2016-07-01 06:26:15 <barthalion> fcolista: is it abi compatible?
2016-07-01 06:26:46 <fcolista> <fcolista> ncopa, are you ok with http://sprunge.us/bRKh ?
2016-07-01 06:26:46 <fcolista> <fcolista> checkapk returns : http://pastebin.com/M3C8HG7X
2016-07-01 06:28:13 <barthalion> (y)
2016-07-01 06:43:42 <barthalion> ffmpeg requires ABI break
2016-07-01 06:43:46 <barthalion> and terrible API changes
2016-07-01 06:43:52 <barthalion> no breakage since 2.0
2016-07-01 06:43:54 <barthalion> that's terrible
2016-07-01 10:26:56 <fcolista> barthalion, ^^^ 
2016-07-01 10:27:39 <barthalion> fcolista: oops, thanks for pinging me
2016-07-01 10:29:10 <barthalion> where is the error there, actually?
2016-07-01 10:33:27 <barthalion> algitbot: retry build-edge-x86
2016-07-04 09:59:28 <clandmeter> ScrumpyJack: next time if you want to move stuff, make sure it builds, and all deps are also moved from testing.
2016-07-04 10:07:19 <clandmeter> smells like an abuild bug...
2016-07-04 10:07:52 <barthalion> nedit or w3m?
2016-07-04 10:07:56 <barthalion> nedit failure looks legit
2016-07-04 10:08:01 <clandmeter> w3m
2016-07-04 10:08:05 <clandmeter> the doc split 
2016-07-04 10:08:13 <clandmeter> there is no custom magic
2016-07-04 10:08:23 <barthalion> the error is about uncompressed man pages
2016-07-04 10:08:42 <clandmeter> i though abuild took care of that?
2016-07-04 10:08:59 <barthalion> Idk
2016-07-04 10:09:07 <clandmeter> ncopa: ?
2016-07-04 10:10:24 <clandmeter> i think i saw a local build also warn my about man pages/dirs in non -doc pkgs
2016-07-04 10:14:41 <ncopa> hi
2016-07-04 10:15:05 <ncopa> whats up
2016-07-04 10:16:07 <clandmeter> i have some pkgs failing on uncompressed man pages
2016-07-04 10:16:57 <ncopa> w3m has no maintainer
2016-07-04 10:17:17 <clandmeter> yeah i also noticed that.
2016-07-04 10:17:25 <clandmeter> will slap ScrumpyJack 
2016-07-04 10:18:32 <clandmeter> ok, nedit has custom doc. that explain part of it.
2016-07-04 10:20:19 <ncopa> w3m lacks subpkackage=-doc
2016-07-04 10:21:24 <clandmeter> yep... i should have checked.
2016-07-04 10:22:18 <fcolista> clandmeter, https://pkgs.alpinelinux.org/flagged reports mosquitto on edge/main that needs to be upgraded to 1.4.9, while mosquitto is already at that version
2016-07-04 10:27:06 <fcolista> ncopa, i'm going to push this: http://sprunge.us/aPfU ... this is the checkapk result, so it's safe to upgrade it: http://tpaste.us/3Jp7 
2016-07-04 10:28:15 <barthalion> minor releases are always safe
2016-07-04 10:28:24 <fcolista> yes
2016-07-04 10:28:29 <fcolista> but is always better check
2016-07-04 10:28:38 <barthalion> so no reason to warn anyone…
2016-07-04 10:28:48 <fcolista> i'm not the maintainer
2016-07-04 10:29:09 <ncopa> fcolista: its safe. please push
2016-07-04 10:29:16 <ncopa> and thanks!
2016-07-04 10:33:14 <clandmeter> lol
2016-07-04 10:33:24 <clandmeter> you beat me with 0.1 second
2016-07-04 10:34:13 <clandmeter> i had gettext-dev in my buildenv...
2016-07-04 10:41:54 <ncopa> i figured :)
2016-07-04 11:01:35 <ScrumpyJack> clandmeter: I checked that each package built, that everything installed and executed and provided advertised functionalilty, and I checked that each package dependancy was met in community. that's why i moved only very few packages, it takes a *long* time to do that. what broke?
2016-07-04 11:02:26 <clandmeter> ScrumpyJack: check git log :)
2016-07-04 11:03:07 <ScrumpyJack> i didn't make any changes to the APKBUILD files that needed tidying up. that was gonig to be my next step once the files are in community (it makes sending patches simpler to apply)
2016-07-04 11:04:33 <ScrumpyJack> so, w3m, h2o dvd+rw-tools and nedit. that's 4 too many :(
2016-07-04 11:04:57 <clandmeter> and maintainership ;-)
2016-07-04 11:05:21 <clandmeter> not sure how many didnt have maintainer
2016-07-04 11:07:38 <clandmeter> barthalion: 3.3.1?
2016-07-04 11:07:52 <ScrumpyJack> as i said, i didn't tidy and APKBUILD files (as it doesn't stop the builds). That's the next step
2016-07-04 11:09:15 <ScrumpyJack> sorry, i don't mean to sound annoying. blame it on Monday.
2016-07-04 11:09:38 <barthalion> clandmeter: yes, mpv likes to cry about this
2016-07-04 11:09:42 <barthalion> see main channel
2016-07-04 11:10:10 <clandmeter> barthalion: i didnt know the version number raises that fast.
2016-07-04 11:10:31 <barthalion> clandmeter: ah, you mean this
2016-07-04 11:10:34 <barthalion> 
ACTION shrugs

2016-07-04 11:10:37 <clandmeter> ;-)
2016-07-04 11:10:38 <barthalion> no harm done, just a commit message
2016-07-04 11:11:08 <barthalion> in some way it does, ffmpeg 0.10 was really cool release 8)
2016-07-04 11:11:42 <ScrumpyJack> clandmeter: w3m build fine for me :( buildozer complains it can't build w3m-lang, but there is no w3m-lang sub pacakge. wassup?
2016-07-04 11:12:01 <clandmeter> ScrumpyJack: no there were other issues with doc or something.
2016-07-04 11:12:15 <clandmeter> the lang part was my buildenv issue
2016-07-04 11:13:18 <ScrumpyJack> no doc issue here :(
2016-07-04 11:14:33 <ScrumpyJack> even after a fresh pull 
2016-07-04 11:15:24 <ScrumpyJack> dvd+rw-tools builds fine too
2016-07-04 11:18:21 <clandmeter> ScrumpyJack: make sure your buildenv is clean
2016-07-04 11:23:27 <ScrumpyJack> i do, and it is
2016-07-04 11:23:47 <ScrumpyJack> http://tpaste.us/2rWd
2016-07-04 11:24:37 <ScrumpyJack> oops
2016-07-04 11:26:09 <ScrumpyJack> removed avanger/db/pcare. should be clean now
2016-07-04 11:26:19 <ScrumpyJack> can't see that it makes a difference
2016-07-04 11:31:44 <ScrumpyJack> can you show me apk info on buildozer?
2016-07-04 12:24:48 <ScrumpyJack> hmm, i don't seem to be getting emails from patchwork
2016-07-04 12:26:05 <ScrumpyJack> clandmeter: does patchwork email notify work for you since you upgraded it?
2016-07-04 12:36:59 <clandmeter> ScrumpyJack: i dont submit patches, so i cannot tell.
2016-07-04 12:41:31 <ScrumpyJack> clandmeter: touche' :)
2016-07-04 12:54:44 <clandmeter> ScrumpyJack: i cannot locate any email regarding status change.
2016-07-04 12:54:56 <clandmeter> maybe it has changed in newer version.
2016-07-04 13:14:15 <clandmeter> ScrumpyJack: i just tried to signup and i do receive emails.
2016-07-04 13:16:54 <clandmeter> ah now i see it starts to email.
2016-07-04 13:17:05 <clandmeter> must have been the tls settings.
2016-07-04 13:21:26 <ScrumpyJack> neat. by the way, thanks to you and ncopa for fixing some of the APKBUILD files of the stuff I moved to community. I guess it's best if i did it at the same time as moving stuff
2016-07-06 06:45:55 <fabled> hum
2016-07-06 06:46:07 <fabled> configure: error: Pthreads are required to build libgomp
2016-07-06 06:46:07 <fabled> make[2]: *** [Makefile:20288: configure-stage1-target-libgomp] Error 1
2016-07-06 06:46:07 <fabled> make[2]: Leaving directory '/home/tteras/aports/main/gcc/src/build'
2016-07-06 06:46:07 <fabled> make[1]: *** [Makefile:22199: stage1-bubble] Error 2
2016-07-06 06:46:08 <fabled> ?!
2016-07-06 06:47:04 <fabled> oh
2016-07-06 06:47:05 <fabled> /home/tteras/aports/main/gcc/src/build/i586-alpine-linux-musl/libgomp/conftest.c:38: undefined reference to `__stack_chk_fail_local'
2016-07-06 06:47:05 <fabled> /usr/i586-alpine-linux-musl/bin/ld: conftest: hidden symbol `__stack_chk_fail_local' isn't defined
2016-07-06 06:47:05 <fabled> /usr/i586-alpine-linux-musl/bin/ld: final link failed: Bad value
2016-07-06 06:47:12 <fabled> ssp
2016-07-06 06:47:16 <fabled> x86 hacks
2016-07-06 07:15:44 <fabled> ok, i have now an idea. should get fix out soon
2016-07-06 08:08:31 <ncopa> seems like we have the memory issue with x86
2016-07-06 08:09:54 <ncopa> i suppose we have to paxmark go binary on x86 too
2016-07-06 09:05:27 <fabled> new musl too, i'll look at that next
2016-07-06 09:57:32 <clandmeter> ncopa: ^ because of recent changes?
2016-07-06 09:58:05 <clandmeter> i dont have these textrel issues on local builder
2016-07-06 09:59:03 <ncopa> hum
2016-07-06 09:59:06 <ncopa> dunno
2016-07-06 09:59:10 <ncopa> the textrels i dont know
2016-07-06 09:59:27 <ncopa> the terraform on x86 is due to bad memory allocator in golang 1.6
2016-07-06 10:05:56 <clandmeter> fabled: new gcc causes textrels 
2016-07-06 10:14:31 <fabled> clandmeter, i don't think gcc is used at all for go programs
2016-07-06 10:14:48 <clandmeter> fabled: i just verified it
2016-07-06 10:14:59 <clandmeter> build it with older gcc and its fine
2016-07-06 10:15:16 <clandmeter> then apk upgraded which only pulled in new gcc
2016-07-06 10:15:19 <clandmeter> then build fails
2016-07-06 10:16:22 <fabled> huh
2016-07-06 10:17:41 <clandmeter> fabled: http://tpaste.us/AyRl
2016-07-06 10:18:13 <fabled> smells there's some bundled .c stuff that is built
2016-07-06 10:19:37 <fabled> maybe it's ignoring CFLAGS or similar
2016-07-06 10:19:58 <fabled> or manually adding some no-pie / no-pic flag
2016-07-06 10:30:48 <clandmeter> fabled: i think i found the offending file
2016-07-06 10:30:51 <clandmeter> https://github.com/mattn/go-sqlite3/blob/master/sqlite3-binding.c
2016-07-06 10:32:13 <fabled> clandmeter, ew. could it use system sqlite3 instead?
2016-07-06 10:34:52 <clandmeter> i have no idea
2016-07-06 10:35:07 <fabled> brb
2016-07-06 10:35:12 <fabled> need to go wake up armhf builders
2016-07-06 10:35:17 <fabled> the edge one that is
2016-07-06 11:37:44 <barthalion> right, new gcc… I did not upgrade my build env
2016-07-06 12:13:28 <ScrumpyJack> barthalion: thanks for that dude
2016-07-06 12:13:40 <barthalion> np
2016-07-06 12:17:16 <ScrumpyJack> barthalion: are you having a look at the older ones too? :)
2016-07-06 12:17:26 <barthalion> going from the top
2016-07-06 12:17:37 <barthalion> not sure how much I will review today
2016-07-06 12:17:51 <ScrumpyJack> i think 2146 has already been applied
2016-07-06 12:19:23 <ScrumpyJack> anyway, really appreciate it
2016-07-06 12:50:27 <clandmeter> fabled: how is it possible im getting textrels with gcc 6 and with 5 I didnt?
2016-07-06 12:50:43 <fabled> various reasons
2016-07-06 12:50:54 <fabled> perhaps the app does not detect gcc6 #defines right
2016-07-06 12:51:06 <fabled> or something changed so -fPIC is not properly used
2016-07-06 12:52:38 <clandmeter> i tried to add -fPIC, but seems not to fix it.
2016-07-06 14:19:30 <nmeum> any idea what could cause the text relocations in gogs?
2016-07-07 07:43:42 <fabled> algitbot: retry master
2016-07-07 09:12:02 <fabled> huh
2016-07-07 09:12:05 <fabled> it built for me
2016-07-07 09:44:52 <fabled> sounds go now has some sort build race condition
2016-07-07 09:44:57 <fabled> now it seemed to build
2016-07-07 09:45:11 <fabled> for x86_64, but not x86
2016-07-07 09:47:35 <fabled> algitbot: retry master
2016-07-07 10:50:34 <fabled> algitbot: retry edge
2016-07-08 06:00:08 <fabled> algitbot: retry master
2016-07-08 06:00:43 <fabled> hum
2016-07-08 06:00:49 <fabled> that is probably race condition in build
2016-07-08 06:10:40 <barthalion> yeah, I couldn't get it to build reliably yesterday, so decided to stash the diff…
2016-07-08 06:14:08 <fabled> it was there for 1.6 already
2016-07-08 06:14:13 <fabled> i saw that fail similarly yesterday
2016-07-08 06:14:26 <fabled> must be something do with bootstrap building
2016-07-08 06:18:31 <fabled> fcolista, why cgmanager is x86_64 only when lxc depends on it and is all ?
2016-07-08 06:19:06 <barthalion> cgmanager should build fine on x86
2016-07-08 06:19:31 <barthalion> but it was disabled by clandmeter on 2015-09-24
2016-07-08 06:20:38 <clandmeter> oh that day, a rainy afternoon.
2016-07-08 06:21:12 <barthalion> yup, Septembers are awful for packagers
2016-07-08 06:21:29 <barthalion> clandmeter: do you have x86 build env to test it?
2016-07-08 06:21:42 <clandmeter> i think so
2016-07-08 06:22:04 <fabled> seems it works
2016-07-08 06:58:57 <clandmeter> \o/
2016-07-08 07:12:40 <fabled> hum. gcc seems to insist putting libgcc_s.so to lib64 in aarch64
2016-07-08 07:12:57 <fabled> ncopa, what was the thing with lib64?
2016-07-08 07:19:10 <fabled> http://lists.opensuse.org/opensuse-arm/2013-01/msg00100.html
2016-07-08 07:24:06 <ncopa> i think it was originally for allowing multilib
2016-07-08 07:24:09 <ncopa> lib32 + lib64
2016-07-08 07:24:17 <ncopa> we dont really support that in alpine
2016-07-08 07:26:52 <fabled> we seem to patch that for x86_64
2016-07-08 07:26:58 <fabled> should we patch aarch64 similarly?
2016-07-08 07:27:01 <fabled> i think that'd be ok
2016-07-08 07:33:32 <ncopa> we probably should
2016-07-08 07:35:51 <barthalion> ncopa: should I split -libs every time it is possible or can I give up on 124KB space savings?
2016-07-08 07:36:54 <barthalion> welp, I guess I will split it
2016-07-08 07:39:43 <ncopa> i dont split -libs every time its possible
2016-07-08 07:39:54 <ncopa> splitting libs makes sense if they libs are shared with other apps
2016-07-08 08:06:08 <barthalion> fabled: mind if I backport http://git.alpinelinux.org/cgit/abuild/commit/?id=97d924df to edge/abuild?
2016-07-08 08:06:55 <fabled> i think we should tag new abuild soon
2016-07-08 08:06:58 <fabled> ncopa, ^
2016-07-08 08:07:26 <fabled> i hope we could still have the delete / move file in $distfiles if checksum failed
2016-07-08 08:10:14 <ncopa> feel free to backport that to edge
2016-07-08 08:11:26 <fabled> ncopa, i'd really want new release soon
2016-07-08 08:11:32 <fabled> there's useful stuff in abuild.git
2016-07-08 08:20:34 <barthalion> redirection of 0r to -r would be useful, but that's more like PEBKAC…
2016-07-08 08:22:12 <fabled> rm 0r ?
2016-07-08 08:22:40 <barthalion> no, abuild
2016-07-08 08:22:57 <barthalion> I keep hitting 0 when I want to do abuild -r
2016-07-08 08:23:10 <barthalion> first world problems
2016-07-08 08:23:43 <fabled> -r might be made the default even
2016-07-08 08:35:35 <barthalion> fixing wbar
2016-07-09 15:12:01 <barthalion> fabled: are you there? Should I enable NEON for llvm?
2016-07-09 16:13:04 <fabled> barthalion, NEON is for armv7+ only; armhf should not have it on by default, armv7 yes
2016-07-09 16:13:18 <fabled> but if it supports it by runtime detection, then yes enable it
2016-07-09 19:58:47 <barthalion> no idea tbh…
2016-07-09 19:58:52 <barthalion> I wonder why did it build before
2016-07-11 19:55:47 <barthalion> \o/
2016-07-11 19:55:54 <barthalion> algitbot: cheers mate
2016-07-12 06:31:47 <ncopa> barthalion: sorry about cpupower issue
2016-07-12 06:32:19 <barthalion> ncopa: no problem
2016-07-12 06:33:50 <barthalion> did you ever see me swinging pitchfork?
2016-07-13 08:27:56 <barthalion> rnalrd: libmilter is in testing
2016-07-13 08:28:00 <rnalrd> y
2016-07-13 08:28:10 <rnalrd> gotta ask first to maint
2016-07-13 09:44:15 <fcolista> algitbot: retry 
2016-07-13 10:15:54 <barthalion> configure: error: pod2man must be installed
2016-07-13 10:15:59 <fcolista> i know
2016-07-13 10:16:01 <fcolista> fixing
2016-07-13 10:16:25 <fcolista> that's due to apkbuild-cpan installed in my buildenv
2016-07-14 07:19:38 <fabled> bah
2016-07-14 07:21:07 <clandmeter> fabled: uboot-tools not good enough?
2016-07-14 07:21:53 <barthalion> lol
2016-07-14 07:22:36 <fabled> we had that too?
2016-07-14 07:22:46 <fabled> i found only uboot-mkimage
2016-07-14 07:23:02 <clandmeter> there is an applet in apk to search for existing packages ;-)
2016-07-14 07:23:11 <fabled> oh, it's in testing
2016-07-14 07:23:22 <clandmeter> yes i told you i added it recently
2016-07-14 07:23:30 <fabled> i somehow missed that
2016-07-14 07:23:33 <clandmeter> else i couldnt convert the images
2016-07-14 07:23:50 <fabled> oh
2016-07-14 07:23:53 <fabled> i searched on wrong box
2016-07-14 07:23:58 <fabled> i searched on your box
2016-07-14 07:24:02 <fabled> not on mine usingedge
2016-07-14 07:24:06 <fabled> lol
2016-07-14 07:24:13 <barthalion> never trust clandmeter
2016-07-14 07:24:44 <fabled> sorry
2016-07-14 07:24:49 <fabled> too early, and no coffee yet
2016-07-14 07:24:51 <clandmeter> dont be
2016-07-14 07:25:04 <clandmeter> it feels good im not the only one making mistakes :)
2016-07-14 07:25:05 <barthalion> yeah, nothing happened
2016-07-15 09:32:42 <barthalion> ncopa: edge armhf were turned off, do you plan to backport this to 3.4?
2016-07-15 09:33:10 <ncopa> yes
2016-07-15 09:33:16 <ncopa> will have to do it on monday
2016-07-16 19:20:35 <barthalion> chris|: ^
2016-07-16 19:41:02 <chris|> hum, odd
2016-07-18 10:37:49 <^7heo> moin jirutka 
2016-07-18 10:42:29 <jirutka> ^7heo: hi, why in #alpine-commits? ;)
2016-07-18 10:44:13 <^7heo> because I didn't sleep yet. Sorry.
2016-07-18 10:45:28 <jirutka> ^7heo: heh
2016-07-18 17:36:22 <fcolista> barthalion, ^
2016-07-18 17:37:14 <barthalion> fcolista: I know, I'm watching
2016-07-18 18:20:17 <fcolista> barthalion, where are you on the list?
2016-07-18 18:20:49 <barthalion> fcolista: I'm jumping randomly now
2016-07-18 18:20:53 <barthalion> trying to build parole
2016-07-18 18:22:15 <fcolista> if gtk+3.0 is not finished, many packages does not build yet
2016-07-18 18:23:09 <barthalion> gtk3 is done now
2016-07-18 18:24:18 <barthalion> looks like we could just bump every pkgrel by 1 and wait
2016-07-18 18:24:25 <fcolista> yes
2016-07-18 18:24:40 <barthalion> fcolista: do you want to do it?
2016-07-18 18:24:58 <fcolista> already doing 
2016-07-18 18:25:03 <barthalion> ok
2016-07-18 18:29:11 <barthalion> fcolista: I'm pushing my last chunk and the rest is yours…
2016-07-18 18:30:05 <fcolista> if you have followed a line i would not become crazy to figure out what i need to do :P
2016-07-18 18:30:39 <barthalion> heh
2016-07-18 18:30:44 <fcolista> np
2016-07-18 18:30:52 <barthalion> it's quite pointless though, I couldn't move without other deps
2016-07-18 18:31:10 <barthalion> just bump each pkgrel, even the one I touched
2016-07-18 18:31:23 <fcolista> what's the command for bump ?
2016-07-18 18:31:27 <fcolista> abump?
2016-07-18 18:31:38 <fcolista> no
2016-07-18 18:31:51 <fcolista> apkgrel
2016-07-18 18:31:55 <barthalion> I don't know, I use some grep-eval-setconf magic
2016-07-18 18:31:59 <fcolista>  -a, --add        Add 1 to current pkgrel
2016-07-18 18:32:00 <barthalion> I don't have abuild installed locally
2016-07-18 19:54:43 <fcolista> bah
2016-07-18 19:54:54 <fcolista> why?!? Here builds :/
2016-07-18 19:54:59 <fcolista> gettext agagin
2016-07-18 19:55:01 <fcolista> *again
2016-07-19 09:05:55 <barthalion> 404
2016-07-19 13:59:32 <barthalion> clandmeter: let's bring it here hoping for less noise
2016-07-19 13:59:56 <clandmeter> im building world for aarch64
2016-07-19 14:00:09 <clandmeter> fabled bootstrapped the needed tools
2016-07-19 14:00:39 <clandmeter> so now it will fail cause some pkg deps on python.
2016-07-19 14:00:56 <barthalion> python2 has provides=python
2016-07-19 14:01:03 <clandmeter> right
2016-07-19 14:01:08 <clandmeter> but pyhthon2 is not yet build.
2016-07-19 14:01:12 <clandmeter> so its not in the index
2016-07-19 14:01:19 <clandmeter> so it cannot provide anything.
2016-07-19 14:01:20 <barthalion> hm, ok
2016-07-19 14:01:34 <clandmeter> im not sure this is really bad.
2016-07-19 14:01:46 <clandmeter> i can also build python2 first
2016-07-19 14:01:53 <barthalion> that would make sense to me, yes
2016-07-19 14:02:05 <clandmeter> but it does sound like a bug.
2016-07-19 14:02:15 <clandmeter> maybe aports should analyze provides
2016-07-19 14:02:38 <barthalion> yeah, I don't see what I can do actually
2016-07-19 14:03:01 <clandmeter> i think ncopa knows if this current way is right or needs a fix.
2016-07-19 14:14:32 <clandmeter> ncopa: ^
2016-07-19 14:22:49 <ncopa> what packages depends on python?
2016-07-19 14:22:56 <ncopa> which is needed for the boot strap
2016-07-19 14:23:33 <clandmeter> i forgot
2016-07-19 14:23:48 <clandmeter> but i can try it again.
2016-07-19 14:24:07 <clandmeter> i just checked out an older version of aports
2016-07-19 14:24:15 <clandmeter> before the python and gettext drama
2016-07-19 14:24:34 <clandmeter> alpine-sdk is ready :)
2016-07-19 14:25:23 <clandmeter> im trying to build lxc, but it pulls in openjade which result in an error.
2016-07-19 14:25:47 <clandmeter> i think their build scripts dont support aarch64
2016-07-19 15:18:08 <barthalion> clandmeter: someone in Fedora/Red Hat (like hrw) might have a patch for it
2016-07-19 15:18:45 <clandmeter> barthalion: fixed aleady :)
2016-07-19 15:18:52 <barthalion> cool
2016-07-19 15:19:01 <clandmeter> fedrora repo :)
2016-07-20 10:51:35 <barthalion> hm, it worked here
2016-07-20 10:53:09 <^7heo> Nevermind, you just got a high entropy build system.
2016-07-20 11:00:26 <barthalion> ^7heo: thank you for your valuable input
2016-07-20 11:04:26 <fabled> /home/buildozer/aports/main/llvm/src/llvm-3.8.1.src/include/llvm/Support/SwapByteOrder.h: In function 'void PrintThreadCommand(llvm::MachO::thread_command, const char*, bool, uint32_t)':
2016-07-20 11:04:26 <fabled> /home/buildozer/aports/main/llvm/src/llvm-3.8.1.src/include/llvm/Support/SwapByteOrder.h:44:33: fatal error: You must enable NEON instructions (e.g. -mfloat-abi=softfp -mfpu=neon) to use these intrinsics.
2016-07-20 11:04:26 <fabled>    return __builtin_bswap32(value);
2016-07-20 11:04:32 <fabled> sigh
2016-07-20 11:04:52 <fabled> why cant builtin_bswap32 work without neon?!
2016-07-20 11:06:05 <fabled> http://gcc.1065356.n5.nabble.com/Bug-target-71056-New-builtin-bswap32-NEON-instruction-error-with-O3-td1262430.html
2016-07-20 11:06:23 <barthalion> https://github.com/archlinuxarm/PKGBUILDs/blob/7615302b2f85404625ad996d1ed12f1f87929b0c/core/gcc/0002-PR-target-71056-Don-t-use-vectorized-builtins-when-N.patch
2016-07-20 11:06:28 <barthalion> how about this patch?
2016-07-20 11:06:35 <^7heo> barthalion: my pleasure.
2016-07-20 11:06:55 <barthalion> fabled: ok, you found it anyway
2016-07-20 11:08:08 <fabled> yes, looks good
2016-07-20 11:10:30 <barthalion> fabled: are you going to apply it or want me to do so?
2016-07-20 11:10:37 <fabled> i'll apply it soon
2016-07-20 11:10:46 <fabled> i may need to add other patches for aarch64 too
2016-07-20 11:10:53 <fabled> then again
2016-07-20 11:10:54 <barthalion> ok
2016-07-20 11:11:01 <fabled> would make sense to make progress
2016-07-20 11:11:05 <fabled> maybe just go ahead and add it
2016-07-20 11:11:16 <fabled> i don't know when i get aarch64 issue figured out
2016-07-20 11:11:33 <barthalion> I see, I'll apply it then
2016-07-20 11:11:53 <fabled> thanks
2016-07-20 11:15:23 <barthalion> this should unblock mesa as well, and then webkitgtk is left to fix
2016-07-20 11:15:28 <barthalion> assuming nothing broke in the meantime
2016-07-20 14:59:55 <fabled> it needs gcc first
2016-07-20 15:00:01 <fabled> i'm fixing that manually
2016-07-22 05:50:09 <fabled> the armhf u-boot failure is https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70830
2016-07-22 05:50:21 <fabled> i wonder if to take gcc-6-stable snapshot, or just cherrypick the fix
2016-07-22 06:28:48 <RieSh2be> snapshot will get you more fixes you did not think about, so why not
2016-07-22 06:37:35 <fabled> well, the snapshots are not exactly built like the release tarballs. though, it seems to be missing only some of the generated doc files.
2016-07-22 06:37:52 <fabled> maybe i'll do what debian does, and generate a cumulative diff from 6.1.0 to latest 6-stable snapshot
2016-07-22 06:38:30 <fabled> it's 10M+; so maybe i add with --exclude po to keep it under 2M
2016-07-22 06:39:06 <fabled> 6.2 is due out soonish too, i was hoping it to come already
2016-07-22 12:04:42 <fabled> algitbot: retry edge
2016-07-22 12:04:47 <fabled> algitbot: retry master
2016-07-25 11:59:48 <fabled> curl: (18) transfer closed with 15610134 bytes remaining to read
2016-07-25 11:59:49 <fabled>  ?!
2016-07-25 12:00:14 <fabled> smells curl bug
2016-07-25 12:00:59 <fabled> wget works on armhf
2016-07-25 12:01:26 <fabled> algitbot: retry edge
2016-07-25 12:02:36 <fabled> algitbot: retry master
2016-07-25 16:35:27 <bpiotrowski> algitbot: retry master
2016-07-26 13:52:29 <bpiotrowski> hrm, again this
2016-07-26 13:52:43 <bpiotrowski> something is off with curl
2016-07-26 13:56:10 <fabled> yes
2016-07-26 13:56:13 <fabled> seems to be broke on arm
2016-07-26 13:56:22 <fabled> it just somehow disconnects in middle of download
2016-07-26 14:12:00 <nmeum> oops
2016-07-27 12:08:05 <fabled> rebased cross-build, and fixed up one commit too
2016-07-27 23:20:18 <nmeum> hm, what could have broken curl on arm? The aport itself hasn't been modified in a while…
2016-07-29 10:00:49 <^7heo> yaaay