IRC Logs for #crux Tuesday, 2017-10-31

jaegerprtverify reports the following for the repo: ERROR  jmf/qtcreator ............ directory not allowed: usr/libexec/00:08
jaegerERROR  jmf/xorriso .............. directory not allowed: usr/share/info/00:08
jaegergood paste, irssi, thanks00:08
saptechhi all00:11
*** g0relike-2 has joined #crux00:11
saptechwas working on getting lxqt basic setup and ran across an lxqt-panel issue00:12
saptechhere's the error log, https://bpaste.net/show/00b3325218a600:12
saptechbuild log00:13
jaegerjohn_cephalopoda: would appreciate if you'd fix those two but I did add your repo. It'll show up after the next sync runs00:13
*** g0relike has quit IRC00:14
jaegersaptech: I'm no lxqt expert but it looks like it libsysstat with qt5 support is missing?00:16
saptechok00:16
john_cephalopodajaeger: Thanks, I'll fix those. I also had an issue with a changed dependency name that I will commit. But I haaven't got the secret key with me rn, so it will take two days.00:16
jaegerok00:17
jaeger:q00:17
jaegerwrong window, heh00:17
saptechI wonder if libsysstat should be a dependency of sysstat? the opt port of sysstat doesn't show any depends00:25
jaegerIt might be one of those packages that's different depending on other stuff installed. Maybe sysstat adds qt5 support if built with qt5 installed00:27
jaegerJust a guess, though00:27
jaegerIt might also be that libsysstat-qt5 is a separate package00:27
saptechI'll keep researching this. I started it Friday and stopped until today00:29
jaegerAre you using timcowchip's lxqt ports? If so, maybe he knows00:30
saptechyes it is his ports00:31
saptechdo he be on irc, is so what name he uses?00:32
saptechif*00:32
jaegerusually by the name I mentioned. I don't see him right now00:32
saptechok00:33
*** TheCephalopod has joined #crux01:04
*** john_cephalopoda has quit IRC01:08
*** SiFuh has quit IRC01:15
*** SiFuh has joined #crux01:16
*** Workster has quit IRC01:48
*** p4cman has joined #crux01:50
*** henesy has joined #crux02:19
*** SiFuh has quit IRC02:42
*** SiFuh has joined #crux02:43
brian|lfsjaeger, the signature is broken on Kodi02:55
jaegerseems fine here03:00
jaeger6b0886e7449fc201e0ec0584b37f9f654c429797a41e6d0b6a4b5a7fd5ec34dc03:00
jaegerdoes the sha256sum from your download match that?03:17
*** _________mavric6 has quit IRC03:24
*** _________mavric6 has joined #crux03:26
brian|lfshmmm not sure too late now03:28
brian|lfshey jaeger someone needs to fix curl03:28
brian|lfsit only resovles for ipv6 or defaults to ipv603:28
brian|lfsI spent an hour with SiFuh early this morning to figure out that was the issue03:28
jaegersteps to duplicate the issue?03:32
*** Workster has joined #crux03:37
*** Workster has quit IRC03:37
*** Workster has joined #crux03:37
brian|lfspretty much do ports -u03:45
brian|lfsand if you have any git repos they will fail03:45
brian|lfsthen we check http https and my network settings all were good03:45
brian|lfsdid curl on one of  my repo urls it failed03:45
brian|lfsI checked the Pkfile for curl and ipv6 was enabled in it03:45
jaegerdo you have a working ipv6 stack? can you, for example, 'ping6 ipv6.google.com' ?03:46
brian|lfstried enabling ipv4 and 6 it wouldn't compiler like that03:46
brian|lfschecking03:46
brian|lfsnah don't have ipv6 enabled03:47
jaegerSo far I'm not finding any curl failures on my systems, got a specific URL that doesn't work?03:48
brian|lfsya03:48
brian|lfshttps://github.com/tsaop/crux-kf5.git03:49
brian|lfsit works on my system after compiling with ipv6 off03:49
brian|lfsand had the same issue on my laptop after updating everything on it03:49
jaegerIt works for me with both curl and git03:50
brian|lfscheck your curl pkgfile03:50
brian|lfsis ipv6 on03:50
jaegeryes. It's the standard core Pkgfile03:50
jaegerboth ipv6 and ipv4 work for me03:50
brian|lfsso yyou saying I need to get with the times and use ipv603:50
jaegerdo you have openldap installed?03:50
brian|lfsbeleive so03:51
jaegerNot if you don't want to, it works both ways03:51
jaegerrebuild openldap03:51
brian|lfsok03:51
jaegercurl links against it if found03:51
brian|lfswell I'm not crazy I had the problem on 2 systems03:52
brian|lfsmy main system it just started acting up way after the fact curl was modified on 10/23 and the problem started eysterday03:52
jaegerThere was a recent glibc update that affected some other ports' name resolution03:52
brian|lfslaptop right after updating03:52
brian|lfsgood point think I noticed it after updating glibc03:53
jaegerrebuilding glib solved some of it, openldap the rest (as far as I'm aware)03:53
brian|lfsnever got glibc to build on either of my systems03:53
jaegerAre you subscribed to the mailing list? It was discussed there03:53
brian|lfsso used Romster's packages03:53
brian|lfsya03:54
jaegerIf you can't build glibc you've got some other problem that should be fixed03:54
brian|lfswill only thing that fails03:54
brian|lfswell cups fails for me avahi disabled03:54
brian|lfsnot sure what I have that uses avahi03:55
brian|lfsheres my glibc compile error https://paste.pound-python.org/show/CJBrJq5joBvYsci9psJf/03:57
brian|lfsmy glibc-32 error https://paste.pound-python.org/show/Q4BSDorS2ToevcbcDKje/03:59
brian|lfsget the same errors on both of my systems03:59
jaegerwhat CFLAGS?04:00
jaegerI suspect this is a -march=native problem04:00
brian|lfsexport CFLAGS="-march=native -O2 -fstack-protector-strong --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2"04:01
jaegerfor science, try it with -march=x86_64 ?04:01
brian|lfsok04:01
jaegerMight also need -fno-stack-protector based on some searching04:03
jaegerbut I think -march=native is more likely the culprit04:03
brian|lfsits x86-64 by the way04:04
brian|lfsrealzied your error after doing _ and it bombing out right away04:04
jaegeryeah, sorry. so used to x86_64 for kernel paths04:05
brian|lfssame error it failed04:05
brian|lfsI'm going to just try witht he x86-64 and nothing else04:05
jaegerwhat about with stock CFLAGS? "-march=x86-64 -pipe -O2"04:05
brian|lfsok can try that04:06
brian|lfsrunning04:07
brian|lfsthat wokred04:08
brian|lfsfor glibc will do glibc-32 next04:09
brian|lfsI don't get many compile failures lately unless I'm missing a dep04:09
brian|lfswould like to know why chromium 62 takes like 4 hours to compile and 61 toook about an hour04:10
prologicMore deps?!04:11
brian|lfsno its crazy slow like its only compiling on one thread04:11
brian|lfsand I checked the Pkgfile04:11
brian|lfsdon't see anything wrong in the Pkgfile04:11
brian|lfsI can only venture maybe one of the patches applied to it04:11
brian|lfsvery weird04:11
brian|lfsbut ya that caused my laptop to take 12 hours to get everything up to date04:12
brian|lfsnot worth it04:12
jaegerIn a container with your original CFLAGS I see the same glibc failure. Going to remove them one by one to find the one that causes the problem04:12
brian|lfsI love CRUX but for me don't make sense on a laptop04:13
brian|lfsI don't use my laptop enough04:13
brian|lfsso to update takes forever04:13
brian|lfsneed to switch to archlinux something bleeding see if my touchpad is fixed yet04:14
brian|lfsAsus uses some weird touchpad and it don't work on linux04:14
*** saptech has quit IRC04:19
jaegerok, it seems to be -fstack-protector-strong04:22
*** pekka20 has quit IRC04:25
*** pekka20 has joined #crux04:25
*** p4cman has quit IRC04:29
brian|lfsglibc must be weak then04:35
SiFuhfrinnst: what does ftw mean?  "frinnst> btrfs ftw"06:00
abenzfor the win06:19
abenzhe probably messed up his filesystem, but had a backup, so he rolled that one and everything worked!06:19
abenzthen expressed his excitement for btrfs here cos it was so easy ;)06:19
abenzgiven you have a backup of course!06:19
*** p4cman has joined #crux06:21
SiFuhWe get a lot bad results with the Seagate 4TB Backup Plus 2.5" Portable Hard Drive. Often dataloss. However, I noticed that the drives peform flawlessy when your change the filesystem to ext4 instead of ntfs.07:01
*** prologic has quit IRC07:42
*** prologic has joined #crux07:49
Romsterbrian|lfs, because you didn't rejmerge pkgmk.conf and see that ther eis a new commented out # JOBS=nproc07:50
*** ngyj has joined #crux08:24
frinnstyeah dont pass silly cflags to glibc08:39
frinnstfuture versions support it tho08:39
*** jue_ has joined #crux09:58
*** jue has quit IRC10:01
*** ivladak has joined #crux10:18
*** john_cephalopoda has joined #crux10:27
john_cephalopodaHi10:27
frinnstg'day10:34
john_cephalopodaYay, my repo is in the ports list :D10:36
frinnst\o/10:36
frinnsthttps://www.youtube.com/watch?v=eELI2J-CpZg&feature=youtu.be&t=37m10s10:53
frinnstawesome10:53
john_cephalopodaHaha! :D11:04
*** p4cman has quit IRC12:30
*** ngyj has quit IRC12:40
*** ngyj has joined #crux12:54
*** saptech has joined #crux13:04
*** ngyj has quit IRC13:14
*** p4cman has joined #crux15:16
*** prologic has left #crux ("Be back later...")16:15
*** john_cephalopoda has quit IRC16:25
*** de11 has joined #crux16:41
*** onodera has joined #crux17:07
*** john_cephalopoda has joined #crux17:26
*** tsaop has joined #crux17:44
*** abenz_ has joined #crux17:47
*** onodera has quit IRC18:00
*** onodera has joined #crux18:01
*** onodera has quit IRC18:01
*** onodera has joined #crux18:02
crash_frinnst: haha stable on the video18:12
*** ivladak has quit IRC18:22
*** de11 has quit IRC19:32
*** j_v has joined #crux20:41
j_vlate last week, i tracked down why gdb build fails when guile-2.2 is installed. i found a simple work around. since we don't have an official guile port, does it make sense to make a flyspray task for that with the fix?20:44
j_vI could also email the maintainer; same basic question for that, is it worth it if there is no official guile port?21:08
pedjaj_v, is the workaround for patch or guile?21:12
pedjas/patch/gdb/21:12
j_vpedja, workaround is for gdb. it's a known gdb issue that has a patch in the devel branch that hasn't been backported: guile > 2.0 isn't supported. it's a simple sed that prevents configure from looking for guile-2.2 versions.21:14
pedjaj_v, I'd open a FS task for gdb with the patch, so even if the gdb maintainer doesn't use it, it can be a reference if someone needs it21:16
*** dlcusa has quit IRC21:16
j_vok, thanks for your input. much appreciated21:17
*** dlcusa has joined #crux21:19
pedjaACTION has the best ideas when slightly drunk, apparently21:24
*** tsaop has quit IRC21:39
*** pedja has quit IRC21:40
*** henesy has quit IRC21:42
*** henesy has joined #crux21:42
*** Lukc_ has joined #crux22:02
*** Lukc has quit IRC22:03
*** Lukc has joined #crux22:07
*** Lukc_ has quit IRC22:07
teK__john_cephalopoda: I'll try to dig up your mail later tonight22:23
john_cephalopodaThanks, teK__22:23
teK__btw, we were fast to rebuild stuff once the server broke, for a change But that was mostly frinnst's work  :-)22:23
jaegerspeaking of fast rebuilds, 28.78% done resilvering the ZFS pool :P22:29
jaeger58 hours to go!22:29
joacimraidz?22:30
jaegeryes22:31
teK__I am attending this Linux Performance Tuning Session  right now, and Ted Tso multiple times recommended not to use RAID most of the times ;>22:31
*** pedja has joined #crux22:45
*** leah2 has quit IRC22:47
jaegerEven for enterprise stuff?22:53
*** leah2 has joined #crux22:53
*** onodera has quit IRC22:56
jaegerunrelated, finally starting to secure erase all the spinning rust sitting in my closet to sell or donate22:56
teK__especially. He acknowledges though, that at Google, they simply use diffrent application+data center layouts vastly outperforming RAID in e.g. rebuild scenarios or rewriting of parity bytes in write loads22:57
jaegerWhat does he suggest to use in its place for people who aren't google? RAID1, for example, seems like a low-hanging fruit for simple failure avoidance22:57
jaegerObviously RAID is no backup22:58
*** pedja has quit IRC22:58
teK__no recommendation, just the warning that RAID is not a silver bullet22:58
jaegerFair enough. It's definitely not a silver bullet, just one piece of the puzzle22:59
teK__it kinda is the unfairish Google took care of that because they can ;-)22:59
jaegerhehe23:00
teK__fwiw: if someone happens to be in san francisco, I am available for $beer until friday local tim ;)23:08
joacimI'll fly right over23:08
joacimbe there in a few days23:08
teK__it IS a nice city from what I can see, so have fun23:09
joacimdoes it really smell like shit?23:09
jaegerI've been there for a few conferences but nothing this time of year23:10
frinnstjaeger did most of the work to be honest23:10
joacimi've heard people say that before23:10
jaegerSome parts of it smell bad but overall it's pretty cool23:10
teK__I do have this oversized US flag right in front of my hotel room ^;)23:11
jaegerhehe23:11
teK__so: business as usual23:11
*** pedja has joined #crux23:11
jaegerteK__: do you like sushi and chinese food?23:12
teK__kinda and def. yes23:12
jaegerside note, you're about 1h16m from most of my team (san jose)23:12
jaegerrecommendations: Ryoko's (sushi), House of Nanking (chinese)23:12
teK__it is bugging me that I do not really have time to roam the bay area23:13
jaegerdoh23:13
jaegergreat city to walk around23:13
teK__will check! Thanks a lot!23:13
jaegerI walked like 18 miles one day last time I was there :D23:13
jaegerNo problem23:13
teK__well I have friday afternoon but that is slated for rain D:23:13
teK__nice23:14
jaegerAnd the city isn't 18 miles wide/long23:14
teK__hehe23:14
teK__I will check out china town for sure23:14
pedjayay, total system lock23:14
pedjafailing hardware, most likely. oh, joy23:16
jaegeruh oh, no good23:16
pedjayeah. when I run tvtime, system works for a while, no sound or image, then locks up hard. no sshing to it, no nothing23:17
pedjaa few 'PCI parity error detected' in dmesg23:20
pedjaand now notebook won't come out of the blank screen...23:26
pedjais it 'every hardware I touch, it breaks' day already?23:31
jaeger:)23:39

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