#archlinux32 | Logs for 2018-11-23

[00:20:00] -!- guys has quit [Ping timeout: 250 seconds]
[00:33:51] -!- guys has joined #archlinux32
[00:47:10] <buildmaster> i686/sn0int is broken (says eurobuild3).
[00:48:43] -!- deep42thought has quit [Quit: Leaving.]
[01:05:03] -!- dbermond has joined #archlinux32
[01:36:17] <buildmaster> i686/haskell-dav is broken (says eurobuild3) - I rescheduled: haskell-xml-hamlet.
[02:20:54] -!- guys has quit [Ping timeout: 250 seconds]
[02:22:33] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[02:22:39] -!- thePiGre1per has joined #archlinux32
[02:34:38] -!- guys has joined #archlinux32
[02:49:25] -!- woshty has quit [Ping timeout: 246 seconds]
[04:21:17] -!- guys has quit [Ping timeout: 244 seconds]
[04:30:36] -!- NoobAlice has quit [Quit: Leaving.]
[04:35:16] -!- guys has joined #archlinux32
[05:10:47] -!- dbermond has quit [Quit: gn]
[05:50:53] -!- thePiGre1per has quit [Ping timeout: 245 seconds]
[06:22:16] -!- guys has quit [Ping timeout: 250 seconds]
[06:36:06] -!- guys has joined #archlinux32
[06:57:54] -!- davor_ has joined #archlinux32
[06:58:42] -!- davor has quit [Ping timeout: 252 seconds]
[06:58:42] davor_ is now known as davor
[07:51:23] -!- abaumann has joined #archlinux32
[07:51:24] <buildmaster> Hi abaumann!
[07:51:24] <buildmaster> !rq abaumann
[07:51:24] <phrik> buildmaster: <abaumann> Working in IT makes working in an asylum for mental people look like holidays.. ;-)
[07:51:57] <abaumann> deep42thought: sorry for taking out the db backup, there is no reason, I was just testing the borg setup and didn't want to wait for the mysql backup to finnish.. it's enabled now. :-)
[07:59:18] -!- deep42thought has joined #archlinux32
[07:59:18] <buildmaster> Hi deep42thought!
[07:59:18] <buildmaster> !rq deep42thought
[07:59:19] <phrik> buildmaster: <deep42thought> arch32 is on the bleeding edge of arch, which is on the bleeding edge of software?
[07:59:31] <deep42thought> Good morning, abaumann
[08:04:15] <abaumann> morning deep42thought
[08:04:40] <deep42thought> regarding the mysql backup: I suspected something like that, but wasn't sure :-)
[08:06:53] <deep42thought> btw: I disabled almost all checks of the web interface regarding tyzoid's package api - maybe we could/should replace it by a local one?
[08:08:06] <abaumann> you mean, on the buildmaster itself?
[08:08:15] <deep42thought> on packages.archlinux32.org
[08:08:26] <deep42thought> e.g. currently my server
[08:09:05] <abaumann> ah. yes. why not.
[08:09:05] <deep42thought> https://git.archlinux32.org
[08:09:07] <phrik> Title: archlinux32/archweb32: Archlinux32 package and build system information interface - Archlinux32 Gitea (at git.archlinux32.org)
[08:09:58] <abaumann> any news from tyzoid?
[08:10:05] <deep42thought> nope
[08:10:56] <deep42thought> maybe my email landed in his spam folder?
[08:23:10] -!- guys has quit [Ping timeout: 250 seconds]
[09:09:58] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-11-23T08:07:42.WXqXQg.stdin"?.
[09:29:47] * buildmaster resumes sanity.
[09:31:25] <abaumann> -rw-r--r-- 1 root root 22 Nov 23 08:09 ../work/build-master-sanity
[09:31:29] <abaumann> is this good?
[09:31:32] <deep42thought> oops
[09:31:36] <deep42thought> I run something as root
[09:31:47] <abaumann> :-)
[09:32:07] <deep42thought> I'm currently trying to get pkgfile running on the buidlmaster
[09:32:12] <deep42thought> so we can use "wtf" again
[09:32:52] <deep42thought> !bug 60893
[09:32:52] <phrik> https://bugs.archlinux.org
[09:33:38] <abaumann> well, all packages have an architecture in them.
[09:33:45] <abaumann> so they should technically not collide.
[09:33:53] <deep42thought> "any" packages?
[09:34:08] <abaumann> mmh. any packages should be the same for all architectures..
[09:34:12] <abaumann> but yeah. I know.
[09:34:17] <deep42thought> but not necessarily in the same repository
[09:34:25] <abaumann> true
[09:34:26] <deep42thought> also: that cache does not store the architecture
[09:34:45] <deep42thought> plus: I'd really like to move the cache away from the "system default location"
[09:35:02] <abaumann> you convinced me :-)
[09:35:06] <deep42thought> :-D
[09:35:14] <abaumann> ..but I'm not the one to _BE_ convinced. :-)
[09:35:22] <deep42thought> I know
[09:35:28] <deep42thought> but it still improved my confidence
[09:36:50] <abaumann> There is a --cachedir argument for pacman, which sets /var/cache/pacman, so why is there none for /var/cache/pkgfile?
[09:37:02] <deep42thought> exactly, what I thought^
[09:37:13] <abaumann> there we have our strongest argument. :->
[09:43:57] <buildmaster> i686/aspell is broken (says nlopc46).
[09:48:24] <deep42thought> buildmaster: wtf libc.so.6
[09:48:25] <buildmaster> deep42thought: testing/glibc /usr/lib/libc.so.6
[09:48:25] <buildmaster> core/glibc /usr/lib/libc.so.6
[09:48:25] <buildmaster> community/aarch64-linux-gnu-glibc /usr/aarch64-linux-gnu/lib/libc.so.6
[10:07:21] <deep42thought> hmm, now we have all the symbols in the database and it looks, like quite a lot packages are linked against unavailable stuff :-/
[10:09:54] -!- dopsi has quit [Quit: ZNC - https://znc.in]
[10:10:58] -!- dopsi has joined #archlinux32
[10:12:01] <deep42thought> PHP message: PHP Fatal error: Allowed memory size of 134217728 bytes exhausted
[10:12:09] <deep42thought> it's even too much for the web frontend :-/
[10:12:27] <abaumann> that's a tat a lot :-)
[10:12:39] <abaumann> maybe everything is sucked into memory.
[10:12:49] <deep42thought> it sure is :-/
[10:12:57] <deep42thought> because, umm, I suck at this
[10:13:01] <abaumann> lol :-)
[10:13:28] <deep42thought> no, the actual reason is, that I wanted to be able to raise 5xx errors in case the query goes wrong
[10:13:32] <abaumann> well. modern webpages need 900 MB to render a text page.
[10:13:41] <abaumann> so, what's 128 MB for a php script.
[10:13:45] <deep42thought> but I can only do this if the http server hasn't started sending content
[10:13:57] <abaumann> ob_start, ob_end?
[10:14:08] * deep42thought googles
[10:14:16] <abaumann> php output buffering
[10:15:28] <deep42thought> why does this decrease load on the memory?
[10:15:39] <abaumann> it buffers to disk.
[10:15:44] <deep42thought> ah, ok
[10:15:55] <abaumann> or tmpfs, which is in memory again. mmh..
[10:16:19] <abaumann> ..but at least, it's somebody's else memory. :-)
[10:16:24] <deep42thought> lol
[10:18:10] <deep42thought> hmm, this does not work if I leave the <?php ?> context again ...
[10:18:19] <deep42thought> ah, spoke too soon
[10:18:20] <deep42thought> nvm
[10:20:47] <deep42thought> well, now "Unknown on line 0" crashes with this oom error
[10:20:49] <deep42thought> :-(
[10:21:59] <abaumann> mixing HTML output and <?php sections is usually not the best idea.
[10:22:49] <deep42thought> I put an "unset($row)" at the end of my loop to improve garbage collection and it seems to have solved it :-D
[10:22:57] <abaumann> oh.
[10:22:59] <abaumann> :-)
[10:23:12] <abaumann> nice catch.
[10:23:23] <deep42thought> actually somewhat strange
[10:23:37] <deep42thought> because $row gets reassigned at the top of that very loop
[10:24:05] <abaumann> scripting languages have strange ideas for garbage collection.
[10:24:21] <abaumann> actually, for a web page the best garbage colleciton is _NOT_ to collect any garbage.
[10:24:45] <deep42thought> yeah, that's probably, what is done: garbage is not collected at all
[10:24:45] <abaumann> why should you do a lot of cleanup while running, when you throw away the whole thing anyway after producing the HTML.
[10:24:53] <deep42thought> yeah
[10:25:31] <deep42thought> ok, the script produces 250k lines of output :-D - it's ok, that it exhausted the memory, then
[10:36:48] <deep42thought> last resort: "LIMIT 10000"
[10:36:58] <deep42thought> (it still failed on my production server)
[11:02:45] -!- abaumann has quit [Quit: leaving]
[11:04:41] -!- woshty has joined #archlinux32
[11:12:09] <deep42thought> abaumann: your firefox patch will not work as you intend
[12:09:34] <buildmaster> i686/firefox is broken (says rechenknecht).
[12:36:04] -!- dbermond has joined #archlinux32
[13:04:35] -!- thePiGrepper has joined #archlinux32
[13:36:10] <deep42thought> we have now an i686 and an i486 package file query - big thanks to falconindy for implementing my featurerequest so fast :-)
[13:36:23] <deep42thought> buildmaster: wtf i686 fluxbox
[13:36:24] <buildmaster> deep42thought: Huh, I don't know that one.
[13:36:28] <deep42thought> *grrr*
[13:36:43] <deep42thought> buildmaster: wtf fluxbox
[13:36:43] <buildmaster> deep42thought: fluxbox [extra], fluxbox [testing]: /usr/bin/fluxbox
[13:36:51] <buildmaster> i686/deepin-file-manager is broken (says nlopc46).
[13:37:38] <deep42thought> buildmaster: wtf i686 fluxbox
[13:37:38] <buildmaster> deep42thought: fluxbox [extra], fluxbox [testing]: /usr/bin/fluxbox
[13:37:41] <deep42thought> buildmaster: wtf i486 fluxbox
[13:37:42] <buildmaster> deep42thought: fluxbox [extra]: /usr/bin/fluxbox
[13:37:48] <deep42thought> there you go
[14:25:44] -!- guys has joined #archlinux32
[14:34:52] -!- NoobAlice has joined #archlinux32
[14:53:49] -!- guys has quit [Ping timeout: 268 seconds]
[14:58:26] <thePiGrepper> deep42thought: hi, what's that package file query feature?
[14:58:46] <deep42thought> you can ask the buildmaster in what package a file can be found
[14:59:06] <deep42thought> this implements "pkgfile" from falconindy under the hood
[15:07:32] <thePiGrepper> oh, nice
[15:07:37] -!- guys has joined #archlinux32
[15:08:54] <thePiGrepper> btw. I think Im almost done with asp32, at least with the main thing which was just adding the changes from the arch32 repo. I havent had much time this month, because of work and stuff.
[15:09:33] <deep42thought> great - if you want me something to test, just let me know
[15:10:12] <thePiGrepper> sure. let me check a couple of things and Ill push it maybe during the weekend.
[15:11:16] <thePiGrepper> btw,do you happen to know if there's a way to use pactree(or something similar) but on the repo, instead of locally?
[15:12:26] <deep42thought> the webfrontend shows dependencies
[15:12:36] <deep42thought> I'm not sure, how pactree shows that stuff
[15:13:48] <deep42thought> why do you need a non-local version of pactree?
[15:13:56] <deep42thought> -b, --dbpath
[15:13:56] <deep42thought> Specify an alternative database location.
[15:14:06] <deep42thought> so just download the databases and voila
[15:15:06] <thePiGrepper> ohh. you're right.
[15:15:55] <thePiGrepper> why? because I wanted to see the overall dependency on some packages
[15:16:25] <thePiGrepper> this came to me a couple months ago when I was looking around for json parsers
[15:16:34] <thePiGrepper> I saw I had several installed
[15:16:42] <deep42thought> there are also some link-graphs on packages.archlinux32.org/buildmaster/ but they're far from complete
[15:16:48] <thePiGrepper> but many of them were used by one or two packages top
[15:17:35] <thePiGrepper> so I thought to myself 'well, using a dependency for just one package on the whole arch repository sure seems too much'
[15:17:49] <thePiGrepper> ofc, that's subjective
[15:17:57] <thePiGrepper> but anyway. just curious
[15:18:14] <deep42thought> yeah, sounds reasonable
[15:18:43] <thePiGrepper> for instance, asp uses one called jq. as far as I know no other package uses it.
[15:19:08] <deep42thought> https://www.archlinux.org has all the info, too
[15:19:09] <phrik> Title: Arch Linux - jq 1.6-1 (x86_64) (at www.archlinux.org)
[15:19:28] <thePiGrepper> it would be interesting to talk to upstream and/or maybe send a PR to make use of a most used one, if there arent specific reason to keep the current one
[15:19:37] <deep42thought> and similarily https://packages.archlinux32.org
[15:19:38] <phrik> Title: Arch Linux 32 - jq 1.6-1.0 (i686) (at packages.archlinux32.org)
[15:19:54] <deep42thought> yes, you're right
[15:22:02] <thePiGrepper> even better, it would be useful to check which is the overlapping of features between these relatively small libraries and if, after running benchmarks, the results prove that it makes sense to migrate, then that's one less package to maintain. everyone wins(sortof)
[15:24:23] <thePiGrepper> and, back to the usefulness of pactree-ing the main database. wouldnt it be a good idea to do this to determine a 'good time' to pass packages from testing to stable? or at the very least to determine build order?
[15:24:53] <deep42thought> the mechainsm is automated currently
[15:25:05] <deep42thought> "only" problem is, that there are bugs :-/
[15:25:17] <deep42thought> we have all the information in some giant mysql database
[15:25:30] <deep42thought> and the buildmaster determins what can be moved and what can be built
[15:25:33] <thePiGrepper> this situation when 'several packages are built but on testing, then one library which is a dependency of the first ones gets built, then both go stable, and thing break, happens sometimes, right?
[15:25:54] <deep42thought> this happens, because I f*ck up stuff
[15:26:09] <deep42thought> e.g. tell the buildmaster to move some stuff against better knowledge
[15:26:21] <thePiGrepper> hmm, but how do you do that, if the buildmaster is automating stuff? what are the criteria it uses?
[15:26:31] <deep42thought> not to break dependencies
[15:26:59] <thePiGrepper> well, it doesnt 'break dependencies' actually, right? it's way more subtle than that
[15:27:04] <deep42thought> a package is moved if it's installable (e.g. all dependencies are available from/in the target repository) and if does not replace a required dependency
[15:27:23] <thePiGrepper> it actually just has more modern dependencies than the ones the actual package was build on, right?
[15:27:25] <deep42thought> my goal was to have all things that matter in the database
[15:27:30] <deep42thought> and then it _breaks_ dependency
[15:27:45] <thePiGrepper> sure, that's why you proposed the linking information to be saved as well, right?
[15:27:54] <deep42thought> if you include the version to the dependency, it's an actual break
[15:27:59] <deep42thought> or linked libraries
[15:28:11] <deep42thought> or even linked symbols (as I'm trying out right now)
[15:28:18] <deep42thought> yes, exactly
[15:30:15] <thePiGrepper> deep42thought: Im still kinda surprised arch doesnt break as often, they have basically the same PKGBUILDs with the same lacking version dependencies
[15:30:32] <thePiGrepper> for most of their packages at least
[15:30:45] <deep42thought> yes, but they have more brains thinking about what may be moved and what not
[15:31:15] <thePiGrepper> well, that's true. the man power comparison is overwhelming...
[15:31:39] <thePiGrepper> it doesnt help that many upstream projects are deprecating i686 at the speed of light.
[15:31:41] <deep42thought> so we implemented some automatism which is ... well, too strict, I'm afraid
[15:31:52] <deep42thought> yes, that, too
[15:32:55] <thePiGrepper> that libreoffice will hit us in a couple years at least, and the firefox thingy is really weird. Ive been talking in the void IRC, and they say they have no such issues, so I want to build their firefox package in a VM to see what happens
[15:34:31] <deep42thought> can you try to make a PKGBUILD from the exact same commands, they're using?
[15:34:41] <deep42thought> maybe it's some subtle linker switch or whatever
[15:37:14] <thePiGrepper> yeah, I was thinking the same thing. first I want to actually reproduce it and see that it works, and then I was planning on doing just that.
[15:37:58] <thePiGrepper> Im having some issues running a void i686 container on my arch host, because of some systemd thingy apparently, but once that's solved, Ill move on
[15:39:09] <thePiGrepper> s/move on/continue/
[15:39:14] <deep42thought> :-D
[16:26:51] -!- deep42thought has quit [Quit: Leaving.]
[16:51:18] -!- abaumann has joined #archlinux32
[16:51:19] <buildmaster> Hi abaumann!
[16:51:19] <buildmaster> !rq abaumann
[16:51:19] <phrik> buildmaster: <abaumann> all slaves busy.. wait.. one slave in the north/western part of.. resists.. :-)
[16:51:47] <abaumann> deep42thought: thanks for fixing my broken sed-fu. This is actually an attempt to use some voidlinux flags to build firefox. :-)
[16:51:54] <abaumann> -Wl,--no-keep-memory -Wl,--reduce-memory-overheads
[16:51:59] <abaumann> sounds like a good idea..
[16:52:13] <abaumann> otherwise I didn't see, why the build should fail..
[16:52:32] <abaumann> ..and voidlinux is not doing something dramatically different, as far as I can see.
[16:52:57] <abaumann> thePiGrepper: looking forward to a build of yours. more eyes see more. :-)
[16:53:23] -!- guys has quit [Ping timeout: 245 seconds]
[17:08:18] -!- guys has joined #archlinux32
[17:52:32] -!- deep42thought has joined #archlinux32
[17:52:32] <buildmaster> Hi deep42thought!
[17:52:32] <buildmaster> !rq deep42thought
[17:52:33] <phrik> buildmaster: <deep42thought> ah, maybe, borg is borged, too
[17:52:57] <deep42thought> abaumann: for one, firefox is broken, because my archiver did some funny stuff (not sure, what exactly)
[17:54:30] <deep42thought> abaumann: lol, I just noticed, that we worked redundantly and created symlinks for /srv/http _and_ /srv/http/build-logs :-D
[17:54:47] <deep42thought> pointing to /data/$something each
[18:00:43] <abaumann> hi deep42thought
[18:00:51] <deep42thought> Hi abaumann
[18:00:52] <abaumann> cool. double symlinks, should hold longer. ;-)
[18:01:19] <deep42thought> I removed one layer now - it caused me to need to hit tab once more than I expected
[18:01:26] <deep42thought> (that's how I noticed it)
[18:13:27] -!- woshty has quit [Ping timeout: 240 seconds]
[18:14:47] -!- woshty has joined #archlinux32
[18:29:46] -!- DepositePirate has quit [Remote host closed the connection]
[18:31:01] -!- DepositePirate has joined #archlinux32
[18:37:38] -!- lodger has joined #archlinux32
[18:38:57] <lodger> hi, firefox pkg needs an update to work with icu 63
[18:41:15] <deep42thought> lodger: news.archlinux32.org
[18:45:47] <lodger> calibre/chromium/libreoffice work good with icu 63
[18:50:27] -!- lodger has quit [Remote host closed the connection]
[18:53:56] -!- lodger has joined #archlinux32
[18:54:13] -!- guys has quit [Ping timeout: 245 seconds]
[18:55:17] <lodger> chromium-widevine doesn't work on 32bit? Netflix needs it to play video, i have installed it from AUR but netflix site still says error.
[18:58:52] <deep42thought> I don't know that package, but have you tried rebuilding it after upgrading chromium?
[19:01:54] <lodger> yes. on PKGFILE of "chromium-widevine" there is only x86_64 on "arch"
[19:02:04] <deep42thought> just add i686, then
[19:02:28] <deep42thought> and if it works, drop the maintainer a note to add it
[19:03:53] -!- lodger has quit [Remote host closed the connection]
[19:06:33] -!- mchasard has joined #archlinux32
[19:08:52] -!- mchasard has quit [Client Quit]
[19:09:08] -!- guys has joined #archlinux32
[19:22:07] -!- abaumann has quit [Quit: leaving]
[20:32:53] <thePiGrepper> deep42thought: one question, regarding building docker for i686, was it too difficult or had any trick to it?
[20:33:17] <deep42thought> thePiGrepper: I was never involved with docker
[20:33:26] <deep42thought> I'm not even sure how to use it :-/
[20:36:44] <thePiGrepper> hehe, ok. np. apparently there isnt any merge file in the arch32 repo, so it should build without issues. I wonder why no other distros has it(besides no official support for i686)
[20:54:50] -!- guys has quit [Ping timeout: 244 seconds]
[21:08:50] -!- guys has joined #archlinux32
[22:17:51] -!- MrBIOS has joined #archlinux32
[22:21:57] <thePiGrepper> n
[22:56:03] -!- guys has quit [Ping timeout: 268 seconds]
[23:02:47] -!- woshty has quit [Ping timeout: 240 seconds]
[23:07:09] -!- NoobAlice has quit [Ping timeout: 268 seconds]
[23:07:25] -!- NoobAlice1 has joined #archlinux32
[23:24:59] <thePiGrepper> deep42thought: hi, Im having an issue with the arch32 iso. I installed a new VM using the iso from july, after installing everything I now realized that it doesnt have the latest packages, it only has packages from around 1 month ago or so.
[23:25:34] <thePiGrepper> 'it' being the installed OS, not the ISO, obviously
[23:26:01] <deep42thought> what mirror do you use?
[23:26:56] <thePiGrepper> the first one in the file is one from canada: arch32.mirrors.simplysam.us/
[23:27:11] <deep42thought> that one is no more
[23:27:40] <thePiGrepper> which one should I leave?
[23:27:47] <deep42thought> mirror.archlinux32.org
[23:27:57] <deep42thought> that one is the master
[23:28:23] <thePiGrepper> and I always thought that pacman look around for the most recent packages in any mirror. so, if the first one is the obsolete one with old packages, it doesnt bother to take a look at the others?
[23:29:53] <thePiGrepper> now it works.. lol. well I learned something new, so It's worth it XD
[23:31:02] <deep42thought> it looks around, but only if the mirrors fail fatally
[23:31:16] <deep42thought> it cannot see if the packages are recent or not
[23:31:41] <deep42thought> but it might be worth a feature request - in the end, the mirrors have the lastsync and lastupdate files available
[23:36:58] <thePiGrepper> I just took a look at them, they are plain files with Unix time. so what would the idea be? check which mirror is has the latest sync/update file? isnt that a reflector feature?
[23:38:06] <deep42thought> yeah
[23:38:10] <deep42thought> I think so
[23:38:28] <deep42thought> but skipping too-old mirrors in-situ would also be nice
[23:38:46] <thePiGrepper> so if I used reflector I woudlnt had had this issue in the first place.hmm
[23:38:56] <deep42thought> probably, yes
[23:40:33] <thePiGrepper> interesting. however, at the end, is there any correlation between last time sync/update and actual current packages? it's not like every mirror has every package, or all the ones that come in the file do?
[23:40:55] <thePiGrepper> well, it makes sense, they are _mirrors_ after all. hehe
[23:45:29] <deep42thought> curl -Ss 'https://mirror.archlinux32.org/i686/core/core.db.tar.gz' | tar -tzv --full-time 2>/dev/null
[23:45:37] <deep42thought> might give a more accurate picture