IRC Logs for #crux Wednesday, 2017-06-21

*** pedja has joined #crux00:53
rmullHmmm, firefox is apparently unable to access the network?01:03
rmullI know there was an update - did I miss something?01:03
rmullLooks like the DNS resolution fails according to wireshark01:04
frinnstdo you get timeouts or just blank pages?01:05
frinnstdid you update glibc in the same round as firefox? telinit U?01:06
frinnsti've had to remount /dev/shm after updating glibc in my testing a few times01:06
rmulllooks like wireshark says the "type" field in the DNS query is "unknown" instead of "A"01:06
rmullif I ping, the type is "A"01:06
rmullI updated glibc after firefox01:06
frinnstdidnt work before you updated either?01:06
frinnstupdated glibc that is01:07
rmullIt worked this morning prior to the update01:07
rmullI'll rebuild ff and see where that gets me01:07
frinnstso the last update before it broke was glibc?01:07
rmullUnsure, may have updated ff, then glibc, with no test in between01:07
rmullwebkit browser has the same issue, so it's not just ff01:09
rmullI'll try remounting shm01:09
rmullAssuming that's just "mount -o remount /dev/shm" that did not fix the issue01:10
frinnstyeah thats what i've used. but my problems didnt result in network issues, rather gfx bugs01:10
rmullI'll proceed with the firefox rebuild01:11
frinnstwhat dns server are you using? some public one or a local recursor?01:11
rmulllocal recursor01:12
rmulland like I said - I don't think the recursor is bad, because the outgoing query has the "unknown" type malformatting01:13
rmullping still queries properly01:13
frinnsta reboot would probably solve it - there were quite a few updates to glibc01:14
frinnsthttps://sourceware.org/git/?p=glibc.git;a=shortlog;h=refs/heads/release/2.24/master <- everything from the 2.24 tag01:14
frinnstoh right, you cant view it right now01:15
frinnstsorry :)01:15
rmullhaha :P no problem01:19
rmullI can reboot. Should I let the firefox build complete first, or do you think that it's not going to solve anything?01:19
frinnsti would be very surprised if it did, but cant hurt01:19
rmullRebooting seems easier, my PC is slow. I'll do that first.01:19
rmullfrinnst: Looks like the reboot did not solve it, continuing with the rebuild01:22
frinnstodd01:22
rmullooh. curl, maybe?01:23
frinnstfirefox doesnt use curl01:23
rmullOh - it was in the depends output01:23
rmulland it doesn't work from the command line01:24
frinnstoh really?01:24
frinnst:>01:24
rmullI wonder if this is somehow going to be one big pebkac. But it was working this morning, and it was not rebooted until just a moment ago01:24
rmulland ping works.01:25
frinnstunless you did something really stupid its our (my fault)01:25
rmullhm, curl rebuild does not fix it01:25
frinnstdo you still have the old glibc package?01:26
frinnstor you could rebuild it with just the patch commented out01:26
j_vwhat about trying to temporarily move your user's firefox configuration directory?01:27
rmullUnfortunately the git driver nukes all the built stuff (still annoyed by that) so I don't have it. But I can rebuild it.01:27
rmullj_v: It affects curl on the command line in the same way, so I don't think that's it01:28
j_vah, good point01:28
frinnstj_v: have you updated glibc?01:31
j_vyes, and ff is working fine... curl too01:33
frinnstso i didnt totally fuck up then :>01:33
j_vi don't think so, personally, my bet is some weird corner case01:34
j_vis resolv.conf handled static or via dhcp?01:34
rmullYeah, kind of curious myself... Not sure why that one particular field in the dns query would have a problem01:34
rmullstatic01:34
j_vhave you checked that it's ok?01:35
rmulllooks normal01:35
*** xcko has joined #crux01:35
j_vok, just wondered if glibc update could have overwritten it, but it's not supposed to, it should be rejected01:36
frinnstwhat about adding a host and ip to your hosts file? crux.nu has address 213.132.101.14001:36
frinnstdoes that work? you'll need to restart firefox01:36
rmullCurl works with the entry to the hosts file01:37
rmulland so does ff01:37
frinnstwhat about if you start nscd?01:38
frinnstjust running /usr/sbin/nscd should be good enough for testing01:38
rmullerm... works?01:39
rmullI think...01:39
frinnstcool01:39
frinnstand odd01:39
frinnstwhat about without nscd and with another dns server? such as googles 8.8.8.801:40
rmullkilled nscd, problem resumes01:41
rmullWill try 8.8.8.801:41
rmullNo luck with 8.8.8.801:42
rmulltype still set to 5100 instead of 000101:43
j_vcould the local recursor need to be rebuilt?01:43
frinnstif it doesnt work with 8.8.8.8 its not his recursor01:44
rmullglibc 2.24-3 finished, I installed it, ff works again01:44
frinnstwtf01:44
rmullSorry man, that's annoying01:46
rmullFeel free to mark it "works for me" until you get another report I guess01:46
frinnsti think this is the only part of the patch that touches resolv: https://sourceware.org/git/?p=glibc.git;a=commit;h=b3b37f1a5559a7620e31c8053ed1b44f798f2b6d01:47
frinnstdo you have DNS_RESOLVER=Y in your kernel?01:49
frinnstah right, we dont use that01:49
frinnstnevermind01:49
frinnsttoo late to think, 3:50 am and the sun is up01:50
rmullThan knob is set to Y fwiw01:50
rmulls/Than/that01:50
rmullif I try updating to glibc 2.25, I'm probably asking for trouble, right?01:54
frinnstupgrade should be fairly fine. but downgrading later - not so much02:03
frinnsti run 2.25 on my desktop02:03
frinnstzzz off to bed02:04
*** abenz__ has joined #crux02:08
rmull2.25 does not work02:08
rmullsame dns issue02:08
rmullthanks, talk to you later02:09
ryu0rmull: did you try using dig?02:10
*** abenz_ has quit IRC02:11
*** onodera has quit IRC02:12
rmullryu0: No, but I just did, and it works fine02:12
rmullProperly formatted A queries02:13
rmullI seem to be able to move back and forth between glibc 2.24-3 and 2.25-1 with no trouble, and firefox works with the old one but not the new one02:16
j_vrmull, could you check the linkage of firefox? 'readelf -d /usr/bin/firefox'02:21
rmullj_v: Here it is when it's broken: http://sprunge.us/VgWK02:24
rmullAnd when it's working: http://sprunge.us/UHfX02:24
j_vhttp://sprunge.us/UHfX02:25
j_voops. wrong window02:25
brian|lfsjust switch to another browser02:25
rmullbrian|lfs: When firefox is broken, so is my webkit-based browser, and so is curl02:26
rmullping and dig aren't broken02:26
rmullI'm fine just using the older version of glibc fwiw02:26
rmullI am just curious about what is going on02:26
brian|lfsI couldn't get the new glibc to compile02:28
brian|lfsor Romster do you have it on your server built yet02:28
j_vrmull, did you test firefox 54?02:28
rmullNo - only using the latest packaged version02:29
j_vrmull, those readelf outputs look normal, not extra libs or weird rpaths02:29
rmullWell that's good at least02:30
rmullI need to step away, thanks all for the help02:30
*** _________mavric6 has joined #crux02:31
j_vdon't blame you, sometimes is best to step back anyways... i need some sleep too, good night02:31
j_vinteresting note, i now find that my claws-mail install is exhibiting the same problems that rmull is having with firefox and curl02:51
*** xcko has quit IRC03:15
*** xcko has joined #crux03:17
j_vfrinnst, did some googling and it appears that debian and ubuntu ran into same/similar issue when applying fix for CVE-2015-5180.03:17
j_vhttps://bugs.launchpad.net/ubuntu/+source/glibc/+bug/167453203:17
*** xcko has quit IRC03:25
*** xcko has joined #crux03:27
*** abenz has joined #crux03:29
*** xcko has quit IRC03:32
*** xcko has joined #crux03:33
*** amonn has joined #crux03:40
Romsterit is now brian|lfs03:49
*** xcko has quit IRC03:51
*** xcko has joined #crux03:53
brian|lfscool whats your url again04:17
brian|lfsAlan finally upda ted pulseaudio today04:18
brian|lfsfirefox don't work for me now that I updated glibc04:34
*** sonophilia has quit IRC04:38
*** tilman_ has joined #crux04:39
*** xcko has quit IRC05:28
*** xcko has joined #crux05:29
*** xcko has quit IRC05:58
Romsterbrian|lfs, http://crux.ster.zone07:50
*** jue has joined #crux07:56
frinnstbrian|lfs: what happens?11:05
dbrookeprobably resolver, I can confirm the glibc update causes me resolver problems with at least FF and curl, as others mentioned overnight, so I've reverted to the package from the ISO for now11:17
*** stratact has joined #crux11:19
dbrookethe patch seems to be related to combining A and AAAA lookups so may only affect people with IPv611:20
dbrookesome of my services, including recursors and http proxy, are IPv6 only and they all seemed to become unresolvable11:23
frinnstwhy can't i reproduce it? :(11:23
dbrookeyou have IPv6?11:24
frinnstyes11:24
frinnstmy lan is mostly ipv611:25
dbrookestrange11:25
dbrookethat you don't see a problem, that is 8-)11:25
dbrookeI'd tried rebuilding FF and also rebooting, reverting glibc fixed it instantly11:26
frinnstwhat about rebuilding nss?11:28
dbrookeI can try ...11:29
dbrookejust running sysup first to check for breakage11:31
joacimworks on my machine (r)(tm)11:41
dbrookeOK it broke, so I rebuilt nss and it's still broken but I'm not sure if nss is actually in use11:44
*** stratact has quit IRC11:44
joacimlooks like the prime x370-pro is missing one pcie slot11:45
dbrookeI can't check if  http://stuff.dbrooke.me.uk/sp-WGxegu  uploaded but that shows that ports -u fails for the non-crux.nu repos so httpup is broken11:48
frinnsthttpup uses curl11:48
frinnstyes it uploaded11:49
dbrookeanyway I'll have to revert for the moment so I can do stuff11:49
frinnstyeah thanks for your help11:49
frinnsti'll try to reproduce it.11:49
dbrookeI can maybe try a bit more investigation later11:49
frinnstgood thing im on vacation and have time to spare :>11:50
dbrookeand just to confirm, pkgadd -u glibc#2.24-3.pkg.tar.xz from the ISO, followed by a restart of FF, fixes it for me11:53
frinnstwhat dns server are you using? self-hosted? (if so, what software) or isp?12:09
dbrookeself-hosted bind (debian), I have 2 and they are also authoritative for my own domains (one master, one slave) as well as being recursive without forwarders12:18
ryu0frinnst: is crux even on 2.25 yet? glibc in core is still 2.24?12:28
frinnstryu0: yes, 2.2412:28
frinnstrmull: do you also use ipv6?12:33
joacimfrinnst: was it you that had an issue with the beta firmware for the prime x370-pro?12:36
frinnstno, but I am running the beta bios12:41
frinnstworks fine, doesnt appear to solve this issue though: https://www.phoronix.com/forums/forum/hardware/processors-memory/955368-some-ryzen-linux-users-are-facing-issues-with-heavy-compilation-loads?p=955498#post95549812:42
frinnstdbrooke: new patch without the resolver stuff: https://crux.guru/glibc-2.24-updates.patch12:51
frinnstif you want to test12:51
*** chinarulezzz has quit IRC12:57
Romsterokay now my DNS has broken and i am purely on IPv4 but i haven't rebooted yet about to reboot13:02
rmullfrinnst: No, no IPv6 on my LAN, though my local ifaces all have v6 addrs in addition to the v4 addrs13:14
rmullI guess what I mean is - I have no v6 route to the wan13:16
rmullnever did anything to explicitly enable any v6 infrastructure13:16
rmulldns queries are v413:17
rmullIt is interesting that having nscd running seems to fix the problem, but I didn't get a pcap during that phase, and I can't until tonight13:17
*** Romster has quit IRC13:27
frinnstif someone can test the above patch i'll push it13:28
*** Romster has joined #crux13:29
frinnstI cant reproduce the problem so i'm a bit lost.13:30
joacimsame. if i tested the patch, i wouldn't be able to tell if it improved anything13:31
frinnstoh wait, I think one of my VMs actually is broken13:31
frinnstwee13:31
frinnstno, curl works but firefox seems somewhat broken13:32
pedjaACTION waits for glibc-32 build to finish to join the fun13:34
pedjaso, for what issues am I looking for? broken DNS ?13:36
frinnstyes13:36
frinnstbut apparantly not system wide13:36
frinnstfirefox and curl are known to have problems13:36
pedjaI use upstream FF13:37
pedjaboth glibc builds have finished, time to sysup13:42
pedjasee you on the other side :)13:42
*** pedja has quit IRC13:42
frinnstRomster: still problems?13:44
*** onodera has joined #crux13:46
*** pedja has joined #crux14:08
pedjawell, this is weird.14:09
pedjacurl was broken, 'dig crux.nu' without explicitly setting the server, too14:12
pedjadig @8.8.8.8 crux nu worked14:12
pedjawget, too14:13
dbrookethat's similar to what I've seen, not everything broke14:15
*** john_cephalopoda has joined #crux14:15
dbrookefrinnst: seems to be working with your revised patch14:15
pedjaand after starting nscd, everything works, dbrooke ?14:16
dbrookeI've not started nscd, don't think I ever have done14:18
pedjait's shipped with glibc14:20
frinnstwohoo i managed to reproduce it on my box at work14:20
frinnstServer not found \o/14:20
*** emmett1 has joined #crux14:21
dbrookepedja: yeah I see it there but I've never used it as I have caching resolvers in my network14:21
pedjafrinnst, so what's different on that box?14:22
pedjaIPv6 LAN side?14:23
frinnstipv4 only14:24
pedjaI have a pcap before and after I started nscd, if you're interested14:25
frinnstdid you have problems before nscd?14:25
frinnstwithout*14:26
frinnstthanks dbrooke14:26
pedjayes, curl was broken, dig without @, Chrome14:26
frinnstcool14:26
pedjaHexchat, too :)14:26
frinnstcan you email them to me?14:27
pedjasure14:27
frinnst<314:27
frinnstright so I'll just push the new patch that unbreaks stuff14:40
frinnstand file a bug upstream perhaps14:40
dbrookelog on my caching resolver from when it was failing ports -u earlier: Jun 21 11:44:52 stroudley named[29733]: error (unexpected RCODE REFUSED) resolving 'crux.ster.zone/TYPE20736/IN': 84.200.74.134#5314:41
dbrookeTYPE20736 I'd expect should be A or AAAA14:42
dbrookeand 20736 is 16 bits worth of their new magic value 439963904 so looks like a type mismatch somewhere14:53
cruxbot[core.git/3.3]: [notify] glibc: reverted CVE-2015-5180 patch - caused resolv problems for some users14:55
cruxbot[core.git/3.3]: [notify] glibc-3: reverted CVE-2015-5180 patch - caused resolv problems for some users14:55
*** lounge has joined #crux15:00
pedjafrinnst, sent (sorry for the delay, I had lunch)15:11
pedjaI extracted just the DNS packets, full pcap is ~14Mb15:14
*** pedja has quit IRC15:17
*** pedja has joined #crux15:23
pedjahm, seems that my dig issue (returns Got bad packet: FORMERR) isn't related to glibc15:32
*** lounge has quit IRC15:34
*** lounge has joined #crux15:39
rmullfrinnst: Reverted glibc works for me, thanks15:44
*** emmett1 has quit IRC16:09
*** Workster has quit IRC16:22
*** edimwsuhgcjfifzs has joined #crux16:22
*** Workster has joined #crux16:22
pedjafrinnst, ignore the dig part of my email, it was an unrelated issue16:44
*** tsaop has joined #crux17:41
*** lounge has quit IRC17:59
*** lounge has joined #crux18:00
*** cippp has joined #crux18:05
*** chinarulezzz has joined #crux18:20
john_cephalopodaHmm, somehow qt(5) is never quite right.18:23
john_cephalopodaSome configurations seem to be a little off.18:23
*** deep42thought has joined #crux18:34
*** ubuuu has joined #crux18:36
*** tsaop has quit IRC18:36
*** ubuuu has quit IRC18:36
deep42thoughtHi, I got a problem: Since this mornings updates, some applications (pidgin, curl, firefox) cannot resolve some host names, while others can (ping, ping6, dig, ssh).18:37
*** tsaop has joined #crux18:38
deep42thoughtMaybe the faulty part is my router, which I also updated, but the my arch box doesn't show these errors :-/18:38
][_R_][Seems to only affect some people, not everyone.  There was an update to glibc 4 hours ago though.18:38
pedjajohn_cephalopoda, open a bug report.18:39
pedjadeep42thought, as ][_R_][ said, that change was reverted. ports -u and update glibc.18:41
deep42thoughtok, then I'll just have to figure out how to do this without name resolution ...18:42
pedjastart nscd, that should fix it temporarily18:42
][_R_][Use pkgadd on an old cached version of glibc18:42
deep42thought][_R_][: good point18:43
deep42thoughtthanks!18:44
*** deep42thought has quit IRC18:44
*** deep42thought has joined #crux18:44
deep42thought\o/18:44
*** tsaop has quit IRC19:06
joacimmy system actually booted with the ram set to DDR4-320019:26
just_funIt's funny how glib issue manifested here: firefox(-rust) worked with external sites (www.) and failed with local ones (avahi/.local).19:37
*** lounge has quit IRC20:07
jaegerjoacim: nice. which RAM did you end up buying? The trident z stuff?20:07
joacimgot a couple of flare x kits20:08
joacimthe second one will come here tomorrow i think20:08
jaegerAh20:08
jaegerMy GTX 1060 should arrive today for the ryzen box20:08
joacimthe trident z stuff was out of stock most places, and the flare x seemed more like guaranteed compatibility20:08
joacimi'm getting the RX 550 =)20:09
jaegerThe flare X is the only kit that's fully intended to run on ryzen at 3200, I believe20:09
joacimhope i can run it off the bottom 8x slot20:09
joacimforgot how huge ATX systems are20:10
jaegercan be, yeah :)20:10
joacimsurprised by how heavy it was too. phanteks used some really thick aluminium pieces. much thicker than lian li20:11
joacimthe glass didnt help either =)20:11
*** deep42thought has left #crux ()20:13
*** Guest1653 has joined #crux20:16
cruxbot[contrib.git/3.3]: openvpn: 2.4.1 -> 2.4.3 fixes several RCE/CVEs20:16
jaegeryeah, it's a heavy bastard20:33
jaegerI really like it, though20:33
joacimyeah feels solid20:36
joacimbooted on first try :D20:36
joacimbut no picture...20:36
j_vwhich phanteks models are you guys using... they look like nice cases20:37
joacimevolv atx here, the glass version20:37
j_vi definitely want to consider one of those for my next build. sound like sturdy, well built20:39
j_vcases20:39
jaegerI've got an evolv atx TG and an evolv ITX20:40
j_vyou're satisfied with quality on both, i take it?20:41
joacimonly downside i can think of is that the holes for cables near the bottom of the motherboard are a little small20:41
joacimsome bigger cables like the usb3.0 one is a bit tricky to get through20:41
*** sonophilia has joined #crux20:42
j_vas issues go,sounds pretty minor. thanks for sharing20:43
joacimyeah20:45
jaegerYeah, I like both cases quite well. The ITX one was a bit of work to fit everything in because I did a custom water cooling loop20:46
j_vnice20:47
joacimi guess it was too early with trying out the amdgpu stuff20:50
cruxbot[contrib.git/3.3]: camlp: new port20:53
cruxbot[contrib.git/3.3]: ocamlbuild: new port20:53
cruxbot[contrib.git/3.3]: wyrd: fix source, add new dependencies20:53
ryu0has water-cooling ever leaked and killed the electronics?21:23
joacimmy NIC comes up as ETH0 in ip link21:29
Romsterupdated rebooted and dns is still broken on firefox curl thunderbird but yet other things work, and i rebuilt nss and nspr no difference21:29
joacimlspci detects it and i'm using the correct driver (igp)21:29
joacimcannot use it to get an IP with dhcpcd eth021:29
Romsterupdating glibc and glibc-32 yet again21:29
joacimany idea of what im doing wrong?21:29
*** xable has joined #crux21:29
Romsterand pidgin don't work.... ahhh waits of rbuild and i got 7 other machines to log in and updte glibc and glibc-32 again -_-21:32
ryu0Romster: and all I have to do is run apt-get dist-upgrade.21:35
ryu0:-)21:35
Romsteryou can keep your ubuntu <_<21:37
ryu0lol21:38
ryu0brb reboot21:38
*** ryu0 has quit IRC21:39
Romsteri didn't have to reboot dns now resolves again21:39
*** ryu0 has joined #crux21:42
Romsterryu0, i didn't have to reboot dns now resolves again21:42
ryu0Romster: i had to reboot for a kernel upgrade.21:43
Romsteri did that before21:43
Romsteri never did bother with hot swapping kernel21:43
*** xable has quit IRC21:52
onoderawhat's going on with glibc?22:03
*** Guest1653 has quit IRC22:04
Romsteri dunno but the last update fixed it again without a reboot22:06
Romstersomething about dns22:06
RomsterA to AAAA records or something i don't quite understand it22:07
john_cephalopodapedja: I don't know what exactly is off about the crux qt5. It's just that applications, that compile perfectly fine on other distros, suddenly get issues with compiling on crux that seem to be related to qt5. And I have no clue if it is due to the way qt5 is packaged or due to misconfiguration on my side.22:08
Romsteri was going to test after work but i seen frinnst already pushed glibc/glibc-32 so i assumed it was fine and updated my 8 machines22:08
pedjajohn_cephalopoda, what applications?22:09
Romsterapparantly the same CVE patch fucked over auto updates on a lot of debian based manchines and broke them. inculding ones in production22:09
Romsterkde works on qt522:11
Romsteryou have to use qmake-qt5 than qmake22:11
ryu0Romster: when it isn't crashing?22:11
Romsterthe rest pickes the paths from qmake-qt522:11
Romsterthere is qt5 5.9.0 but i haven't pushed that yet22:12
pedjaRomster, I would suggest you wait for 5.9.1 :)22:13
Romsteryeah i'd rather keep the well tested than some major changes that aren't quite well thought out yet22:14
pedjajohn_cephalopoda, open the bug, attach the ports and error logs to it.22:15
pedjaI use a couple of qt5 apps, haven't had any issues with them22:16
Romsteralready 2 bugs known in gt5 5.9.0 could be more22:16
RomsterQTBUG-61140 xcb: Don't destroy foreign windows22:17
RomsterQTBUG-60558 QInternal::unregisterCallback: don't crash on unregistering22:17
john_cephalopodaI once reported a bug for those applications.22:17
pedjaFS# ?22:17
john_cephalopodaThe thing that happened now for me is, that the krita appimage isn't working.22:17
Romsteri don't recall seeing it?22:17
Romsterwhat's the FS# number22:18
Romsteri did see reports about needing qt-switcher or chooser but i worked around that22:18
Romsterkf5 repo is using opt/qt5 with no reported problems.22:18
pedjayeah, that one was mine :)22:18
pedjaqt-chooser one22:18
Romsterah i didn't like that solution22:18
pedjawell, it's upstream's preference, so22:19
Romsterqmake sets the paths i don't see why qt-chooser needs to exist, qt3 and qt4 coexised before qt4 and qt522:20
Romsterupstreams preference probably because nitwits can't package?22:20
john_cephalopodaWait, I'll report that Krita issue.22:20
RomsterACTION shrugs22:20
pedjato work around some apps broken qt detection, I guess22:20
Romsterif you are going to report make sure i can get the Pkgfiles from some repo so i can reproduce your build22:21
Romsternothing more annoying than foo breaks bar but i can't get bar because it's not published.22:22
pedjajohn_cephalopoda, where did you get that krita appimage?22:22
Romsterof if it is sometimes missing the patches that went with it22:22
john_cephalopodapedja: On their official website.22:22
Romsterhttps://crux.nu/portdb/?a=search&q=krita22:23
Romsterno results22:23
john_cephalopodapedja: Krita devs told me that the appimage was tested on several distros (Ubuntu, CentOS and some others) without problems, gives me an error though, that looks like it could be related to qt5.22:23
john_cephalopodahttps://download.kde.org/stable/krita/3.1.4/krita-3.1.4-x86_64.appimage22:24
john_cephalopodaCould you test if it works for you? Else I simply might have some broken configuration. I had qt5 issues before that just resolved themselves magically, so it might be something like that again.22:25
Romsterok so where is your attempt of the Pkgfile22:25
Romsterthat is a binary file22:25
john_cephalopodaI know. I am trying to execute it and it fails.22:26
Romsterseriosuly john_cephalopoda  when you report things you do a harlf asrsed job at it.22:26
Romstersource + build instructions to22:26
Romsterhttps://git.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/krita oh look a refference to go by22:28
Romstermake a Pkgfile for it22:28
Romsterthat binary is not going to be compatible with crux22:28
john_cephalopodaRomster: I am not talking about building a package. I am talking about trying to execute this appimage, which appears to work on all tested platforms but not on Crux. I don't know if it is due to a misconfiguration on my side or a general Crux problem, that's why I ask.22:28
john_cephalopodaAppimages are made to be compatible with most distros.22:29
Romsterhave you straced it?22:29
Romsterand the last ting it bombs out on patebin22:29
Romsterthing*22:29
RomsterJoin us on #AppImage on irc.freenode.net22:32
Romsterperhaps they can help you pinpoint what in crux is not compatible22:33
Romsteri haven't got the time to dig though it to find the issue.22:33
john_cephalopodahttps://bpaste.net/raw/2efd804b737922:34
Romsterthey say they package all the dependencies other than the base system.. what do they define as /base system/22:35
pedjaexport QT_DEBUG_PLUGINS=1, then ./krita.appimage will reveal the issue22:35
pedjathis is interesting https://github.com/AppImage/AppImages/blob/master/recipes/krita/Recipe22:37
joacimfrinnst: did you have to do anything special to get the onboard NIC to work?22:38
pedjatl;dr is that krita appimage will have to be tweaked to run on Crux22:38
john_cephalopodapedja: Thanks.22:38
pedjanp22:38
john_cephalopodahttps://bpaste.net/raw/4d5626346abb22:38
joacimI just get the default eth0 device, udev should rename it to enp*, right?22:39
john_cephalopodaIt looks like xcb wants a special zlib.22:39
pedjanot special, just an older version libpng was comiled against22:40
pedjacompiled*22:40
pedjaand that's xcb plugin that ships with krita, not from qt5 port22:41
pedjaso, it's an upstream bug22:42
Romsterwe have /usr/lib/libpng16.so.16 symlink in our libpng22:42
john_cephalopodaYeah. When I asked on #krita, I was told that it could possibly be a misconfiguration that forces it to use qt5.22:42
john_cephalopodaBut apparently not.22:42
Romsteroh older zlib i see it22:42
Romsterthere libqxcb.so is compiled against ZLIB_1.2.9 e have zlib 1.2.1122:44
Romsteryou'll need to make a compatible zlib-compat port22:44
pedjaif you look at that recipe for krita appimage, you'll see a bunch of hacks/workarounds for various distros22:44
Romsterthey should provide support for older zlib versions if they don't want to use current stable in krita22:46
john_cephalopodaI could try to make a Krita Pkgfile again. I made one once and it worked for a short time, but then I cleaned up my system and deleted some dependencies and decided that using the appimage would be easier.22:46
Romsterthis is not a opt/qt522:46
pedjafew of the appimage'd apps I tried worked without issues on Crux, iirc.22:48
cruxbot[contrib.git/3.3]: libconfig: new port22:49
cruxbot[contrib.git/3.3]: sslh: 1.17 -> 1.1822:49
Romsteryou could try ln -s /lib/libz.so.1.2.11 /lib/libz.so.1.2.9 but more than likely that wont be ABI compatible. but it should offer then earlier symbols it requires in that later version as the major version and minor version are in the same series. 1.222:50
Romster1.2.x22:51
Romsterwelcome to binary packages trying to be built to work on all distros that don't use the most current stable version22:52
cruxbot[contrib.git/3.3]: thc-hydra: fix source22:54
john_cephalopodaRomster: As long as the big distros use outdated packages, appimages will have to resort to those. Plus it's more likely that a lib installed on a random distro is backwards-compatible than that it is forward-compatible.22:55
Romsterhttp://www.zlib.net/zlib-1.2.9.tar.gz doesn't even exist on the site even22:55
Romsterso yeah just make a symlink and try that22:55
john_cephalopodaTried it, still the same error.22:56
john_cephalopodaIt probably compares a version string.22:57
john_cephalopodaOr it's not compatible.22:57
cruxbot[contrib.git/3.3]: thc-hydra: use github for source22:58
Romsteri can't even find that file on there site but i see it here http://zlib.net/fossils/22:58
john_cephalopoda"fossils" :D22:59
ryu0ACTION adds john_cephalopoda to the fossil record.22:59
john_cephalopodaI'm not really old :รพ22:59
ryu0john_cephalopoda: true, you're just old enough that people are only interested in carbon dating you. =p23:00
cruxbot[contrib.git/3.3]: synergy: fix source23:01
john_cephalopodaryu0: "I usually only go on carbon dates" would be the perfect pick-up line for an archaeologist :D23:02
john_cephalopodaOr "You seem to have a low-carbon diet. Can I still try to date you?"23:03
Romsterjohn_cephalopoda, https://gist.github.com/therealromster/5fcdce2754d49e5b894b2fd2536170bc23:06
Romsterdo that for every outdated library it needs23:06
john_cephalopodaRomster: I just ldd'd the libpng16 in my /usr/lib/ and it links against an outdated libz.23:08
Romsterthen fix your system23:09
frinnstjoacim: no, just use the correct driver :-) igb23:11
frinnsthave you received your stuff?23:11
joacimno idea what i did wrong, but i make a fresh kernel config23:11
joacimit works now23:11
joacimyeh =)23:11
joacimassembled it today. this was the last attempt for today. was getting ready for playing video games for an hour if i couldnt figure the ethernet device out23:12
john_cephalopodaThis looks odd... I compiled it anew several times but it keeps telling me that - could also be a misinterpretation though...23:12
john_cephalopodaEh, I should just go to sleep.23:12
john_cephalopodaNight23:12
joacimlater23:13
*** john_cephalopoda has quit IRC23:13
*** duncaen has quit IRC23:32
*** duncaen has joined #crux23:33

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!