IRC Logs for #crux Wednesday, 2015-11-18

*** SiFuh has quit IRC00:22
*** BitPuffin|osx has joined #crux00:29
*** xeirrr has quit IRC00:36
cruxbot[opt.git/3.1]: wine: 1.7.54 -> 1.7.5500:47
*** |narcos| has quit IRC01:24
cruxbot[compat-32.git/3.1]: gdk-pixbuf-32: 2.32.1 -> 2.32.201:40
cruxbot[compat-32.git/3.1]: glib-32: 2.46.0 -> 2.46.201:40
cruxbot[compat-32.git/3.1]: libpng-32: 1.6.18 -> 1.6.1901:40
*** sdfgsdfg has joined #crux01:48
*** ivs has joined #crux02:09
*** BitPuffin|osx has quit IRC02:51
*** ___mavric61 has quit IRC03:27
*** ___mavric61 has joined #crux03:29
*** user_ has joined #crux03:40
*** SiFuh has joined #crux03:41
*** ivs has quit IRC03:42
*** xor29ah has quit IRC03:45
*** user_ has quit IRC04:01
*** SiFuh has quit IRC04:02
*** Workster has quit IRC04:07
*** fktkfidbjmhautpa has joined #crux04:07
*** Workster has joined #crux04:07
*** bfxzzvxshgtijrdi has joined #crux04:11
*** Workster has quit IRC04:11
*** Workster has joined #crux04:11
*** Workster has quit IRC04:12
*** rgrxoursmswfnuwl has joined #crux04:12
*** Workster has joined #crux04:12
*** Ghost_r00t has quit IRC04:17
*** xor29ah has joined #crux04:25
*** SiFuh has joined #crux04:25
*** ivs has joined #crux04:42
*** tilman_ has joined #crux05:22
*** tilman has quit IRC05:24
*** sudobaal has joined #crux06:18
*** sq271 has joined #crux06:29
*** Romster has quit IRC07:53
*** pejman has quit IRC07:58
*** pejman has joined #crux08:33
*** pejman has quit IRC08:33
*** pejman has joined #crux08:33
*** sq271 has quit IRC08:41
*** sq271 has joined #crux08:53
*** Romster has joined #crux08:57
*** matcz has quit IRC09:08
frinnst <- that probably makes me immortal10:25
Romsterbrewing coffee right now10:36
*** nwe has joined #crux10:50
*** Ghost_r00t has joined #crux11:00
Ghost_r00thello crux people. ;)11:00
*** nwe has quit IRC11:01
*** nwe has joined #crux11:03
Ghost_r00tin a normal desktop do we need cache and/or general file system local cache?11:03
*** |narcos| has joined #crux11:16
RomsterGhost_r00t, for what program?11:24
Ghost_r00tRomster: I am asking the same. I want to enable that in the kernel OR not.11:29
Romsterwhich option exactly?11:30
Romstereep the lighening is getting closer.11:31
Romster A community collaborative real-time lightning location network11:46
Ghost_r00tfile system > cache > general file system local caching manager11:50
Romsternever used it11:51
Romsterwhat is this some other bcache?11:51
Romstera generic filesystem caching manager that can be used by various network and other filesystems to cache data locally.11:52
Romsterso unless you use nfs or samba or something, you don't need it.11:53
*** xeirrr_ has joined #crux12:10
Ghost_r00twhat do they have over FSCACHE? (samba nfs ...)12:15
Ghost_r00twhat does floating point do in modern CPUS?12:16
rmullGhost_r00t: It does a lot of things. Can you rephrase your question?12:20
Ghost_r00trmull: would you like me to paste the exact text that I have problem with?12:22
*** ziggypup has joined #crux12:22
*** ziggypup has left #crux ()12:22
Ghost_r00tThe pipeline units in vector computers are not integrated floating point uints but attached vector units to a CPU that itself has a floating point unit12:24
Ghost_r00twhat is the floating point units job?12:25
rmullthe FPU performs computation of IEEE754 floating point numbers, whereas the rest of the CPU's arithmetic logic unit can only compute things that are whole integers12:26
rmullIn processors without an ALU, floating point math can still be done, but it's a lot slower12:28
rmull*without a FPU12:28
rmullAs far as your problem sentence - it's a little weird. Do you have a link for context?12:29
rmullWikipedia explains vector computing reasonably well:
rmullYour sentence seems like it was written by a non-native english speaker or something12:32
Ghost_r00tit is better now.12:38
Ghost_r00tbut what difference does it make that these FPU units have register or direct-to-memory arch?12:39
*** john_cephalopoda has joined #crux12:40
*** BitPuffin has joined #crux12:42
rmullI would suppose those terms are describing how the CPU exchanges data with the FPU. In x86 for example there are dedicated floating point registers that can be maniuplated with a specific set of assembly instructions. Registers are the lowest-level fastest "working" memory of a CPU - in order to perform any computation the CPU typically has to load data from elsewhere into its registers, do the math there, and load the data back to cache/ram.12:45
rmull"Direct-to-memory" isn't a term I'm familiar with12:45
rmullPossibly an external memory-mapped bus?12:45
rmullThat doesn't seem likely though12:45
Romsteryour talking about DMA12:50
Romsterdirect memory access12:50
Romsterwhere different devices can talk directly to memory and not hog bandwidth on the cpu12:51
*** shaggy814 has joined #crux12:59
*** xeirrr_ has quit IRC12:59
rmullDMA still hogs bandwidth, though, because the data still goes on the bus. It just cuts down on interrupt overhead.13:02
rmullexternal DMA might not go on the bus, but that's probably not what's being used in the case of a FPU since it's used for moving data between two external memory-mapped locations13:03
rmullgotta run13:05
*** xeirrr has joined #crux13:05
frinnstpro tip: dont install a syslog config that forwards everything to the syslog server - on the syslog server13:17
frinnstloops are fun13:18
frinnstEvent: 'threshold of high load exceeded: current value 12.32' <- on lo :D13:21
shaggy814This may be a stupid question but I am trying to write a pkgfile for a package that uses capital letters in the url and package name. Is there any reason to not name the package with caps as well?13:23
frinnstwe try to use non-caps for package names/ports13:26
frinnstI dont think anything uses caps13:26
shaggy814hmm ok, do you happen to know how to strip case of the downloaded package in the pkgfile?13:27
shaggy814maybe rename?13:28
frinnstdont strip anything. just dont use caps in name=13:28
frinnstcheck contrib/virtualbox for example13:29
shaggy814ok I will take a look there, thanks :D13:29
shaggy814I cant believe how easy that was. Thanks for the help13:34
*** xeirrr has quit IRC13:40
rmullI haven't noticed anyone ever complaining too much about hardcoded URL elements13:45
*** g0relike has quit IRC14:00
Romsterjust try not to hard code the version if you can help it.14:01
Romsterthat's the only bit that is annoying if it is.14:01
Romsteraka boost doesn't hardcode it.14:02
*** tierd891 has joined #crux14:06
shaggy814Are there any repos that use git? or is that frowned upon?14:10
*** vlnx has quit IRC14:10
*** tired890 has quit IRC14:10
frinnstshaggy814: its not as efficient as, say, rsync14:13
frinnstbut easier to set up these days14:13
frinnstbut there seems nobody has bothered to document it on the wiki14:15
rmullshaggy814: You can use all the crux repos via git if you want14:15
frinnst(but please dont)14:15
rmullfrinnst: Why not?14:15
frinnstbecause it will download *everything*14:15
frinnst--depth=1 or whatever in the port would be useful14:16
rmullMy /usr/ports/core is 17M14:16
rmull(via git). How big is it without git?14:16
rmullcontrib -> 65M14:17
frinnstcontrib is 19mb14:17
rmullSo if the difference is just a few MB, is it really that bad?14:18
rmullIn terms of server CPU, maybe?14:18
frinnstyeah its pretty cpu-expensive. but as long as not everybody uses git it should be fine14:19
rmullIs git more CPU-expensive than rsync?14:19
frinnstyes, a lot more14:19
frinnston the client side too14:19
rmullgit is definitely a lot faster to sync14:19
rmullThis suggests that there is significantly less CPU time for git as well
frinnsthm, maybe its something with our git version then, or something14:21
rmullTo clarify, by "sync" I meant a ports -u on a mostly-up-to-date ports tree14:21
rmullOh - you've taken measurements?14:21
rmullI have not, but maybe it's worth doing14:22
frinnstnot in a scientific way or anything. just counted 1001, 1002 etc :)14:22
frinnstUpdating collection core14:23
frinnstFinished successfully14:23
frinnstthats with rsync14:23
rmull0.459 with git14:23
rmull(was already up to date)14:23
frinnstyeah, same here14:23
*** xeirrr has joined #crux14:24
rmullsys time was 0.035s14:25
frinnstsys0m0.011s here14:25
rmullRsync wins there14:26
rmullAny idea how the server side could be measured?14:26
rmulloh, and user: 0.026s14:26
frinnsti'll look at it when I get home, dont have ssh access from work14:28
frinnst(unless I forget)14:28
rmullLooks like CPU time is pretty equivalent at least14:31
rmullon the client14:31
*** xeirrr has quit IRC14:35
frinnstRomster: do you play minecraft?14:35
frinnstI know you use docker so this might be something for you!:
Romsternah i haven't frinnst14:36
*** xeirrr has joined #crux14:36
Romsteri've seen that for docker though already.14:37
Romstermight be a game i might get, been watching it on twitch.14:37
frinnst not without issues :)14:38
RomsterJust don't clean up after yourself for a month or two.14:39
Romsteri mightnto hit that issue with my thin-provisioning snapshots in lvm214:40
Romsterso i don't have a overhead in docker doing the work for me, instead it's moved to dm-mapper in the kernel14:41
Romsterfunny comment though14:45
john_cephalopodacontrib youtube-dl is outdated. Latest version is 2015.11.15 - see
Romsteri'll fix it after i wake up14:55
*** sdfgsdfg has quit IRC14:56
frinnstthis reminded me of my customers:
*** ivs has quit IRC16:15
*** ivs has joined #crux16:41
*** g0relike has joined #crux16:43
cruxbot[opt.git/3.1]: [notify] poppler: update to 0.38.017:04
cruxbot[opt.git/3.1]: poppler-glib: update to 0.38.017:04
cruxbot[opt.git/3.1]: poppler-qt4: update to 0.38.017:04
*** sudobaal has quit IRC17:05
cruxbot[opt.git/3.1]: nano: update to 2.4.317:10
Ghost_r00trmull: about thoses issue -"your last 3 lines answering my question"- do you have any text or website in mind for further reading. I have to cement my base so to speak.17:30
*** xeirrr has quit IRC17:37
rmullGhost_r00t: Hmm, not sure about a single source, but you could try an online computer architecture course:
rmullNot sure how much time you want to put into it if you just have a few specific questionts17:42
Ghost_r00trmull: all in;17:43
rmullI have a computer engineering degree fwiw17:43
rmullmost of my practical experience is with embedded stuff though17:43
Ghost_r00trmull: anything that would help is appreciated. even keyword for search. pdf chm IDN even TXT17:44
rmullI would start by looking for resources on computer architecture/organization, and drill down depending on which aspects interest you17:45
*** vlnx has joined #crux17:49
Ghost_r00trmull: any starting point for that? any specific web sites besides that one?18:37
*** tierd982 has joined #crux18:52
rmullThis is kind of neat - old (1948) computer that only had 8 instructions consisting of subtraction, negation, and jumping.
*** tierd891 has quit IRC18:55
*** john_cephalopoda has quit IRC19:05
cruxbot[core.git/3.1]: btrfs-progs: updated to 4.3.119:06
*** hhhhhhhh has joined #crux19:17
*** |narcos| has quit IRC19:34
*** kyubiko has left #crux ("WeeChat 1.3")19:37
cruxbot[core.git/3.1]: wget: update to 1.1719:39
cruxbot[core.git/3.1]: kmod: update to 2219:39
cruxbot[opt.git/3.1]: git: footprint update for perl 5.2219:42
cruxbot[opt.git/3.1]: imagemagick: footprint update for perl 5.2219:42
cruxbot[opt.git/3.1]: irssi: footprint update for perl 5.2219:42
cruxbot[opt.git/3.1]: p5-xml-parser: footprint update for perl 5.2219:42
cruxbot[opt.git/3.1]: mysql: disable tcp_wrappers19:42
*** cruxbot has quit IRC19:42
rmullSo many updates!19:44
*** BitPuffin has quit IRC19:47
*** cruxbot has joined #crux19:50
tierd982exciting times!19:57
*** Stealth has quit IRC20:00
*** kori has quit IRC20:01
Ghost_r00trmull: yes but it is the second chapter of the book I am reading and it is directly related to pipelining vector processor superclusters20:04
juethe last messages from cruxbot for opt/3.1 are wrong, the mails as well, sorry for that20:06
*** Stealth has joined #crux20:07
*** ivs has quit IRC20:16
*** kori has joined #crux20:17
Wildefyrer is there any way to make prt-get update stop if it fails on upgrading one of the packages given to it?20:30
jueprt-get grpinst20:39
juesorry, that only works for fresh installs and not for updates20:45
rmullp5-xml-parser footprint is all wacky20:52
shaggy814is there a way to generate footprints for packages?20:58
Wildefyryeah just had to update it myself20:58
Wildefyrshaggy814, -uf20:58
shaggy814prt-get -uf?20:59
Wildefyrprt-get <option> -uf20:59
Wildefyrso normally update21:00
shaggy814is there a way to do it without installing the package? I want to include a footprint in my repo but I dont necessarily want to install everything haha21:00
Wildefyrpkgmk -uf21:00
shaggy814ahhhh thanks man21:01
Wildefyryou still have to make the package21:01
shaggy814thats what im looking for21:01
shaggy814i dont mind building it21:01
*** john_cephalopoda has joined #crux21:40
frinnstopt is currently a bit messed up at the moment21:42
frinnstfor 3.1.. we are working on it :)21:42
*** tierd982 has quit IRC21:48
*** casucci has joined #crux21:49
casucciHello everyone! New Crux user with two questions: how to generate /etc/ssl/certs/ca-certificates.crt? and Why can't I connect wifi with wpa_supplicant/dhcpcd on boot?21:50
casucciMore info: I have openssl and ca-certificates installed and up to date. I can connect manually by using ip link set wlp3s0 up, wpa_supplicant -B -i wlp3s0 -c /etc/wifi.conf, dhcpcd wlp3s0. When put these commands in /etc/rc.d/net they fail.21:57
*** hhhhhhhh has quit IRC22:05
*** casucci has quit IRC22:16
*** casucci has joined #crux22:17
*** casucci has quit IRC22:48
cruxbot[opt.git/3.1]: wine: 1.7.54 -> 1.7.5523:16
cruxbot[opt.git/3.1]: [notify] poppler: update to 0.38.023:16
cruxbot[opt.git/3.1]: poppler-glib: update to 0.38.023:16
cruxbot[opt.git/3.1]: poppler-qt4: update to 0.38.023:16
cruxbot[opt.git/3.1]: nano: update to 2.4.323:16
cruxbot[opt.git/3.1]: gtk3: updated to 3.18.523:16
*** casucci has joined #crux23:34
casucciHello Everyone!23:35

Generated by 2.11.0 by Marius Gedminas - find it at!