#archlinux32 | Logs for 2018-09-14

[00:42:43] -!- prstoetzer has quit [Quit: prstoetzer]
[00:44:04] -!- prstoetzer has joined #archlinux32
[01:53:05] -!- isacdaavid has quit [Quit: Leaving.]
[02:33:22] <buildmaster> haskell-yesod-test is broken (says buildknecht2) - I rescheduled: haskell-yesod-form, haskell-yesod-persistent.
[03:00:09] -!- KitsuWhooa has quit [*.net *.split]
[03:03:20] -!- KitsuWhooa has joined #archlinux32
[03:03:26] -!- rdap has joined #archlinux32
[03:04:30] <rdap> has anyone else been having problems running firefox on arch 32?
[03:32:06] <buildmaster> shellcheck is broken (says buildknecht).
[03:32:58] -!- woshty has quit [Ping timeout: 245 seconds]
[04:59:33] <rdap> thanks, i'll have a go at building from upstream, then.
[05:12:47] <elibrokeit> rdap: what problems?
[05:12:56] <elibrokeit> building it has been... interesting lately
[05:13:46] <elibrokeit> like, have fun with exhausted RAM when not cross-compiling fun, but a side order of SSE2 everywhere etc.
[05:48:15] <rdap> specifically, this:
[05:48:39] <rdap> XPCOMGlueLoad error for file /usr/lib/firefox/libxul.so:
[05:48:39] <rdap> libicui18n.so.61: cannot open shared object file: No such file or directory
[05:48:40] <rdap> Couldn't load XPCOM.
[05:48:40] <rdap>
[06:01:10] <elibrokeit> well, that's just because firefox needs to be rebuilt for the new ICU package, but firefox cannot be rebuilt right now for other reasons
[06:28:59] * buildmaster goes insane.
[06:39:44] * buildmaster resumes sanity.
[06:56:54] -!- rdap has quit [Ping timeout: 264 seconds]
[07:35:11] -!- abaumann has joined #archlinux32
[07:35:12] <buildmaster> Hi abaumann!
[07:37:07] <abaumann> rdap: rebuilding firefox on 32-bit is challenging at the moment, alternatives is to build it outside of the build system (I had some success when patching it for non-LTO and fiddling with some memory settings of the machine.
[07:37:42] <abaumann> A good alternative is various AUR packages with binary versions of the 32-bit firefox from mozilla diretly, see AUR firefox-bin or https://bbs.archlinux32.org
[07:37:43] <phrik> Title:x86_64 Extra firefox 61.0.1-1 Standalone web browser from mozilla.org (Page 2) / Installation / Arch Linux 32 Forums (at bbs.archlinux32.org)
[07:46:28] <abaumann> deep42thought: I have two unsigned firefox/thunderbird packages 62.0 built on a real machine without chroot with 4 GB RAM. We could maybe sign them and put them into the system.. (basically manual huge_memory builds from my point), http://archlinux32.andreasbaumann.cc
[07:46:30] <phrik> Title:Index of /stuff (at archlinux32.andreasbaumann.cc)
[07:47:29] <abaumann> buildmaster: !wtf dub
[07:47:37] <buildmaster> abaumann: [community] dub (1.10.0-1.0): /usr/bin/dub /usr/share/bash-completion/completions/dub
[07:54:44] <abaumann> buildmaster: why don't you stabilize dub?
[07:54:45] <buildmaster> abaumann: Cannot find "dub?"
[07:54:49] <abaumann> buildmaster: why don't you stabilize dub
[07:54:50] <buildmaster> abaumann: Cannot find "dub"
[07:56:44] <abaumann> Ok. db-update didn't want to do anything with dub.. I rescheduled it for building..
[08:00:54] <abaumann> mmh. dub got built as 1.11.0, but cannot be tested or moved? weird..
[08:15:18] -!- deep42thought has joined #archlinux32
[08:15:19] <buildmaster> Hi deep42thought!
[08:15:26] <abaumann> ah. dub has no link to dmd whatsoever..
[08:17:13] <deep42thought> abaumann: I'll put firefox and thunderbird into the repositories
[08:17:24] <deep42thought> for the future, we might want to have some script which does that?
[08:17:53] <abaumann> yes. maybe a good idea.
[08:18:09] <abaumann> but generally we should not build packages outside the build system..
[08:18:52] <deep42thought> yeah
[08:19:13] <deep42thought> but we wanted to put manually built packages into [build-support] once - so that seems reasonable
[08:19:21] <abaumann> yes.
[08:19:38] <deep42thought> can you sign the packages, please?
[08:19:45] <abaumann> I have a workaround.. but it means: start the build and be fast and fiddle with /var/lib/archbuild/xxx while the package starts building. :-)
[08:19:49] <abaumann> ok.
[08:20:05] <abaumann> are detacheds sig's ok?
[08:20:16] <deep42thought> yes
[08:20:32] <deep42thought> gpg --detach-sign --local-user=$your-build-key $pkg.tar.xz
[08:20:39] <abaumann> done.
[08:20:48] <abaumann> on 'stuff'
[08:21:13] <deep42thought> thx
[08:21:18] <deep42thought> this is your build key, right?
[08:21:25] <abaumann> I hope so. :-)
[08:21:26] <abaumann> yes.
[08:21:41] <deep42thought> at least, it's known to my pacman ;-)
[08:21:56] <abaumann> The same I use on the build slave eurobuild3 and on the i486 VM
[08:22:26] <deep42thought> perfect
[08:22:51] <deep42thought> apropos i468vm: I still need your ssh key (and a name for the vm)
[08:22:58] <abaumann> not sure what the permance impact will be of dropping LTO in Gecko, but let's see :-)
[08:23:04] <abaumann> ah. sorry. yes.
[08:23:10] <deep42thought> no hurry
[08:23:18] <deep42thought> I need to fix my glibc patch first, anyways ;-)
[08:24:12] <abaumann> mmh, we have tyzoid-srv0-bs0, so tyzoid-srv0-vm486 maybe
[08:24:21] <deep42thought> ok
[08:24:56] <deep42thought> operator is you (not tyzoid)?
[08:24:57] <abaumann> mmh. as for the key..
[08:25:12] <abaumann> where is the key of eurobuild3? *puzzle*
[08:25:20] <deep42thought> should I use the same?
[08:25:43] <abaumann> you can.
[08:25:50] <abaumann> most likely I just copied it.
[08:25:52] <deep42thought> no, I can't
[08:25:53] <deep42thought> sry
[08:26:05] <deep42thought> the build master distinguishes the slaves by ssh-key
[08:26:06] <abaumann> which is actually a good thing :-)
[08:26:07] <deep42thought> I forgot that
[08:29:28] * buildmaster goes insane.
[08:30:12] <deep42thought> that was fast
[08:37:46] <deep42thought> building packages by hand has the disadvantage, that all the automatic dependency detection stuff _after_ having built the package does not happen
[08:38:17] <abaumann> true
[08:38:36] <abaumann> so we have to manually schedule all firefox- things..
[08:38:48] <abaumann> preferably not firefox itself :-)
[08:38:49] <deep42thought> we do?
[08:38:54] <abaumann> I don't know. Do we?
[08:39:14] <abaumann> i18n firefox..
[08:39:16] <deep42thought> I was refering to "link dependencies"
[08:39:20] <abaumann> ah.
[08:39:22] <abaumann> sorry.
[08:39:51] <deep42thought> these packages were either already built or they were not yet built and will be built now/later
[08:39:55] <abaumann> ah is81n stuff is on 62.0 anyway already.
[08:40:00] <abaumann> *i18n
[08:40:06] <abaumann> yep.
[08:40:08] <abaumann> all fine.
[08:40:11] <deep42thought> :-)
[08:40:50] <deep42thought> btw: I changed the build-assignment-block logic
[08:40:56] <deep42thought> we do not wait for makedepends, now
[08:41:02] <abaumann> ah.
[08:41:11] <deep42thought> we only wait, that _any_ package is available providing the makedepends
[08:41:30] <abaumann> is binutils a makedepend of gcc?
[08:41:33] <deep42thought> e.g. if you have makedepends=(git), you only need _some_ git, not necessarily the latest
[08:41:48] <abaumann> git is a good working example.
[08:41:53] <deep42thought> binutils,glibc,gcc,linux-api-headers is a totally different logic
[08:42:00] <abaumann> they are depends?
[08:42:07] <deep42thought> they are "toolchain" packages
[08:42:14] <deep42thought> which will be built before everything else
[08:42:18] <deep42thought> and only one at a time
[08:42:28] <abaumann> this surely helps unblocking some builds.
[08:43:21] <deep42thought> and in the following order: linux-api-headers -> glibc -> gcc -> binutils -> gcc -> glibc
[08:44:03] <abaumann> mmh. now. I just start a build-packages on the i486 VM?
[08:44:12] <abaumann> and on master?
[08:44:16] <deep42thought> probably not
[08:44:18] <abaumann> or on my branch?
[08:44:27] <deep42thought> somehow the right pacman.conf needs to be involved
[08:44:50] <abaumann> ok. this I patched in devtools32 locally
[08:44:50] * buildmaster resumes sanity.
[08:44:52] <deep42thought> e.g. your staging-i686-build and staging-with-build-support-i686-build needs to use them
[08:44:56] <deep42thought> ok
[08:44:58] <deep42thought> then, yes
[08:45:00] <deep42thought> that's it
[08:45:24] <abaumann> ok.. no risk.. no fun :-)
[08:47:00] <abaumann> oh. packages should be on master..
[08:47:11] <deep42thought> no, doesn't matter at all
[08:47:32] <deep42thought> can even be a bare repository
[08:47:41] <abaumann> ah, I meant the packages git repo.
[08:47:44] <deep42thought> it just needs to see the correct commit
[08:47:47] <abaumann> I was still on the i486 branch
[08:47:52] <deep42thought> no problem
[08:47:55] <abaumann> ah.
[08:48:07] <deep42thought> it does a "git archive $commit"
[08:48:19] <abaumann> then yes. no problem.
[08:48:38] <deep42thought> only question is if the "git remote update" at the beginning fetches the correct branches
[08:48:42] <deep42thought> tehre I'm unsure
[08:55:10] <abaumann> tzdata-2018e-1-any vs. tzdata-2018e-2.2-i486.
[08:55:23] <abaumann> tzdata has a zic binary in it, so it's hardly any.
[08:55:36] <abaumann> and now the tzdata-2018e-1 takes precedence over tzdata-2018e-2.2-i486
[08:56:02] <deep42thought> strange - it's not "any" upstream O.o
[08:56:32] <deep42thought> it's not "any" in our repos either
[08:56:40] <abaumann> maybe it was
[08:56:45] <abaumann> in bootstrap/i486/core/tzdata-2018e-1-any
[08:56:53] <abaumann> so, I have to remove this file and redo the db..
[08:56:55] <deep42thought> that's yours :-D
[08:57:13] <deep42thought> no
[08:57:19] <deep42thought> just reschedule it for i486
[08:57:24] <deep42thought> and rebuild it
[08:57:37] <abaumann> the problem is, it misses the zic binary and the fails building
[08:57:37] <deep42thought> then the one in our repos should take precedence
[08:57:52] <deep42thought> tzdata needs tzdata to build?
[08:57:55] <abaumann> yes.
[08:57:57] <deep42thought> :-(
[08:58:00] <abaumann> don't ask..
[08:58:07] <deep42thought> I won't
[08:58:26] <abaumann> I really not questioning this anymore.. I know it's wrong..
[08:58:42] -!- oaken-source has joined #archlinux32
[09:09:38] <abaumann> reporting upstream. eli will like it :-)
[09:10:27] <deep42thought> if it's a valid bug report, he will like it - if not, he wile just "like" it
[09:10:38] <abaumann> :-)
[09:12:52] <abaumann> tyzoid-srv0-vm486 abaumann i486/bzip2
[09:12:54] <abaumann> sweet
[09:13:01] <deep42thought> :-)
[09:13:21] <abaumann> let's be happy as long as it runs and doesn't crash..
[09:17:28] <abaumann> mmh. build-support takes the i686 staging script..
[09:17:35] <deep42thought> yes
[09:17:38] <deep42thought> that's what I meant
[09:17:49] <deep42thought> maybe, I should just fix that ...
[09:18:09] <abaumann> I fear, it's more..
[09:19:45] <abaumann> build_command='staging-i686-build'
[09:19:51] <abaumann> aha. a little bit hard-coded. :-)
[09:20:33] <abaumann> no problem. I'll hack it in the builder/i486 branch, and when it works, we can merge properly.
[09:20:43] <deep42thought> I thought, you patched the *-i686-build versions
[09:20:47] <deep42thought> but it's fixed, now
[09:21:06] <abaumann> ah. I was never running via build-packages, only via a test-build-package
[09:23:45] <buildmaster> bzip2 is broken (says tyzoid-srv0-vm486).
[09:24:01] <deep42thought> buildmaster: you should tell us the architecture to that, too!
[09:24:20] <abaumann> buildmaster can program itself now? ;-)
[09:24:58] <deep42thought> no, but I'm used to being ignored on educational advices (by my children)
[09:25:07] <abaumann> lol
[09:33:21] <abaumann> mmh. bzip2/i486 got built, but the database is not updated..
[09:33:21] <abaumann> -rw-r--r-- 1 http http 29 Sep 12 11:35 i486/staging/staging.db.tar.gz
[09:36:03] <deep42thought> I'll take a look
[09:48:28] <deep42thought> abaumann: I see only error reports for the bzip2 build from your slave
[09:48:31] <deep42thought> https://packages.archlinux32.org
[09:48:33] <phrik> Title:ssh-log (at packages.archlinux32.org)
[09:48:51] <abaumann> mmh.
[09:49:07] <deep42thought> maybe some post-build action, that failed?
[09:49:26] <abaumann> I'll check.
[09:49:40] <abaumann> I think. I saw a bzip2 package on my mirror, but it's one in bootstrap.
[09:49:53] <abaumann> Funny thing. I see no error in the shell..
[09:50:12] <deep42thought> https://buildmaster.archlinux32.org
[09:50:33] <deep42thought> did it print "... ok" or "... failed"?
[09:50:36] <deep42thought> probably the latter
[09:51:09] <abaumann> I just rerun it.
[10:23:06] <deep42thought> somehow, I broke the loop detection logic
[10:23:21] <deep42thought> the buildmaster didn't see, that i486/bzip2 is in a loop
[10:28:02] <abaumann> oh.
[10:28:23] <deep42thought> because it is not, if you're allowed to use the i686 packages as makedepends
[10:28:36] <deep42thought> ... which makes no sense (except for the database)
[10:41:55] <abaumann> ./build-packages: line 484: sponge: command not found
[10:42:08] <abaumann> cool. I have to build one for i486 :-)
[10:45:05] <deep42thought> oops
[10:45:54] <abaumann> it might even build :-)
[10:46:16] <abaumann> oh. it needs half of perl.. sweet..
[10:46:33] <deep42thought> damn
[10:46:45] <deep42thought> I make heavy use of sponge in the build tools :-/
[10:47:09] <abaumann> no problem. I'll build in into bootstrapping manually.
[10:50:42] -!- ofara has quit [Ping timeout: 252 seconds]
[10:55:14] -!- rdap has joined #archlinux32
[10:55:38] -!- rdap has quit [Remote host closed the connection]
[11:29:47] <abaumann> I knew it, I shouldn't have reported: https://bugs.archlinux.org
[11:29:49] <phrik> Title:FS#60049 : [tzdata] install depends on a pre-installed zic binary (at bugs.archlinux.org)
[11:30:18] <abaumann> IMHO an unfair comparision to the toolchain.. but he.. that's me.
[11:45:22] -!- woshty has joined #archlinux32
[11:47:14] <abaumann> I really love packages who insist to build their documenation, like moreutils :->
[12:15:44] -!- ofara has joined #archlinux32
[12:31:58] -!- abaumann has quit [Quit: leaving]
[14:49:53] -!- ofara_ has joined #archlinux32
[15:00:59] <deep42thought> looks, like we have a working thunderbird and firefox again in our repositories
[15:01:03] <deep42thought> thanks, abaumann!
[15:02:48] -!- deep42thought has quit [Remote host closed the connection]
[15:03:15] -!- deep42thought has joined #archlinux32
[15:03:15] <buildmaster> Hi deep42thought!
[15:05:37] -!- oaken-source has quit [Ping timeout: 252 seconds]
[15:36:30] -!- oaken-source has joined #archlinux32
[15:40:42] -!- ofara_ has quit [Ping timeout: 264 seconds]
[15:43:20] -!- oaken-source has quit [Ping timeout: 252 seconds]
[15:47:41] -!- ofara_ has joined #archlinux32
[16:36:39] -!- deep42thought has quit [Remote host closed the connection]
[16:36:55] -!- isacdaavid has joined #archlinux32
[17:28:23] -!- isacdaavid has quit [Quit: Leaving.]
[17:31:11] -!- MrBIOS has joined #archlinux32
[18:03:57] -!- MrBIOS has quit [Write error: Connection reset by peer]
[18:04:58] -!- MrBIOS has joined #archlinux32
[20:54:58] -!- {levi} has joined #archlinux32
[20:58:08] -!- {levi} has quit [Remote host closed the connection]
[20:59:30] -!- {levi} has joined #archlinux32
[21:39:11] -!- oaken-source has joined #archlinux32
[21:54:52] -!- prstoetzer has quit [Quit: prstoetzer]
[21:55:30] -!- ofara_ has quit [Ping timeout: 252 seconds]
[21:55:59] -!- prstoetzer has joined #archlinux32
[21:57:33] -!- ofara_ has joined #archlinux32
[22:13:04] -!- oaken-source has quit [Ping timeout: 272 seconds]
[22:24:54] -!- MrBIOS has quit [Read error: Connection reset by peer]
[22:46:11] -!- belsk_tay has joined #archlinux32
[22:46:37] <belsk_tay> I wanted to file a bug but can't seem to do it through the website.
[22:47:33] <belsk_tay> Last full upgrade (-Syu) changes icu to 0.62. Firefox is still dependent on icu 0.61 so fails to execute with missing library
[22:48:50] <belsk_tay> firefox libxul is linked to libicui18n.so.61 and libicuuc.so.61
[22:49:41] <belsk_tay> [nada@roam ~]$ firefox XPCOMGlueLoad error for file /usr/lib/firefox/libxul.so: libicui18n.so.61: cannot open shared object file: No such file or directory Couldn't load XPCOM. [nada@roam ~]$
[22:50:30] <belsk_tay> I would attempt to fix this myself, but traveling on an underpowered netbook makes it challenging.
[22:51:36] <belsk_tay> Thanks for maintining a great distro that enables these old devices to still function.
[22:56:29] <{levi}> You probably need to update firefox. Use Andreas's AUR package from https://aur.archlinux.org
[22:56:49] <{levi}> Or just use Chrome
[23:00:19] <belsk_tay> unfortuantly, firefox has all my saved passwords..
[23:00:36] <belsk_tay> thanks for the AUR, I'll give it a shot.
[23:14:15] <belsk_tay> Thanks, that worked. Functioning firefox again
[23:22:44] -!- belsk_tay has quit [Quit: Page closed]
[23:32:52] -!- ofara_ has quit [Ping timeout: 272 seconds]