IRC Logs for #crux-devel Monday, 2015-06-15

*** Feksclaus has quit IRC01:51
*** ___mavrick61 has quit IRC02:33
*** ___mavrick61 has joined #crux-devel02:35
*** jue has quit IRC07:22
teK___and backupexec was like: oops I think I ate that kernelbase.dll again *thihi*07:53
teK___+: I am not the windows admin08:01
teK___-: I was the only person "reachable" today at 7 am08:01
teK___but who needs sleep anyway :>08:02
teK___(if you can debug fucking horseshit^W^WWindows)08:02
frinnstno you cant08:02
frinnstyoure not supposed to08:02
teK___to what? Sleep?08:03
teK___I am watching the support comp. trying the exact same things I did 2 hours ago08:03
teK___great that our windows admin does not seem to be telling them :D08:04
*** maldoror has joined #crux-devel12:14
*** maldoror has quit IRC12:15
*** jue has joined #crux-devel14:29
frinnstthoughts on creating 3.2 branches?17:37
teK___but I will keep my mouth shut just a little longer :)17:40
frinnstfeel free to add to it17:41
frinnst*ahem* luks *ahem*17:41
juefrinnst: good idea ;)17:44
jaegerI need to add dosfstools to the ISO and replace grub2-efi with syslinux, too17:44
jue3.1 is over a year old, so a new version kinda overdue17:45
jaegerdefinitely agree on that17:45
jaegerDid any of you guys object to 3.2 being a transition release for /usr/man -> /usr/share/man?17:46
jaegerjue: unrelated to crux, got a cool bike carrier for my car:
juewow, looks great, I'm sure you have tested it above 100 mph :)17:49
jaegerI tested it at about 45 mph, only got it yesterday :)17:50
juewrt man pages, no objections from my side. mandb works with both locations so a transition is easy.17:51
jueteK___: please comment on the plan to remove tcp_wrappers17:53
teK___may I ask about the reasoning behind usr/man vs usr/share? LSB?19:10
teK___wtf freenode.19:10
teK___any way. I always felt tcpwrapper to be nice door sign asking intruders not to come in19:11
teK___comparing tcpwrappers  and gperf I'd weigh in in favour of gperf.19:12
jaegerhow are gperf and tcpwrappers related?19:12
teK___one gets in, one  is thrown out?19:13
jaegeras for the usr/share/man thing, it seems to be the default for most software now19:13
jaegerI don't really see any benefit to us continuing to override it in every Pkgfile19:13
teK___so we will switch to share for  core and opt, waiting for contrib to follow sllowly?19:14
teK___the argument against a symlink is the notorious /usr on a shared partition debate, right?19:14
jaegerwell, since mandb seems to support both, we could use the 3.2 release as a transition period like I mentioned earlier19:14
jaegertell people to switch to the default for 3.3+19:14
jaegera symlink doesn't help us, in my opinion19:15
jaegerthe point of switching to usr/share/man is to save work19:15
jaegeryes, it means a chunk of up-front work to revert to it19:15
jaegerin the future, though, we won't need to add --mandir=/usr/man to everything19:15
teK___once we "asked" everybody to remove the --mandir directive19:16
jaegerre: gperf, gperf is for making hashes, yes? tcpwrappers is for blocking access to services. am I missing something obvious here19:16
teK___the one is to be swapped out so I simply compared 'importance'19:16
teK___noting fancy or number-based.19:17
jaegerare they mutually exclusive for some reason?19:17
jaegersorry, just confused19:17
teK___no, not at all19:17
teK___this means we  will have man pages in two different  pplaces for quite some time19:17
teK___for the benefit of  removing a line from Pkgfiles. Hm...19:18
jaegerpossibly, yes. but if we tell people via the ML or IRC or whatever to switch, and make prtverify report /usr/man, that should help19:18
teK___I told mike_k to fix php-gd, which is in contrib and probably full of bugs atm19:18
teK___I doubt he will update it any time soon19:18
teK___lacking a rather grave update. More grave  than --mandir..19:19
jaegeralso consider jue's comment about 3.1 being more than 1year old :)19:19
jaeger3.2 -> 3.3 might be plenty of time, heh19:19
teK___how is that a good thing? ;)19:19
teK___21:17 < teK___> this means we  will have man pages in two different  pplaces for quite some time19:19
jaegerIt isn't, that was a joke19:19
jaegerI don't see a downside, personally19:20
teK___sorry, I am a bit slow today19:20
teK___yeah, it will not really hurt anyone..19:20
teK___(as you said, mandb supports it transparently)19:20
jaegerI don't even feel like it would be a requirement for releasing 3.219:20
jaegerwe could remove --mandir as ports are updated19:21
teK___I liked the clear ruling on that matter in our guidelines19:21
jaegerIt'd still be clear, I think. Just using /usr/share/man instead of /usr/man and it just happens to be often default :)19:21
teK___for the sake of simplicity, we could go for it, yeah. Still, I like /usr/man better because it's shorter, the indirection of  /usr/share brings almost no benefit, and Pkgfiles are really simple already19:23
teK___so ..let's do it19:23
jaegerIt's not so important to me that I want to fight over it, for what that's worth. :) Just one less thing to change from upstream was my thought process19:24
teK___well, if shorter  Pkgfiles are a target, we ought to do it19:24
teK___I feel exactly the same way, it's not worth a huge discussion19:24
teK___wrt tcpwrappers. Are  you a cups user?19:25
teK___# Only listen for connections from the local machine.19:25
teK___seems to be the default, right?19:25
jaegerI believe so19:25
teK___cups seems to be the only service one would shut off of the rest of the world by default19:26
teK___we are going to force users to install iptables even for the simplest restriction policy19:27
teK___basically, I'm ok with that.19:28
teK___in the same light, one could dispute the necessity of perl in core19:31
teK___I only had a quick glance but groff seems to be the only port that could require it seriously19:31
jaegerIf everything supported tcpwrappers I'd maybe appreciate it more... but iptables can handle pretty much any service regardless of whether or not it has support19:32
jaegerI'd also be fine with removing anything from core that doesn't need to be there. we also talked about zip and unzip in the past19:32
jaegerAFK a few minutes, gotta walk to another building19:32
teK___< walking to another computer :)19:32
jaegerthanks. fighting with our tape library still19:44
teK___the power thingy was not the cause of trouble?19:44
jaegerit still seems to be power-related, the drives don't power up19:44
jaegerbut they're stumped as to why19:44
jaegerthey sent me 2 new drives, I plugged them in today and nothing19:45
teK___at least they got nobody else to blame it on19:45
teK___how are backups/data access organized atm?19:46
jaegerwell, can't backup or access anything on tape right now, so yeah19:47
jaegerour network storage appliances keep daily snapshots, though19:47
teK___so, there's no manual workaround19:47
jaegernone currently, no19:47
teK___RAID aint no backup, sure :)19:47
jaegerMight be nice to have a machine that could hold one LTO-6 drive in case of emergency19:47
teK___we were told to make a vmware snapshot before the next full backup to disk is done on Friday. That's for an important mission ciritcal server m\19:48
teK___if that's sufficiently large enough?19:48
jaegerwe still technically have offsite replication, though that will go away soonish19:49
teK___sounds like luxury19:49
jaegerso the tape library being broken right now isn't a big deal. It will be once our offsite storage replication becomes onsite storage19:50
teK___dont tell the vendor ;)19:50
jaegerthey already know that but since none of their tier 1, tier 2, or engineers have seen this problem, they're interested in it19:51
jaegerI was on the phone with them thursday and the tier 2 guy had his whole team around his desk19:51
teK___I rethought the (g|)zip issue in core. If you remove these, you'd have to remove GNU tar if you were strict19:51
teK___sounds nice19:51
teK___we once had an escalation to some Fujitsu Escalation bla bla team wrt RAID-10 performance without any substantial result..19:52
jaegerthey're very helpful and friendly so far, I don't have any big complaints19:52
teK___they are as long as you have a valid support contract. :D19:54
jaegerof course. :)19:54
teK___We experienced this when the Lead Buyer (!) terminated things without telling us19:54
teK___they have never been as helpful as before that time again19:55
teK___they got 2000€ for installing two ESX hosts and four naked VMs + BackupExec for the final VMs20:00
teK___oh and a CISCO switch. ~ two days of work so they must not complain.20:01
jaegerpretty decent return for a couple days20:01
teK___indeed. But now they have to fix the file corruption issues we are seeing with 2k1220:01
jaegerfile integrity is overrated :)20:03
teK___but I digress. I say let's dump tcpwrappers20:03
teK___yeah especially for DLLs like kernelbase.dll20:03
teK___there is a proverb..20:04
teK___only a fool will take identical action twice and expect a different outcome20:04
teK___after they 'fixed' things today (just as last monday), IE crashed until we restarted the whole VM again. Just as it did last week.20:05
teK___I dare predicting, that the server will be screwed again next Monday20:05
jaegerwasn't that an albert einstein quote? "insanity: doing the same thing over and over again and expecting different results." :)20:05
jaegerI must admit I'm glad I don't do windows systems administration20:06
teK___yeah, quite possible20:06
jaegerthe tape library's front panel runs FreeDOS, that's close enough :D20:06
jaegerI did have to teach the windows guy about powershell's "while" loop feature this morning20:06
teK___I usually dont either. I did project management for this particular ERP introduction project20:06
teK___windows sucks as hell as soon as clicking some options does not suffice anymore. Albei the existance of powershell20:07
teK___PS is madness20:07
jaegerI actually love telling him "powershell can do X" because he's afraid of the command line20:07
jaegerright after I told him about while loops he found a different way to do what he needed that had a system tray icon20:07
teK___methadone for the heroin junkie!!120:07
teK___Romster: bumped llvm-32 just after I did it with llvm. Respect! :)20:13
jaegerhigh five!20:23
teK___I am scretely planning on keeping up with upstream updates until some time mid July20:23
teK___but don't tell anyone.20:23
jaegeryour secret is safe20:24
*** Feksclaus has joined #crux-devel22:39

Generated by 2.11.0 by Marius Gedminas - find it at!