IRC Logs for #crux Saturday, 2012-03-03

*** jdolan has joined #crux00:43
*** ChanServ sets mode: +o jdolan00:43
*** nogagplz has joined #crux01:05
Romsterout of pascal swirls buy more pascal swirls02:00
*** j^2 has quit IRC02:14
*** j^2 has joined #crux02:14
*** vaddi has joined #crux02:53
*** vaddi has quit IRC03:09
cruxbot[opt.git/2.7]: epdfview: fixed bug #796.04:04
cruxbot[opt.git/2.7]: [notify] libvorbis: fixed buffer overflow.04:04
cruxbot[opt.git/2.7]: freeglut: updated to 2.8.0.04:04
cruxbot[opt.git/2.7]: zsh: updated to 4.3.17.04:12
cruxbot[opt.git/2.7]: vorbisgain: fixed the build.04:20
*** Rotwang has joined #crux04:24
cruxbot[opt.git/2.7]: vorbisgain: updated to 0.36.04:37
*** Rotwang has quit IRC04:45
*** jdolan has quit IRC04:47
*** Jessica_Lily has quit IRC04:51
*** Jessica_Lily has joined #crux04:52
*** Jessica_Lily has joined #crux04:52
*** vaddi has joined #crux05:25
*** mike_k has joined #crux05:50
*** prologic has joined #crux06:06
*** prologic has joined #crux06:06
frinnsthttp://arstechnica.com/science/news/2012/03/robots-swarm-the-stage-at-ted.ars06:07
frinnstpretty fucking awesome06:07
teK_you seen terminator, right?06:07
frinnsthaha, dont be a downer06:07
teK_i'm not06:08
teK_and the talk even is about autonomous robots06:08
teK_didn't see that06:08
teK_what could possibly go wrong06:08
teK_they opened german airspace for drones (military + police)06:09
teK_yay!06:09
teK_but they are cool, no doubt :-)06:17
Romsterterminator like more than 5 times on 1 2 and 306:20
mike_kwhile there are some of you here... why don't we have SYS_GID_MIN/SYS_GID_MAX, etc in /etc/login.defs? seems handy to use with groupadd/useradd in pre/post-install scripts07:03
*** jdolan has joined #crux07:10
*** ChanServ sets mode: +o jdolan07:10
frinnstwhat io-scheduler are you guys using on your desktop?07:11
*** jse has joined #crux07:50
juemike_k: you've asked that already a few days ago, I guess you missed my reply the last time?07:59
mike_kjue, yep. sorry. I'll search the backlog08:00
juemike_k: here it is:08:01
jueFeb 29 18:00:54 <jue> mike_k: no idea, but as always, suggestions and patches are greatly welcome ;)08:01
mike_kjue: ah, thanks. I'll use bugtracker08:01
juemaybe it would be better to use the default login.defs that comes with shadow instead of our own?08:03
mike_kI've never touched it, and those missing defaults are the only thing I'd like to change.08:07
mike_kthe upstream's one is too verbose, as for me. and as we already have changed the upstream's defaults, it would be better to stick to our current values. I'd only add a few.08:10
cruxbot[contrib.git/2.7]: exiftool: update to 8.8008:14
cruxbot[contrib.git/2.7]: gajim: update to 0.14.408:21
juemike_k: ok, awaiting your report in our tracker ;)08:25
juebbl08:25
cruxbot[contrib.git/2.7]: jedit: update to 4.5.008:31
*** pitillo has quit IRC08:57
*** pitillo has joined #crux08:59
jaegerfrinnst: noop and cfq09:21
jaegerfrinnst: er, I take that back, I used noop exclusively09:32
entehttp://www.youtube.com/watch?v=qhKUJrsFX5I here, enjoy some quality music09:35
mike_kdoes "attach a file" link have disappeared from bugtracker?09:54
tilmanente: are you moritz or am i confusing with another guy?09:58
enteno, that's me09:59
tilmanhave you ever tried to build/port the haskell platform?09:59
enteno, just what's in the haskell collection10:00
tilmanok10:00
cruxbot[contrib.git/2.7]: logrotate: update to 3.8.110:30
*** Rotwang has joined #crux10:32
*** Rotwang has joined #crux10:32
*** mike_k has quit IRC10:47
*** Lily has joined #crux10:51
*** Jessica_Lily has quit IRC10:53
*** Lily is now known as Jessica_Lily10:53
*** Jessica_Lily has joined #crux10:53
*** jdolan has quit IRC11:14
tilmanbuilding ghc makes me want to build a new computer11:21
teK_:X11:21
tilmanbeen running for 1.5 hours11:21
thrice`:D11:22
tilmanchances are it's a huge waste of time/power because i probably won't accomplish anything w/ haskell after all :p11:22
thrice`gotta start somewhere :-)11:24
tilman=======> Build result:11:27
tilmanfinished :o11:27
enteI remember ghc building quite fast in comparison to some other stuff11:37
entemind you, my machine was a 2GHz P411:37
teK_it's a sign11:38
entegcc, for instance11:38
entebut I might remember that wrong11:39
enteat least it was faster than firefox :-)11:39
tilmanente: dunno, i build gcc very rarely11:39
enteI still have some binary packages for that haskell stuff :-)11:39
entethey're old and probably don't work anymore11:40
frinnstjaeger: yeah, cfq fucking sucks11:47
jsebut, but the tin says completely fair!11:47
frinnstit worked ok with less RAM11:47
jaegerfrinnst: all of my drives are either SSD or hardware RAID so the logic is that the controller knows best :)11:47
frinnstjse: some processes are more equal than others :)11:48
frinnstdunno, but since i put in 16gigs, it got worse11:48
jseThat phrase is so tiring to hear.11:48
*** j^2 has quit IRC12:22
*** nthwyatt has quit IRC12:26
*** nthwyatt has joined #crux12:26
*** j^2 has joined #crux12:29
*** Rotwang has quit IRC12:31
*** j^2 has quit IRC12:42
*** j^2 has joined #crux12:58
*** lasso has joined #crux13:49
*** lasso has quit IRC14:33
* frinnst is having an altbier called "control alt delete" :)14:38
*** Jessica_Lily has quit IRC14:42
*** Jessica_Lily has joined #crux14:46
*** Jessica_Lily has joined #crux14:46
*** vaddi has quit IRC15:26
*** jdolan has joined #crux15:36
*** ChanServ sets mode: +o jdolan15:36
teK_encrypted youtube? O_o15:53
Romsterfrinnst, SFQ17:35
Romstertilman, i got ghc in my romster repo it builds17:37
*** tilman has quit IRC17:53
*** tilman has joined #crux17:54
*** ChanServ sets mode: +o tilman17:54
frinnstRomster: sfq?18:04
tilmanRomster: yes. it's outdated tho btw18:13
tilmanRomster: 'haskell platform' is ghc+lots of often needed libraries18:13
joe9but, once you have any version ghc, you can just build ghc from source18:13
joe9I do that.18:13
tilmanjoe9: that's what i did. just saying18:13
joe9oh, ok.18:14
Romsterjoe9, i made it bootstrap http://romster.dyndns.org/linux/ports/crux/romster/ghc/Pkgfile18:18
Romsterbut i haven't done the other libs that make the platform18:19
joe9i think once you have ghc, the others can be installed pretty easily.18:19
joe9I doubt there are enough haskell users on crux to justify putting the other ports.18:20
Romsteryeah unless i come across some nice program that needs those18:20
Romsteri did save wilhelmy-ports.git if i should need it in future18:23
*** jse has quit IRC18:43
*** jdolan has quit IRC19:09
*** vaddi has joined #crux19:31
Romsterbumping ghc once the files finish downloading.19:31
* nogagplz bumps Romster out of a window19:32
Romstertilman, libpthread-stubs mesa3d xorg-font-alias xorg-font-util xorg-glproto xorg-libpixman xorg-libxcb xorg-libxcursor xorg-libxi xorg-libxrandr xorg-mkfontdir xorg-mkfontscale xorg-server xorg-sessreg xorg-util-macros xorg-xcb-proto xorg-xcb-util xorg-xcompmgr xorg-xdpyinfo xorg-xev xorg-xinit xorg-xinput xorg-xprop xorg-xrandr xorg-xwininfo need version bumps if your not too busy sometime. trying to make it easier for everyone to see what needs bum19:41
Romsterping unless there is a reason why your holding off due to some bug or something? source is from http://romster.dyndns.org/version_sort/up2date-xorg-Tilman-Sauerbeck19:41
Romsternogagplz, i submitted emulators and nogagplz to portsdb19:43
Romstertilman, if it's too much work for you how about turning xorg into a group of a few active devs to keep things up to date?19:45
joe9rmull: are you planning a release of ghc 7.4?20:01
joe9Romster: planning on bumping ghc to release 7.4?20:01
joe97.4.1, to be precise?20:02
Romsteryep bumping as we speak waiting for the files to finish downloading.20:03
joe9ok, cool. thanks. I am also trying to see if it builds.20:03
Romsteralmost done downloading20:03
Romsteri'm fetching the bootstrap binary too and testing in 64bit and 32bit chroots20:04
joe9oh, ok.20:04
Romsterwonder how many are on 32bit still?20:04
joe9I think I am on 32 bit.20:05
Romsteri know only a few are on multilib including myself and the rest be on pure 64bit20:05
Romstermy firewall pc is on 32bit but i plan to migrate that later when i get around to it.20:05
joe9http://codepad.org/PNRftR6v20:06
joe9you probably consider the machine to be ancienty.20:06
joe9s/y/.20:06
Romsterbzip2: Compressed file ends unexpectedly;20:07
Romsterperhaps it is corrupted?  *Possible* reason follows.20:07
Romsterwhat the hell20:07
joe9i seems to be running fine on mine.20:07
RomsterCeleron those CPU's are so gutless i see a few at work and theya re so slow for there GHz clock speeds.20:10
RomsterI know my phenom II isn't quite as good as a I7 but it uses less power and isn't far behind a I720:11
joe9haha, best, I got. will upgrade at some point in the next year. but, works fine now. so, why bother. except firefox with too many tabs.20:11
Romsterthat would make a good web server/ppp firewall box though20:11
joe9do you know if "cabal install world" works after a ghc upgrade?20:11
Romsterbut as a desktop it's like driving a 4 cylinder car around20:12
joe9haha..20:12
Romsterno idea20:12
joe9ghc build working fine?20:12
Romsterseriosuly even a pentium D was better than a celelron20:12
Romstercorrection is better20:12
*** BitPuffin has joined #crux20:12
Romsteryou could change the CPU if you got your hands on one.20:12
BitPuffinHey!20:12
joe9I bought it used for around $50, I think.20:12
joe9I will, in the next few months or so.20:13
BitPuffinI'm interested in becoming a crux user some day20:13
joe9BitPuffin: try it now.20:13
Romsterhi BitPuffin20:13
joe9it is cool. will be enlightening.20:13
BitPuffinYeah I think it would give me a better understanding of linux and all20:13
BitPuffinright now I'm on arch for that reason20:13
Romstersocket 775 joe9 ?20:13
joe9Romster: i think so.20:13
Romsterhttp://www.ebay.com.au/itm/Intel-Pentium-D-945-CPU-3-4GHz-800-4M-Socket-775-SL9QQ-/140598875836?pt=AU_Components&hash=item20bc5896bc20:14
BitPuffinCRUX seems like you learn even more about how things work20:14
Romsterworthwhile upgrade on that.20:14
RomsterBitPuffin, or forced to learn as there is no hand holding.20:14
BitPuffinYeah20:14
joe9ok, thanks. I feel speed to be an issue when I use firefox, ghc or uzbl-browser.20:14
BitPuffinEven Arch has some hand holding20:14
joe9but, other than that, no issues at all.20:15
BitPuffinI guess I'd have to compile kde from scratch to get 4.8 haha20:15
joe9AU 29 is pretty cheap.20:15
joe9BitPuffin: still on kde? ditch gnome and kde. use xmonad or some such window manager.20:16
BitPuffinI've thought about that20:16
BitPuffinBut I just got in to kde and it's got a nice ecosystem20:16
jaegerhere's an idea: use whichever you like best, not that which an IRC channel tells you to :)20:16
joe9i started off on kde almost a decade ago. moved to gnome. and now, am happy with none of them polluting my workspace.20:16
joe9BitPuffin: I agree. jaeger's is the voice of wisdom.20:17
BitPuffinHaha seems like it20:17
jaegerIt's just that it's a completely subjective choice20:17
BitPuffinActivities are nice too20:17
BitPuffinWhen looking through the ports, how do I see what version is supported (in web browser)20:18
Romsterjoe9, and a decent amount of ram?20:19
BitPuffinAnother fun question, is anyone running crux as a server?20:20
RomsterBitPuffin, yeah arch has some hand holding it's more linux from scratch with a bolt on package manager here, plus i don't think kde is maintained currently. nore is gnome, xfce or some window manager is what we all use or not even use xorg at all.20:20
Romsteryes i have crux as a home server/gateway20:20
Romsteryou can taylor crux to any specific task20:21
BitPuffinAh I see, well if I jump into crux you'll see KDE and Blender being updated frequently :)20:21
joe9Romster: i have 2GB in RAM.20:21
BitPuffinat least blender20:21
joe9i think that should be good.20:21
Romstertar -tvv foo.tar.bz2 seems to take forever to test.20:21
joe9I think the cpu is the bad one in my case.20:21
BitPuffinWould be cool to throw crux on my linode somehow..20:21
Romsterjoe9, that's not bad yeah i'd defently make sure it's a socket 775 and grab that cpu20:22
Romsteris linode using vserver?20:22
BitPuffinNot sure20:22
BitPuffin"Xen hosting"20:22
Romsterprologic, is making a vserver image IIRC20:23
jaegerIt would be pretty easy to put crux on a linode server20:23
Romsterjoe9, some nice cpu compound like artic silver too20:23
Romsteri use that on all my systems20:24
Romsterjust watch out as it is conductive20:24
joe9oh, really, that makes a difference?20:24
Romsterelectronicly20:24
Romsterwell depends how hot is it currently?20:24
BitPuffinI'd totally wanna run it then haha20:24
joe9i am just trying to find what cpu socket my system has, without having to open it up.20:24
Romsterto it's max cpu temp.20:24
joe9PLGA77520:25
Romsteryep 77520:25
jaegerhttp://library.linode.com/linode-platform/custom-instances/custom-distro-howto <-- one method20:25
Romsterjaeger, how did you get the crux-multilib/* in gitweb is that same layout in gitosis or do you just have gitosis/{opt,core,xorg}.git gitosis/ports.git and then you git clone them to there respective directories in www/....20:27
BitPuffinneat!20:27
joe9romster, what do you think of the "Pentium Dual Core" cpu's?20:28
BitPuffindoes crux use udev?20:29
*** Rotwang has joined #crux20:29
Romsterjaeger, this is what i do currently in post-receive hook http://pastebin.com/SkUZAQRg20:29
Romsterjoe9, even better if your motherboard can handle it. look up your motherboards website check cpu compatibility to your bios version20:30
jaegerBitPuffin: yes, it does20:30
joe9Romster: ok, thanks.20:30
Romsteri had to flash my bios to use this cpu i got now.20:30
jaegerRomster: gitweb needed no special configuration for the layout, I just pointed it to the repositories directory20:30
BitPuffinallright20:30
joe9haha..20:30
Romsterjoe9, be sure to look at the revision number of the motherboard20:30
joe9oh, ok.20:30
Romsterif you really can't find the label fire up hirens cd or windows and look at cpuz20:31
Romsterit has a motherboard tab should list the details.20:31
Romsterbut not all boards will give the details.20:31
joe9i think I can. just lazy to open up the box.20:31
joe9ghc still building on your m/c?20:32
Romsterjaeger, so in your setup you haven't got all your gitosis repos in one directory?20:32
BitPuffinCRUX is a lo like freebsd ain't it?20:32
RomsterBitPuffin, it borrows bsd like init system that's about all20:32
jaegerRomster: I don't understand what you're asking; they're all in a repositories directory under the vcs user's homeidr20:32
jaegerhomedir20:32
BitPuffin+ a ports system!20:33
jaegerports could be said to be rather similar20:33
jaegerthere are differences but the idea is similar20:33
Romsterok i have20:33
Romsterls /home/git/repositories/20:33
Romstercrux-fixups.git  emulators.gitgitolite-admin.git  nogagplz.git  romster.git  testing.git20:33
Romsterbut you have sub-directories but i see in the gitolite that they all reside in /home/git/repositories/20:33
jaegerAre you asking about the nesting?20:33
Romstercorrect20:33
jaegerAh, you can make subdirs20:33
jaegergitosis and gitweb have no trouble with them, nor do they need any special configuration. I assume gitolite is the same, though I don't know that for sure20:34
Romsteri'm thinking of separating the 64bit and 32bit by using a sub-directory20:34
Romsterthough in my config i have in ~/gitolite-admin/conf/gitolite.conf20:35
Romsterrepo    romster20:35
Romster        RW+     =   DannyRawlins20:35
Romsterso don't know if i can do repo foo/romster20:35
jaegergitolite's config is definitely different so I can't say... try it and see20:38
Romsterah i just talked to a person in #gitolite and he just tested that yes i can do repo foo/bar20:39
Romsterto nest.20:39
jaegergood deal20:40
*** vaddi1 has joined #crux20:42
Romsteri didn't see it in the manual so i was wondering.20:45
Romsterelse i'm gonna confuse 32 and 64bit users on ports trees.20:45
Romsteri like your nesting better20:45
*** vaddi has quit IRC20:46
Romsteri'm also doing the crux-fixes wrong, going to move that to fixes/{core,opt,xorg,contrib} for my -flto stuff then i can fork them edit and commit locally but later when there is updates i can merge them in, less effort for me.20:47
BitPuffinHahahahah this distro is sweet. Okay I should definitely try it out20:50
Romsterdon't be suprised if it fails to boot the first time it took me quite a few turns to get it to work20:53
BitPuffinlike if I have more than one cd drive?20:54
Romsteralso if your keen you can try the unoffical pure 64bit or even the multilib20:54
BitPuffinor how do ya mean20:54
BitPuffinI think I'm aiming for 64bit20:54
BitPuffindon't see any multilib though20:55
Romsterhttp://jaeger.morpheus.net/linux/crux/tmp/20:55
*** Rotwang has quit IRC20:55
BitPuffininteresting20:55
Romsterit's expermental jaeger nogagplz and myself run it currently mostly for the reason of wanting 32bit for wine and some emulators20:55
BitPuffinhow does i differ from regular 64?20:55
jaegermultilib is unofficial but so far works well20:55
Romsteron the 64bit system, if oyu don't need that go with pure 64bit20:56
BitPuffinDo want wine20:56
BitPuffin:)20:56
Romsteronly difference is it has a gcc version compiled for multilib 32bit suport along wth 64bit support else it's the same a pure 64bit20:56
Romsteryou'll want multilib then :)20:57
Romsterit's stable i haven't had any major issues.20:57
Romsterexcept a few teething issues when it was first available.20:57
BitPuffinAllright20:57
BitPuffinlike what?20:57
Romsterjaeger, mind moving the multilib iso out of tmp and putting it with your updated 32bit iso type location and getting it added to contributed isos on crux.nu?20:58
Romsteroh nothing much when it was first done i had some compiling issues with CFLAGS etc on some ports, that's all been fixed now.20:58
jaegerIn the future I'd like to but at the moment still considering it experimental20:58
Romsterk20:58
BitPuffinI see20:59
Romsterif you hit a bug talk to jaeger or myself we both have commit access to multilib.20:59
Romsteroh i do see one issue jaeger we need to add prt-get port to core and add in the multilib repos to it's prt-get.conf21:00
Romsterelse you have to maually edit prt-get.conf to add the multilib repos.21:00
RomsterBitPuffin, one word of caution in etc/prt-get.conf don't set preferhigher to yes, you'll run into issues of non 64bit ports when you do a prt-get sysup. other than that i know of no other issues.21:01
BitPuffinpreferhigher means it.. goes for a newer version? or what?21:02
Romsterthan once you install it aditionally edit /etc/prt.get.conf and add http://paste.pocoo.org/show/560618/ thsoe misisng lines to it when you enable the contrib repo too.21:03
Romstercorrect21:03
Romsterthe order of prtdir in prt-get.conf will dictate which same name port gets installed.21:03
Romsterbut if preferhigher is yes, then any of them same name ports that has a higher version is preferred.21:04
BitPuffinI'll have to log this or something lol21:04
Romsterthat's fine for 32bit crux but for multilib and 64bit crux it will cause problems of breaking compiles.21:04
Romsterthat's the only known catch i know of.21:04
Romsterit is experiential for a reason :) just watch out for that and you'll be fine.21:05
BitPuffinI see21:06
BitPuffinI'll try and watch it then21:06
BitPuffinbut now I have to go to bed21:06
jaegerI've been idly thinking about a port repo aggregator to work around that21:06
BitPuffinsee you all!21:06
jaegerhaven't coded anything, though21:06
jaegertake care21:06
Romsterlater BitPuffin21:06
BitPuffinI'll be back soon21:06
BitPuffintake care21:06
*** BitPuffin has quit IRC21:06
Romsterjaeger, yeah i've been thinking about that too merge all ports into one prtdire for core opt xorg contrib etc...21:07
Romsterhope i didn't put him off multilib21:07
Romsterbut it's best to be aware of the gotchas before you start.21:07
Romsteroh i should of pointed out this is logged.21:07
Romsterfor his notes21:08
jaegerI would probably do it on the server side to make it easy for the client side21:08
Romsterif you make one i'd be keen on it for doing the same but to add in my fixes for -flto etc. for myself.21:09
Romsterand i'm sure frinnst could use it too.21:09
Romstermain problem to solve is not waiting for a set time to pass to update. like the older 5 minute cron job we had for git to rsync.21:10
Romsterused a git hook there, don't think there is any way around that though21:10
jaegerIt's a dead simple concept, really, I'd probably just do it in bash with rsync21:10
Romsterperhaps it could be triggered to repoll when someone tries to rsync from it? is that doable?21:11
jaegerIf a repo failed to update, scrap the entire process, SQL rollback style21:11
jaegermaybe with gamin or similar but that's far more involved than I was thinking21:11
Romstercheck for new updates on every rsync invocation or something with a rolling 1 minute timeout if someone else rsyncs in that 1 minute period it's extended to 1 minute again.21:12
Romsteronly an idea feel free to do it how you think21:13
Romsterbe cool if we get this BitPuffin on multilib too.21:14
jaegeryou know, maybe it's not even something that needs to be scripted. I have another idea21:14
Romsteroh why not a patch to prt-get.conf to handle this?21:14
jaegerI like the idea of the server doing it more than the client, less CPU time involved21:15
Romsterlike say group prtdir groups to not use preferhigher21:15
Romsterbut it's kind of a pina not being able to do cd /usr/ports/core ; pkgmk -r -u or something now.21:16
Romsteronly needed when you really mess something up :D21:16
jaegerThe more I think about it the more it looks like a perfect use case for unionfs/aufs21:17
Romsterhmm indeed that's not a bad idea.21:19
jaegerthe git hooks are already in place for the repos, no need to reinvent that wheel21:19
Romsterbut is that the crux way of doing it simple21:19
jaegerI don't care about simplicity on the server side, really :) The client side won't see it21:20
Romsteri would prefer client side rsync all them core opt xorg type repos into one tree for each like 32bit crux has now. actually forget patching prt-get why not have ports command merge it all for us?21:21
jaegerI try not to overcomplicate things but in thise case it makes more sense to me to have the work on the server's side of things21:21
Romsteryeah that's a better idea i got make a core xorg opt group line in httpup and rsync files and let ports merge it into one common /usr/ports/{core,opt,xorg}21:23
Romsterno polling time issues then.21:23
Romsterand /usr/bin/ports is shell so easy to do.21:24
Romsterjaeger, thoughts?21:26
Romsteror you still prefer server side.21:26
jaegerI see no valid reason to make every client do it, do you?21:29
jaegerIf there are 5 multilib installs that's 5 times as much processing21:30
jaegeretc.21:30
Romsternot really but then you don't have to deal with update issues like waiting 5 minutes for newly pushed updates to git.21:30
jaegerIf it's done on the server only it's just a normal rsync on the client21:30
Romsteri guess doing it server side is far simpler for the end user.21:31
Romsteri just hate that time lag part.21:31
jaegeryeah, that's what I'm getting at21:31
Romsterthat's the only reason i was thinking client side to avoid the time lag on update. but i guess i can live with that.21:32
Romstersmall price than more complexity on client side.21:33
Romstero i'm sold21:33
Romsterok*21:33
jaegerwell, it's a subjective thing, you could always do it on the client side if you really want to, I just prefer the simple user experience21:33
Romsterjaeger, ok server side why not merge in to one git tree or is that too much?21:34
Romsterthen git to rsync that21:34
Romsterforce merge in the correct order removing older files ah no forget that that's too much effort.21:35
jaegerThat seems like an unnecessary complexity to me21:35
Romsteri guess rsync them all remove duplicates rm -r foo in the correct priority order then merge into one foo.rsync21:36
jaegerI'm thinking aufs/unionfs them and then clients can just rsync21:37
Romsteri'l leave that to you to write, i got many things todo as it is now. unless your pressed for time then i can have a go at server side merging.21:37
jaegerwe'll see21:37
Romsterfusefs just seems unnecessary21:37
Romsterunless there is something like a dupefs that exists21:38
jaegeraufs isn't fuse, as far as I know21:38
jaegerI may be wrong21:39
Romsterhmm so you can add branch's to aufs21:41
jaegeryeah21:41
jaegerthe idea is to take 32-bit trees as the base, then add pure64 on top, then multilib on top of that21:41
Romsternever used aufs but that seems like a neat idea than to re-implement a step21:41
Romstersounds good and i can use your code on my server to overlay my crazy -flto on top of that again :D21:42
jaegerindeed21:42
jaegerI can't work on it quite yet but might be able to in a couple hours21:43
Romstersweet yeah do that i can put up with the cron 5 minute wait21:43
Romsterno rush this would also avoid having to add in prt-get to core to point to all these layers, just edit ports to take advantage of this new merged branch.21:44
Romsteryou used aufs before?21:45
Romsterif you have you got more experience than me at this.21:45
Romsterouch just hit my knee on my desk...21:46
*** __mavrick61 has quit IRC21:50
*** [2m]Dudde has quit IRC21:50
jaegerI used unionfs once for a livecd21:50
jaegerthe overlaid filesystems were compressed modules that installed various things21:51
jaegersorta like slax does it21:51
*** __mavrick61 has joined #crux21:51
*** [2m]Dudde has joined #crux21:51
*** jdolan has joined #crux21:59
*** ChanServ sets mode: +o jdolan21:59
Romsteryeah i'm aware of unionfs jsut haven't used it though but it's used on live cds22:00
jaegeraufs is basically its successor, I believe22:06
Romsterjoe9, ah i see where i went wrong they are building now.22:35
*** gb_away has quit IRC22:52
Romsterhahah oh i must be slow at that quit message :D22:56
*** gb_away has joined #crux22:57
Romstergcc: Internal error: Segmentation fault (program as)22:58
Romsteroh nice22:58
Romstertrying again it was probably some random race condition22:59

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!