IRC Logs for #crux Saturday, 2011-05-07

*** ross49 has joined #crux00:03
ross49Hey all, I copied the crux .img file to USB using unetbootin. When I boot, I get that it can not find the installation media, and asks me to mount the installation media to /.tmpfs/.media when I mount /dev/sdb1 to /.tmpfs/.med.. I still get that installation file not found00:05
ross49any ideas on how I can fix this?00:05
*** ThePub has quit IRC00:16
*** ThePub has joined #crux00:16
*** Rotwang has quit IRC00:30
*** Rotwang has joined #crux00:31
* ThePub sighs00:54
RomsterThePub, changing init system is fun tho :D00:54
ThePub"pkgrm pkgutils"00:54
ThePubnot exactly the smartest command :)00:54
Romsterum you can untar the pkgutils archive and reinstall it that way.00:56
ThePubyeap, did right afterwards00:57
ThePubbut I issued a series of shell commands:  pkgrm pkgutils; pkgadd pkgutils-versioned00:57
ThePubso like.. I'm tired and acting idiotic.. more than normal anyways.00:57
Romsterbut why did you pkgrm that in the first place O_O00:58
ThePubbecause I rewrote pkgmk :)00:58
Romsterah well i've done some rewriting myself so what did you change?00:59
ThePubadding versioning support to packages00:59
Romsterhmm got a diff of it?00:59
ThePubI'd also like to integrate some concepts from gobo into the Pkgfile too01:00
Romsterdiff -pruN pkgutils.orig pkgutils01:00
ThePubmaybe when it's a bit further along :)01:00
ThePubjust started messing my machine up.  if I get something workable it'll go up on github for sure01:00
Romsteri've got some hackie stuff in mine but i've got some other ideas to redo some of them in a cleaner way.01:00
Romsteri'm into new ideas.01:00
Romstersounds to me your doing something like gentoo's slots01:01
ThePubI think versioning has merit, if it doesn't make building a package anymore difficult.01:01
Romstermajor thing is i don't like the whole prefix the entire path idea. i much prefer to use /usr/bin/foo-xx for the number etc.01:02
ThePubthe only problem I see with that is generic scripting becomes more difficult01:03
Romster/usr/share/$name-$version /usr/include/$name-$version01:03
Romsteryou could do what i do with distcc, is have a directory of symlinks to the other foo-xx names, then perfix that path to run it01:04
ThePubI've had a few ideas sitting around for awhile I want to play with01:06
ThePubone simple idea is this:  why bother with any of that at all?  why not just keep the archived packages sitting in a repo, and instead install stub scripts under /usr/bin?01:07
ThePubwhat they would do is initialize that programs environment and mount the required packages into a view of that program01:07
ThePubkind of like how puppy linux does its thing, but on a per-program basis01:07
ThePubor even a per-"class" method.  for example a program could be classified as "system" and be mounted in/under a whole system heirarchy that runs the system at startup01:08
ThePubthen wierd-program-x needs to start and we pluck out a few libraries and add another in to a custom view, which allows it to run in a tailored environment01:09
ThePubbut really I was just going to try and mimick some of the concepts from Gobo under the Crux environment for now01:10
ThePubthere are some novel ideas in that project but it suffers from under-representation.  and, quite frankly, I think while the ideas are sound their tools are a bit afunctional in many cases01:11
ThePubalthough, sleep first.  really have to get on a good and reliable schedule of that.  prefer staying off the meds if it just means regular sleep :)01:12
Romsteri have a similer idea to monitor what desn't get used much and pkgrm it off the system but monitor if later on it's needed then pkgadd it.01:17
Romsterbut all that doesn't follow the KISS idea01:17
ThePubno, it doesn't01:17
Romstertilma_n and ju_u would beg to differ.01:18
ThePubbut I think it can be implemented using the KISS ideals, it just wouldn't follow the CRUX concepts anymore.01:18
Romsterthey would say well it don't use much disk space, disks keep getting bigger, if you need to prune you do that manually.01:19
Romster  VG   #PV #LV #SN Attr   VSize VFree01:19
Romster  data   3  11   0 wz--n- 3.63t 1.09t01:19
Romsteri have a bit of space now, not counting the space in the FS too.01:20
ThePubI have too much space.  a whole 1tb drive that never gets touched01:20
ThePubeven my main driver which is only 300gb, really only has crux, swap space, and my home folder on, which is a "whopping" 20gb *yawn*01:21
Romsteri can never have enougth space01:21
ThePubI just don't keep stuff around.  If it came from the cloud, it can stay imo.01:22
ThePubeverything else of mine is usually a text file.. so that takes hardly anything.01:22
Romsterdistfiles alone take up around 42G01:23
*** ross49 has quit IRC01:26
*** mike_k has joined #crux02:21
cruxbot[opt.git/2.7]: mysql: update to 5.5.1203:10
*** childintime has joined #crux04:19
*** pedja has quit IRC04:52
*** deus_ex has joined #crux05:21
*** deus_ex is now known as pedja05:23
*** Evil_Bob has joined #crux05:47
*** lasso has joined #crux06:44
*** prologic has quit IRC07:31
*** pitillo has quit IRC07:57
*** pitillo has joined #crux07:59
*** pitillo has quit IRC09:02
*** pitillo has joined #crux09:04
*** aubic has quit IRC11:00
*** aubic has joined #crux11:03
*** Evil_Bob has quit IRC12:28
Rotwangopening ssh default port for world access is not the best idea12:58
Rotwangthere seems to be lots of zombie computers out there12:58
*** pedja has quit IRC13:08
*** deus_ex has joined #crux13:11
*** andarius has joined #crux13:35
andariusgreetings and salutations13:36
*** jse has joined #crux14:08
*** conbot has joined #crux14:09
conbotI tried editing lilo.conf, but another vfs problem occured. I even installed the ata drivers; nothing was different when i booted again.14:13
*** mike_k has quit IRC14:14
conbotill use google14:16
*** conbot has left #crux14:16
jseconbot strikes again...14:16
*** Evil_Bob has joined #crux14:42
Evil_Bobjust wanted to note download of contrib/ntfs-3g gives a 40414:43
Evil_Bobalso ntfs-3g and ntfsprogs is merged:
Evil_Bobso opt/ntfsprogs can probably be replaced / updated14:44
Evil_Bobill file a bug report15:05
*** ephemeros has joined #crux15:40
*** jse has quit IRC16:14
*** Evil_Bob has quit IRC16:30
*** lennart has quit IRC16:35
*** lennart has joined #crux16:36
*** prologic has joined #crux17:22
*** prologic has quit IRC17:22
*** prologic has joined #crux17:22
*** andarius has quit IRC17:31
*** lasso has quit IRC18:11
ThePubso what about deleting /usr18:20
ThePubinstall everything in /18:20
prologicthat's precisely what I want to know too18:30
prologicwhy on earth ? :)18:30
prologicI mean you can18:31
prologicbut why ?18:31
ThePubthis really only counts for a local device18:36
ThePubin those cases what is the purposeful seperation of /bin and /usr/bin?  or /sbin and /usr/sbin.  ultimately they're just all programs which run on the local machine, why are we seperating them all out?18:37
ThePubyou could even go so far as to say, "if it's executable put it in /bin, permissions will determine who can execute it"18:38
*** ephemeros has quit IRC18:43
entewe don't need /opt and /usr/local.. pardon /local either18:57
enteand /home/user could be /user for single user machines, after all root isn't in /home either (because it might not be mounted)18:57
entebut why?18:58
joacimresistance from people who think changing the directory layout is heresy18:58
prologicyou can go and change the directory layout all you want19:02
prologicbut no distro is going to adopt such a new layout19:02
prologicbecause there are standards we like to follow19:02
joacimi dont think i know anyone who likes those standards19:05
entewell, most of them are historical19:05
entehaving a separate /usr does make sense though19:05
enteis IMO for things required to bring the system up to a usable state19:06
enteI mean /19:06
entebut whatever19:07
joacimfound some hate against homebrew (a package manager for OSX)19:07
enteyou could go the gobo way and make a separate /Programs/bash-4.x/... for every program19:07
joacimit likes to keep /usr/local writable by all19:07
ThePubente: that is actually the historical reason for / too, no?  "/" is usually what is local while /usr could be mounted later.19:07
enteif I had a mac, I'd probably use pkgsrc19:07
enteThePub: yes19:08
joacimit is optional ofc. but people didnt like that everyone had as much access to /usr/local as they did to /Applications19:08
entethe historical reason for /usr (and the reason for the name) is that /usr was intended for users19:08
enteand in the beginning also contained user home directories19:08
entei.e. /usr/bin was probably the home directory of "bin" at some point19:09
joacimbin laden is dead now19:09
enteearly UNIX even had a man user, I could imagine /bin/man being setuid man (to keep a cache of cat-pages)19:09
* ente shuts up19:09
ThePubI was re-reading the stali page and thought I agreed with the statement, "Generally, /usr will be removed, what a useless directory, for non-base system stuff we might consider /local"19:11
entewhat benefits do you get from doing so, other than "I do it different than everyone else" and breaking people's assumptions about $PATH?19:12
ThePubactually, how does it modify $PATH?19:12
ThePubif you're use a user /bin is in your path, and if you're root then you have it too.19:12
enteit does not modify $PATH19:13
entebut it breaks scripts expecting things in /usr/bin19:13
enteexcept if you symlink /usr to /19:13
ThePubthat cements the point that extra directories for a straight-up local installation are pointless doesn't it?  :)19:13
enteI don't get it19:14
joacimthose scripts are bad then =)19:14
enteenv is always /usr/bin/env19:14
joacimthey shouldnt break due to assumptions about the path19:15
entethat's how #! works19:15
enteit only works with absolute paths19:15
entebut anyway19:15
enteyou don't care about anything I'm saying19:15
entejust do it19:15
joacimlots of stuff dont want to compile on a mac because the scripts cant find things where linuxes usually put them19:15
entethere's more important stuff to do19:15
thrice`what is #!?19:15
entethrice`: head -1 /bin/zcat19:16
thrice`ah, that :>19:16
ThePubente: I just think it's an interesting thought, it isn't personal!  ^_^19:17
joacimwho dont care about anytnhing you're saying?19:17
entenobody does :P19:17
joacimofc they care, just have a different opinion =)19:17
enteI think there's more important things than refactoring the file system standards. init as well.19:17
joacimi'm sure flamewars on irc are caused by irc itself =)19:17
thrice`what do you find more important? :)19:18
joacimcant really discuss anything with short blurbs19:18
entethat AJAX bloat growing over our heads19:18
entepeople better bring some sanity into the web instead of their init system19:18
joacimi guess it is more important to make another layer on top of alsa instead of using oss, or making another hal =)19:19
entewhy another layer?19:19
joacimisnt that what PA is?19:19
enteI am not informed about PA, so I will not make any statement :P19:19
entegenerally, I have no idea about audio programming19:20
entepeople get pissed when you make false statements, so I try not to make any19:20
joacimindeed =)19:20
joacimi always get my panties in a twist when someone says macs are more expensive19:20
entethey are19:21
joacimcompared to what =)19:21
entecompared to my 3rd-hand computers from 200219:21
joacimi wouldnt mind one of those19:21
joacimwould still be too fast for grim fandango maybe19:21
entethese are incredibly cheap and have enough computational power19:21
enteI rarely buy new hardware19:22
joacim4-5 years since my last shopping spree19:22
joacimstarting to feel the need for a bit more ram19:22
*** childintime has quit IRC19:31
ThePubcompared to my emachines macs are more expensive ;)19:33
ThePubplus I can upgrade my emachine and am not really forced into the next iterative apple cycle19:33
entejoacim: how much more is a bit more?19:33
* ThePub used to be an Apple VAR19:33
enteI dislike apple, but they probably produce relatively high-quality hardware19:33
enteit doesn't need to be that expensive though19:33
enteapple has always done that19:34
joacimi have 2GB now, another 2GB should be have me covered for another few years19:34
entesince the very beginning19:34
joacimi have to restart my browser once in a while because it loves ram and never gives it up for other applications19:34
ThePubyou're paying for the experience, not the hardware or software but the combination.19:34
enteI have 1GB at most in all my machines and they do not feel slow19:34
entebrowsers are memory hogs, that's true19:34
entebut the reason it is this way is once again (surprise) software bloat19:35
entethat's what I meant by bringing sanity to the web19:35
enteeverything interfacing it is broken19:35
enteeither utterly broken or completely unusable19:35
joacimparts sourced from newegg would give me a computer that performs the same as a mac mini, but also at the same price of a mac mini19:35
joacimit really depends on what you want19:36
ThePubyou can buy a better emachine for $300, just in a larger package ;)19:36
entewhat I want is breaking Wirth's law. We can do so much better than that19:36
joacimi doubt it19:36
entewhat exactly do you doubt?19:37
joacimsame goes for the imacs, show me a similar i5 machine with a 27" display =)19:37
joacimthat the 300 USD emachine is better ;)19:37
entedisplays are cheap19:37
joacimdells 27" display was at the same price point as apples 27" display19:38
joacimsame panel, different bezel19:38
enteI remember when I bought my first 17" screen, and it was like 370€19:38
enteit has an s-video in though, which is missing at most newer displays :P19:39
enteI can hook my C64 up without any adapters :P19:39
joacimis hard enough to find a tv with both hdmi/dvi and analog component video19:40
enteI guess I'm another 10 years away from HDMI19:40
entei.e. when it's successor is released19:40
joacimi quite like displayport =)19:40
enteI have heard of displayport twice or so19:41
entejoacim: macfag :P19:41
joacimvesa-fag in this case =)19:41
jaegermaybe thunderbolt :)19:43
enteI dislike the optional encryption "feature" in both HDMI and DisplayPort19:43
entethis seems highly redundant19:44
joacimsome people have a use for it i guess. but i dont think anyone has a good reason to add such encryption to consumer electronics19:45
jaegertheir reason is to protect their content, probably nothing more19:46
ThePubidk, lookup the emachine stats.  they're pretty good for all the desktops.19:47
ThePubonly difference of course is Win7 vs osX19:48
ThePubbut we run Linux, right?  :)19:48
joacimoh good lord. they used makepasswd to name their computers :(19:48
ThePubmy ET116105 has served me well, lol19:49
joacimthey dont really have c2d or i5-based computers19:51
ThePubit's all amd stuff19:51
ThePubso there isn't any direct comparison to apple's intel offerings19:51
joacimthats how i compare. by compiling a replica of a mac through newegg or dell =)19:51
ThePubunfortunately if the only comparison is the exact same hardware, piece for piece, then no vendor option will compare to apple.19:55
joacimi think cost is still the best argumetn against getting a mac19:55
joacimi know a few people whose main gripe with macs are that they are pretty19:55
joacimthey hate that19:55
joacimthey think being pretty means theyre bad at something19:55
joacimlots of simple and high quality items are pretty, i'd say most are.19:56
ThePubarguing against beauty is always stupid :)19:57
jaegerhrmm... I can't build nfs-utils at the moment, it seems to think it's missing libtirpc, which is installed19:59
jaegeranyone else run into this?19:59
jaegerchecking for clnt_tli_create in -ltirpc... no20:00
jaegerconfigure: error: libtirpc not found.20:00
jaegerah, this looks more helpful: /usr/lib/gcc/i686-pc-linux-gnu/4.5.2/../../../ undefined reference to `_des_crypt_call'20:00
cruxbot[opt.git/2.7]: nss: updated to 3.12.1020:03
cruxbot[opt.git/2.7]: nspr: updated to 4.8.820:03
*** prologic_ has joined #crux20:38
*** prologic has quit IRC20:42
*** Dudde has quit IRC21:34
*** Dudde has joined #crux21:35
RomsterRotwang> opening ssh default port for world access is not the best idea <-fail2ban or denyhosts with iptables -m recent is a good idea.22:03
Romsterjaeger, recompile revdep?22:29
Romsteri dunno why ntfsprogs is in opt for only dep for contrib/testdisk22:33
*** dinx has joined #crux22:39
jaegerthat was the first thing I checked, linkage seems fine22:52
jaegerdoes it build properly for you?22:53
Romsteror perhaps some lib is too old and need a version bump22:53
Romsteri haven't tried...22:53
Romstertrying it now.22:53
Romsterin a crux chroot.22:53
Romsteri still want to make a buildbot thing for every git commit. to test for breakage and report it.22:56
Romsterbut if i ever get to doing it is another question.22:56
Romster=======> Building '/usr/ports/packages/nfs-utils#1.2.3-1.pkg.tar.gz' succeeded.22:57
Romsterprt-get: installing nfs-utils 1.2.3-122:57
Romsterfine here i'd sugest rebuilding the dep tree to it.22:58
jaegerI did that already, too22:58
Romsterheh what could it be then...22:58
jaegersometimes I hate computers22:58
jaegernss doesn't build, either, lovely22:59
Romsterpkginfo -i |egrep '(libtirpc|libnfsidmap|libevent|rpcbind|nfs-utils)' |xargs22:59
Romsterlibevent 2.0.11-1 libnfsidmap 0.24-1 libtirpc 0.2.1-1 nfs-utils 1.2.3-1 rpcbind 0.2.0-122:59
Romstertry those versions.22:59
Romsteri don't think i've bumped that chroot in a week23:00
jaegerI can't even find where _des_crypt_call is defined23:02
cruxbot[contrib.git/2.7]: ntfs-3g_ntfsprogs: renamed from ntfs-3g, 2010.8.8 -> 2011.4.1223:06
cruxbot[contrib.git/2.7]: testdisk: update changed dependency, ntfs-3g -> ntfs-3g_ntfsprogs23:06
Romsteri don't see that either anywhere.23:12
Romsterperhaps libtool or pkg-config files are refferencing something old.23:13
Romsteri'd okgrm them ports and then try to rebuild them again23:13
Romstersometimes i wish computers were never invented.23:28
*** dinx has quit IRC23:28
Romsterjaeger, i got an idea i could upload these packages and you could pkgadd them then try to rebuild them all then?23:32
Romsteronly thing is i've bumped gcc on that chroot to 4.5.3 but i don't think that would be any issues.23:32
jaegerI'm gonna mess with it more tomorrow, too late for now23:37
jaegerthanks, though23:37
cruxbot[contrib.git/2.7]: libva: 1.0.6 -> 1.0.1223:51
*** dinx has joined #crux23:54

Generated by 2.11.0 by Marius Gedminas - find it at!