#archlinux32 | Logs for 2019-01-21
Back
[00:02:55] -!- alexandre9099 has quit [Read error: Connection reset by peer]
[00:03:07] -!- alexandre9099 has joined #archlinux32
[00:07:59] <buildmaster> i686/hedgewars are broken (says rechenknecht).
[00:20:02] -!- AndrevS has quit [Remote host closed the connection]
[00:26:21] -!- yans has quit [Ping timeout: 244 seconds]
[00:32:43] -!- isacdaavid has quit [Quit: Leaving.]
[01:03:52] -!- oaken-source has quit [Ping timeout: 272 seconds]
[01:34:38] -!- woshty has quit [Ping timeout: 268 seconds]
[01:58:18] -!- autofsckk has quit []
[03:29:40] -!- davor_ has joined #archlinux32
[03:30:29] -!- davor has quit [Ping timeout: 268 seconds]
[03:30:30] davor_ is now known as davor
[03:32:38] <buildmaster> i686/gitlab is broken (says rechenknecht).
[03:40:17] -!- MrBIOS has joined #archlinux32
[06:02:53] -!- T`aZ has quit [Ping timeout: 246 seconds]
[06:04:50] -!- T`aZ has joined #archlinux32
[06:21:05] -!- thePiGrepper has quit [Ping timeout: 246 seconds]
[07:23:16] -!- MrBIOS has quit [Quit: MrBIOS]
[07:24:54] -!- thePiGrepper has joined #archlinux32
[07:34:04] -!- MrBIOS has joined #archlinux32
[07:34:26] -!- MrBIOS has quit [Client Quit]
[07:34:52] -!- MrBIOS has joined #archlinux32
[07:35:12] -!- MrBIOS has quit [Client Quit]
[07:35:41] -!- MrBIOS has joined #archlinux32
[07:35:58] -!- MrBIOS has quit [Client Quit]
[07:36:29] -!- MrBIOS has joined #archlinux32
[07:36:44] -!- MrBIOS has quit [Client Quit]
[07:37:03] -!- MrBIOS has joined #archlinux32
[07:37:30] -!- MrBIOS has quit [Client Quit]
[08:02:12] <buildmaster> i686/stratisd is broken (says rechenknecht).
[08:06:00] -!- yans has joined #archlinux32
[08:14:32] -!- thePiGrepper has quit [Ping timeout: 272 seconds]
[08:22:16] -!- woshty has joined #archlinux32
[08:50:56] -!- deep42thought has joined #archlinux32
[08:50:56] <buildmaster> Hi deep42thought!
[08:50:56] <buildmaster> !rq deep42thought
[08:50:57] <phrik> buildmaster: <deep42thought> ah, maybe, borg is borged, too
[08:51:32] -!- thePiGrepper has joined #archlinux32
[08:55:13] -!- oaken-source has joined #archlinux32
[09:20:04] -!- thePiGrepper has quit [Ping timeout: 244 seconds]
[09:58:44] -!- thePiGrepper has joined #archlinux32
[10:17:20] -!- thePiGrepper has quit [Ping timeout: 250 seconds]
[10:49:04] -!- oaken-source has quit [Ping timeout: 272 seconds]
[10:51:15] <buildmaster> i686/python-pip-bootstrap is broken (says buildknecht).
[11:11:38] <buildmaster> i686/libreoffice-fresh is broken (says buildknecht).
[12:06:18] -!- thePiGrepper has joined #archlinux32
[12:17:13] -!- oaken-source has joined #archlinux32
[12:25:39] -!- thePiGrepper has quit [Quit: leaving]
[12:45:41] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2019-01-21T11:43:30.718xuV.stdin"?.
[13:10:05] -!- lodger has joined #archlinux32
[13:36:57] * buildmaster resumes sanity.
[13:41:28] -!- T0MuX has joined #archlinux32
[13:41:37] <T0MuX> hi
[13:41:41] <deep42thought> Hi
[13:41:51] <T0MuX> I just encoutered an issue with riot-desktop on archlinux32
[13:42:10] <T0MuX> I updated the entire system last night
[13:42:57] <T0MuX> and two things : first, I had to reinstall ffmpeg-compat57, and second, electron was asking for libhttp_parser.so.2.8
[13:43:18] <deep42thought> yeah, electron fails to build for quite some time
[13:43:18] <T0MuX> but, didnt exist anymore next to the update. I sniffed arround in /lib, and saw libhttp_parser.so.2.9
[13:43:22] <deep42thought> https://packages.archlinux32.org
[13:43:22] <phrik> Title: Arch Linux 32 - electron 1.8.4-1.0 (i686) (at packages.archlinux32.org)
[13:43:29] <T0MuX> so, I did sudo ln -s libhttp_parser.so.2.9 libhttp_parser.so.2.8
[13:43:55] <T0MuX> and now, riot-desktop runs. but I think this not a long term solution ^^
[13:44:09] <T0MuX> deep42thought: ah ok !
[13:44:49] <deep42thought> buildmaster: wtf libhttp_parser.so
[13:44:49] <buildmaster> deep42thought: http-parser [community]: /usr/lib/libhttp_parser.so
[13:45:01] <T0MuX> so it is for this reason there a little late for the electron version in archx32
[13:45:14] <deep42thought> yes, that's the reason
[13:45:36] <deep42thought> you can try to unpack the real library from the archive package: https://archive.archlinux32.org
[13:45:38] <phrik> Title: Index of /packages/h/http-parser (at archive.archlinux32.org)
[13:45:44] <deep42thought> this should be better than the symlink solution
[13:46:48] <T0MuX> but, I already update http-parser
[13:46:55] <deep42thought> yes
[13:46:58] <T0MuX> by `yay -Sy http-parser`
[13:47:00] <deep42thought> but you need an _older_ lib
[13:47:46] <deep42thought> e.g. download the latest file in the archive and then 'sudo bsdtar -C / -xf http-parser-2.8.1-1.1-i686.pkg.tar.xz /usr/lib/libhttp_parser.so.2.8 /usr/lib/libhttp_parser.so.2.8.1'
[13:47:56] <T0MuX> hah, you say I have to download an archive of http-parser to setup it ?
[13:48:01] <deep42thought> yes
[13:48:12] <deep42thought> well, the archived version of the package
[13:48:28] <deep42thought> but if the symlink works, you can also stick with that
[14:29:24] -!- lodger has quit [Remote host closed the connection]
[14:39:40] -!- phrik has quit [Read error: Connection reset by peer]
[14:39:45] -!- City-busz has quit [Quit: http://quassel-irc.org - Csevegjen kényelmesen. Bárhol.]
[14:39:46] -!- elibrokeit has quit [Quit: A random quit message]
[14:51:00] -!- City-busz has joined #archlinux32
[14:51:03] -!- elibrokeit has joined #archlinux32
[14:51:06] -!- phrik has joined #archlinux32
[14:51:07] <buildmaster> Hi phrik!
[14:51:07] <buildmaster> !rq phrik
[14:55:20] <phrik> buildmaster: <phrik> ayy
[14:55:57] -!- elibrokeit has quit [Ping timeout: 252 seconds]
[14:56:15] -!- elibrokeit has joined #archlinux32
[15:42:42] -!- MrBIOS has joined #archlinux32
[16:42:07] -!- deep42thought has quit [Quit: Leaving.]
[16:42:44] <buildmaster> i686/mysql-workbench is broken (says rechenknecht).
[17:43:09] <T0MuX> deep42thought: yes this is my wish ^^ I just keep aware of it, and will remove it when necessary
[18:10:51] -!- deep42thought has joined #archlinux32
[18:10:51] <buildmaster> Hi deep42thought!
[18:10:51] <buildmaster> !rq deep42thought
[18:10:52] <phrik> buildmaster: <deep42thought> my cluster of 1k 4/86's will be happy to iterate Maxwell-Vlasov-equations :-D
[18:10:54] -!- Fear has quit [Quit: ZNC - https://znc.in]
[18:29:34] -!- Fear has joined #archlinux32
[18:33:40] -!- Piraty has quit [Quit: No Ping reply in 180 seconds.]
[18:34:19] -!- Piraty has joined #archlinux32
[18:35:50] -!- thePiGrepper has joined #archlinux32
[19:50:34] -!- woshty has quit [Ping timeout: 272 seconds]
[19:50:45] -!- autofsckk has joined #archlinux32
[19:50:48] <autofsckk> hello guys
[19:50:53] <deep42thought> Hi autofsckk
[19:51:03] <autofsckk> i have problems with qbittorrent, getting this error
[19:51:04] <autofsckk> qbittorrent: error while loading shared libraries: libboost_system.so.1.69.0: cannot open shared object file: No such file or directory
[19:51:34] <autofsckk> any ideas? i think that libboost_system is wrong, almost sure the same erro i have with deluge which i have been using for a long time
[19:51:57] <deep42thought> there are some broken links to libboost_* known to the database
[19:52:01] <deep42thought> https://packages.archlinux32.org
[19:52:02] <phrik> Title: More and less critical issues with the database (at packages.archlinux32.org)
[19:52:12] <deep42thought> probably one of those needs to be moved from testing to stable
[19:52:22] <deep42thought> can you do a "lddtree" on the binary, that fails?
[19:52:35] <deep42thought> and tell me which library is actually linking against the missing lib?
[19:54:14] <autofsckk> sure, give me a second
[19:54:29] <autofsckk> i dunno how to use lddtree
[19:54:33] <autofsckk> let me read about it
[19:54:44] <deep42thought> lddtree $failingbinary
[19:54:56] <autofsckk> lddtree: warning: qbittorrent: did not match any paths
[19:54:56] <deep42thought> e.g. lddtree /usr/bin/firefox
[19:55:07] <deep42thought> you need to specify the full path
[19:55:46] <autofsckk> where do you want me to paste it?
[19:56:04] <deep42thought> !give autofsckk ptpb
[19:56:07] <phrik> autofsckk: "<command> |& curl -F c=@- https://ptpb.pw" OR "curl -F c=@path/to/a/file https://ptpb.pw"
[19:56:30] <deep42thought> I think, it's libtorrent-rasterbar
[19:57:54] <autofsckk> curl: (35) OpenSSL SSL_connect: SSL_ERROR_SYSCALL in connection to ptpb.pw:443
[19:58:09] <autofsckk> i got that when trying to use ptpb jajajaja
[19:58:09] <deep42thought> oh
[19:58:26] <deep42thought> well, anyways, I see, that libtorrent-rasterbar is wrongly linked
[19:58:32] <deep42thought> I'll schedule it for a rebuild
[19:58:44] <autofsckk> https://bpaste.net
[19:58:47] <phrik> Title: show at bpaste (at bpaste.net)
[19:59:04] <deep42thought> yup
[19:59:07] <deep42thought> thank you :-)
[19:59:23] <deep42thought> I'll move it to stable asap
[19:59:37] <deep42thought> ... but first it needs to be built ;-)
[19:59:46] <autofsckk> thanks deep42thought
[19:59:52] <autofsckk> i also have deluge broken
[19:59:58] <autofsckk> do you want to see the output=?
[20:00:01] <deep42thought> buildmaster: wtf /usr/lib/libtorrent-rasterbar.so.9
[20:00:02] <buildmaster> deep42thought: libtorrent-rasterbar [extra]: /usr/lib/libtorrent-rasterbar.so.9
[20:00:02] <buildmaster> libtorrent-rasterbar [extra]: /usr/lib/libtorrent-rasterbar.so.9.0.0
[20:00:11] <deep42thought> yeah, post it :-)
[20:00:18] <autofsckk> https://bpaste.net
[20:00:19] <phrik> Title: show at bpaste (at bpaste.net)
[20:01:06] <deep42thought> do you have python-service_identity installed?
[20:01:15] <autofsckk> i dont know, let me see
[20:01:18] <deep42thought> ah, wait
[20:03:09] <deep42thought> some packages falsely linking against libboost_python27.so.1.69.0
[20:03:28] <deep42thought> but it looks like I don't see the one which breaks deluge
[20:03:56] <autofsckk> im installing python-service_identity right now
[20:04:09] <autofsckk> isnt it libboost_python27?
[20:04:11] <deep42thought> but probably also the libtorrent one
[20:04:25] <deep42thought> that one is not linked against the missing lib
[20:04:39] <deep42thought> ah, sry
[20:04:55] <deep42thought> it's the other way round: libboost_python27 is missing (in the correct version)
[20:05:01] <deep42thought> e.g. your version is newer
[20:05:11] <deep42thought> and something needs to be recompiled against the newer version
[20:38:34] -!- woshty has joined #archlinux32
[21:00:48] <deep42thought> autofsckk: libtorrent-rasterbar landed in [extra] - can you update and check that the before-broken apps work?
[21:23:00] <tyzoid> deep42thought: I'm planning on finshing up https://accounts.archlinux32.org - anything needed from your part?
[21:23:01] <phrik> Title: Arch Linux 32 (at accounts.archlinux32.org)
[21:23:10] <tyzoid> Afaik, all of the token stuff should still be functional
[21:23:33] <deep42thought> I can't think of anything I need
[21:25:52] <deep42thought> you're going to interconnect that with bugs and git, too?
[21:26:36] <deep42thought> the buildmaster needs no connection to that, neither does packages or pkgapi
[21:26:52] <tyzoid> The idea is to get bugs connected to it, yeah
[21:26:57] <tyzoid> I'm not sure git will be as easy
[21:27:17] <tyzoid> since that also stores SSH and GPG public keys
[21:27:52] <tyzoid> Though the question also extends to things to add to the accounts.archlinux32.org site itself
[21:27:57] <deep42thought> hmm, but authentication is not connected to those, right?
[21:28:11] <tyzoid> It is for actually ssh-ing to the repos
[21:28:44] <tyzoid> Though I think giving git.archlinux32.org a pass is okay, since it's mostly just the devs that need access to that
[21:29:04] <deep42thought> maybe I think too abstract about that, but as I see it, you can authenticate by different means: ssh key, gpg key, user+pw or some external service that confirms the identity
[21:29:22] <deep42thought> yeah, it can be added later if we like to
[21:29:28] <deep42thought> or will this be a big change?
[21:29:35] <tyzoid> I agree - thinking abstractly is nice, but I'm building on top of the existing auth system from fluxbb
[21:29:49] <deep42thought> yeah, well, that's my problem :-)
[21:30:02] <tyzoid> In theory, there's not difference between theory and practice
[21:30:09] <deep42thought> !grab tyzoid
[21:30:10] <phrik> deep42thought: foo
[21:30:21] <tyzoid> :)
[21:30:26] <tyzoid> One of my favorite expressions
[21:36:13] <tyzoid> deep42thought: Another thing to consider - cgit uses the pgp auth keys set from gitea
[21:36:37] <tyzoid> since we do have two separate git repo viewers configured
[21:36:49] <deep42thought> ah, hmm
[21:37:09] <deep42thought> I guess, one would need to set up the authentication for each one separately
[21:37:41] <tyzoid> My question for you, since I do believe you have one thing that's integrated with the sso system
[21:37:50] <tyzoid> I can pull data from bbs.archlinux32.org into accounts
[21:38:01] <tyzoid> but I don't have an easy way to pull data from other connected systems
[21:38:15] <deep42thought> I think, I have connected nothing
[21:38:27] <deep42thought> I tried around a little, but actually never needed any authentication
[21:38:41] <deep42thought> at least I cannot remember :-/
[21:38:47] <tyzoid> I thought you hooked it up so only admins can view some of the restricted pages
[21:39:08] <deep42thought> let me grep for it ...
[21:39:30] <deep42thought> nope
[21:40:00] <deep42thought> IIRC, that was once the idea for something, but I think, I never used it, because nothing was actually worth protecting it ;-)
[21:40:07] <tyzoid> Gotcha
[21:58:10] <buildmaster> i686/libgexiv2 is broken (says buildknecht).
[22:23:56] -!- isacdaavid has joined #archlinux32
[22:38:04] -!- lodger has joined #archlinux32
[22:38:31] -!- lodger has quit [Client Quit]
[23:00:17] <buildmaster> i686/firefox is broken (says buildknecht2).
[23:01:49] <thePiGrepper> hi deep42thought
[23:01:55] <deep42thought> hi thePiGrepper
[23:02:26] <thePiGrepper> did you found the issue with qbittorrent? I just updated yesterday and Im having that same issue now. does it have to do with boost?
[23:02:39] <deep42thought> it should be fixed, now
[23:02:52] <thePiGrepper> really? ohh, maybe the mirrors then?
[23:03:05] <deep42thought> err, wait
[23:03:25] <deep42thought> libtorrent-rasterbar was rebuilt, but it still links against the wrong library O.o
[23:04:31] <thePiGrepper> maybe that's true as well, but even if that wasnt the case, Im not seeing it available for update?
[23:04:44] <deep42thought> https://packages.archlinux32.org
[23:04:47] <phrik> Title: Arch Linux 32 - libtorrent-rasterbar 1:1.1.11-2.1 (i686) (at packages.archlinux32.org)
[23:04:57] <deep42thought> which mirror do you use?
[23:05:06] <thePiGrepper> let's see..
[23:05:21] <thePiGrepper> http://archlinux32.vollzornbrot.de
[23:05:36] <thePiGrepper> dont ask me why. I just noticed two days ago
[23:05:51] <thePiGrepper> which one is the most official(?)
[23:06:08] <deep42thought> the master is mirror.archlinux32.org
[23:06:12] <thePiGrepper> thx
[23:07:30] <thePiGrepper> now I do have something in [extra] apparently
[23:07:42] <thePiGrepper> so, that package is still bad?
[23:08:04] <deep42thought> dunno
[23:08:14] <deep42thought> it shows broken stuff in the mysql database
[23:08:16] <thePiGrepper> I'll download it anyway and see what happens.
[23:08:29] <thePiGrepper> yeah, still broken
[23:08:33] <thePiGrepper> same issue
[23:08:33] <deep42thought> :-(
[23:08:50] <deep42thought> let's see who built that ...
[23:10:14] <deep42thought> buildmaster: wtf libboost_chrono.so
[23:10:14] <buildmaster> deep42thought: boost-libs [extra], boost-libs [testing]: /usr/lib/libboost_chrono.so
[23:10:39] <deep42thought> hmm, the build slave installed the correct version of the dependency
[23:10:43] <deep42thought> I don't understand it ...
[23:11:18] <deep42thought> ah, wait
[23:11:28] <deep42thought> now it needs even _newer_ libs :-(
[23:12:03] <deep42thought> let's move boost and break some more stuff :-D
[23:12:07] <thePiGrepper> yeah, I have currently libboost_system.so.1.68.0
[23:12:17] <thePiGrepper> apparently 69 is required
[23:12:38] <thePiGrepper> next version of boost-libs
[23:13:51] <thePiGrepper> it seems like a very simple issue, I have .69 installed in my arch setup
[23:13:59] <thePiGrepper> can you build it?
[23:14:05] <deep42thought> it is already built
[23:14:10] <deep42thought> I just moved it from testing to stable
[23:14:16] <deep42thought> another update should do the trick
[23:14:18] <thePiGrepper> let's see
[23:14:56] <deep42thought> oh, damn
[23:15:01] <deep42thought> that broke quite a lot of packages
[23:15:09] <deep42thought> ... let's move the rest from testing to stable, too
[23:15:21] <thePiGrepper> wait. isnt it simpler to go back?
[23:15:32] <deep42thought> na, updates are never bad
[23:15:35] <thePiGrepper> if it's just one package(qbittorrent) and it's been 2 seconds
[23:15:40] <thePiGrepper> ok
[23:15:45] <deep42thought> I just need to push all the packages which are linked against the new boost version, too
[23:15:51] <thePiGrepper> btw, what broke? and how do you know that?
[23:15:57] <deep42thought> https://packages.archlinux32.org
[23:16:01] <phrik> Title: More and less critical issues with the database (at packages.archlinux32.org)
[23:16:31] <thePiGrepper> ohh, I see
[23:17:25] <deep42thought> "but can be replaced by the one in community-testing" - that's, what I'm currently doing
[23:22:24] <thePiGrepper> there's no way that updating all community with community-testing will create a dependency issue with the 'old' core and extra repos, right?
[23:22:39] <deep42thought> of course, there is :-D
[23:22:56] <deep42thought> we just move as many packages as possible and see what is broken in the end and rebuild that
[23:23:59] <thePiGrepper> there must be a solid solution for this issues. something that once implemented deals with this specific kind of issues
[23:24:09] <deep42thought> db-update -p
[23:24:15] <deep42thought> a.k.a. "progressive"
[23:24:29] <deep42thought> = move everything, that does not break more than what is already broken
[23:25:43] <thePiGrepper> the packages get rebuilt automatically after one of its dependencies get a new version built, is that right?
[23:25:58] <deep42thought> no
[23:26:05] <deep42thought> but there's a tool for that, too
[23:26:08] <deep42thought> seed-build-list -a
[23:26:42] <deep42thought> which puts everything on the buildlist which will not work with the newest version of its dependencies and which itself is the newest version of that package
[23:27:38] <thePiGrepper> hmm, that seems like a different kind of usual issue, but not exactly the same
[23:28:01] <thePiGrepper> that deals with not-yet ready for its dependencies newest version- kind of problem
[23:28:36] <thePiGrepper> while mine was more the 'let's build again because our dependencies are getting replaced soon' kinda issue
[23:28:59] <thePiGrepper> it's like mine would kick first, and then yours would be a backup plan instead things dont work as expected
[23:29:01] <deep42thought> or not-ready-for-newest-dependencies-anymore packages
[23:29:12] <thePiGrepper> right
[23:33:06] <thePiGrepper> well. the good news is that qbittorrent is working! the 'bad' news is that if something broke, it doesnt show ... yet :-)
[23:33:12] <deep42thought> the problem with your approach is that not everything needs to be rebuilt if a dependency was updated and that there may be cycles which you would need to detect and break separately
[23:35:00] <thePiGrepper> ok, but those would be the case for which an 'exception' system would be put in place to deal with those. for the former, hmm, well that would be more an optimization of the solution, to avoid rebuilding endlessly
[23:35:11] <deep42thought> actually, that's like of the haskell approach ;-)
[23:35:57] <deep42thought> e.g. if you rebuild a haskell package, then all dependent packages _must_ be rebuilt, too
[23:36:13] <deep42thought> that's how we end up with so ridiculous high pkgrels for some packages ...
[23:38:02] <deep42thought> https://www.archlinux.org
[23:38:02] <phrik> Title: Arch Linux - haskell-yesod-default 1.2.0-462 (x86_64) (at www.archlinux.org)
[23:39:22] * deep42thought wonders how often one needs to recompile a haskell package before getting a hash collision ...
[23:40:09] <thePiGrepper> that's an issue for upstream as well?(I wouldnt know, I dont use haskell..)
[23:41:01] <deep42thought> yes, thats inherent to how haskell compiles stuff
[23:43:01] <thePiGrepper> I see. I was wondering, wouldnt the other number be more fit for that kind of repackaging? what was the purpose of that last number btw
[23:43:04] <thePiGrepper> ?
[23:43:21] <deep42thought> $epoch:$pkgver-$pkgrel.$sub_pkgrel
[23:43:21] <thePiGrepper> it tends to remain in .0
[23:43:41] <deep42thought> yeah, when _we_ rebuild (but upstream does not), we increase the sub_pkgrel
[23:44:06] <deep42thought> and when upstream rebuilds (but the real upstream does not), they increase pkgrel (and we set sub_pkgrel to 0)
[23:44:37] <deep42thought> and when the /real/ upstream releases a new version, well pkgver is set accordingly, pkgrel is set to 1 and sub_pkgrel to 0
[23:46:23] <thePiGrepper> ok, but those rebuilts in upstream can be either for dependency update-sake as well as for PKGBUILD update-sake, right? there's no way to tell which one of those is it
[23:46:37] <deep42thought> yes
[23:46:44] <deep42thought> and it does not matter to us either
[23:46:49] <deep42thought> we just rebuild, too :-)
[23:47:00] <deep42thought> ... hopefully in the right order
[23:47:07] <deep42thought> but apparently we do that wrong some times
[23:47:46] <thePiGrepper> well, yeah. but if you think about it, is there actually any use in keeping upstream versioning, besides some kind of unused tracking?
[23:48:20] <thePiGrepper> maybe internally, at build time it _might_ be useful, but after that it actually isnt
[23:48:34] <deep42thought> yes, there is 'seed-build-list -m' which seeds the buildlist from an upstream mirror
[23:48:44] <deep42thought> e.g. it schedules everything, that lacks behind
[23:49:23] <deep42thought> and additionally it makes looking for errors easier
[23:49:45] <deep42thought> e.g. "1.2.3-4.x" worked, but "1.2.3-5.y" does not
[23:49:57] <deep42thought> so we can check if upstream changed something from 4 to 5
[23:50:09] <deep42thought> this gets hard if you mangle the x and y into the pkgrel, too
[23:50:48] <thePiGrepper> hmm, I see
[23:51:16] <deep42thought> and besides: what would be the advantage of dropping the sub_pkgrel and putting it into the pkgrel, too?
[23:51:41] <thePiGrepper> well, then that was the reason for the sub_pkgrel number. thx for sharing
[23:51:56] <thePiGrepper> makes sense.
[23:51:56] <deep42thought> yeah, np :-)
[23:52:16] <thePiGrepper> about the other thing though, I still think there must be a better way to avoid all this breakage from time to time
[23:52:36] <thePiGrepper> instead of 'pass everything from testing to stable and see what happens'
[23:52:39] <deep42thought> I just recently fixed a bug in db-update
[23:52:53] <deep42thought> well, maybe my statement was a little bold
[23:53:05] <deep42thought> it will break nothing, that did not appear broken to the database before
[23:53:24] <deep42thought> but unfortunately, some packages _look_ broken but actually work
[23:53:32] <thePiGrepper> how is the linkage database project going btw?
[23:53:40] <deep42thought> I try to filter those out by some exception rules, but I certainly did not find all
[23:53:50] <deep42thought> well, it's in the database
[23:54:04] <deep42thought> and the buildmaster honores the link dependencies
[23:54:09] <thePiGrepper> is all that info going upstream as bug reports so they update the PKGBUILDs?
[23:54:24] <thePiGrepper> they should right?
[23:54:31] <thePiGrepper> that would be the right solution
[23:54:31] <deep42thought> they seem not too keen to insert many dependencies :-D
[23:54:36] <thePiGrepper> reason?
[23:54:45] <deep42thought> "it works as-is"
[23:55:04] <deep42thought> and "it breaks for people who want to build from aur"
[23:55:12] <thePiGrepper> except it doesnt, they just have more eyes, so they dont see the flaws as much..
[23:55:19] <deep42thought> yeah
[23:55:28] <deep42thought> more eyes and more brains ;-)
[23:55:41] <thePiGrepper> haha
[23:56:01] <thePiGrepper> I dont think that's true, the average is higher here I'd bet
[23:56:14] <deep42thought> !bug 61018
[23:56:14] <phrik> https://bugs.archlinux.org
[23:56:41] <deep42thought> ... and that's only the "provides" part of the link stuff
[23:57:39] <thePiGrepper> wow, they closed it because of 'not wanting that many issues opened'
[23:57:49] <thePiGrepper> that's the official reason
[23:57:54] <deep42thought> it is :-)
[23:58:26] <deep42thought> I understand, that it's a bad idea to have the provides thing somewhere _and_ the depends thins in an aur package
[23:58:42] <deep42thought> ... but also only if you don't compile the aur packages in a clean chroot
[23:59:08] <deep42thought> s/thins/thing/
[23:59:33] <thePiGrepper> right