#archlinux-ports | Logs for 2018-05-23

Back
[00:18:32] -!- bork has quit [Read error: Connection reset by peer]
[00:24:35] -!- bork has joined #archlinux-ports
[00:32:23] -!- ebrasca has quit [Ping timeout: 252 seconds]
[00:45:57] -!- ebrasca has joined #archlinux-ports
[01:21:15] -!- belanthor has joined #archlinux-ports
[03:14:01] -!- belanthor has quit [Remote host closed the connection]
[03:23:39] -!- oaken-source has joined #archlinux-ports
[03:29:34] -!- oaken-source has quit [Ping timeout: 260 seconds]
[04:23:32] -!- p71 has quit [Remote host closed the connection]
[04:44:48] -!- p71 has joined #archlinux-ports
[05:40:09] -!- hringriin_ has joined #archlinux-ports
[05:43:19] -!- hringriin has quit [Ping timeout: 252 seconds]
[05:43:19] hringriin_ is now known as hringriin
[06:35:40] -!- titus_livius has joined #archlinux-ports
[09:57:18] -!- oaken-source has joined #archlinux-ports
[10:27:35] -!- oaken-source has quit [Ping timeout: 240 seconds]
[10:59:12] -!- oaken-source has joined #archlinux-ports
[11:19:19] <ebrasca> oaken-source: Hi
[11:35:07] <oaken-source> ebrasca: hi
[11:35:26] -!- insecta has quit [Quit: WeeChat 2.1]
[11:37:00] -!- insecta has joined #archlinux-ports
[11:54:53] <ebrasca> oaken-source: It is copiling with only 1 thread.
[12:00:32] <ebrasca> oaken-source: Can I make to compile with more threads? (stage3)
[12:03:23] <oaken-source> ebrasca: take a look at build/stage3/config/makepkg.conf
[12:03:42] <oaken-source> what does the MAKEFLAGS="" line say?
[12:09:28] <ebrasca> oaken-source: MAKEFLAGS="-j9"
[12:10:01] <eschwartz> if the sources don't support multiple jobs, not much you can do
[12:10:37] <ebrasca> I am sure gcc suport multiple jobs.
[12:10:56] <eschwartz> gcc definitely should, yes
[12:10:58] <oaken-source> do you have maybe set MAKEFLAGS in the environment? that would override the value in makepkg.conf
[12:11:10] <ebrasca> oaken-source: How to check?
[12:11:22] <oaken-source> env
[12:12:07] <eschwartz> htop => "e" while selecting the make job itself => look to see what MAKEFLAGS is defined as
[12:12:08] <ebrasca> oaken-source: but I need to check it in host or chroot of stage3?
[12:13:01] <oaken-source> ^ the idea by eschwartz is better than mine :)
[12:13:58] <eschwartz> I guess you could use ps to figure out the PID, then cat /proc/$PID/environ, but only if you hate tools that do things for you...
[12:15:24] <ebrasca> I think it is MAKEFLAGS=w --
[12:16:07] <eschwartz> then you could inspect the process tree going upward to see where that env first came into being
[12:24:35] <ebrasca> Here screnshot of htop in host. https://ibb.co
[12:24:36] <phrik> Title:2018 05 23 122255 1920x1080 scrot — imgbb.com (at ibb.co)
[12:30:43] <ebrasca> Up it is "MAKEFLAGS="
[12:31:45] <ebrasca> And "MAKEFLAGS=w --" I can't see other MAKEFLAGS.
[12:47:42] <ebrasca> Is someting bad?
[12:48:08] <oaken-source> I don't know
[13:00:40] <eschwartz> MAKEFLAGS="" implies something is explicitly setting it to empty
[13:00:58] <eschwartz> otherwise the var would simply not exist
[13:18:43] <ebrasca> I have no idea what it can be.
[13:42:51] -!- guys has quit [Ping timeout: 265 seconds]
[14:04:57] -!- e1z0 has quit [Ping timeout: 240 seconds]
[14:07:51] -!- e1z0 has joined #archlinux-ports
[14:35:30] -!- oaken-source has quit [Ping timeout: 260 seconds]
[14:40:16] -!- guys has joined #archlinux-ports
[15:18:11] -!- oaken-source has joined #archlinux-ports
[15:39:30] -!- isacdaavid has joined #archlinux-ports
[16:11:24] -!- guys has quit [Ping timeout: 252 seconds]
[16:15:40] -!- insecta has quit [Ping timeout: 268 seconds]
[16:17:33] -!- insecta has joined #archlinux-ports
[16:27:07] -!- guys has joined #archlinux-ports
[16:55:38] -!- abaumann has joined #archlinux-ports
[17:22:24] -!- isacdaavid has quit [Quit: Leaving.]
[17:30:14] -!- e1z0 has quit [Ping timeout: 252 seconds]
[17:31:45] -!- e1z0 has joined #archlinux-ports
[17:57:45] -!- isacdaavid has joined #archlinux-ports
[18:15:59] -!- bill-auger has quit [Remote host closed the connection]
[18:18:45] -!- bill-auger has joined #archlinux-ports
[18:26:47] -!- sunshavi has quit [Quit: restart]
[18:32:36] -!- sunshavi has joined #archlinux-ports
[19:12:39] -!- sunshavi has quit [Ping timeout: 268 seconds]
[19:37:50] -!- isacdaavid has quit [Quit: Leaving.]
[20:17:57] -!- abaumann has quit [Quit: leaving]
[20:29:58] <ebrasca> oaken-source: How is your copiling going in stage3?
[20:37:06] <oaken-source> segfaults in glibc package stage.
[20:37:52] <ebrasca> oaken-source: I have no idea what it is doing now.
[20:41:25] <ebrasca> oaken-source: Are you sure it some day is going to end? ( I think my PC have done this 4 "bash-decross gcc-libs make-decross ncurses" )
[20:42:53] <oaken-source> ebrasca: yeah, it takes a long time to compile
[20:46:05] <ebrasca> oaken-source: I hate it only using 1 thread of 8 to compile while in other arquitecture.
[20:52:05] <oaken-source> I don't know why that is
[20:52:18] <oaken-source> it should honor the setting in makepkg.conf
[20:53:34] <girls> ebrasca: do you build in a chroot and does that have a different /etc/makepkg.conf?
[20:54:09] <eschwartz> [06:03:23 AM] <oaken-source> ebrasca: take a look at build/stage3/config/makepkg.conf
[20:54:09] <eschwartz> [06:03:42 AM] <oaken-source> what does the MAKEFLAGS="" line say?
[20:54:09] <eschwartz> [06:09:28 AM] <ebrasca> oaken-source: MAKEFLAGS="-j9"
[20:54:33] <eschwartz> I'm assuming this is what's being cp'ed into the build chroot
[20:55:02] <girls> there is some sedding happening with the /etc/makepkg.conf (at least with archbuild, there is)
[20:55:15] -!- bork has quit [Ping timeout: 245 seconds]
[20:55:20] <girls> so it's worth to look at the actual file
[20:56:19] <eschwartz> oh, true
[20:56:26] <eschwartz> sed -e '/^MAKEFLAGS=/d' -e '/^PACKAGER=/d' -i "$copydir/etc/makepkg.conf"
[20:56:55] <eschwartz> for x in ... "MAKEFLAGS='${MAKEFLAGS:-}'"; do echo "$x" >>"$copydir/etc/makepkg.conf"; done
[20:57:39] <eschwartz> But, I did assume this is set in the makepkg.conf during the build too, and that ebrasca looked at the file to double-check
[20:57:42] <ebrasca> It is MAKEFLAGS=''
[20:57:51] <eschwartz> Welp
[20:58:01] <eschwartz> if the conf in the chroot is empty...
[20:58:29] -!- bork has joined #archlinux-ports
[20:59:18] <eschwartz> I did neglect to realize makechrootpkg actually explicitly sets it in that case
[20:59:47] -!- bork has quit [Excess Flood]
[20:59:57] -!- bork has joined #archlinux-ports
[21:08:40] <ebrasca> How to fix it?
[21:15:05] <girls> ebrasca: what command do you invoke to build?
[21:15:12] <girls> This takes some makepkg.conf from somewhere
[21:15:18] <girls> and there you need to edit
[21:16:07] <girls> e.g. staging-x86_64-build takes it from /usr/share/devtools/makepkg-x86_64.conf
[21:18:29] <ebrasca> girls: I run it with "sudo bash create.sh"
[21:18:52] <girls> ah, and what does that call?
[21:20:01] <ebrasca> girls: stage3.sh
[21:20:36] <girls> where is that script? Is there some code base?
[21:21:51] <ebrasca> Here is link to git https://github.com
[21:21:53] <phrik> Title:parabola-riscv64-bootstrap/stage3.sh at ppc64le · oaken-source/parabola-riscv64-bootstrap · GitHub (at github.com)
[21:24:56] <girls> there is 'MAKEFLAGS="-j$(($(nproc) + 1))"' in makepkg.sh
[21:26:49] <eschwartz> well, makechrootpkg gets this from your $HOME/.makepkg.conf
[21:27:43] <oaken-source> right, ~/.makepkg.conf might override this as well
[21:27:50] <oaken-source> good thinking
[21:27:57] <ebrasca> oaken-source: but
[21:28:07] <ebrasca> oaken-source: Someting overwrite it later ...
[21:28:46] <oaken-source> did you manually call librechroot sometime?
[21:29:03] <oaken-source> that would override the file, unless you pass -M
[21:29:06] <eschwartz> makechrootpkg derives select values from /etc/makepkg.conf and $HOME/.makepkg.conf... like PKGDEST, and also MAKEFLAGS
[21:29:43] <eschwartz> it then uses these select values to overwrite the values in the chroot
[21:31:10] <ebrasca> eschwartz: Thanks I have write to /etc/makepkg.conf and now it have correct value in chroot.
[21:34:19] <eschwartz> The idea is that makechrootpkg creates clean chroot builds, but the settings which don't impact the build, merely the packaging metadata and output directories, are things which should respect your makepkg configs
[21:34:37] <eschwartz> MAKEFLAGS is something which doesn't impact builds
[21:34:50] <eschwartz> (==> products)
[21:35:16] <girls> eschwartz: only true, when you don't have race conditions ;-)
[21:36:17] <ebrasca> girls: I have never have any race conditions. ( I think )
[21:36:37] <girls> lucky you
[21:36:40] <girls> :-D
[21:36:53] <eschwartz> that's broken packages
[21:36:58] <girls> I know
[21:37:08] <eschwartz> and they should explicitly set options=(!makeflags)
[21:37:23] <eschwartz> for this precise reason
[21:37:44] <girls> my approach is to report upstream and get it fixed there :-D
[21:38:32] <eschwartz> yes, but the PKGBUILD should still work around this while the upstream bugreport is open...
[21:45:58] <ebrasca> Yea it work 100% of my cpu copiling XD.
[21:50:27] <oaken-source> finally progress :)
[21:50:48] <ebrasca> I was thinking it is only afected by contend in scripts of oaken-source.
[21:51:23] <eschwartz> nope, for example the PACKAGER variable is set from there
[22:50:28] -!- belanthor has joined #archlinux-ports
[22:50:28] <oaken-source> ebrasca: I have unified the ports on the master branch of the repo and will continue working on there. It might be incompatible with the ppc64le branch now
[23:22:43] -!- sunshavi has joined #archlinux-ports
[23:48:16] -!- oaken-source has quit [Ping timeout: 252 seconds]