i'm guessing my build env is setup wrong somehow
2017-08-27 18:34:06 <foxkit> hmm.  I'm using 2.29 here, not 2.28
2017-08-27 18:34:35 <foxkit> also have three patches that alpine master doesn't because mips and ppc 6_9
2017-08-27 18:35:19 <mitchty> this git checkout might be a bit outdated now, but for x86 I assume it would be ok
2017-08-27 18:36:28 <foxkit> I don't know if Alpine master has 2.29 yet, either.  I haven't sent a PR for it yet
2017-08-27 18:36:31 <foxkit> mainly because I totally forgot
2017-08-27 18:36:58 <mitchty> i also am testing out a patch for binutils arm, ghc is behaving erratically
2017-08-27 18:37:02 <foxkit> probably should update the PR and merge in the changes ncopa requested
2017-08-28 04:54:54 <foxkit> as just a piece of fun: I was able to successfully upgrade a Gentoo-based Adélie 1.0-alpha2 box to Alpine-based Adélie 1.0-alpha3 using apk upgrade -la
2017-08-28 04:55:00 <foxkit> and it all just works
2017-08-28 04:58:16 <danieli> cool
2017-08-28 07:40:22 <vkrishn> ncopa, I am trying to upgrade to v3.6, seems udhcpc is not able to set network, any help would be nice
2017-08-28 07:41:04 <vkrishn> also v3.1 used 74Mb at startup, now its 127Mb
2017-08-28 07:41:23 <vkrishn> tested with v3.6.0 and v3.6.2
2017-08-28 07:43:39 <vkrishn> how to disable it ?
2017-08-28 07:46:49 <vkrishn> I can pre-set ip address for net devices if needed
2017-08-28 07:52:29 <okeuday_bak> it appears that main/zlib isn't being built on the architecture armhf and it should be based on the package file
2017-08-28 08:14:55 <vkrishn> I wonder why this result, http://tpaste.us/oZJY
2017-08-28 09:35:26 <vkrishn> ok, now, re-arranged startup and cron scripts, thanks!
2017-08-28 15:05:02 <^7heo> Shiz: you told me there was a difference between gnupg1 and gnupg-2.X.X-rX but I don't remember
2017-08-28 15:05:12 <^7heo> or maybe I should search the logs
2017-08-28 15:06:49 <^7heo> <Shiz> i'll jump in and say gpg1 > gpg2
2017-08-28 15:06:51 <^7heo> Good enough for me.
2017-08-28 15:06:55 <^7heo> That's why I have it installed then.
2017-08-28 15:07:11 <Shiz> gpg2 needs a daemon running in the bg at all time i believe
2017-08-28 15:07:17 <^7heo> oh, okay.
2017-08-28 15:07:19 <Shiz> that;s why i prefer gpg1, as it's also still maintained
2017-08-28 15:07:23 <^7heo> yeah
2017-08-28 15:07:27 <^7heo> Thanks for stating why
2017-08-28 16:13:08 <ncopa> foxkit: congrats!
2017-08-28 16:22:06 <chambbj> Question about GitHub PRs (specifically new aports in the testing repository). I've had several sitting in the queue for weeks now, with very little attention. Do I need to advocate for those somewhere to move them along? Or is this fairly normal for new aports, especially in testing?
2017-08-28 16:22:36 <^7heo> I think that the guy who did most of the PR reviews is on (much deserved) holdidays right now
2017-08-28 16:22:47 <^7heo> so that might just be unfortunate timing.
2017-08-28 16:23:36 <chambbj> Fair enough. Just a new contributor to this project, so I'm not familiar with what is "customary".
2017-08-28 16:24:03 <^7heo> no worries; just trying my best to answer your question ;)
2017-08-28 16:37:16 <ncopa> chambbj: we do have problems keeping up. sorry
2017-08-28 16:37:28 <ncopa> chambbj: which PR is it?
2017-08-28 16:40:16 <PureTryOut[m]> ncopa: you mentioned running `default_prepare` instead of manually patching files, do you have any docs for that?
2017-08-28 16:41:47 <ncopa> PureTryOut[m]: not much docs, but: https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#prepare.28.29
2017-08-28 16:42:45 <PureTryOut[m]> yeah I saw that, but I don't really understand that. should I run `default_prepare || return` from within my prepare function, and will that automatically apply the patches? or what did you mean exactly
2017-08-28 16:52:11 <ncopa> running `default_prepare` will automatically apply the *.patch
2017-08-28 16:52:19 <ncopa> using `patch -p1 ...`
2017-08-28 16:54:04 <PureTryOut[m]> aah, good to know
2017-08-28 16:55:10 <PureTryOut[m]> does it have to be run in the directory where the patches have to be applied? so most of the time `"$srcdir"/$pkgbase-$pkgver`?
2017-08-28 17:00:24 <okeuday_bak> anyone know why zlib isn't being built for edge at http://build.alpinelinux.org/buildlogs/build-edge-armhf/main/
2017-08-28 17:00:27 <okeuday_bak> ?
2017-08-28 17:00:55 <okeuday_bak> oh, nm, it is built now
2017-08-28 17:01:30 <okeuday_bak> can someone trigger a rebuild of http://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/cloudi/cloudi-1.7.2_rc2-r0.log ?
2017-08-28 17:03:34 <chambbj> @ncopa No problem. PRs are #1891 (testing/eigen), #1893 (testing/nitro), #2063 (testing/libgeotiff), #2064 (testing/hexer), and #2066 (testing/lazperf).
2017-08-28 17:52:49 <ncopa> chambbj: i have reviewed them for you, sorry for not doing so earlier :-/
2017-08-28 17:52:57 <foxkit> fabled: thanks for merge!  I hope it was not too much pain.  next up will be a similar set, but much less, and about X11 stuffs
2017-08-28 17:53:20 <foxkit> think I already have 17 or so packages... none of them have any weird external deps, just 'expect' which is in main.
2017-08-28 17:53:26 <ncopa> did fabled merge the august 2017 PR with the zillion commits?
2017-08-28 17:54:03 <foxkit> yep... need to find a way to make it less painful for both sides... it took me two hours to prepare that branch yesterday :/
2017-08-28 17:54:12 <foxkit> maybe weekly instead of monthly
2017-08-28 17:54:39 <ncopa> thanks to both of you!
2017-08-28 17:54:52 <chambbj> Thanks @ncopa. I should have these updated this afternoon!
2017-08-28 17:55:01 <ncopa> i wanted to merge it last week but didnt reach it
2017-08-28 17:55:04 <foxkit> we are releasing alpha3 tomorrow, the first 'release' (alpha) post-Alpine switch
2017-08-28 17:55:10 <foxkit> it looks pretty good.  we have four arches supported
2017-08-28 17:55:14 <ncopa> nice!
2017-08-28 17:55:22 <foxkit> x86_32/64, and ppc32/64
2017-08-28 17:55:34 <ncopa> what is the genera feeling compared to base it on gentoo?
2017-08-28 17:56:46 <foxkit> generally good.  there are some weird things.  for instance, the live CD is about 40 MB smaller, despite having /more/ packages
2017-08-28 17:57:04 <foxkit> but lvm2 doesn't work at all and I don't know why.  it works fine on local computer but on live CD it just dies
2017-08-28 17:57:26 <ncopa> lvm2 code is nasty
2017-08-28 17:57:31 <foxkit> being able to have real triggers without nasty hack is a very good thing and it's solved a lot of problems on the x11 side
2017-08-28 17:57:36 <foxkit> yeah it is
2017-08-28 17:58:08 <ncopa> they do some really nasty hacks to prevent mlockall
2017-08-28 17:58:29 <ncopa> with alpine those hacks are not needed so we change the config to enable mlockall by default
2017-08-28 17:58:38 <ncopa> not sure if you use the alpine config for lvm or have your own
2017-08-28 17:59:29 <ncopa> foxkit: i am very happy to have you with us
2017-08-28 18:01:05 <foxkit> I haven't made any changes to lvm2 APKBUILD.  I don't think I even added test suite (can't remember though)
2017-08-28 18:01:19 <foxkit> ncopa: I'm happy to be here, in a community that is not designed to be as abusive as possible..
2017-08-28 18:01:31 <foxkit> gentoo got bad .-.
2017-08-28 18:01:47 <foxkit> they also started mass-removing support for arches like ppc, arm, etc and focus solely on x86_64 now
2017-08-28 18:02:01 <foxkit> though some people are 'fighting' that and putting arm back..
2017-08-28 18:02:21 <foxkit> a real mess.  no cohesion or attempt to solve things as a community.  it seems much better here
2017-08-28 18:18:00 <PureTryOut[m]> what does "fatal: No annotated tags can describe 'git hash here'. However, there were unannotated tags"
2017-08-28 18:18:04 <PureTryOut[m]> *mean?
2017-08-28 18:18:54 <ncopa> maybe ask in #git?
2017-08-28 18:19:07 <PureTryOut[m]> ooh it's just a git thing?
2017-08-28 18:19:29 <PureTryOut[m]> actually, maybe that's not why this apkbuild fails...
2017-08-28 18:19:50 <PureTryOut[m]> I ran `abuild -r`, the end of the log says the APKINDEX.tar.gz has an untrusted signature
2017-08-28 18:21:14 <ncopa> you probably need create or install local keys
2017-08-28 18:21:20 <ncopa> abuild-keygen -h
2017-08-28 18:24:18 <PureTryOut[m]> nope, installed the keys, keeps saying that
2017-08-28 18:27:55 <foxkit> PureTryOut[m]: it's just trying to add a git tag to the APKINDEX
2017-08-28 18:59:40 <PureTryOut[m]> ncopa: made the changes you requested to the consolekit2 PR
2017-08-28 19:00:19 <PureTryOut[m]> (I'll squash everything once it's deemed ready for merging)
2017-08-28 20:18:48 <ncopa> great
2017-08-28 23:51:26 <okeuday_bak> ncopa: is it possible to trigger a new build of http://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/cloudi/ the main/zlib dependency didn't exist until earlier today
2017-08-28 23:52:14 <ncopa> okeuday_bak: ok
2017-08-28 23:52:56 <ncopa> i think any push should trigger rebuild
2017-08-28 23:53:17 <ncopa> i think the arm build is blocked by other failure
2017-08-28 23:53:39 <ncopa> this one: http://build.alpinelinux.org/buildlogs/build-edge-armhf/main/icu/icu-59.1-r1.log
2017-08-28 23:53:43 <okeuday_bak> ncopa: k, then it is ok, if it isn't ready yet
2017-08-28 23:53:50 <ncopa> will try look at it tomorrow
2017-08-28 23:54:10 <okeuday_bak> ncopa: wasn't sure if that architecture failing was a problem
2017-08-29 06:24:00 <fcolista> morning all
2017-08-29 06:24:13 <danieli> good morning fcolista
2017-08-29 06:24:28 <fcolista> hi danieli
2017-08-29 06:26:21 <danieli> what's up?
2017-08-29 06:33:57 <scadu> hi
2017-08-29 08:55:59 <fabled> i filed a bug on libgcrypt/ppc64le issue at https://dev.gnupg.org/T3375
2017-08-29 08:56:13 <fabled> i disable tests on ppc64le for now
2017-08-29 13:36:22 <danieli> why is the alpinelinux.org website marked with a copyright?
2017-08-29 13:36:33 <danieli> just curious. it's already on github, and open source is nice
2017-08-29 13:42:28 <mosez> danieli: and why should it not be marked with a copyright?
2017-08-29 13:44:24 <clandmeter> its related to branding, not code.
2017-08-29 13:47:19 <_ikke_> Open source does not mean there is no copyright involved
2017-08-29 13:49:02 <danieli> clandmeter: gotcha, that's all i wanted to know
2017-08-29 13:49:14 <danieli> _ikke_: means the content is copyrighted but not the actual code behind it? or?
2017-08-29 13:49:21 <danieli> not sure in this case
2017-08-29 13:49:40 <_ikke_> Both can still be copyrighted
2017-08-29 13:50:11 <_ikke_> The license usually grants you permission to use, modify and distribute the code
2017-08-29 13:50:18 <danieli> can, of course, but what is it in this case?
2017-08-29 13:50:20 <clandmeter> most of our code is under MIT or related license.
2017-08-29 13:50:58 <danieli> clandmeter: didn't find info about that on your website, should probably be added
2017-08-29 13:51:16 <clandmeter> its not possible to list all licenses
2017-08-29 13:51:23 <clandmeter> its not a single product we ship
2017-08-29 13:51:37 <danieli> the primary ones, of course
2017-08-29 13:51:47 <danieli> iirc mostly GPL and MIT
2017-08-29 13:51:47 <clandmeter> well its a single product but build upon multiple projects.
2017-08-29 13:51:52 <danieli> well yeah
2017-08-29 13:52:00 <clandmeter> i talk about our own projects
2017-08-29 13:52:13 <danieli> that's good enough, your website is for alpine, not all the projects it incoroprates
2017-08-29 13:52:53 <_ikke_> http://pkgs.alpinelinux.org has info about the licenses for each package
2017-08-29 13:53:29 <danieli> the info I'm suggesting to put on your website would more or less be about the alpine parts of alpine, to put it like that
2017-08-29 13:54:02 <clandmeter> the license should be included in the repo's of those projects
2017-08-29 13:54:15 <clandmeter> not sure how you want us to change that?
2017-08-29 13:54:35 <_ikke_> https://git.alpinelinux.org/cgit/alpine-mksite/tree/ does not contain a license
2017-08-29 13:54:42 <danieli> I see
2017-08-29 13:54:48 <clandmeter> right
2017-08-29 13:54:50 <_ikke_> Not sure if it makes sense for that project though
2017-08-29 13:54:54 <clandmeter> because we are sloppy
2017-08-29 13:55:09 <danieli> yeah, it probably doesn't, but being consistent is usually nice
2017-08-29 13:55:19 <clandmeter> ncopa mostly puts it in MIT
2017-08-29 13:55:44 <danieli> does anyone know the reasons, or did he just choose the one he likes?
2017-08-29 13:57:48 <clandmeter> I always choose  MIT cause i mostly dont care about copyright.
2017-08-29 13:58:06 <danieli> define "don't care"?
2017-08-29 13:58:10 <danieli> licenses are pretty important
2017-08-29 13:58:37 <clandmeter> i dont care what other ppl do with it.
2017-08-29 13:58:53 <clandmeter> and im not an expert on licenses.
2017-08-29 13:59:44 <danieli> do you develop software only because you use it yourself?
2017-08-29 14:00:11 <danieli> luckily there's a lot of nice info online, and neat sites like https://choosealicense.com/
2017-08-29 14:01:43 <clandmeter> mit fits my profile, i never read up on other licenses cause its of no interest to me.
2017-08-29 14:01:57 <clandmeter> i prefer spend my time on other things :)
2017-08-29 14:02:25 <danieli> i've read up on licenses because i know the importance, my projects have different purposes too
2017-08-29 14:02:35 <danieli> and point taken - MIT is a nice and permissive license anyway
2017-08-29 14:04:10 <danieli> what do you guys estimate the size of the alpine community to be? user count wise
2017-08-29 14:08:52 <clandmeter> we are not activly collecting statistics so we dont have any numbers.
2017-08-29 14:12:28 <PureTryOut[m]> must be quite a lot, since Alpine seems to be popular in Docker containers
2017-08-29 15:48:29 <danieli> think doing geoip round robin DNS for mirrors would be a good idea?
2017-08-29 15:48:46 <danieli> like dl-{location}.alpinelinux.org or similar
2017-08-29 15:49:41 <_ikke_> danieli: you can add _compdef statements
2017-08-29 15:49:42 <danieli> and dl. to the closest one
2017-08-29 15:50:43 <danieli> _ikke_: compdef? only compdef i know of is in a zsh context
2017-08-29 15:50:45 <danieli> what do you mean?
2017-08-29 15:51:01 <_ikke_> danieli: sorry, wrong channel
2017-08-29 15:51:05 <danieli> oh hahaha
2017-08-29 16:02:33 <danieli> how much bandwidth does serving an alpine repo generally consume?
2017-08-29 16:41:52 <Shiz> danieli: a fair bit, but less so if you're non-primary
2017-08-29 16:41:57 <Shiz> like cz/nl
2017-08-29 16:42:12 <Shiz> i don't have any concrete numbers myself though
2017-08-29 17:12:18 <danieli> Shiz: got a number?
2017-08-29 17:12:29 <Shiz>   @Shiz : i don't have any concrete numbers myself though
2017-08-29 17:12:31 <Shiz> :p
2017-08-29 17:12:36 <Shiz> ncopa or clandmeter are more likely to know
2017-08-29 17:12:40 <danieli> oh sorry, i missed that line - my bad
2017-08-29 17:12:46 <danieli> i'm going back to my nap
2017-08-29 17:13:16 <danieli> hm
2017-08-29 17:13:20 <danieli> know anyone who has numbers?
2017-08-29 17:13:47 <skarnet> people generally offer me a drink before they ask for my number
2017-08-29 17:14:25 <danieli> lol
2017-08-29 17:19:39 <Shiz> i also just told you two guys who may have numbers :p
2017-08-29 17:19:52 <Shiz> jirutka should have the numbers for cz. which is a primary mirror
2017-08-29 19:26:28 <clandmeter> Daniel, cz mirror generates around 25Tb per month
2017-08-29 19:28:39 <clandmeter> A regular mirror around 1tb
2017-08-29 21:08:59 <Shiz> danieli: ^
2017-08-30 04:08:37 <danieli> Shiz: jesus christ, that's quite a bit. around a tb a month.
2017-08-30 05:53:10 <clandmeter> danieli, if you setup a new mirror it will probably take a year or more to generate that amount of traffic
2017-08-30 05:54:16 <clandmeter> the mirrors i have access to are older/master mirrors, im sure there are user mirrors with much less traffic.
2017-08-30 07:00:28 <danieli> clandmeter: gotcha
2017-08-30 09:20:28 <clandmeter> ah i didn't notice busybox has its own (limited) tls implementation in the latest version.
2017-08-30 09:20:40 <clandmeter> and it doesnt issue any warning about its limitations.
2017-08-30 09:22:05 <danieli> clandmeter: i did not know that
2017-08-30 09:23:17 <clandmeter> tls: a tiny TLS1.2 code, wire it up for wget
2017-08-30 09:25:28 <clandmeter> https://git.busybox.net/busybox/tree/networking/wget.c#n61
2017-08-30 10:27:31 <fabled> we should probably change github travis hook to use 'abuild rootbld' instead to detect if the package build() is trying to use internet
2017-08-30 12:04:36 <clandmeter> That would be nice yes
2017-08-30 12:04:54 <clandmeter> I guess a lot of packages will break
2017-08-30 12:07:34 <fabled> clandmeter, probably some; but I'm hopeful it's not too many :)
2017-08-30 12:07:44 <fabled> i know we have few Go packages that download stuff from net
2017-08-30 12:07:56 <fabled> and few others too; e.g. dahdi iirc.
2017-08-30 12:08:23 <fabled> net can be exposed via options="net" if really, really needed
2017-08-30 12:47:32 <grharry1> I am getting a "Segmentation fault occurred at   (nil)" when configuring openvpn with LDAP authentication caused by grsec. I Wonder if this can be fixed via some magic sysctl.conf parameters ?
2017-08-30 13:41:18 <clandmeter> fabled, i know of a few more then just go
2017-08-30 13:41:30 <fabled> rust too
2017-08-30 13:41:35 <clandmeter> yep
2017-08-30 13:41:36 <fabled> and yes, more also
2017-08-30 13:41:40 <fabled> ruby possibly
2017-08-30 13:41:40 <clandmeter> node
2017-08-30 13:42:07 <danieli> probably a lot of things
2017-08-30 15:41:23 <ncopa> grharry1: can you: ulimit -c unlimited
2017-08-30 15:41:28 <ncopa> and then get a coredump
2017-08-30 15:41:49 <ncopa> and create a backtrace with gdb?
2017-08-30 15:42:03 <ncopa> gdb --core ./core /path/to/openvpn
2017-08-30 15:42:05 <ncopa> bt
2017-08-30 15:42:14 <ncopa> and create an issue on bugs.alpinelinux.org?
2017-08-30 17:13:34 <ncopa> faffolter: im trying to pm you
2017-08-30 17:25:02 <TBB> there's that change in Freenode that might prevent you from doing that
2017-08-30 17:25:24 <TBB> altho, he's doing a pretty good job avoiding messages :)
2017-08-31 09:47:09 <ScrumpyJack> morning
2017-08-31 17:54:35 <Ganwell> I it possible to see the alpine builders logs somewhere? If I submit a package and something fails on non-x86, someone else has to fix it, because I don't learn about it.
2017-08-31 17:55:08 <Ganwell> I mean of course the failed one.
2017-08-31 17:56:13 <Ganwell> Of course, just as I ask I find it: http://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/py3-pyo/py3-pyo-0.8.6-r0.log
2017-08-31 17:58:00 <Ganwell> But that means something else is broken. The package has been built, but it isn't on the package index.
2017-08-31 19:41:56 <opendata> has anyone got the nvidia binary drivers working?
2017-08-31 19:53:13 <skarnet> not even Nvidia can pretend having achieved that
2017-08-31 20:03:03 <ollieparanoid[m]> ;D
2017-08-31 22:12:21 <skazz> Hello, is the repo 'https://git.alpinelinux.org/cgit/alpine-iso/' used to create the current 3.6.2 images? It seems to be a bit out of date still referencing grsec
2017-08-31 22:12:54 <_ikke_> skazz: yeah, iirc, they are using a different method now, but not sure what method
2017-08-31 22:19:49 <okeuday_bak> skazz: https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage is the current way
2017-08-31 22:20:42 <skazz> okeuday_bak: Thank you, having a look now =]
2017-08-31 22:24:33 <skazz> It says useradd when it should be adduser by default
2017-08-31 22:25:50 <skazz> and it needs alpine-sdk for the abuild-keygen
2017-08-31 22:29:35 <okeuday_bak> skazz: yeah, alpine-sdk was there in the past but someone must have removed it
2017-08-31 22:30:45 <skazz> Also for speed purposes and saved bandwidth, why not recommend in some places 'git clone --depth 10 --branch master git://git.alpinelinux.org/aports' ?
2017-08-31 22:33:01 <okeuday_bak> perhaps it wasn't there yet, will fix it
2017-08-31 22:36:35 <skazz> okeuday_bak: Creating the iso on that page, it refers to 'mkimage.sh', where is it? I'm just trying to build aports/scripts/mkimg.standard.sh
2017-08-31 22:37:01 <skazz> nvm
2017-08-31 22:37:03 <okeuday_bak> skazz: it is in aports/scripts/
2017-08-31 22:37:31 <skazz> okeuday_bak: Sorry, I thought it was installed, forgot the ./
2017-08-31 22:37:40 <okeuday_bak> skazz: I am not the expert on this, I just was able to make custom ISOs in the past, but the wikipage was worse then
2017-08-31 22:38:32 <skazz> okeuday_bak: I appreciate your help alot, thank you
2017-08-31 22:38:40 <okeuday_bak> np
2017-08-31 22:41:26 <okeuday_bak> skazz: the directions I use are at https://github.com/CloudI/live_cd/tree/master/alpine#alpine-linux-cloudi-livecd which I think is a simpler sequence, but it doesn't seem appropriate to make the central wikipage more complex, I am not sure
2017-08-31 22:49:11 <skazz> okeuday_bak: Thanks for that, I have all I need now, also need to change the workdir because a memory install does not have enough space
2017-08-31 23:19:18 <skazz> It is saying 'No space left on device' although /var is mounted to a disk with plenty of space and I have changed the outdir and workdir to be /var/mkiso/ (with correct permissions)
2017-08-31 23:19:31 <skazz> This is trying to use aports/scripts/mkimage.sh
2017-08-31 23:23:12 <okeuday_bak> skazz: you likely need to add something like https://github.com/CloudI/live_cd/blob/master/alpine/mkimg.cloudi.sh#L7-L8
2017-08-31 23:25:51 <skazz> okeuday_bak: I'm just trying to create the standard iso, it gets some of the way (42%) but then says failed to create boot/vmlinuz-hardened: No space left on device and linux-hardened-4.9.44-r0
2017-08-31 23:26:36 <okeuday_bak> skazz: that sounds like you are running out of filesystem space building the iso, and you may have filled up the /tmp/ dir with mkimage temp directories
2017-08-31 23:30:04 <skazz> okeuday_bak: Ah, ok, it was still using /tmp even though --workdir was set, seems to be running fine now, thanks again