IRC Logs for #crux Wednesday, 2015-01-28

diverseleo-unglaub: too bad he argues with you00:00
hnode_mhe: none, only the one I have created now: 20-intel.conf00:00
hnode_with content copy from
mheyes, good idea00:01
hnode_should I use glamor or uxa00:01
mhetry disabling the accel-stuff00:02
mheand are you sure you have intel and not fbdev? On my eee-pc I need fbdev00:04
*** leo-unglaub has quit IRC00:04
greduanAnyone here use vimb as their browser?00:06
mhedamn, it is getting too late, going to bed here00:09
greduangood night00:09
tired890nighty night00:11
Workstergreduan, either with sudo or polkit00:16
*** paul772 has quit IRC00:16
greduanWorkster: I tried sudo but it doesn't work, with su it does though00:16
Worksterxfce4 uses polkit00:16
greduanI'll try with polkit once I'm a little bit more setup, it's not that big a deal atm00:17
Worksterleo ...ah he is gone was gonna ask what file to patch00:17
Worksterpersonally i just do sudo /sbin/shutdown ...00:18
hnode_mhe: good night and thanks00:18
Workstereither -r now or -h now, but it's just my desktop no other users so now is ok there00:18
greduanI see00:18
Worksterxfce4-power-manager probably need that00:19
greduanIt seems I also don't have access to wpa_supplicant or dhcpcd00:19
diversepolkit is just unnecessary abstraction mess from redrat, you would think it's cousin, systemd, would learn its lesson, but nooooooo...00:20
greduanas a non-root user00:20
greduanI still don't quite understand what all the fuss is about when people talk about whether to use systemd or not00:21
*** jdolan has quit IRC00:22
diversegreduan: if it doesn't bother you as a user, don't worry about it00:22
greduanwon't worry bout it then00:22
greduandamn this this keyboard, it's so hard to type with a keyboard that sometimes doesn't register keys00:23
diversethere are specific reasons, but it depends what type of user you are00:23
greduanI assume for a non-techy user it doesn't solve any particular problem?00:25
Workstersytemd to a power user is a lot of hate00:25
greduanI do know it's WAY more complex than what CRUX has, for example. Which is BSD style init, ja?00:25
Worksterend uses couldn't care less over systemd00:26
diversesomeone that just wants something to work out of the box without thinking about how the system works, would not care about systemd00:27
greduanThey would probably prefer it since it abstracts a lot of stuff, no?00:27
prologicdidn't know we had a tag on unix & linux stach exchange00:28
diverseprologic: that must of been the work of kori00:28
prologicperhaps :)00:28
prologicI edited it's Tag Wiki00:28
prologicI don't mind00:29
prologicI prefer StackExchange as a forum compared to mailing lists and forums00:29
greduanI think I'm gonna start hanging around there, see if I can help someone00:29
prologicwe should put a link to the above uri on's wiki somewhere00:29
diversegreduan: do it because you want to, you have no obligation.00:29
prologicor somewhere more obvious00:29
prologicStackExchange is a good tool IHMO00:30
greduandiverse: Of course. :)00:30
*** jdolan has joined #crux00:30
greduanCompilin webkit, expecting this to take a _long_ while00:31
korigreduan: nope00:32
koriit won't take a long while00:32
koriit'll take a long long while00:32
diversekori: was my assumption correct? ;)00:33
koriabout what?00:33
diversekori: look at my last highlight00:33
kori21:28 diverse: prologic: that must of been the work of kori00:34
korithis, yes?00:34
koriaye, that was not my work, I don't use stackexchange00:34
korialso, I didn't use linux in 201100:34
prologicdoesn't matter who started the tag00:35
diverseI was just curious :)00:35
koribut I'd totally start some form of community00:35
*** hnode_ has quit IRC00:37
koriI love CRUX so much00:37
Worksterprologic, i can add stackexchange to links on wiki when i get home00:37
Worksterwont be offical though just other crux stuff links00:38
koriI'd like 6c37 to be acknowledged by the general CRUX community one day00:38
koriconsidering our CRUX "division" is fairly healthy :P00:38
*** hnode has joined #crux00:38
prologicWorkster: great :) be sure to puyt it in an obvious place or two00:39
prologicdon't make it hard to find :)00:39
diverseprologic: it's his github group00:39
Worksterkori, its on the portsdb page now00:40
*** jdolan has quit IRC00:40
greduanI'm guessing 6c37 is Japanese?00:40
Worksteryou'd need a REPO file for httpup.00:40
diversegreduan: it's unicode for a Japanese kanji character00:40
Worksteri know you got a git driver00:40
diversegreduan: take a guess00:40
greduanwell I guessed it's Unicode, I meant the group itself. lol00:41
rmullsyncing all crux ports with git is a no-brainer00:41
greduanI really like how Japanese use way less avatars with real photos.00:41
greduanrmull: agreed00:41
diversegreduan: it's the hexademical number of unicode number it represents00:41
tired890wish my laptop had an atheros WLAN.. those broadcoms are a nightmare00:42
tired890I compiled the kernel with b43 driver.. first it loads then it says not supported sigh00:42
diverseokay it's a code point00:43
greduantired890: I've got one of those00:43
tired890greduan, and? got it working?00:43
greduanYou still need to install the firmware blob with fwcutter00:43
greduanI use it so often I have a file with the necessary firmware, want a copy?00:43
greduanfolder, not file00:44
greduanAnd it comes with a script in case you don't trust me. ;)00:44
greduanIf you do want a copy, where should I upload it?00:45
tired890greduan, fwcutter?00:45
tired890greduan, an alternative would be to use the broadcom-sta driver00:46
tired890most people seem to recommend this over b4300:46
greduanI see00:46
greduanI wonder what the difference is...00:47
greduanStill, I don't mind the blob.00:48
tired890this is a full fledged prop driver00:48
tired890not just a FW for an open source alt00:48
tired890but it seems better performance/consistency is expected from this as opposed to b4300:49
greduanI'll be looking into that00:49
tired890how is your wireless anyways? stable and fast?00:49
greduanStable yes, as fast as my internet speed at least (2MB/s)00:49
greduanThat's the fastest speed I've tried it at00:50
tired890that sounds good enough00:50
greduanSometimes, like in some edge cases so rare I forget what they are, I do have to reboot for the wifi to start working again.00:51
tired890can you share your exact card model ?00:51
tired890dmesg || grep b4300:51
tired890dmesg | grep b4300:51
greduanBroadcom 4312 (core revision 15)00:52
greduanI've a Dell Vostro 1500, two of them actually.00:53
greduanThis wifi card is the only thing I don't like about these laptops00:53
tired890I see00:54
greduankori: webkit is taking a long long while as you said. Already taking more than compiling my kernel00:55
korigreduan: it takes longer than firefox00:56
koriat least 2x as much00:56
koriso don't sit around waiting00:56
greduanYeah I'm gonna leave this compiling into the night00:57
korialso, that glibc vuln00:57
diversekori: we're not affected00:57
koridiverse: still00:58
korithe fact it exists is dumb00:58
greduankori: Thankfully I'm running it in a terminal in DWM, so I can do other stuff meanwhile.00:58
diversethe fact that other distros still use the old glibc without the fix is dumb00:58
greduanI would switch to musl if some things didn't break becauses of it00:59
diversemusl != a GNU libc replacement01:00
koridiverse: see the amount of 'D' and 'd' in that smiley01:00
koriit was jokes01:00
greduanIf one develops with musl in mind then it's automatically compatible with libc, no?01:00
koriI dislike GNU a lot01:01
diversekori: use more 'P' and 'p' next time :P01:01
greduanI was about to say webkit finished compiling. But no, a download failed so it stopped. Dangit!01:02
greduanjaeger: Are binary packages something being considered for CRUX?01:03
greduanOptional, of course01:03
*** tilman has quit IRC01:03
prologicgreduan I think so01:03
prologicmusl is POSIX compliant afaik01:03
prologicso it should run fine on libc too01:03
prologicassuming the libraries are both installed the same way I guess :)01:04
greduanYeah, AFAIK the reason it breaks some libc stuff is because libc is NOT 100% POSIX compliant, so some features some programs use aren't supposed to be there, so they aren't in musl01:04
prologicgreduan> jaeger: Are binary packages something being considered for CRUX? <-- several of us are trying to work on this -- it's not trivial to a) implement and b) support01:04
prologicit requires resources and many power which we simply don't have01:05
korican we say glibc when we're talking about GNU libc01:05
prologicso if we want it (even if optional) it has to be quite automated IHMO01:05
*** tilman has joined #crux01:05
greduanI see01:05
greduanSome server somewhere compiling stuff automatically and stuff?01:05
prologicit's not as simple as that :)01:05
prologicbut for your own purposes that would work fine01:05
prologicas long as your machines are all the same01:06
prologicthat's why we have pkg-get01:06
*** jdolan has joined #crux01:06
prologicso if you have 10s of vultr/hetzner/linode vms for instance go the pkg-get route01:06
koriI still have the 3char-get trauma01:07
*** jdolan has quit IRC01:07
koridang apt-get removing the entirety of GNOME when I just wanted to remove cheese :<01:07
*** jdolan has joined #crux01:07
greduanprologic: I do not. lol, I wish I did.01:08
greduanWho is currently working on an infrastructure for binary packages?01:08
prologicon-one technically01:09
prologicwe're toying with the idea01:09
prologicat least I am01:09
prologicWorkster/Romster is interestd as well01:09
prologicas I said, it's not trivial :)01:09
greduanI've a friend who has like $1000 or $100,000 in DigitalOcean credits, cause one of the owners was nice and gave it to him. He said he'll donate servers to non-profits01:09
prologicat best it would support only one architecture and default builds01:09
greduanMaybe I can approach him about this?01:10
prologicyou'd still have to recompile if you changed your system in wierd ways01:10
prologicwow that's mighty nice of the DO co01:10
prologicI'd happily accept some credits for crux use01:10
greduanIndeed, he was his whatever thousandth Twitter follower01:11
prologicbut alas DO doesn't support custom ISO(s) or Image(s)01:11
prologicso running CRUX on it is quite hard01:11
prologicI'm probably going to close my DO account at some point myself01:11
prologicin favour of vultr.com01:11
greduanWould CRUX be required for this infrastructure?01:11
greduanI mean of course I can see why you would want it01:11
prologicprobably not01:12
prologicit could be done in Docker containers quite easily01:12
prologicso dunno :)01:12
prologicyou'd still want some part of CRUX though01:12
prologiceven if it's just a container doing ghte builds01:12
greduanI see01:13
greduanFound the tweet
greduan.@ahdinosaur Hey Thanks for being my 20,000th follower. I've gone a head and given you a dollar in DO credit from everyone who follows me.01:17
hnodeone thing that makes me choice Crux was becouse is source based,01:18
greduanhnode Indeed, me as well01:18
greduanBut for some stuff it is nice to have a precompiled binary, for example Webkit, which simply takes too long to be compiling yourself01:18
hnodeyes :)01:19
greduanAnd this is my friend's tweet regarding not-for-profits
greduanAnyone have a precompiled webkit you can share? lol01:19
greduanprologic: I posted the tweets regarding DO credits01:20
prologicI'll consider his offer :)01:24
jaegergreduan: not officially but as mentioned some users keep binary packages around. I'm not one of those, generally :)01:24
prologicmaybe it would help run some infrastructure up on DO01:24
prologicI keep binary packages around in Gopherspace :)01:24
greduanMaybe we don't have to have an automatic build thing, just, users upload binaries they've taken the time to compile (if they want) for other users to save some time.01:25
prologicthat could work01:27
prologicwe could also collectively do this01:27
prologicwhere each crux user participcates a Tahoe-LAFS node01:27
prologicand perhaps one of us devs/mcintainers (maybe me) could run the intorducer node01:28
prologiccall it a "distributed binary package cdn"01:28
greduanThat would be cool01:28
greduanAlthough that's the first time I've heard of that kind of node01:28
prologicand users that participate just point their "package dir" to the mount point01:28
prologicyeah look it up01:29
prologicit's quite nice01:29
Romstergreduan, PKGMK_SOURCE_MIRRORS=(
diverseRomster: he wants to pull prebuilt packages instead01:29
prologicwe're talking binaries :)01:29
greduanNot for everything, but some things would benefit from this for sure.01:30
greduanI say we set this up01:30
diverseRomster: do you have a prebuilt webkit package?01:31
greduanOnce I'm more accustomed to CRUX, you guys know me better and I learn me some C, I'd really like to contribute to CRUX's development / community.01:31
Romstergreduan, i manaually compile in docker (previously chroots) and upload to
greduanis z3bra on often here? He's the one who introduced me to this distro through his blog posts01:32
greduanRomster: awesome!01:32
Romstergreduan, said about a webkit build failure due to a download failure. i suggested using the crux mirror.01:33
greduanOh I see01:34
greduanIt's gtk3 that can't download01:34
Romsterin /etc/pkgmk.conf01:34
greduanWill do01:34
greduanAll right, did that01:35
greduanLet's see if that works...01:35
greduanWell it downloaded now01:36
greduanIt's running into some trouble with dependencies though...01:36
greduanThanks for the mirror advice Romster01:37
diversegreduan: z3bra hangs here all the time :)01:37
greduanCool. :)01:38
greduanCan I just say, this IRC is so much nicer than the #suckless one01:39
greduanRomster does the webkitgtk#2.6.0-1.pkg.tar.xz file contain its dependencies as well? I imagine it doesn't but I just want to make sure...01:40
diverseif you already pulled all of them before building webkit, you should be fine01:40
greduanWell I'm running into some trouble with some gtk3 dependencies.01:41
greduanI'll try to post a full log01:41
greduanAnyone know a paste service that supports the CLI?01:42 is my preference01:42
jaegeruses curl01:42
diversegreduan: you can use the wgetpaste script01:42
Romstergreduan, no just that port itself01:44
Romsterprt-get depinst webkitgtk --ignore=webkitgtk01:44
Romsterthen wget that and pkgadd webkitgtk#2.6.0-1.pkg.tar.xz01:44
greduanThanks. :)01:44
diverseI'm sure you have them already pulled, but do that just in case ;)01:45
Romsterit wont hurt if there installed or not01:45
greduank :)01:45
Romsteri just package the huge long time to compile ports manually01:46
greduanYou mean you package what you already compiled yourself?01:46
diversejust the big stuff01:46
greduanI see01:46
diverseas convenience pkgs01:47
Romsterjust those few.01:47
Romsterbecause i can do make -j20 here on distcc01:47
Romsterand build it in like 45 minutes01:47
greduanOh god01:47
greduanIf it's 45 mins, I wonder how many hours it would take on one or two cores01:48
Romsterwhere it could take 3 hours for some if there got old hardware01:48
diverseRomster has a nice small server cluster which he can use to build stuff in parallel01:48
greduanI've an Intel Core 2 Duo01:48
Romstergreduan, what cpu?01:48
Romsterouch probably a few hours01:48
greduan2.0GHz each core I believe, two cores01:48
Romsteri have 4 of them in my distcc cluster lol01:49
greduanI see01:49
Romsterone is 2.6Ghz though the rest are 2Ghz01:49
greduanI see01:49
Romsterold hardware i collected01:49
greduanI see01:49
greduanOld hardware to the rescue. lol01:49
greduanso I'm running `sudo prt-get depinst webkitgtk --ignore=webkitgtk > log.txt` but that doesn't include all the log.01:50
Romsterdo have 3 quad cores too so they probably do the bulk of the work01:50
greduanI see01:50
Romsteryou can turn on logging in /etc/prt-get.conf01:50
Romsterand your not redirecting stderr &2> or something i forget now redirect stderr to stdout01:50
Romsterbut don't do that use logging in prt.get.conf01:51
Romsterand logs are in /var/log/pkgbuild/01:51
Romsterwhen its on01:51
greduangot it01:53
greduanhttp not https01:54
greduan fixed link01:55
greduanjaeger: is pretty sweet, thanks for the link. :)01:58
Romstercontrib/wgetpaste is also handy01:59
greduanI have the pkgmk user configured01:59
greduanI mention it since that seems to be part of the problem01:59
Romstergreduan, run this prt-get sysup02:00
greduanwith sudo?02:00
greduanOut of curiosity, what does that do?02:01
Romsterthen after do : prt-get depinst pkgutils ; su to root... prt-get update -fr `revdep` ; exit02:01
Romsterahnot pkguitls... i meant prt-utils02:02
Romsterprt-utils has a revdep command.02:02
Romsterafter that run sudo rejmerge02:02
Romstercarefuyl go thoug your configurations edit merge keep skip... as you see fit. then you should be right to get gtk3 after.02:02
Romsterthere the most used commands for updating.02:03
greduanI see02:03
Romsterrevdep checks for broken libraries after major prt-get sysup02:03
Romsterrevdep takes time few minutes.02:03
cruxbot[opt.git/3.1]: samba: removed cruft02:03
cruxbot[opt.git/3.1]: ldb: fixed man path for systems with docbook-xsl02:03
cruxbot[opt.git/3.1]: ntdb: fixed man path for systems with docbook-xsl02:03
Romsterthat was fast.02:04
greduanI'll be going through this. Thanks! :)02:04
Romsterafk work02:06
greduanwouldn't sysup update my Firefox which AFAIK takes a lot to compile?02:07
greduanJust gonna add an --ignore=firefox02:07
diversegoing to get some sleep, ttyl02:14
greduangood night02:15
*** hnode_ has joined #crux02:15
*** BitPuffin has joined #crux02:15
*** pekka10 has joined #crux02:19
*** hnode has quit IRC02:19
greduanOh yeah, something that I'd like to know is, how often is a new CRUX release released?02:31
greduanIt seems it's a yearly thing roughly02:31
greduanI imagine it's something like a release when ready thing?02:32
greduanProbably should mention you jaeger.02:33
korigreduan: pretty much yeah02:34
koriI think a new release comes out when there are sufficient breaking changes02:34
greduanI see02:35
greduanso basically whenever you feel like a new release is necessary to make a new install not a pain to catchup with the changes?02:35
korithat might be it02:36
koriI installed @ 3.1 :D02:36
greduanI installed today, so same here. :P02:37
*** _root_ has quit IRC02:55
*** _root_ has joined #crux02:55
*** alancio has joined #crux02:56
*** BitPuffin has quit IRC03:13
*** drijen has quit IRC03:22
*** hnode_ has quit IRC03:26
*** drijen has joined #crux03:35
greduanI'll be going to bed, see you tomorrow everybody03:36
*** greduan has quit IRC03:40
*** mavrick61 has quit IRC03:41
*** tierd891 has joined #crux03:42
*** mavrick61 has joined #crux03:42
*** tired890 has quit IRC03:44
*** tierd982 has joined #crux04:05
*** tierd891 has quit IRC04:08
*** fengshaun has joined #crux05:05
*** SiFuh has quit IRC05:28
*** SiFuh has joined #crux05:30
nwegood morning07:38
*** Feksclaus has joined #crux07:44
*** pal_bth has joined #crux07:57
tierd982morning nwe08:27
tierd982any of you guys got cpu scaling (cpufreqscaling) setup correctly and working?08:32
tierd982or ya'll running at full Mhz all the time? :)08:32
Romstermorning nwe tierd98208:38
tierd982morning Romster08:38
nwetierd982: I dont using cpufreqscaling, I used it before like 5 years ago :P08:38
Romstertierd982, i just let the kernel do ondemand most of the time my cpu runs at 800Mhz08:38
nwegood morning Romster how are you ?08:38
Romsteri'm good after cooking a really nice spaghetti bolognese08:40
tierd982Romster, let me get this right pls, so you only go to :08:40
Romstermy cpu has 4 power states from 800mhz 2.2ghz 2.7ghz and 3.4 ghz ondemand seems to offer little latency as it switches states.08:41
tierd982Kernel (menuconfig) > Power Management and ACPI options > CPU Freq. Scaling >08:41
Romsterand set it to ondemand08:41
Romsteras default08:41
tierd982did that08:42
tierd982what about "scaling drivers"08:42
frinnstwhat cpu do you have?08:42
tierd982AMD FX835008:42
Romsterphenon II 965 black edition08:42
Romsteri know right.08:42
tierd982single core performance is same I'd say tho08:43
Romsteri just ordered a gtx 750 ti 2GB black editon graphics card to replace my aging old gts25008:43
tierd982all black edition hardwarez08:43
Romstersinge core on this sucks at 800mhz i did that before turn off 3 cores and stuck it on 800mhz08:44
tierd982Romster, what about the drivers mate08:45
tierd982right now none are ticked for inclusion..08:45
Romsternvidia binary blob08:45
tierd982I assume none is needed?08:45
Romsteroh yu just set the one ondemand it'll build it in.08:46
Romsteri don't see a point in making that a module. but you can if you like.08:46
frinnstno, you need an actual driver to control the hardware too08:47
tierd982thats what I thought08:47
frinnstI have a amd cpu In my server, let me check what I use08:47
Romsteri set mine up long ago...08:47
tierd982thanks guys08:47
Romstermotherboard chipset driver08:47
frinnst ACPI Processor P-States drive08:47
frinnstyou have a amd powernow! driver too, but its only for older hardware08:48
frinnstsame with intel, speedstep is deprecated in most cases08:48
tierd982what about "AMD frequency sensitivity feedback powersave bias" ?08:48
tierd982that sounds useful :)08:48
frinnstI dont have it enabled. not sure what its for08:49
Romsterread the help on it08:49
Romstersome of them are for touchy hardware or older tech.08:50
tierd982frinnst, its specifically for ondemand08:50
tierd982enable it guys it sounds helpful08:50
*** diverse has quit IRC08:50
tierd982oh wait08:51
tierd982only for newer CPUs tho08:51
RomsterAMD Family 16h and above08:51
*** pal_bth has quit IRC08:51
tierd982bulldozer (mine) is 15h08:52
Romsteri don't think this is even 16h08:52
*** pal_bth has joined #crux08:55
tierd982phenoms are k10/ 10h08:56
Romsteryeah a old thing now08:56
Romsteris amd ding anthgn better than FM2 boards.. think i'm going for an I7 on my next build08:57
Romsteri was all ahh an I9 will be out then.... nope they made it an I7 chip -_-08:57
tierd982Romster, Zen architecture is coming 201608:57
Romsteroh it's only a hint08:58
tierd982no its not08:59
tierd982it was announced by CEO09:00
tierd982head of design team is Jim Keller09:00
tierd982he is like a celebrity in processor design world09:00
tierd982so it sure sounds it will be something09:01
Romsterhired an apple person...09:01
tierd982apple hired an amd person, amd took him back09:01
tierd982he designed your phenom btw!09:01
tierd982at the time phenom IIs were serious competitors to intels09:02
Romster we went out and got Jim Keller, we went out and got Raja Koduri from Apple, Mark Papermaster, Lisa Su,' Read explained, referring to his company's recent hiring spree of big-name industry specialists. '09:02
Romsterthat's why i got one.09:02
Romsterit wasn't quite as powerful as an i7 back then but it was close and a fraction of the price.09:03
tierd982like my FX, in highly parallel tasks it almost catches up to an i7 xxxxk !09:03
Romsterthen they went to crap and i was like oh.... looks like i'll go jump into intel next build09:03
tierd982price however, is a fraction09:04
tierd982not to mention the whole platform (motherboards..etc) is all much cheaper with amd..09:04
Romsternot a fan of this HT crap they do in intel09:04
Romstermake it real cores. make the software bette.09:04
Romstereh mobos don't cost alot09:04
Romstera lot...09:04
Romsterthis mobo was like $299 aud when i got it.09:05
Romsterwasn't a low end one.09:05
Romsteri never use low end mobos on my builds09:05
Romsterit's like the heart of the system. for interfacing.09:05
tierd982relatively yes, but compare an amd board with an intel board (similar feature-set in terms of ports/audio/etc), you'll see intels are much more expensive09:05
Romsteryeah around double09:06
tierd982just looks like a big ripoff circus with intel.. you get ripped off by everyone on the block09:06
Romsterthe whole idea of backing amd was to make intel more competitive09:07
Romsterbut amd messed up majorly on bulldozer09:07
Romsterand then i've had crap luc with any ATi video card in linux that i've given up trying.09:08
tierd982much better now i gotta say..09:08
tierd982but anyways congrats on your 750ti09:08
tierd982solid choice09:08
frinnst<3 amd these days in linux09:09
Romsteryeah i didn't wanna shell out on a 900 series.09:09
Romsterand a750 will be more than enough for my needs09:09
frinnstjust not having to bother with binary driver crap makes it all worth it09:09
Romsteronlything bianry is the nvidia blob when will noveau have 3d acceleraton09:10
Romsteri could live if some of the stuff fell back to the cpu though mesa3d09:10
Romsterbut as it stands it's still too inmature.09:11
frinnstand it will probably always remain that way09:11
frinnstunless nvidia gets its act together and start helping out09:11
frinnstyou might get good support on ancient hardware but they will never be able to keep up09:12
tierd982alrighty guys09:13
tierd982one more tip pls09:13
Romsterwell nvidia has submitted patches09:13
Romsterbut not a lot09:13
tierd982how can I verify cpu scaling is working fine?09:13
frinnstcat /proc/cpuinfo ?09:14
tierd982ah on idle now it should show lower than stock, got it09:14
Romsteri have conky it displays the cpu frequency09:14
frinnstcpu MHz: 800.000 <- thats not the native speed :)09:14
tierd982alrighty will reboot into new kernel09:15
tierd982right now:09:15
tierd982cpu MHz: 4026.89609:15
Romsteryou OC it?09:15
tierd982no thats the stock speed on 835009:16
*** tierd982 has quit IRC09:24
*** tired890 has joined #crux09:27
tired890cpu MHz: 1400.00009:27
tired890Romster, frinnst, thanks guys got it sorted09:27
tired890now let me compile gimp and see if it actually scales back to 4ghz09:28
Romsterit will09:28
Romsterwhat is that a dual core or a quad?09:28
Romsterwth i only have 4...09:30
tired890I think your general performance is almost exactly identical to mine.. except "make -j8" of course :P09:31
Romsterbe sure to set export MAKEFLAGS="-j8" in /etc/pkgmk.conf09:31
*** jue has quit IRC09:31
tired890haha yes I do that before I compile the kernel when I install crux :P09:31
Romsteryeah power wise it's probably the same09:31
Romsteri'd probably have an advantage on single threaded apps than your cpu09:31
Romsterother than that. not much difference09:32
Romsteri don't need to go too big but i'm always using tons of cpu mostly compiling09:32
Romstercpu reviews can be skewed. so i'd take that with a grain of salt09:33
tired890a 16 core opteron will work beautifully as a compile box09:35
*** doomicide has joined #crux09:35
tired890Romster, I kept checking the Mhz while its compiling09:35
tired890it does scale09:35
*** jue has joined #crux09:36
Romsterhmm i built my last pc in 201309:40
Romsterah that was my brothers pc... man i cant find my parts for this build09:41
Romster2009 was the last build i did to this and it's rock solid.09:43
tired890Romster, my prev build was in 200609:44
tired890I keep my systems for as long as they are usuable..09:44
Romsteri may have gone a bit over board on the cpu cooler.09:44
tired890so one day I was using my athlon x2 4200+09:45
Romsteri got my older quadcore that was  done in 2006 thats in my distcc cluster.09:45
tired890and it was lagging (I later realized both HDD and ram were failing).. didnt fail, just slow as crazy09:45
tired890so I rage left the house09:45
tired890went and bought a new one09:45
Romsteralso doubles up as a second pc for media for the dinging room tv09:45
tired890its almost always like this..09:46
Romsterman i seen some really slow seagate hdd's in the past...09:46
Romsterthey wernt slow they were snail paced.09:46
tired890I know that feeling09:47
tired890comes complete with the soothing creaky noise09:47
Romsterbrothers pc had a very slow hdd... when i rebuilt it i didn't realise the hdd be that slow. he only plays wow so i opted to keep the hdd and his vista install...09:47
Romstersecond upgrade a ssd and win 7 for his machine.09:47
Romsterthe difference it made.09:47
Romsterchatter chatter it was nearly as bad as a dot matric printer09:48
tired890my good friend09:48
tired890you compare a failing HDD to an __SSD__ ?09:48
Romsteri replaced that slow ass hdd with a ssd.09:48
alancioI recently had a 3TB seagate drive that was failing09:48
Romstermy wd greens even crapped over the speed of that old seagate hdd.09:48
alancioit was full of data, and I could read & write perfectly, but at 1MB/s09:48
tired890its like riding a bicycle all your life then getting a motorcycle.. you cant go back09:49
Romsteromg that's painful.09:49
Romstermust of had a ton of ECC corrected errors in smart.09:49
tired890I have not met one person who got an SSD then settled for anything less09:49
tired890its addictive09:49
tired890as in becomes some kind of "main component"09:49
tired890not luxury anymore09:49
Romsteryeah its required... have a ssd for root as well.09:50
tired890cheery song I wish to share:
cruxbot[opt.git/3.1]: ntdb: added python as a dependency10:02
cruxbot[opt.git/3.1]: samba: fixed wrong path for perl modules10:02
alancioBTW, I'm still using a regular HD at work10:05
alancioI run several VMs there10:05
alancioso SSDs are not good for that10:05
alancioexpensive + no block level trim support in linux (afaik)10:06
*** darfo has quit IRC10:07
tired890alancio, agreed.. too expensive10:12
*** khanku has joined #crux10:12
tired890as for trim and whatnot, I just put discard/noatime in fstab.. and get an HDD with 5 (now samsungs come with 10 years) warranty!10:13
tired890ops, HDD = SSD10:13
frinnstmy intel dc s3500 ssd just arrived \o/10:20
frinnstenterprise ssds ftw10:20
frinnstnot that expensive10:23
alanciois anybody here interested in a port of iguanair for the iguana infrared receiver?10:38
*** doomicide has quit IRC10:51
prologicfrinnst, for home use?10:56
prologicare they really *that much better*?10:56
frinnstyeah, for home. They are more durable. no more buildnig ports in tmpfs for me }o/10:57
alancioI prefer using tmpfs to avoid wearing the SSD out, and its faster anyway10:59
frinnstworks well until you try to build firefox11:04
alanciohaven't had problems yet11:05
alancioit all depends on the size of your tmpfs ;)11:06
frinnst8gb is usually not enough11:07
prologichow big if your tmpfs alancio ?11:09
alancio16 Gb11:11
prologicah wow11:11
prologichow much total ram have you got?11:11
alanciowell actually I have 16 Gb of ram11:13
alancioyou can specify the size of the tmpfs with mount -o size=xxx, I think I use 15 gb to be safe11:13
alancioonly when building monsters11:13
prologicanyone have a clue what I might be missing here?11:16
prologicvultr vps11:16
prologicstill can't get it to boot11:16
prologicafaik the sata/sci hw is supposed to be all virtio based11:16
prologicat least lspci -k says so when the crux 3.1 iso boots up11:16
prologicI have *all* CONFIG_VIRTIO* options enabled (y)11:17
prologicalancio, ahh yeah I have 16GB here too but didn't fiddle with the defaults :)11:17
nweprologic: do you have support for you fileystem?11:23
alanciohey what happened with Cezar, the guy asking questions in the mailing list?11:23
*** leo-unglaub has joined #crux11:24
prologicext4? yes11:24
Romsteroh yeah i bumped my stupid small docker images form 10GB to 50GB11:31
prologicholy hell11:32
Romsterand because i use lvm now, instead of using loop mounted lvm i made real LVs for docker.11:32
prologicyou know image are meant to be small and transportable :)11:32
RomsterDOCKER_OPTS="-d -p /var/run/ --storage-opt dm.basesize=50G --storage-opt dm.datadev=/dev/data/data --storage-opt dm.metadatadev=/dev/data/metadata"11:32
Romsteryes i build stuff in docker containers thats the only reason for the bigger size.11:33
Romsterand the bigger size is all done in a COW image in sparse extents in lvm already11:33
prologicI'd still do it differently myself11:34
prologicbut oh well :)11:34
Romsteri cold et away with 20GB i supose11:34
Romsteri wanted enough to build firefox and other crap in docker.11:34
Romsterrunning images will still be kept small11:34
prologicsurely you can build firefox int he default 10GB?11:34
Romsterit just gives me headroom11:34
prologicreally? that's nuts11:35
cruxbot[opt.git/3.1]: ldb: updated footprint11:35
cruxbot[opt.git/3.1]: samba: removed redundant dependencies11:35
cruxbot[opt.git/3.1]: subunit: added missing dependency11:35
Romstersoruces + built objects + making the tar archive and storing all the distfiles for it11:35
Romster16GB may just be enough11:35
Romsterbut 10 GB haha no.11:35
prologicthat's still nuts :)11:36
Romsterif you kept the distfiles elsewhere and build in tmpfs then maybe 10GB will cut it.11:36
Romsterbut i hadn't tried that.11:36
alancioI am sure last time I did it with 15 Gb11:37
alanciofor the workdir only11:37
Romsterit probably can do it in 15GB11:38
Romsterbut i failed with 10GB11:38
*** Romster has quit IRC11:42
*** jdolan has quit IRC12:14
*** leo-unglaub has quit IRC12:21
*** jdolan has joined #crux12:29
*** Romster has joined #crux12:34
Romsterthat wasn't fun ah well i fixed it my xorg-server was linked to something else i remvoed some old craft. reboted had to rebuild xorg-srver forgot the evdev so i had to sysrq to reboot. and then i had to downgrade nvidia drivers...12:36
Romsterand rebuild input drivers.12:36
Romsteri should know better -_-12:36
teK_does anybody in here use qemu instances with vlans, connected by sockets?13:05
jaegeralancio: trim works fine in many filesystems in linux now. ext4, xfs, btrfs, nilfs, probably more13:09
jaegerhas for quite some time, actually13:11
*** pal_bth has quit IRC13:39
*** jdolan has quit IRC13:39
*** dougl has joined #crux13:41
*** greduan has joined #crux13:42
greduanGood morning all13:43
EmoSpicegood morning.13:43
*** dougl has quit IRC13:43
*** Feigrim has joined #crux13:43
greduanHere I thought the sysup would finish compiling by the morning it didn't cause the internet got cut out in the middle of the night (we disconnect our modem to sleep) so what's left is all the X11 stuff to compile.13:50
greduanI have compiled that already in another distro and it didn't take more than like 30 mins in total, so I'm not that disappointed.13:52
greduanI'm glad it wasn't webkit or something13:52
greduanGotta download webkit as a binary13:52
EmoSpiceYeah. Xorg's never ben my problem. Webkit or FF or the like's MUCH worse :P13:52
EmoSpice...I built it >_<13:52
greduanOh gosh13:53
greduanAny significant upgrades in performance after compiling it yourself?13:53
EmoSpiceJust left the machine on overnight and went to the bar.13:53
EmoSpiceAt least not that I've noticed (not that I'm looking013:53
jaegerhmm, wonder how long webkit takes on my main box13:54
EmoSpiceIt didn't take nearly the whole night, but I cba to go back and watch it :P13:54
greduanWell apparently it takes round 45 mins on 20 cores, all around 2.0GHz13:54
greduanRomster: shared some binaries with me yesterday and mentioned how long it took to compile it13:55
jaegerwhich one should I test? webkitgtk, webkit-gtk2, or webkit-gtk3?13:55
greduanI used webkitgtk since that's the one that's required for what I'm gonna compile13:55
greduanNot entirely sure what the difference is...13:55
EmoSpicegreduan: what project are you building on top>13:56
greduanMaybe webkitgtk is just both gtk2 and gtk3?13:56
EmoSpicedwb? luakit? midor? surf?13:56
jaegerI'll give it a shot, I'm curious13:56
greduanEmoSpice: vimb13:56
EmoSpicegood browser :)13:56
greduanI'd use dwb but it has crashed on me whenever I visit BitBucket, and it seems to not be active anymore13:56
greduanIndeed, good browser13:56
EmoSpiceit's not.13:56
EmoSpicehave you used luakit?13:57
Feigrimdwb is nice but yeah, not developed anymore :/13:57
greduanI think so13:57
greduandwb is the only one I gave a fair chance though13:57
EmoSpicedwb was always...weird for me to configure. I liked the idea but could never grok the configuration completely13:57
FeigrimI have used luakit, if you want to hack lua code it's fun I guess13:57
Feigrimit's like dwb but way more complicated to customize13:57
EmoSpiceFeigrim: I had like...3 configurations to change.13:57
FeigrimEmoSpice: I just did it in the browser13:58
greduanThere was one similar to vimb13:58
Feigrimto change things like hint characters13:58
EmoSpice"PyQt5 and QtWebKit"13:58
EmoSpiceis why I haven't tried that one.13:59
greduanI do not want to compile that13:59
greduanWhat shell do you guys use?14:01
*** jdolan has joined #crux14:01
greduanI'm currently trying out mksh, which feels aster than bash for sure but I wonder how much I can customize.14:01
jaegerI use bash14:01
EmoSpicegreduan: I love mksh, but I'm currently addicted to zsh14:05
EmoSpice(rprompt is critical for me >_>)14:05
juebash for interactive use, /bin/sh is a symlink to dash14:05
greduanIn CRUX, is /bin/sh a symlink to dash?14:05
jaegernot yet, we've been considering it14:06
jueno, not yet14:06
juebtw, I ran in more problems with mksh as /bin/sh than with dash14:07
EmoSpiceI imagine you did.14:07
greduanWhy is that?14:07
greduanAren't they very similar in syntax and everything?14:07
EmoSpicemksh is a ksh derivative14:07
juedidn't looked at that deeper, I just recompiled everything installed and some ports failed with mksh but not with dash14:08
EmoSpiceyes, they're similar, but similar != the same14:08
greduanI see14:08
jaegeruh oh, my "paypla" informations have been leaked14:10
*** diverse has joined #crux14:12
greduanLike password?14:13
jaegerpresumably. I ignored the email because it obviously wasn't from paypal and they couldn't even spell paypal correctly14:14
*** diverse has quit IRC14:14
*** leo-unglaub has joined #crux14:14
teK_must have been a young and hip competitor14:15
teK_remember when new services just did 's/er$/r/g' on their names just to show how cool they were? :}14:15
*** diverse has joined #crux14:17
diversejaeger: talk about phishing phail14:20
diversewell, I was rebuilding my system after prtwashing just for kicks14:23
diverseone interesting thing is, prt-get's url is dead, and before you tell me about PKGMK_SOURCE_MIRRORS, yeah I know, but I still felt like bringing it up.14:25
jaegerok, it took me 24m31.5s to build webkitgtk14:25
jaeger-j8 @ 4.4GHz14:25
diversethat's how long it takes me to build firefox14:26
diverse(I heard webkit is slower to build)14:28
jaegerI'll time firefox as well14:28
diverseshould be probably 18m at that metric14:28
diverse*probably be14:29
*** Feigrim has quit IRC14:29
*** Feigrim has joined #crux14:30
diverseI should probably lower that -j9 to -j814:31
diverseI thought the amount of jobs would be n_cores + 114:32
EmoSpiceThere's some discussion on which is "better"14:32
diverserelated to the number of compiler jobs?14:33
EmoSpiceI use n_cores, but that's because I haven't done any research and it pegs the cpu wide open.14:33
greduanjaeger: that's faster than I thought it would be.14:34
diverseEmoSpice: back when I used gentoo a long time ago, it was recommended to do n_cores + 1, but it seems to be the concensus now to just use n_cores14:36
EmoSpiceI've heard both. Like I said - I'm not informed on the specifics, so I just use n_cores14:36
jaegerI doubt the difference is worth much14:37
diversewell I wonder what that extra job is for, considering the all of the cores would be taken up14:38
jaegerfirefox took 19m57.6s14:59
greduanThat was pretty fasts15:00
greduanI'm still compiling llvm here15:01
greduanllvm is pretty big15:01
diversejaeger: cool15:02
diversegreduan: llvm shouldn't take long, compared to firefox15:03
diversebut of course, if you could upgrade to 4 cores or higher, it will make your building experience that much better :)15:04
*** BitPuffin has joined #crux15:04
diverseor cluster15:05
greduanI'm not familiar with those though15:05
greduanHas anyone owned a Lenovo X220?15:05
greduanI'm thinking that'll be my next laptop. After 7-8 years with these I'm thinking of changing.15:06
diversegreduan: I'm planning on doing a 2-cluster, but if I can find a more general way of doing it, not just for compiling, that would be nice :)15:07
greduanCan a cluster be made using two laptops?15:09
diverseyou just got to treat that other laptop like a server15:10
greduanI see15:10
greduanRight now I'm using Ubuntu one one (this one and technically my brother's) because it "just works" and compiling stuff on CRUX on the other one (mine).15:11
jaegergreduan: I have an X220, it's rock solid15:12
greduanjaeger: did you buy it with the best specs or no?15:12
jaegernot the best, medium-high specs I guess. I mainly wanted the IPS panel15:13
jaegerit has an i5-2520m if I remember correctly, 8GB RAM15:13
jaegerIt's at home right now but I can check this evening15:13
greduan*compiling cairo*15:14
greduanjaeger: Don't worry about it, just wanted to hear some opinion(s) on it. :)15:14
jaegerI'm very happy with it15:16
greduanBeing stuck with a dual core (2.0GHz) with only 4GB RAM is tiring, so I'm really liking how it's got 8GB RAM and a 2.7GHz dual core (for the i7).15:16
EmoSpiceThe X220 looks pretty similar to the machine I bought last year: the E430, which is amazing.15:17
greduanYeah it's very close15:18
jaegera good chunk of the latest round of crux's multilib development happened on my X220, incidentally :)15:19
greduanthe processor isn't quite as nice in the E430 (2.1GHz vs 2.7GHz with the i7).15:19
greduanCool. :)15:19
greduanMultilib in this case means compiling with 32-bit compatibility, ja?15:19
EmoSpiceMine's an i5 @ 2.4GHz, iirc.15:20
EmoSpicemaybe 2.515:20
greduanThat is pretty good15:21
jaegergreduan: yes. previously crux was 32-bit only. There were some unofficial multilib offerings before 3.015:21
diverseEmoSpice: cat /proc/cpuinfo15:21
EmoSpicediverse: yes. I'm not on that machine currently :P15:21
greduanI'm thinking of getting the i7, just so that it'll last me another couple of years before I have to replace it15:21
greduanA 2007 laptop has lasted me until 2015 though, a 2011 laptop, or 2010, don't remember, will probably last me 7 years _at least_ before I feel the need to replace it.15:22
greduanUnless one of these years we move on to a better technology.15:23
*** kori has joined #crux15:25
*** kori has quit IRC15:25
*** kori has joined #crux15:25
*** dougl has joined #crux15:43
teK_greduan: bear in mind that the mobile i7 cpus are )_really_ noticeably slower than their desktop coutnerparts15:56
greduanteK_: Indeed, although still better than the i3 or i5 mobile. :)15:58
diverseteK_: at the same freqencies you mean?15:58
diverseI guess having slower laptop ram would bottleneck it16:00
teK_my desktop i7 has a higher frequ., tbh. Still there is a huge penalty during compiling ports on my laptop compared to my desktop16:00
jaegermobile CPUs just aren't meant to be powerhouses16:01
teK_I wanted: power efficient if under-used, and power-eequal during full utilisation ;-)16:01
teK_I admit that I have to leave my purple bubble and enter cruel life from time to time16:02
diverseteK_: that's what being mean to frinnst is for, right? ;P16:04
teK_that's because leaving the bubble makes me sad16:06
*** leo-unglaub has quit IRC16:06
*** onodera has joined #crux16:17
*** onodera has quit IRC16:32
greduanrunning into some trouble building xorg-server
greduanNote that logging is in append mode, so all the previous tries will be there as well not just the last one.16:45
greduansomething to do with an "epoxy" package16:46
jaegercheck xorg-server's deps with 'prt-get depends xorg-server'16:47
greduan[ ] libepoxy16:48
greduanInstalling now16:48
greduanthanks jaeger. :)16:48
jaegernp. It was added after 3.1's release16:49
*** dougl has quit IRC16:52
*** lnds has joined #crux16:56
greduanbest way to handle this?17:08
greduanI read in the manual not to use -f so I wanna check if it would matter in this case17:09
teK_what package besides  xorg-server owns these?17:09
teK_(find out by running pkginfo -o usr/lib/xorg/modules/
greduanteK_: all three files are owned by the xorg-glamor-egl package.17:11
teK_hey xorg-Team should this package include a README on what to do about conflicts?17:12
teK_i.e. jaeger frinnst tilman17:12
greduanSince this probably matters, at install I just selected every package and installed it, since I thought the experience of installing would be worse I thought it'd be better to just avoid installing a lot of stuff.17:22
greduanSo core, opt and xorg all the packages were installed (those available in the 3.1 ISO).17:22
*** Pingax has quit IRC17:26
teK_for test and first time installations that's a good approach17:28
teK_I don't remember when I installed my desktop but I did more than one major update with it, so it's been some years. In this case it's better imho to trim the base installation and kernel to set of things you really need17:29
greduanNext time I re-install (if I do) I'll choose exactly what I need.17:30
greduanDependencies are managed automatically at install, yeah?17:30
teK_prime example could be jfsutils .. they won't hurt but you probably never gonna need these :)17:30
teK_greduan: by setup, yes. After that you have to use prt-{get,cache}17:31
greduanYeah. And also I've now learnt that anything to do with the hardware is mostly managed by the kernel, so just configure it correctly and you'll be fine.17:32
greduanOh yeah, what's the kernel upgrade strategy for CRUX? Wait for another ISO?17:32
teK_do as it fits your needs17:33
teK_from 2.6 to 3.19 you can virtually use any version with our current glibc17:33
greduanAwesome! I'll be upgrading soon then17:34
greduanIt'll be another hour of reading the help of every option in the kernel. lol17:34
teK_the kernel mediates between userland (apps) and the hardware, but it knows some additional tricks like the security subsystem that just regulates access to ressources17:34
teK_greduan: been there, done that ;)17:35
greduanNow I know at least what is needed for my hardware to work properly.17:35
teK_keep exploring :)17:36
greduanOh yeah, something that wasn't clear to me. Can I use my old .config with a new kernel (just making sure to properly reconfigure)? I'd like to not have make all the decisions again if possible. :)17:36
teK_good point and indeed you can17:37
teK_copy the current config from /proc/config.gz to the new kernel directory17:37
teK_like zcat /proc/config.gz > .config17:37
teK_then run make oldconfig17:37
teK_(in the kernel src directory)17:38
teK_it will ask for changed options only17:38
greduanI also have backup of the .config in several places. :P17:38
teK_the kernel usually is configured to include the config in /proc17:38
greduanYeah, I'll make sure that's selected as well.17:38
greduanDunno if to continue managing packages or if I should wait to properly update xorg-server...17:39
*** retard has joined #crux17:42
teK_hm? I am involved with xorg. The people I mentioned earlier are. Feel free to  open a bugreport if there's no response17:44
greduanI'll open a bug17:46
greduanIs the account opened to submit a bug the same account used on the rest of
teK_regretably not17:49
greduanI see17:54
tilmanxorg-glamor-egl should be nuked17:56
tilmanafaik these days glamor lives in xorg-server17:57
tilmanin the early days it was developed in its separate package (while it was immature/crashy/sucky/we)17:57
greduanSo I uninstall and report a bug for its existence?17:57
teK_greduan: you can wait for one of the devs (with commit access) to tune in17:59
tilmani don't even have a xorg-glamor-egl port17:59
teK_tilman: did wenuke your ssh keys?17:59
tilmanteK_: not sure what we did, but they don't work anymore18:00
teK_I guess we didnt because _are_ coming back18:00
tilmangreduan: is your system "old", or did you just set it up?18:00
teK_portdb does not count if the port is on our 3.1 ISO ;)18:00
greduantilman: just installed it yesterday18:00
tilmanit's on the iso?18:00
tilmanoh moy18:00
teK_greduan: just remoce the port, don't commit a bug report and reinstall / rebuild xorg-server.18:01
tilmanpkgrm xorg-glamor-egl18:01
teK_this does worK? :P18:01
teK_I think always did that with prt-get. :)18:02
greduanxorg-server updated successfully18:02
greduanThanks tilman, teK_ :)18:02
teK_your shell was set to /bin/false, so you're one command away from commit access :))18:07
tilmangood to know18:08
teK_just say you don't care :P18:10
tilmani won't say anything for now18:10
greduanSo yesterday Romster gave me a URL to some pre-compiled binaries that would probably take forever to compile on my computer (like webkit)
greduanWhat is the proper way to download those?18:11
greduanI tried wget and it didn't work18:11
greduanalthough I used a # instead of %2318:11
greduanWith a %23 it works18:12
teK_are you a ZSH user?18:12
teK_if so, include in your setup:18:13
teK_autoload -U url-quote-magic18:13
teK_zle -N self-insert url-quote-magic18:13
greduanRight now I'm using mksh18:13
teK_oh ok18:13
teK_then your are DOOOOMED18:13
teK_ZSH can autoescape # as it has a special meaning in it :)18:13
greduanI see18:13
tired890so yesterday I updated my kernel to 3.18.4 .. now virtualbox wont run (I'm guessing the vboxdrv module needs to be rebuilt)18:13
teK_i.e. if you paste the url in the shell it will add a \ in front of each #18:14
teK_tired890: yes, there's a script for that18:14
teK_vbox_rebuild_module or something18:14
greduanteK_: I see. I think the problem is related to how # points to an ID in the DOM. You know what I mean?18:14
tired890teK_, many thanks! was hoping to hear that :D18:15
teK_that's not the problem18:15
tired890find is unable to locate it.. where are those scripts usually?18:15
teK_greduan: the # is interpreted by your shell that's all18:15
greduanI did try a \# though18:15
greduanwget was giving me a 40418:15
diverseteK_: oh yeah, speaking of zsh, I notice there was something wrong with completing the port name everytime I tab complete it for prt-get18:15
teK_diverse: nothing coming or or something false?18:16
teK_if you get no response, check permissions of the dbfile prt-get.. prt-get does not work well with restrictive umasks18:16
*** hhhhhhhh has joined #crux18:16
tilmanTIL that the CTO of joyent if the guy who tried to insult miller back in the 90s with that "did you ever kiss a girl" line18:16
diverseteK_: when I type "fire[TAB]" I get this: "% prt-get install fireCan't open cache file: /var/lib/pkg/prt-get.cache"18:17
teK_greduan: try running echo to see what your shell does to the string18:17
teK_diverse: well then run prt-get cache :P18:17
teK_tilman: wtf :p18:17
tilmanor rather, i was surprised the jerk from back then is a somewhat known guy now ;)18:17
tired890where would one find these rebuild scripts guys?18:18
greduanteK_: it outputs it as-is18:18
teK_tired890: have a look at virtualboxe's footprint18:18
diverseteK_: could you note that in the readme file :)18:18
teK_i.e. pkginfo -l virtualbox  (and maybe | grep bin)18:19
teK_diverse: sure18:19
greduanteK_: HOWEVER. the file autocompletes with a \#18:19
cruxbot[opt.git/3.1]: zsh: hint on the relation between prt-get completion and prt's cache18:21
teK_greduan: yeah, \ escapes the sign coming after it :)18:22
teK_[tek@basra][/tmp]% touch =nosuchbinary                                         :(18:22
teK_zsh: nosuchbinary not found18:22
teK_[tek@basra][/tmp]% touch \=nosuchbinary                                        :(18:22
teK_i.e. = has a special meaning in ZSH, too18:23
tired890do_dkms sounds like it..18:24
*** leo-unglaub has joined #crux18:29
tired890alrighty, for future reference to souls googling this irc log18:37
tired890After updating your kernel on crux you need to rebuild vboxdrv18:38
tired890simply go to /usr/share/virtualbox/bin/src18:38
tired890make clean && make all && make install18:38
tired890modprobe vboxdrv18:39
tired890and voila!18:39
greduanThanks! :)18:39
jaegertired890: tired890 /etc/rc.d/rc.vboxdrv setup18:58
jaegeroops, double name there, sorry18:58
tired890jaeger, I don't have that file..19:01
tired890ls /etc/rc.d/19:01
tired890alsa   dbus  inetd  net     saslauthd  sshd wifi19:01
tired890crond  exim  lo     rsyncd  slapd      sysklogd19:01
jaegerAh, guess that's not installed with the OSE version or whatever it's called. I use the binary version. Ignore me. :)19:01
tired890thanks anyways :)19:02
tired890glad you did, irc logger now has solutions to both source and binary installs ;)19:02
frinnstwhy the fuck wont grub use /boot/grub/grub.cfg ?19:04
frinnstI just get a standard prompt.. no meny entries or anything19:04
jaegergrub or grub2?19:05
frinnstgrub2, sorry19:05
jaegerin the menu can you load your config file manually?19:05
jaegerI wonder if it can't read that partition/filesystem19:05
jaegeryou can use ls or find with tab completion to see what's available19:06
frinnstI can boot it fine, yeah19:06
frinnsttab completion works19:06
frinnstits pretty much my old grub.cfg but with new UUIDs19:06
jaegerah, ok. that's weird19:06
diversesounds like grub2 did a mysterious breaking change19:07
frinnsthah. stupid if statement19:08
frinnst(that was a new addition when stuff didnt work at all)19:08
greduanDoesn't grub need a command run so that it registers the changes in the boot code? I find that's the case if I update the kernel and afterwards don't run lilo in order to update lilo's idea of where the kernel is.19:11
jaegerlilo needs that, grub doesn't19:11
frinnstoh well, everything seems to work as it should finally19:14
tired890I've decided to run office inside virtualbox as opposed to libreoffice.. not the most elegant but hey, need to work and those weird symbols popping up inside libreoffice (when opening a word doc) are a dealbreaker19:16
teK_seamless mode might be an option for you.
tired890hopefully they get it sorted soon (M$ vandalizing on purpose?)19:17
tired890teK_, interesting!19:18
*** leo-unglaub has quit IRC19:19
*** kInOzAwA_ has joined #crux19:20
*** kInOzAwA has joined #crux19:20
*** kInOzAwA_ has quit IRC19:27
dxlr8ris it possible to roll back, or to see the Pkgfile of an older version?19:31
dxlr8rthe new samba requires a lot more, and I was really satisfied with the last version. would like the new one, but not with all the bells and whistles :)19:31
joacimcan get the old one through git19:37
joacimshould be somewhere on crux.nu19:38
dxlr8rfound one that fits in the search ports section of the site19:39
dxlr8rsame version too19:39
greduanYay installed Vimb19:52
greduanI am so glad I could download a binary of webkitgtk, cause it turns out I actually needed webkit-gtk2. I would have killed myself if I had to compile both. lol19:52
tired890compile baby! march=native !20:01
greduanI haven't the power for that20:02
*** BitPuffin has quit IRC20:03
tired890you know last night I installed crux on my laptop20:04
tired890it has an amd E-450 apu (1.6Ghz dual core)20:04
tired890literally about 30 minutes ago it finished installing xfce20:05
tired890prt-get depinst xorg (then I went to sleep)20:05
tired890then prt-get depinst xfce420:05
greduanwhat time is it for you?20:05
tired890hence the nickname20:06
greduanI installed it yesterday in the morning with an Intel Core 2 Duo (2.0GHz dual core).20:06
greduanI pre-installed anything that came in the ISO and yesterday (and today) I re-compiled for a newer version.20:07
greduanI only installed dwm though, not xfce20:07
greduanAnyone have any advice for the firsts Pkgfile?20:10
tilmanport an autoconf-using program first20:12
tilmanthose usually follow the same recipe, giving insta-success20:12
greduantilman: good point20:16
greduanAlthough I'm fairly familiar with the Arch PKGBUILD process20:16
tilmanif you are inexperienced with building *nix software from sources, then pkgmk is a piece of cake20:17
*** leo-unglaub has joined #crux20:17
greduaninexperienced or experienced?20:17
tilmanthe latter ;]20:17
greduanI am experienced. :)20:18
greduanI just find making a file from a template a little bit of a pain.20:18
greduanhey leo-unglaub. :)20:18
teK_greduan: checkout prtcreate from prt-utils20:21
greduanteK_: will do. Thanks :)20:22
teK_mkdir newport && cd newport && prtcreate && vim Pkgfile && pkgmk -um && pkgmk20:22
teK_s/vim/$EDITOR/  .. old habit :)20:23
tired890greduan, making your first crux port?20:23
greduantired890: that's right20:23
greduanAlthough I've made several PKGBUILDs in the past20:23
*** leo-unglaub has quit IRC20:24
greduanteK_: This is exactly what I was looking for. lol20:24
greduanWhat's the standard way to do Git sources? In Arch I just set a git://... as a source and it'll pull it20:25
dxlr8rusually with pkgmk it is smart to use -kw to in case something goes wrong20:25
dxlr8rto see if you need patching etc.20:26
greduandxlr8r: good to know, thanks20:26
teK_greduan: there's no builtin support for that, ports usually run git clone --depth 1... in build()20:26
teK_we could add support for that but then you would have to list git as a dependency, otherwhise building will fail if git is not installed20:27
greduanteK_: OK, and I imagine git is added to Depends on line?20:27
teK_it will fail either way ;-)20:27
greduanArch has a make dependency thing20:27
teK_compare opt/mplayer.. I pack up tarballs for each release20:27
greduandependencies only needed at compile-time20:27
teK_we have a poicy on  that too, see our development wiki  page20:27
diversegreduan: all dependencies for compile-time and runtime of the app in question20:29
tired890hi diverse20:29
teK_prologic: ping!20:30
diverseteK_: was that towards me?20:30
teK_read up our dev page20:31
diversewell aren't some dependencies not make dependencies?20:31
teK_of course they are :)20:31
teK_you missed the case when the  dependency is in core20:31
teK_any Python coders presen?20:31
diversewell I was just saying in general outside of core...20:32
tilmanwhat's your actual question?20:32
teK_tilman: I have a python port of a library.. it provides athin  layer above  a C++ lib. The C++ version has some methods consumedata(FILE* fd).20:33
tilmanyeah, i never wrote python bindings for c/c++ lib ;)20:33
tilmanwhy call it 'fd' when it is a FILE*20:33
teK_So naturally the port does that too. The thing is.. I receive data via HTTP and do not flush things to disk just to read them back in20:33
teK_tilman: I am just sketching so I'd be the idiot :)20:34
teK_the point is..20:34
teK_I do know cStringIO20:34
diversegreduan: of course, don't don't need to specify extra dependencies if the other dependencies pull those extra dependencies.20:34
teK_the thing is.. it does not implement the  fileno attributeso  I cannot push my data in a cStringIO.String() and feedthatto consumedata()20:35
greduandiverse: Right. :)20:35
teK_the is a from_bytes() routine but, of course, it has different and  incomplete semantics compared to consumedata()20:35
diversetry to keep it simple20:35
greduanwill do20:35
teK_*there is20:35
diversetired890: sorry for ignoring you earlier, hey there20:36
tilmanteK_: just bite the bullet and use a temporary file20:36
tilmanif you cared about performance you wouldn't use python in the first place?20:36
tired890no worries mate :)20:36
diversetired890: teK_ totally caught me off guard20:36
teK_that's my thing :P20:37
diverseteK_: go back in your bubble20:37
teK_tilman: well I don't think the performance is hurt THAT much by using python. In my scenario, the client may be computationally week, not the server  :)20:37
tilmanteK_: how about a pipe then?20:38
tilmancan you consume the data while it is being received?20:38
teK_omg :\20:38
teK_but good idea20:38
tilmani love pipes20:38
tilman<3 unix20:39
teK_<3 unix20:39
tired890I'm glad to hear this20:39
tired890that my good friends mean this dist will never switch to non-unix-complying paradigms20:39
tired890read: systemDeath!20:39
tilmanteK_: i'd recommend to make your C api use fds rather than stdio (FILE*)20:40
teK_the MacOS version would be20:40
teK_❥ unix20:40
teK_tilman: it's not my library. I even looked at the python base code that gets packed up as module20:40
teK_I have a giffor that, too20:40
tilmanoh, so you aren't the idiot after all?20:41
tilmanwell... then you need to make sure that $library accepts a non-seekable fd20:41
teK_actually I lost it :|20:41
tilmanyour keyboard is breaking btw20:42
tilmanspace bar gonna RIP soon20:42
teK_cStringIO _is_ seekable. it just has no freaking fileno20:42
teK_I know. Has been since the cleaning before the last cleaning..20:42
jaegerprologic: if you haven't seen this already,
tilmanyou're also talking gibberish20:42
tilmanget some sleep man :p20:42
diverse+1 on tilman's statement20:43
teK_I am not20:43
tilmanteK_: tbh, the fact that the python api offers some seekability doesn't mean anything when it doesn't implement the API required by your C lib ;)20:43
tilmanyou could write an adaptor of course20:44
teK_no it doesn't, thats right..20:44
tilman1.) make $library not use FILE*, but make it take separate read/write/seek function pointers20:44
tilman2.) implement those functions for cstringio20:44
tilman3.) profit!20:44
tilmanteK_: e.g., libvorbisfile does it that way20:45
teK_you just want to annoy me into using a temp file^W^Wpipe20:45
tilmanbut it's a very common scheme :)20:45
teK_it's out of my control. You do realize cString is part of the standard python distro.. I could send a patch  and hope for the best20:46
teK_I have a feeling 1-3 are not going to happen20:46
tilmanthat won't work20:46
teK_but that may be GIBBERISH coming out of me once more20:46
diverseteK_: it is20:46
teK_stfu diverse20:46
tilmanif there's an fd, then people expect that you can read(2)/write(2) it20:47
tilmanacutally there's fmemopen()!20:49
teK_I stumbled upon that one.. yes20:49
tilmaniirc it has portability issues though20:49
teK_I *love* portabiblity as windows kernel and user-land and UNIX userland is involved20:50
tilmanyou're writing a windows kernel driver?20:53
tilmannot sure if impressed20:53
teK_not sure if listening at all20:53
tilmani guess you know about msys/mingw?20:54
tilmanprobably the fastest way to get unix-ish software up and running on windows. and unlike cygin, it lets you build binaries for the mscrt as opposed to cygwin20:54
teK_you are right it was amazing how well that worked20:55
tilmantoo bad there's no decent terminal emulator for windows20:55
teK_I wrestled with the native lib-version for MSVS20:56
teK_then tried cygwin20:56
diverse(there isn't a decent virtual terminal in windows to begin with)20:56
teK_then had insta-success with ming on linux.. it's always the last thing you try ;)20:56
tilmanyou should have asked in the temple of knowledge _first_20:56
tired890I can setup and maintain a forum dedicated for crux.. this will give us better exposure I reckon, since google will index content.. what do you guys think?20:57
teK_not sure if allowed to enter with my shoes still on20:57
teK_tired890: we have avoided exposure ever since ;>20:58
tired890nice and cozy :)20:59
*** nrxtx has joined #crux21:08
*** tek__ has joined #crux21:11
*** openfbtd has joined #crux21:12
*** tek__ has quit IRC21:12
teK_tilman: jfyi: switched to a machine with an intact keyboard <321:24
teK_trying the pipe approach now ;)21:24
tilmangood luck21:25
teK_Return a pair of file descriptors (r, w) sounds awesome :P21:25
teK_the C++ library does indeed use funky(int fd). Yet I still did not fully understand if you are OK with that or not ;)21:25
tilmanfd good21:26
tilmanFILE* bad21:26
teK_teK thankful21:27
teK_teK programming now21:27
*** lnds has quit IRC21:27
prologicteK_, pong!21:27
teK_prologic: thanks for reporting back. Do you know a more elegant than pipes in python to parse data from a variable to a function requiring a file descriptor? cStringIO won't work21:28
diverseteK_ smash keyboard?21:28
teK_hahhaha, we'll see about that21:28
prologicteK_, need more context :) code? samples?21:32
prologicwhere's the variable come from?21:32
teK_does it matter? I don't think so21:33
teK_I use http something from python21:33
nwehmm any idea about this.. on my built in soundcard.. I get sound if I put my headphone in "sub" output21:33
teK_there's a POST request coming in with serialized data and I have to pass it to a library(C++)-wrapper that needs a fd21:33
nwebut If I have it in the right output I dont get any sound...21:34
teK_nwe: google for alsa config switch output21:34
prologicteK_, ahh21:34
prologicuse a memory mapped file?21:34
diverseprologic: what's a memory mapped file, out of curiousity?21:35
prologicexactly what it sounds like :)21:35
diverseas in "cache"?21:35
prologica fd where it's contents are mapped to memory21:35
prologicbit like a tmpfs :)21:36
tilmanhe wants to avoid using files though21:36
teK_isn't this the wrong direction, prologic? :)21:36
diverseprologic: ah, okay21:36
prologicPython just warps around the C library anyway21:36
tilmanteK_: yes21:36
prologicteK_, surely it can be written to?21:36
prologicmaybe you're right21:36
prologicyeah I'm not sure teK_21:37
prologicI think you're stuck with a temporary file21:37
prologicor a pipe21:37
teK_pipes are fine for now, I guess21:38
diverseprologic: so a swapfile would be a good example?21:40
diverses/a good/another/21:41
greduanIt seems the norm for the Maintainer/Packager line is first and last name comma boop at host dot tld21:41
greduanIs this a set convention or does it change?21:41
*** doomicide has joined #crux21:41
teK_it's best to stick with
teK_so it's more like a .. guideline21:42
greduanCan a username be used?21:44
prologicdiverse, eh? :)21:44
diverseprologic: mmap file example21:44
prologicdunno :)21:45
tilmangreduan: it's your port; do what you feel. there's positive karma if you're following the guidelines in the handbook though21:46
tilmanthere is no code in pkgmk that checks for invalid names in Pkgfiles and refuses to build them ;>21:46
greduanOK, good karma good karma. :)21:47
greduanLooks like not a lot of people visit #crux-ot22:09
*** darfo has joined #crux22:28
teK_tilman: I just tested the whole chain from kernel to DB and things look okayish22:34
teK_thanks highest priest of wisdom :D22:34
teK_teK glad22:34
nwewhy is it so much job to get alsa to work correctly..22:34
teK_it usually isn't; just ask, maybe someone happens to know22:34
nweIf I want to use surrsound51 what have I miss in my asoundrc file ...
nweright know I got sound from my subwoofer output, and if I connecting my headphone into the output in front of my chassi...22:36
greduanso rejmerge... What is the difference between Keep, Upgrade and Merge?22:37
teK_Keep: do nothing, Upgrade: replace with update, merge: interactively user-based merging22:37
greduanteK_: Cool thanks22:38
*** Stealth has joined #crux22:38
greduannwe: Are you sure you have compiled your kernel with support for all the hardware you want to use?22:39
jaegernwe: <-- try something like this, perhaps. you might have to remap channels or something22:39
jaegermy older HTPC (zotac mag hd-nd01) had the channels arranged incorrectly over HDMI, my newer one works without intervention22:40
jaegeryou can use speaker-test to test the mapping, too22:40
nwegreduan: yes I have all that stuff for my soundcard.22:40
greduanthis is a weird diff...22:44
greduanoh I see22:45
nwejaeger: I try that but still have the same problem.22:45, it's asking me whether to keep the old one or the new one, what should I do here?22:46
*** dougl has joined #crux22:48
greduanIt's a cache file so it probably doesn't matter but I want to make sure it doesn't have something important for the program to run22:50
jaegergreduan: you can update it safely, or recreate it with ldconfig22:54
greduanjaeger: cool, thanks22:54
*** Stealth has quit IRC22:58
*** Stealth has joined #crux23:02
*** hhhhhhhh has quit IRC23:04
*** doomicide has quit IRC23:09
greduanCan't start Xorg :/23:40
teK_[ 85956.922] (EE) module ABI major version (15) doesn't match the server's version (18)23:41
greduanWhat does that mean?23:41
teK_recompile the vesa module (so it gets built against your updated xorg-server)23:41
greduanAh OK23:42
teK_prt-get update -fr xorg-...23:42
greduanThank you23:43
greduanCompiling now...23:43
greduanOnly xorg-xf86-video-vesa or also intel and stuff?23:45
greduanCompiling only vesa didn't solve the problem, it seems23:46
teK_tilman: os.pipe() has size limits, though. So I atm the sending app just gets stuck if the limit is hit (as the write() is ablocking operation)23:47
*** cinolt has joined #crux23:51
greduanHad to do xorg-xf86-video-intel as well23:53

Generated by 2.11.0 by Marius Gedminas - find it at!