IRC Logs for #crux-devel Thursday, 2014-01-09

*** mavrick61 has quit IRC03:25
*** mavrick61 has joined #crux-devel03:26
*** mavric268 has joined #crux-devel04:49
*** pitillo_ has joined #crux-devel04:50
*** mavrick61 has quit IRC04:55
*** pitillo has quit IRC04:55
*** Romster has joined #crux-devel08:10
pitillo_hey Romster, would be possible for you to add printproto and xorg-libxp to the xorg collection? (they are e18 deps)11:07
*** pitillo has joined #crux-devel11:07
Romsterpitillo, yeah i saw that earlier, i'll get to that shortly11:16
pitilloRomster: great, thank you11:19
pitilloteK_: seems the repo isn't exported to rsync (not listed in portdb neither)11:22
Romsterjust ate dinner and compiling11:30
pitillothank you Romster, there is no hurry. when you have some time, if you can take a look11:35
Romsterseriously this needs printproto?11:47
pitilloI think so... more deps added (gstreamer for emotion, the video backend, used by terminology too)11:56
pitillo# grep -i Print.h /var/log/pkgbuild/efl.log11:57
pitillochecking for X11/extensions/Print.h... yes11:57
pitillonice, thank you very much :)11:59
Romsteri had printproto in my romster repo goes to nuke that.11:59
Romsterlast i read it was depreciated11:59
pitilloyeah, I picked it from your repo12:00
Romsterhehe i think i got the widest selection...12:04
Romsterjsut been relaxing the past few days and gaming after work. but i do need to get serious on testing crux 3.1 so i got another hdd to install it too and tweak anything necessary.12:05
Romsterincluding testing lvm2 in udev rules.12:06
Romsterwhile not breaking my existing ssd crux 3.112:06
Romsterssd crux 3.012:06
Romsterand i always thought e17 was a DE too. I read it's a WM.12:07
pitilloit's always good to have time for all things and probably disconnect a bit between job and hobby stuff12:11
RomsterlibXfont bug O_O was added since 199112:12
Romsterand only fixed now in libXfont 1.4.712:12
pitilloit's considered a wm but it's more than that... and with a full support (not made in e18 by my side) it's nearest a DE than a WM from my pov... (econnman, bluz and linking with gstreamer make it close to DE than a WM)12:13
pitilloyeah... I was surprised when I readed the notify on ML... and when checked it... more surprised12:14
Romsteroh gstreamer as well no issues there?12:15
Romsterseems to heavy to be jsut a WM to me too.12:15
pitilloyeah... they tell it's a WM with lots of libs... probably what a DE is12:16
Romsterwell i guess it's a WM unless they have aplets that cover all the areas that a DE does.12:18
Romsterone of those technicalities12:19
pitillothat's the point... there are widgets which do those kind of things12:19
Romstersorry for the big word hope you can translate12:19
pitilloyeah, I think I undestand you right :)12:20
Romsteryou use arm too don't you pitillo12:20
pitilloyes, but on ARM I prefer to keep E17... E18 it's prettier and it's a bit weight for them (with fullHD res it's smoothly but a bit slow)12:21
Romsteri started writing some Pkgfile generation for spitting Pkgfiles with differences with ifdef blocks.12:21
pitilloyes, I've readed about it some weeks ago12:21
Romsteri haven't done much more yet but i plan too.12:21
pitillobut as sepen said, I prefer to keep all things splitted and more when it's covered by overlays... in this way, all things are sorted and Pkgfiles are easier to read (probably those ifdef blocks haven't sense on x86)12:22
Romstermainly want to deal with foo and foo-32 ports but can adopt to arm too.12:22
Romsteryeah but the generated Pkgfiles are what the user will see.12:23
Romsterthe script to spit them out is what i'll tweak when need be.12:23
Romsterand use it to auto generate foo and foo-32 ports at the same time.12:23
pitilloif we talk about x86 and x86_64 it has sense... but mixing more arches like ARM (lot of ABI's, from armv5 to future armv8) or PPC (which has 32b/64b too) will make Pkgfiles very big and probably will break one of the nicer things they have (simplicity and easy to read/understand)12:23
Romsterso i can spit out pure 64 32 and multilib Pkgfiles12:24
pitillodeeping in other arches can make that task in a big pain... for example on ARM there are some things unsorted yet making small differences on footprints... and this seems hard to handle without overlays12:25
Romsteryour missing the point, you commit the Pkgfiles that are generated. so parts that are the same are always in sync.12:25
Romsteri haven't dug into arm much yet.12:25
pitilloprobably I don't understand it right12:25
Romsteri wound't mind some arm hardware myself too.12:25
Romsteri'm basicly making a m4 type pre-processor12:26
Romsterbut not with m412:26
pitillonow it's pretty easy to get a checp ARM devices and it will be cheaper and easier in a future12:26
Romstermuch like a bunch of tools to make a configure script but i'm spitting out Pkgfiles.12:26
pitillosounds like a hard job12:27
Romsternot sure yet i only done a few so far.12:27
pitillooverlays are a very strong solution to maintain different arches12:27
Romsterit's just an experiment atm.12:28
pitilloprobably the bigger change between upstream ports and arm ones are footprints12:28
pitillo(toolchain components on a side)12:28
Romsterit's pretty much a overlay in one file that will spit out different flavours of the Pkgfile alone... patches footprints md5sums all still reside in there own repos12:28
Romsterya know half my problem is finding where i save everything...12:33
Romsteri so need a clean up12:33
pitillosort and organization...12:34
Romsteri'm basicly got some core 2 duros 3 of them i'll see if i can grab 2 more. then i'm gonna get a sas card for each of them and a sas enclosure. and fill them with wd green drives and setup moosefs over it.12:35
Romsterthen i'm gonna move all my stuff over to that.12:35
Romsterbut i've been trying to find a itx board that has sas or a pcie x8 socket at least for moosefs use.12:36
Romsterthe core2's be ok to get me started12:37
frinnstcareful with wd green12:37
Romsterbut i'm kinda looking for a 12 volt solution small board.12:37
frinnst(if you are going to use them 24/7)12:37
Romsteri got a bunch of them here now 24/7 running.12:38
pitillothat sounds great and funny... but wd green HD... not sure if it's a good choice12:38
Romsterin raid112:38
frinnstme too, but i'd use red12:38
pitilloummmm here I haven't good experience with them12:38
frinnstthere are some firmware limitations with the green ones12:38
Romsterthink reds be a better idea? i'm not using NAS or raid with that many hdd's12:38
Romsterand i'm aware of TLER12:39
frinnstoh, not gonna raid them? i misunderstood then12:39
Romsteri've only had 2 hdds fail out of about 1512:39
Romsteri got raid1 now with greens12:39
Romsterbut i ownt raid in moosefs12:39
frinnstyeah just disable the powersaving stuff then12:39
Romsteryeah i never let them park12:40
Romsterjournaled fs solves that as there is a limited park count.12:40
Romsterso i'm on the hunt for a itx board that's lower power but enough for moosefs and able to take a few gb of ram.12:41
Romsterbut in no rush as i can use what i got now. just looking for a good sas card.12:42
Romsterbut i'll start using the on board 4 sata sockets on each mobo to get me started12:42
Romsteri'm not looking for speed just mass storage.12:42
Romsterbasicly move the mass storage off my desktop12:43
Romsterand sort it out better12:43
Romsterof course i'll be running crux on each box.12:44
Romsterseems atmoms only come with pci slots.12:44
Romster dunno maybe something like that is my only option for pcie and at least 8gb ram each machine12:46
Romstertryin to avoid high end video waste of power wont be used.12:46
Romsterdual gigabit lan be interesting though for distcc moosefs use.12:47
Romsteri know arm can be very power efferent though12:48
frinnstis distcc bandwith intensive? i'd think it was more a latency issue12:48
Romstermostly latency12:48
Romsterbut dual lan for moosefs use would be interesting as most of these itc boards have 2 lans.12:48
Romstererr nics12:48
Romstershrugs dunno if it's worth messing with 2 seperate gigabit switches because i can.12:49
Romsterrouting would be a pain12:49
Romsterwas thinking one set for distcc the other for file transfers between moosefs nodes as it ballances itself.12:50
Romsterjust ideas atm.12:50
Romsterbut i do have machines here already on a switch to start my project off.12:51
Romstermaybe i should ask on the moosefs mailinglist on what sort of small boards they use too.12:53
Romsteronly x86 damn12:57
Romsterneed 64bit12:57
Romsterhas PCIe x8 so that's good12:58
Romsternone have power consumption ratings.12:58
Romstermore researching for the perfect mobo.13:02
Romster=======> ERROR: Footprint mismatch found:13:24
RomsterMISSING   -rwxr-xr-x      root/root       usr/lib/
Romster=======> ERROR: Building '/var/ports/packages/openssl#1.0.1f-1.pkg.tar.gz' failed.13:24
Romsteron every single one of my machines does the same thing.13:25
frinnstbah, doesnt look like a important file anyways13:54
frinnst=======> Building '/usr/ports/pkg/openssl#1.0.1f-1.pkg.tar.gz' succeeded.14:02
frinnstsilly distcc makeflags?14:03
Romsterhaha not important...14:11
Romsterleaves a dangling symlink too.14:12
Romsterjust -j2014:12
Romsternothing over the top and everything else compiles just fine.14:12
Romsterover 5 machines14:12
Romsterand that openssl has it hard fixed to -j1 anyways14:14
frinnstquite expensive :/14:37
jaegerI tried to buy a NUC yesterday, out of stock at a couple places14:38
jaegermy HTPC is dying, going to replace it with one of those14:39
Romsterlooks like a mac mini14:49
Romsterif only they had a 4x sas socket on the back14:51
frinnsthp microserver <314:58
frinnstno sas though14:58
Romsteri'l have to settle for a pcie slot for a card14:59
Romsteri5's though will use a bit of power. that thing be ok for a desktop15:00
jaegerIt uses a lot less than you might expect15:01
jaegerIt's not a fullblown desktop chipset15:02
frinnstRomster: i got the same ssl footprint error on my rpi21:12
frinnsti use ccache21:12
frinnstmaybe we need to override custom makeflags completely in openssl21:13
frinnstunless someone feels the urge to fix the bug21:13
frinnstgoddamnit, debugging build issues without ccache on rpi will suuuuuck21:18
jaegerccache and distcc have caused numerous problems when debugging, so annoying21:18
jaegerespecially when multilib development was going on21:18
frinnstdont think its a link or build issue, just sucky makefiles21:19
teK_Romster: did you answer Margaret's email?21:23
teK_we do have a 3.0 branch in englightenment, dont we?21:29
pitilloyeah, I've tested a commit on it yesterday21:30
teK_ @pitillo21:41
teK_case closed. :>21:42
pitilloteK_: perfect, thank you :)21:43
frinnstwtf, still failed with unset makeflags22:49

Generated by 2.11.0 by Marius Gedminas - find it at!