November 18, 2010 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >
November 18, 2010 [00:00:14] <nrubsig> alanc: No no.[00:00:18] <nrubsig> alanc: please calm down[00:00:26] <Meths> lblume: s/acting/standing/[00:00:37] <alanc> heres a ksh93 P1 bug: 6942280 ksh93 failure to write to process at other end of pipe[00:00:41] <nrubsig> alanc: I was just trying to figure out which bug you were referring to ?[00:01:01] <alanc> (I don't think that's the one that makes X builds randomly hang though)[00:01:17] <nrubsig> alanc: that has been fixed - see http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/tests/unfixed_tests/sun_solaris_cr_6942280_print_p_cannot_find_coprocess.sh[00:01:52] <nrubsig> alanc: the "unfixed_tests/"-directory refers to the part that it is not fixed in OS/Net[00:01:57] <nrubsig> alanc: Which random hang ?[00:02:22] *** ams_ has quit IRC[00:02:51] <alanc> sometimes the autoconf/automake/libtool scripts, when run with ksh93, get stuck waiting for a pipeline to complete, with defunct processes in the pipeline[00:03:16] <nrubsig> alanc: do you have a stacktrace floating around ?[00:03:26] <alanc> no[00:03:47] <alanc> don't see any stuck on our build machines at the moment, but they are rebooted every two weeks[00:03:53] *** glagasse has joined #illumos[00:04:38] *** glagasse has quit IRC[00:04:47] <nrubsig> alanc: The only two issues with being stuck with pipelines involved are the command substitution issue. The shell doesn't see the end of the pipe coming because |I_PEEK| randomly shows data from other processes pipes.[00:05:43] <nrubsig> alanc: ask kupfer[00:06:07] <nrubsig> alanc: basically we snoop on other pipelines, regardless whether they are in another zone or whatelse[00:06:14] * alanc has got to learn to not to try to give concrete examples to some people...[00:06:28] *** glagasse has joined #illumos[00:06:35] <nrubsig> alanc: ?![00:08:34] <alanc> I really wasn't trying to debug ksh93 right now, was just using it as an example of the difference between "____ has a bug right now that broke ___" and "____ ELIMINATED ALL SUPPORT FOR ____! THE HEARTLESS EVIL SLIMESUCKING BASTARDS!"[00:08:57] * lewellyn eliminates support for irc[00:09:35] <nrubsig> alanc: give me 15mins to lookup the rant in the cygwin users list.[00:09:44] *** vitaliy_gusev has quit IRC[00:17:43] *** gebi has quit IRC[00:18:08] *** Tenzer is now known as Tenzer|off[00:22:56] <jbk> i'm surprised that's not a higher priority issue[00:23:03] <jbk> since it sounds like it could effect trusted solaris as well[00:23:10] <nrubsig> can anyone please check on Solaris 11 Express what ksh93 says for $ ~/bin/ksh -c 'print "( compound -A a1=( [4]=( typeset -a x=( functions ) ) ) ; compound -A a2=( [4]=( typeset -a x=( ! ! ! alias ) ) ) )" | read -C c ; print -v c' 1>/dev/null # , please ?[00:23:14] <jbk> espeically if it's cross-non-global zones[00:23:29] <jbk> i would imagine some outfits might freak out on that :)[00:24:53] <alanc> nrubsig: /home/alanc/bin/ksh: Command not found.[00:25:04] <alanc> if I drop the ~[00:25:08] <alanc> 1[1]: read[1]: typeset: c.a1[4].x: compound assignment requires sub-variable name[00:26:50] <nrubsig> jbk: I've send comments about the issue to a couple of people at Sun and noone. Technically we would need a reduced testcase which doesn't drag-in half of ksh93 but somehow I'm not exactly keen to look again at the old command substitution code... it took a few hundred hours to rewrite it to avoid fifofs and somehow I'm not going to crawl into it unless someone waves with money.[00:27:36] <nrubsig> alanc: replace "functions" with "xfunctions" and try again.[00:27:42] <nrubsig> (please)[00:28:01] * nrubsig wants a function key for the word "please"[00:28:07] <alanc> nrubsig: no output[00:28:19] <nrubsig> ok[00:28:45] <nrubsig> so in some cases "read -C" tries to do keyword+alias expension[00:28:49] <nrubsig> hurray[00:28:54] <nrubsig> what a mess[00:29:02] * alanc goes back to ripping the CDE gate to shreds[00:29:12] <nrubsig> heh[00:29:34] *** samuelyounge has quit IRC[00:29:36] <alanc> sadly, the private copy of ksh93 survives this round[00:30:12] * nrubsig puts up a sign "!!HIRE ME!!"[00:31:51] <alanc> sorry, no budget for making dtksh suck less[00:32:59] <lewellyn> that sucks.[00:33:23] <lewellyn> it would be kind of neat to see /usr/bin/ksh93 have the dt bindings... ;)[00:33:42] <nrubsig> alanc: well, I could also do janitor work on OS/Net, do command modernisation, work down the list of open issues in i18n land (Ienup has lots of open issues), maybe I can look at fifofs or I could just do some performance work[00:34:44] <alanc> I don't think there's a janitorial budget, and you already know about the job openings in modernization[00:35:03] *** dnm has quit IRC[00:35:09] <alanc> I think performance work right now is centered on Exadata tuning[00:35:35] <nrubsig> alanc: I applied to those but my submissions have all been closed without further comments.[00:40:51] *** vitaliy_gusev has joined #illumos[00:52:01] *** vitaliy_gusev has quit IRC[00:53:39] *** xmikus01 has quit IRC[00:55:16] *** samuelyounge has joined #illumos[00:57:05] *** vitaliy_gusev has joined #illumos[01:02:39] *** glagasse has quit IRC[01:05:47] *** vitaliy_gusev has quit IRC[01:08:11] *** vitaliy_gusev has joined #illumos[01:09:47] <samuelyounge> Why does the illumos-gate build invoke javac? Are there java dependencies in illumos? If so what?[01:10:50] *** postwait has quit IRC[01:11:39] <e^ipi> http://src.opensolaris.org/source/search?q=javac&project=/onnv&path=/onnv/onnv-gate[01:12:52] *** ams_ has joined #illumos[01:13:19] <samuelyounge> Neat tool, I had not see that before[01:14:23] <e^ipi> it's a little inaccurate b/c it doesn't follow illumos, it "follows" the locked sun opensolaris tree[01:14:30] <e^ipi> but close enough[01:14:37] <samuelyounge> yea[01:14:43] <e^ipi> ( i don't know if anyone's hooked up grok to illumos-gate )[01:14:51] <gwr> Also: http://src.illumos.org/source/[01:15:05] <LeftWing> e^ipi: http://http://src.illumos.org/source/search?q=javac&project=illumos-opengrok&defs=&refs=&path=&hist=[01:15:13] *** vitaliy_gusev has quit IRC[01:15:27] <e^ipi> so they have[01:15:29] <e^ipi> neato[01:15:38] <samuelyounge> yea, very cool[01:16:19] <e^ipi> grok's got a bit of frustration but overall good stuff[01:16:38] <e^ipi> certainly beats the pants off cvsweb & trac & whatever else[01:17:56] <richlowe> grep or ack leave considerably more time for coffee breaks.[01:20:42] *** vitaliy_gusev has joined #illumos[01:22:42] *** _ams_ has quit IRC[01:23:10] *** Aram has quit IRC[01:25:01] *** vitaliy_gusev has quit IRC[01:26:14] *** McBofh has quit IRC[01:29:16] *** meineerde has quit IRC[01:30:04] *** ShadowHntr has joined #illumos[01:30:05] *** ShadowHntr has joined #illumos[01:31:29] *** meineerde has joined #illumos[01:31:34] *** vitaliy_gusev has joined #illumos[01:38:48] *** vitaliy_gusev has quit IRC[01:39:01] *** edude03 has joined #illumos[01:41:36] *** vitaliy_gusev has joined #illumos[01:45:50] *** randw has quit IRC[01:47:49] *** McBofh has joined #illumos[01:57:46] *** vitaliy_gusev has quit IRC[02:10:18] *** ams_ has quit IRC[02:12:19] *** vitaliy_gusev has joined #illumos[02:17:20] *** gber has quit IRC[02:17:29] *** vitaliy_gusev has quit IRC[02:28:27] <lewellyn> anyone feeling sparky? http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6891268[02:29:06] *** vitaliy_gusev has joined #illumos[02:29:20] <alanc> isn't that in sfw?[02:29:21] *** _ams_ has joined #illumos[02:29:29] *** Gervystar_ has joined #illumos[02:30:13] <nrubsig> lewellyn: seriously: zlib should be compiled with largefile support[02:30:31] <lewellyn> alanc: shh! ;)[02:30:35] <nrubsig> richlowe: not grep --threaded[02:30:39] <lewellyn> it seemed like debate-fodder![02:30:40] <nrubsig> er[02:30:54] <alanc> nrubsig: but wouldn't that break ABI of existing linked applications?[02:31:25] *** aszeszo has quit IRC[02:31:47] <nrubsig> alanc: Mhhh... maybe we need to invent something... lets think... we need an invention... maybe we can use special words or numbers to distinguish between the versions... man.. this is sooo hard[02:31:52] <nrubsig> =:-)[02:32:20] <nrubsig> alanc: bump shared library version. Maybe shoot the person who did the original non-largefile setup[02:32:42] <alanc> or just EOL 32-bit, and be glad 64-bit has only one file size choice[02:32:47] *** Gervystar__ has joined #illumos[02:33:15] *** Gervystar has quit IRC[02:33:15] *** Gervystar__ is now known as Gervystar[02:33:30] <nrubsig> alanc: 32bit userland, too ?[02:33:53] <nrubsig> The ARC case... may be... interesting... =:-)[02:34:05] <samuelyounge> Are there any know issues with the current illumos ftp package? I am downloading some files from gnu.org and keep getting corrupt files.[02:34:13] <alanc> that would end the silly questions about "Why isn't my /bin/true compiled as a 64-bit app?"[02:34:31] <alanc> "Because true only needs one bit, the other 63 are wasted"[02:34:33] <nrubsig> erm[02:34:34] <lewellyn> alanc: because it should be a 10 byte file ;)[02:34:56] <nrubsig> lewellyn: AFAIK the record for space is around 90 bytes[02:35:03] <nrubsig> er[02:35:06] <lewellyn> #!/bin/sh^M[02:35:08] <nrubsig> s/space/SPARC/[02:35:24] <nrubsig> lewellyn: ETOOMUCHOVERHEAD[02:35:29] <ShadowHntr> LOL :)[02:35:29] <lewellyn> nrubsig: so? ;)[02:36:06] <lewellyn> it's trivial and obvious as to what it does and eliminates lots of other issues![02:36:24] *** Gervystar_ has quit IRC[02:36:36] <lewellyn> false, of course, should be 17 bytes[02:37:14] * nrubsig implements it with 2 bytes and special code in intpexec[02:37:29] <lewellyn> that's not trivial and obvious.[02:37:54] <nrubsig> lewellyn: nah, but fast _and_ bloated[02:37:59] <lewellyn> i wonder what color the bikeshed is today...[02:38:03] <lewellyn> smrt: bikeshed color[02:38:03] <smrt> gray70[02:38:08] <lewellyn> very good color, that[02:38:13] <nrubsig> smrt: bikeshed color[02:38:14] <smrt> gray70[02:38:27] <lewellyn> nrubsig: it uses $RANDOM so it's unique across invocations of smrt atm[02:38:29] <nrubsig> lewellyn: what the fsck is a bikeshed color ?[02:38:33] <lewellyn> smrt: say goodbye[02:38:33] *** smrt has quit IRC[02:38:42] <alanc> isn't that the color of the new S11X login/grub backdrop?[02:38:50] *** smrt has joined #illumos[02:38:51] <lewellyn> smrt: bikeshed color[02:38:53] <smrt> magenta4[02:38:59] <alanc> nrubsig: http://www.bikeshed.com/[02:39:08] <lewellyn> alanc is faster[02:39:25] <bradend> Oh you guys.[02:39:39] <alanc> surprised it never came up during one of your ARC fastttracks[02:39:40] <lewellyn> nrubsig: if i could re-seed $RANDOM or get a new $RANDOM in a subshell, it'd be a different color each time.[02:39:44] <lewellyn> alanc: haha[02:40:46] <alanc> heh, just noticed smrt's $RANDOM loginid for this session[02:41:03] <lewellyn> though, smrt's behavior atm is pretty consistent with the average bikeshed argument participant. stick to a side until something happens (like sleep) then choose a new side.[02:41:13] <lewellyn> alanc: yes. so did another channel :)[02:41:37] <alanc> wonder if that's smrt's way of telling us it will play with either s11ex or illumos[02:41:39] <lewellyn> i should submit that script to whichever bug is related to identd[02:41:57] <lewellyn> it solves the issue of no identd ;)[02:42:07] *** N2Deep has quit IRC[02:42:40] <Meths> gdamore: Reading, thanks.[02:42:58] <lewellyn> i wonder if roland's seen it yet. he probably laughed at my approach.[02:43:21] <lewellyn> alanc: 17:42 bikcmp: smrt: i love your ident![02:43:30] <lewellyn> that was just before yours (my clock is fast)[02:44:36] *** vitaliy_gusev has quit IRC[02:44:49] *** vitaliy_gusev has joined #illumos[02:45:02] <lewellyn> nrubsig: http://www.greenviolet.net/paste/4ce485b176f2e.html[02:45:13] <gdamore> Meths: yw[02:45:27] *** N2Deep has joined #illumos[02:46:23] <nrubsig> $(($RANDOM % 25144 + 1)) should be $((RANDOM % 25144 + 1))[02:47:02] <nrubsig> lewellyn: I could replace the "sed" but right now I'm in a job interview with test.[02:47:57] <lewellyn> why RANDOM vs $RANDOM?[02:48:24] <lewellyn> and i could replace the sed too, but i got inconsistent results.[02:48:46] <nrubsig> The $ forces a conversion from interna value to string... which is unneccesary within (( ))[02:48:57] <nrubsig> s/interna/internal/[02:48:59] <lewellyn> ah[02:49:20] <lewellyn> at least you didn't find fault in how i choose a line number ;)[02:51:46] *** alfism has quit IRC[02:51:48] <nrubsig> lewellyn: ENOTIME, job interviwe[02:51:53] <lewellyn> hahaha[02:52:13] <nrubsig> (unfortunately half via IRC)[02:56:40] *** herzen has joined #illumos[02:58:58] <samuelyounge> illumos ftp no worky.[03:01:15] <ottom> lewellyn: you could have that 'sed' quit immediately after it prints the chosen line[03:01:35] <lewellyn> ottom: meh. it's not significantly different :)[03:01:58] *** vitaliy_gusev has quit IRC[03:03:31] <ottom> man, first #!/bin/sh for 'true' and now this. It's almost like every cycle isn't sacred to you :-)[03:04:40] *** vitaliy_gusev has joined #illumos[03:05:19] <lewellyn> if every cycle were sacred, i'd print a static string. ;)[03:09:00] <snuff-work> anyone have a guesstimate of how long a zpool scrub of 5tb worth of zvols takes.. mine has been going for 19hrs.. and is sitting on 100% still..[03:09:36] *** Stellar has quit IRC[03:11:08] <lewellyn> go sleep :)[03:12:05] *** descipher has joined #illumos[03:13:03] <snuff-work> well.. i guesstimate that it got to 100% after about 10 hrs.. so its been almost extra 10 hrs sitting at 100%[03:32:31] *** _ams_ has quit IRC[03:33:23] *** aszeszo has joined #illumos[03:34:23] *** vitaliy_gusev has quit IRC[03:34:23] *** aszeszo has joined #illumos[03:37:34] *** smrt has quit IRC[03:37:48] *** smrt has joined #illumos[03:43:29] *** ShadowHntr has quit IRC[03:45:26] *** ShadowHntr has joined #illumos[03:45:59] *** zynox has quit IRC[03:48:35] *** N2Deep has quit IRC[03:49:01] *** ShadowHntr has quit IRC[03:49:03] *** vitaliy_gusev has joined #illumos[03:49:03] *** _ams_ has joined #illumos[03:49:40] *** dreh has quit IRC[03:52:04] *** lattera has quit IRC[03:55:35] *** dreh has joined #illumos[04:03:00] *** stallion has joined #illumos[04:03:11] <stallion> jbk: ping[04:10:13] *** aszeszo has quit IRC[04:11:39] *** kohju has quit IRC[04:13:05] <bradend> pong.[04:14:35] *** _ams_ has quit IRC[04:15:36] *** vitaliy_gusev has quit IRC[04:15:42] *** samuelyounge has quit IRC[04:16:49] *** vitaliy_gusev has joined #illumos[04:16:50] *** kohju has joined #illumos[04:18:04] *** _ams_ has joined #illumos[04:24:44] <jbk> stallion: pong[04:24:59] <jbk> man i wish there was some text parsing libs in ON :)[04:25:01] <freetown2> man in middle attack![04:25:17] <bradend> \o/[04:25:34] <jbk> all my communications are signed with an invisible hash :)[04:25:50] <bradend> Corned beef or potato?[04:26:14] <e^ipi> *bubblebubble* uhhh... what?[04:26:22] <freetown2> they actually make a good combination so both bradend[04:26:40] * bradend is hungry for invisible hash.[04:27:21] <freetown2> 35 mins till lunch...sigh[04:29:09] <stallion> jbk: bleh. any chance I could pull that bundle rather than having you upload it?[04:29:47] <jbk> i don't have anywhere where i could host it[04:31:01] <stallion> hrmm[04:31:08] <bradend> I recently discovered that S3 is dandy for that sort of thing.[04:31:21] <bradend> Easy to place non-stupid access controls, and cheap.[04:31:52] *** POloser has joined #illumos[04:32:30] *** zynox has joined #illumos[04:32:52] <alanc> jbk: maybe if you wrote code in a language with builtin string handling, like perl or python[04:32:54] * alanc ducks[04:33:13] <jbk> i can kick too :)[04:34:09] *** _ams_ has quit IRC[04:34:51] *** vitaliy_gusev has quit IRC[04:35:42] <lewellyn> so, are there any illumos-based sparc distros yet?[04:36:23] <lewellyn> i got my poor man's lom back up :)[04:36:29] <lewellyn> great little device[04:36:58] <freetown2> lewellyn, maybe after gdamore beats up the OI crowd[04:37:00] <lewellyn> i should change the output of ttya to 115.2K though. 9600 is slow :P[04:37:07] *** vitaliy_gusev has joined #illumos[04:37:10] <lewellyn> freetown2: i'm underimpressed with OI still[04:37:15] <freetown2> OR he gets beaten by the OI crowd[04:37:48] <lewellyn> i'm thinking of shelving what i'm working on atm to install a distro and poke at Xsun[04:38:05] <freetown2> lewellyn, yeah...building me own samba as the one that is packaged for 147 has issues[04:38:32] <lewellyn> SFEphp is 3 years since its last update, today...[04:38:55] *** _ams_ has joined #illumos[04:39:44] *** samuelyounge has joined #illumos[04:41:19] <richlowe> alanc: everyone needs their hair shirt.[04:41:48] <richlowe> we can't _all_ have freetype, after all.[04:41:56] <alanc> heh[04:42:52] <alanc> though I was talking with another rich today about how much some of the ON cmd code could be improved using the recent additions to libc (asprintf, getline, and even older additions like strlcat)[04:43:48] <richlowe> a lot of the ON command code could be improved by the not-so-recent additions, too.[04:44:16] <richlowe> I was going to say it hadn't been maintained since Arsenio was on TV, but I think _some_ of it would follow up with "Arsenio _who_?"[04:44:20] <alanc> but then I know how much X code could be shrunk by replacing len = strlen(a) + strlen(b) + strlen(c).... ; buf=malloc(len) ; snprintf(buf, len, ...) with asprintf()[04:46:19] <alanc> some of the code hasn't been maintained since Archie Bunker was on TV[04:46:38] <bradend> Those were the daaaays,[04:47:36] <lewellyn> richlowe: that's sad :([04:47:43] <alanc> though some of that was finally purged this year - graph, spline, FMLI...[04:51:45] <samuelyounge> Is there a way to resume a compile of illumos-gate after a system crash or reboot?[04:52:53] <McBofh> samuelyounge: not without a bit of futzingf[04:53:10] <McBofh> you can try an incremental build, but you'll have to know which objects need cleaning up first[04:53:20] <McBofh> safer for your build to restart it from scratch[04:53:29] <samuelyounge> Yea[04:53:47] <samuelyounge> screen will not resume after a system crash either[04:54:11] <samuelyounge> I thought that was its purpose.[04:54:37] <lewellyn> no[04:54:47] <lewellyn> its purpose is to be a terminal multiplexer[04:54:54] <lewellyn> if the terminals go away, so does it.[04:54:55] <samuelyounge> ok[04:55:35] <samuelyounge> it crashed while executing the `ir2hf` command. Anyone know what that does?[04:55:44] <jbk> alanc: when was getline added?[04:56:20] <McBofh> samuelyounge: it's an internal part of the compiler[04:56:55] <Meths> jbk: 147[04:57:00] <lewellyn> so yeah. no illumos sparc distros, then?[04:57:03] * lewellyn grabs SX[04:57:08] <samuelyounge> I have these illumos lockups when I am building illumos-gate quite often. It runs for many hours then hangs.[04:57:23] <samuelyounge> Sometimes it makes it through other times it does not.[04:57:33] <samuelyounge> A terrible waste of time.[04:57:33] <bdha> Metal or virtual host?[04:57:36] <samuelyounge> VM[04:57:39] <bdha> Which?[04:57:42] <samuelyounge> VMware[04:57:44] <bdha> What host?[04:57:51] <samuelyounge> Mac OS x[04:58:00] <bdha> How much RAM assigned to vmware?[04:58:06] <samuelyounge> 2GB[04:58:13] <bdha> Compiling in a VM is horribly slow regardless.[04:58:14] <samuelyounge> 4GB swap[04:58:17] <bdha> Get some metal.[04:58:22] <alanc> jbk: 147[04:58:26] <samuelyounge> No metal cash[04:58:32] <bdha> Craigslist.[04:58:42] <bdha> Or freecycle.[04:59:05] <samuelyounge> The VM works great as long as you pre-allocate HD space, if you don't it is unbearably slow.[04:59:14] <richlowe> samuelyounge: the other trick is, "do an incremental, expect it to bitch about object files, do clobber in every directory about which it bitches, do another incremental"[04:59:17] <samuelyounge> And this lockup problem with the illumos-gate builds.[04:59:27] <richlowe> whether that actually saves time depends on the width of the tree at the time you dropped.[04:59:29] <McBofh> samuelyounge: building ON or Illumos is a memory and swap-hungry process[04:59:35] <richlowe> it's mostly ctfconvert that screws up here[04:59:45] <richlowe> since it molests objects in place, it can leave them lacking both stabs/dwarf _and_ ctf[04:59:49] <richlowe> so the next run will choke[05:02:56] <samuelyounge> Here goes another 6 hours. F@#^%$^%$&![05:05:50] <McBofh> add more ram, add more swap, don't run lint, or wsdiff[05:06:09] <samuelyounge> ok, what is wsdiff?[05:06:15] <samuelyounge> I don't run lint[05:07:07] <McBofh> nroff -man /opt/onbld/man/man1/wsdiff.1 |less[05:07:25] <samuelyounge> thanks[05:07:49] <richlowe> if you don't know what it is, you aren't running it.[05:08:00] <samuelyounge> Even better :)[05:08:09] <samuelyounge> Ignorance is bliss.[05:08:21] * McBofh goes back to breaking his head with a new API and library[05:09:39] <samuelyounge> does anyone know the optimization equlivate of -mtune=i686 for the Sun complier?[05:09:47] <samuelyounge> gcc[05:10:59] <samuelyounge> many thanks[05:11:03] <alanc> or -xchip=<something>[05:11:31] <alanc> xarch affects ISA used (extensions like MMX, SSE, etc.), chip is other tuning, like cache sizes & branch prediction[05:11:46] <samuelyounge> right[05:12:43] <lewellyn> there's a page smrt knows about with descriptions too[05:12:48] <alanc> (though gentoo is far saner than the "Linux from Scratch" crowd)[05:13:05] <lewellyn> alanc: and tend to get a working system[05:14:09] <richlowe> LFS is harder to form into a pejorative noun.[05:15:23] *** xapiens has quit IRC[05:15:37] <lewellyn> or adjective[05:25:19] <samuelyounge> could someone download any .tar.bz2 from a ftp site using illumos ftp client and then test the tar, and let me know if it is corrupted or worked. When I download using the illumos ftp client I get corrupt files. I would like to know if the problem is particular to my system or illumos in general before I file a bug report.[05:28:15] <samuelyounge> sftp works fine.[05:35:48] *** stallion has quit IRC[05:39:51] *** zynox has quit IRC[05:43:09] <nrubsig> e^ipi: try -xalias_level ... much fun... =:-)[06:15:43] <jbk> ok.. here's a bit of an interesting question[06:17:06] <jbk> if I define _XPG6 and _XOPEN_SOURCE=500 to get posix definitions for struct msghdr (for use with sendmsg/recvmsg)[06:17:22] <jbk> can I do that _just_ for that file[06:17:33] <jbk> and not enable that for the rest of the stuff (linked into the same binary)[06:17:48] <jbk> or are there extra linker tricks going on?[06:18:15] <jbk> the problem i'm running into is that enabling those (even with __EXTENSIONS__ defined) is hiding some prototypes used in other files[06:18:47] <jbk> since there's only one file that calls recvmsg/sendmsg, i'm hoping i can localize it to just that file[06:19:40] *** zynox has joined #illumos[06:21:37] <movement> jbk: I think you might be able to get away with it. specifying those flags ends up adding a flag to the system call that lets it know it's using the posix defs[06:26:17] *** vitaliy_gusev has quit IRC[06:26:22] *** vitaliy_gusev has joined #illumos[06:27:07] <jbk> i'll cross my fingers :)[06:27:52] *** _ams_ has quit IRC[06:28:07] *** _ams_ has joined #illumos[06:47:32] *** freakazoid0223 has quit IRC[06:48:42] *** nrubsig has quit IRC[06:54:07] <e^ipi> lblume: you're chinese IIRC. the initial 'x' in Xian... pronounced like a z or a zh ?[06:54:22] <McBofh> "seee"[06:54:27] <McBofh> and he's French, but living in China[06:54:38] <McBofh> so Xi'an is Seee Ahn[06:54:41] <McBofh> long A[06:55:02] <e^ipi> cool, thanks man[06:55:06] <McBofh> yw[07:03:43] *** Zauberpo1y has joined #illumos[07:05:53] *** Zauberpony has quit IRC[07:06:08] * lewellyn sics e^ipi on #Mandarin[07:32:48] *** edude03 has quit IRC[07:57:54] *** lgtaube has quit IRC[08:00:17] *** zynox has quit IRC[08:05:13] *** samuelyounge has quit IRC[08:05:49] *** hajma has quit IRC[08:06:16] *** dnm has joined #illumos[08:11:14] *** zynox has joined #illumos[08:14:34] *** Bleubear_ has quit IRC[08:20:29] *** vitaliy_gusev has quit IRC[08:22:16] *** joffe has quit IRC[08:22:21] *** vitaliy_gusev has joined #illumos[08:32:06] *** joffe has joined #illumos[08:38:57] *** descipher has quit IRC[08:41:47] <gdamore> so, the sparc build seems to be going nicely, after fixing a few Makefile problems.[08:42:18] <gdamore> (an interesting problem in libc's sparc targets, and a mistake I made in the FIPS code cleanup)[08:42:31] <lewellyn> gdamore: can i build illumos from SX?[08:42:38] <gdamore> No idea.[08:42:43] <gdamore> *Probably*.[08:42:54] <lewellyn> i never got 134 installed on my sparc, 134b is repo-only. so SX is my only choice for bootstrapping atm.[08:42:56] <gdamore> But since nobody really knows what is in SX11, its hard to say[08:43:03] * lewellyn is finally watching it install.[08:43:11] <lewellyn> 30% done transferring contents[08:43:34] <lewellyn> maybe i'll get another sparc or two up this weekend if this goes well[08:43:42] <gdamore> If there were a SPARC version of OI, you'd have hope. That's yet another rant....[08:44:17] <lewellyn> well, i didn't have much luck with the x86 version shutting down properly, so i'd have less hope than you for a sparc version working properly[08:47:30] <gdamore> Let me guess... Nvidia crash?[08:47:47] <gdamore> Nvidia framebuffer driver crashes during shut down on my desktop.[08:47:55] <gdamore> Too bad I can't fix it, but ENOSOURCE.[08:50:05] *** RoyK has quit IRC[08:56:32] <lewellyn> gdamore: no. happened on i945 and in vbox and in vpc2007.[08:56:38] <lewellyn> i gave up after 3 strikes.[08:59:35] *** schily___ has quit IRC[08:59:35] *** hsp has joined #illumos[08:59:59] <gdamore> weird.[09:02:18] <lewellyn> yeah. not worth my time to investigate since they seem to tell people to wait for the next build when they complain.[09:03:55] *** sancho has joined #illumos[09:04:53] *** schily___ has joined #illumos[09:07:01] *** hajma has joined #illumos[09:10:16] *** anilg has joined #illumos[09:10:30] *** th has left #illumos[09:11:20] *** samuelyounge has joined #illumos[09:12:22] *** lgtaube has joined #illumos[09:15:25] *** anilg has quit IRC[09:17:47] *** sancho has quit IRC[09:20:21] *** anilg has joined #illumos[09:21:56] *** alhazred has joined #illumos[09:24:56] *** enderst has quit IRC[09:26:01] <lblume> Lucky there are others on call for Chinese pronunciation when I'm away![09:26:46] *** enderst has joined #illumos[09:29:08] *** taemun has quit IRC[09:30:05] *** taemun has joined #illumos[09:32:38] *** sancho has joined #illumos[09:44:44] *** gebi has joined #illumos[09:44:45] *** gebi has joined #illumos[09:50:59] *** lgtaube has quit IRC[09:54:35] *** samuelyounge has quit IRC[09:54:45] *** jgrafton has quit IRC[10:00:39] *** jgrafton has joined #illumos[10:02:47] *** DocPheniX has quit IRC[10:03:16] *** DocPheniX has joined #illumos[10:04:22] *** frankS2 has quit IRC[10:11:37] *** Tenzer|off is now known as Tenzer|work[10:13:03] *** RoyK has joined #illumos[10:13:44] *** frankS2 has joined #illumos[10:17:17] *** jamesd_laptop has joined #illumos[10:17:17] *** jkimball4 has quit IRC[10:19:12] *** alhazred_ has joined #illumos[10:20:02] *** jamesd2 has quit IRC[10:20:11] *** alhazred has quit IRC[10:25:40] *** jurikm has joined #illumos[10:48:10] *** jamesd_laptop has quit IRC[10:48:39] *** jamesd_laptop has joined #illumos[10:50:51] *** xmikus01 has joined #illumos[10:51:55] *** jkimball4 has joined #illumos[11:03:43] <trochej> Coffee[11:16:30] *** gber has joined #illumos[11:22:52] <estibi> hi[11:39:04] *** jurikm has quit IRC[11:39:47] *** anilg has left #illumos[11:48:03] *** alhazred_ has quit IRC[11:51:44] *** McBofh has quit IRC[11:55:42] *** naOdpis is now known as nettezzaumana[11:56:11] *** dnm has quit IRC[11:58:02] *** alhazred has joined #illumos[11:59:10] *** McBofh has joined #illumos[12:06:18] *** sancho has quit IRC[12:19:28] *** alhazred_ has joined #illumos[12:19:59] *** alhazred has quit IRC[12:21:07] *** sancho has joined #illumos[12:46:31] *** hsp has quit IRC[12:52:22] *** Aram has joined #illumos[13:07:04] *** jurikm has joined #illumos[13:07:10] *** jurikm has left #illumos[13:17:32] *** DerSaidin has quit IRC[13:24:56] *** DerSaidin has joined #illumos[13:24:56] *** DerSaidin has joined #illumos[13:25:44] *** edp_work has quit IRC[13:27:21] *** merzo has joined #illumos[13:29:39] *** alhazred_ has quit IRC[13:34:31] *** edp_work has joined #illumos[13:34:37] *** joffe has quit IRC[13:34:48] *** alhazred has joined #illumos[13:36:39] *** gber_ has joined #illumos[13:39:00] *** gber has quit IRC[13:42:48] *** alhazred has quit IRC[13:43:03] *** alhazred has joined #illumos[13:45:19] *** gber_ is now known as gber[13:45:34] *** merzo has quit IRC[13:47:41] *** jamesd2 has joined #illumos[13:50:33] *** alhazred_ has joined #illumos[13:50:33] *** alhazred has quit IRC[13:50:53] *** jamesd_laptop has quit IRC[13:51:42] *** Gervystar has quit IRC[13:55:44] *** PatrickS has quit IRC[13:56:58] *** alhazred_ has quit IRC[13:57:28] *** alhazred_ has joined #illumos[14:09:48] *** alhazred_ has quit IRC[14:09:55] *** Gugge_ has joined #illumos[14:10:16] *** joffe has joined #illumos[14:13:05] *** Gugge_ has quit IRC[14:13:12] *** Gugge has quit IRC[14:13:20] *** axisys has quit IRC[14:16:12] *** spanglywires has joined #illumos[14:29:36] *** alhazred has joined #illumos[14:33:30] *** POloser has left #illumos[14:58:56] *** Gervystar has joined #illumos[15:14:32] *** cmbntr has joined #illumos[15:14:34] *** pablomh has joined #illumos[15:15:20] *** cmbntr has quit IRC[15:19:08] *** spanglywires has quit IRC[15:22:20] *** merzo has joined #illumos[15:23:29] *** xmikus01 has quit IRC[15:27:49] *** edude03 has joined #illumos[15:34:17] *** Gugge has joined #illumos[15:42:27] *** vitaliy_gusev_ has joined #illumos[15:42:27] *** vitaliy_gusev has quit IRC[15:42:28] *** _ams_ has quit IRC[15:42:44] *** _ams_ has joined #illumos[15:52:20] *** pizdec has joined #illumos[15:52:42] *** lattera has joined #illumos[16:04:09] *** postwait has joined #illumos[16:10:58] *** dajhorn has joined #illumos[16:18:05] *** jamesd2 has quit IRC[16:28:34] *** alhazred_ has joined #illumos[16:29:01] *** alhazred has quit IRC[16:29:11] *** nikolam has joined #illumos[16:29:44] *** sancho has quit IRC[16:36:13] *** stallion has joined #illumos[16:37:43] *** RoyK has quit IRC[16:39:00] *** dmarker has quit IRC[16:39:49] *** dmarker has joined #illumos[16:47:43] *** Gervystar has quit IRC[16:48:48] *** Gervystar has joined #illumos[16:50:22] *** jamesd2 has joined #illumos[16:56:15] *** xapiens has joined #illumos[17:03:40] *** edude03 has quit IRC[17:04:56] *** dnm has joined #illumos[17:07:06] *** merzo has quit IRC[17:13:45] *** _ams_ has quit IRC[17:13:59] *** anilg has joined #illumos[17:15:35] *** _ams_ has joined #illumos[17:17:25] *** carlomagno has quit IRC[17:24:00] *** nikolam has quit IRC[17:31:28] *** carlomagno has joined #illumos[17:31:55] *** Gervystar has quit IRC[17:32:45] *** alhazred_ has quit IRC[17:33:34] *** Gervystar has joined #illumos[17:45:04] *** alhazred has joined #illumos[17:46:06] *** Tenzer|work is now known as Tenzer|off[17:46:08] *** axisys has joined #illumos[17:48:18] *** hajma has quit IRC[18:03:37] *** smrt has quit IRC[18:03:52] *** smrt has joined #illumos[18:10:12] *** xmikus01 has joined #illumos[18:12:21] *** samuelyounge has joined #illumos[18:13:01] *** cmbntr has joined #illumos[18:15:29] *** randw has joined #illumos[18:18:24] *** jamesd2 has quit IRC[18:18:52] *** jamesd2 has joined #illumos[18:19:42] *** jamesd_laptop has joined #illumos[18:23:23] *** jamesd2 has quit IRC[18:31:42] *** PatrickS has joined #illumos[18:37:49] *** cmbntr has quit IRC[18:38:15] *** jamesd_laptop has quit IRC[18:38:37] *** jamesd2 has joined #illumos[18:39:02] *** sancho has joined #illumos[18:46:37] *** glagasse has joined #illumos[18:55:45] *** jamesd_laptop has joined #illumos[18:55:50] *** alfism has joined #illumos[18:58:48] *** jamesd2 has quit IRC[18:59:27] *** anilg has quit IRC[19:02:51] *** edp_mobile has joined #illumos[19:07:05] *** held has quit IRC[19:07:28] *** Tenzer|off is now known as Tenzer[19:08:03] *** alfism has quit IRC[19:08:26] *** alfism has joined #illumos[19:11:15] *** held has joined #illumos[19:12:39] <CIA-75> illumos Garrett D'Amore <garrett at nexenta dot com>: 368 snmpdx is obsolete and neesd to go Reviewed by: gwr at nexenta dot com Reviewed by: rbg at openrbg dot com Approved by: richlowe at richlowe dot net[19:13:25] *** dmarker has quit IRC[19:13:33] <lewellyn> yay![19:13:38] *** smrt has quit IRC[19:13:39] <lewellyn> cia finally looks sane.[19:13:49] *** dmarker has joined #illumos[19:13:52] *** smrt has joined #illumos[19:15:05] *** DocPheniX has quit IRC[19:16:20] *** jamesd_laptop has quit IRC[19:20:50] <gdamore> thanks lewellyn.[19:21:04] <lewellyn> it just took a few commits. ;)[19:21:05] <gdamore> too bad the only person doing anything for CIA to notice seems to be me. :-([19:21:12] <lewellyn> they just don't happen often enough :([19:21:21] *** DocPheniX has joined #illumos[19:21:46] <lewellyn> gdamore: well, if i can get illumos bootstrapped on sparc, maybe i can fix that soonish :P[19:23:03] *** Gervystar has quit IRC[19:23:10] <lewellyn> i expect that S11X -> illumos isn't going to be happy, though.[19:23:20] <gdamore> me either.[19:24:01] <lewellyn> mainly because i don't know how to tell onu that "yeah, really, install these lower-versioned things."[19:24:29] <gdamore> we need to crank a version number.[19:24:36] <gdamore> or change the version numbering altogether.[19:24:45] <lewellyn> i suggested the latter previously[19:26:08] *** Thrae has joined #illumos[19:26:50] *** hajma has joined #illumos[19:28:11] <richlowe> upgrade S11X->illumos is just a plain bad idea[19:28:34] <richlowe> any expectation that the non-ON bits would work is unfounded.[19:28:43] <richlowe> the zpool versions won't match (on a fresh install)[19:28:43] <richlowe> etc.[19:28:45] <alanc> wouldn't a fresh install of S11X have a newer zfs version than illumos can read/[19:28:53] <alanc> rich wins[19:29:54] <gdamore> Yes, I think S11X -> illumos == fail.[19:30:09] <gdamore> although....[19:30:24] <gdamore> a boot pool probably won't have any of the newer ZFS features in it....[19:30:29] <Aram> actually it works quite well.[19:31:05] <lewellyn> sadly, though, <= 134 gives me hassles to install on sparc. 134b is repo-only. so that really limits what i can upgrade to illumos from.[19:31:09] <Aram> I agree that we should not expect it to work, but it seems that it does, and we should at least make a way for it to be possible.[19:31:22] <Aram> (change version algorithms).[19:31:40] <Aram> Triskelios suggested something. Triskelios?[19:31:54] <lewellyn> i was going to see if i can't figure out distro-constructor later today/tonight[19:32:41] <lewellyn> at least once i get SFE working on S11X. pkgbuild's unhappy with me. :([19:32:57] *** RoyK has joined #illumos[19:33:03] <lewellyn> btw, totally non-topical: http://goo.gl/wf00Q[19:35:17] <richlowe> oooh, an error message[19:36:00] <lewellyn> oh?[19:36:01] <Triskelios> Aram: the "built-on" and "branch" version which appears at the end of the package FMRIs and I believe is used by the incorporation to apply its constraint is probably what you want to change. the branch version is currently "0.<ON build number>", incrementing the major version might work[19:36:21] <lewellyn> stupid aws[19:36:52] <lewellyn> http://www.spinthebottle.com/blog/how-to-build-the-ikea-hadron-collider[19:36:57] <lewellyn> that's the first link on google[19:37:12] * lewellyn wanders[19:37:54] <Aram> Triskelios: I think it's the best idea to do now, although it can't work on the long run. what if Joerg forks illumos? then he will increment that number to be 2, and we will be `less' then him. also, oracle can change that number any time.[19:38:08] <Aram> breaking us.[19:39:12] <lewellyn> forking a not-yet-popular fork... that seems like a waste of effort.[19:42:26] *** glagasse has quit IRC[19:42:35] *** alfism has quit IRC[19:46:10] *** jamesd2 has joined #illumos[19:52:48] *** nikolam has joined #illumos[19:52:59] <gdamore> so.. weird thing going on.[19:53:21] <gdamore> I've got sparc illumos built, except that there is a problem with n2cp -- the niagra2 crypto driver.[19:53:41] <gdamore> that used to be in the closed source, but its delivered as part of the crypto tarball...[19:55:34] <richlowe> and?[19:56:21] <richlowe> that state of affairs is just fine, putting it in the closed-bins and the crypto tarball would hardly be logical.[19:57:13] *** mavhc has quit IRC[19:58:28] <Triskelios> Aram: since the concern is entirely over image-update, I believe you can have an additional illumos metapackage, taking advantage of the fact that image-update is too conservative to remove that[19:59:05] <gwr> richlowe: did you look at Garrett's "no SEA" webrev?[19:59:40] <Triskelios> Aram: that looks like a cheap way to get a namespace[20:00:56] <gdamore> richlowe: the problem is our directions elided the crypto tarball...[20:01:21] <gdamore> so I'm going to have to go back and reinstate the crypto tarball for now.[20:01:37] <gwr> Is the n2cp driver required?[20:01:43] <gdamore> I wonder if the hardware specs for N2 crypto are available.[20:01:46] <gdamore> gwr: no.[20:01:49] <richlowe> gdamore: Not particularly[20:01:58] <richlowe> part of the reason the code is closed.[20:02:00] <gdamore> I'm half inclined to just remove the manifest file.[20:02:14] <richlowe> if you're going to do that, obsolete the package properly.[20:02:20] <gwr> If it's just h/w acceleration, that seems like something for "vendor value-add" (i.e. Oracle:)[20:02:28] <gdamore> Um... I don't think I should obsolete it "properly".[20:02:36] <gdamore> If people can get the driver, they can use t.[20:02:41] <gdamore> Maybe.[20:02:59] <gdamore> (I wonder if there are dependencies that are now unmet though, as a result of the changes I made to crypto a while back)[20:03:04] *** nikolam has quit IRC[20:03:28] <gwr> one could "stub out" any unmet dependencies, probably.[20:05:07] <gdamore> Well, I can either leave it around, and require the crypto tarball (and hopefully *that* doesn't screw things up), or I can remove the manifest files. Obsoleting it is IMO the wrong thing to do, as some other consolidation *could* provide these deliverables.[20:05:41] <richlowe> Then they'd provide the package.[20:05:47] <richlowe> and your obsoletion wouldn't exist anymore.[20:06:07] <richlowe> of course, if stuff depends on the package, you're stuck with leaving it dangling anyway, I guess.[20:06:27] <gdamore> I dont think anything depends on it. Nothing *should* at any rate.[20:07:22] <gdamore> of course, ncp was delivered with the sun4v platform package instead of in its own manifest. (yech)[20:08:01] *** eldee has joined #illumos[20:08:36] <gwr> richlowe: Oh pkg guru, speak to us of package obsoletion, and the pkg versions in which that happens, and how those interact with future versions... (please?:)[20:08:55] <gwr> (I did read pkg/README :)[20:08:57] <gdamore> so, opinions here.[20:09:24] <gdamore> Should I yank these two drivers from illumos... or should allow them to persist in closed form for now (reinstating the crypto tarball for SPARC) ?[20:09:31] *** eldee has quit IRC[20:10:01] <gdamore> At the moment I don't know that we'll ever be able to open these up properly.[20:10:15] <gwr> If there's a way for end users to go get the driver for h/w acceleration and use it,[20:10:27] <gwr> then it seems a "no brainer" to remove the driver.[20:10:34] <gwr> Not sure about that "if" though...[20:12:38] *** bradend has quit IRC[20:13:15] *** bradend has joined #illumos[20:14:24] *** bradend has quit IRC[20:14:57] <gdamore> what do you mean?[20:15:00] *** bradend has joined #illumos[20:15:31] <gdamore> the only way for users to use the h/w acceleration is with the driver. They can potentially set up their own consolidation or install manually from the crypto tarball....[20:15:59] <gdamore> I don't know if the crypto bits are redistributable though... I think they might not be.[20:16:21] <gwr> That's what I'm after: If they download the crypto tarball, could they copy that driver in to where it belongs and be "done"?[20:17:05] <gwr> Providing a little script to do that for them might be nice.[20:17:21] <gdamore> Probably.[20:17:35] <gwr> Do that in the "class action script" for the package? (joking:)[20:18:05] *** ivo_ has joined #illumos[20:18:19] <gdamore> I'm not sure where we'd deliver such a script... and I'm not entirely confident that the binaries will work. I *think* they will, but since I don't have source....[20:18:55] <gwr> So then just an FAQ entry with instructions, or something. done.[20:19:49] <gdamore> rich: do you concur?[20:21:18] <gwr> We still have a question re. what to do with/about the package that used to deliver n2cp...[20:21:27] <richlowe> the crypto bits are redistributable.[20:21:31] <richlowe> at least the ones you have.[20:21:37] <richlowe> (closed-bin tarball and crypto tarball)[20:21:53] <richlowe> I have no idea what I'm meant to concur with.[20:22:28] <gwr> I suggested not delivering that driver.[20:22:48] <gwr> (and maybe tell users how/where to get it)[20:23:58] <e^ipi> anyone good with sed ?[20:24:20] <samuelyounge> a little[20:24:23] <richlowe> gwr: sounds reasonable,[20:25:00] <gdamore> Ok, I'm going to nuke the manifest for driver/crypto/ncp2 and remove ncp from the sun4v manifest then. Right?[20:25:20] <gwr> So, maybe related to this; Garrett and I were discussing what's the "right" way to obsolete packages.[20:25:23] <samuelyounge> e^ipi:what are you trying to do?[20:25:29] <e^ipi> samuelyounge: i'm trying to turn a mac address potentially in the format of 00:01:ba:db:00:21 in to 0:1:ba:db:0:21[20:25:37] <gwr> What version should the obsoleted one show?[20:25:37] *** Jondice1 has quit IRC[20:25:44] <e^ipi> ie, strip the leading 0 from each piece[20:26:56] <e^ipi> what i'm using right now is s/00/0/g;s/:0/:/g;s/::/:0:/g;s/:$/:0/ but it's ugly and is 3 regexs instead of one[20:27:44] <gber> e^ipi: sed -e 's|0\([0-9]\)|\1|g' ?[20:28:30] *** Jondice1 has joined #illumos[20:28:41] <e^ipi> hmm... that seems to work[20:28:42] <e^ipi> thanks[20:41:38] *** HyperJohnGraham has quit IRC[20:44:55] *** postwait has quit IRC[21:05:51] *** vitaliy_gusev_ has quit IRC[21:09:22] *** psychicist has joined #illumos[21:14:31] *** pjfloyd has joined #illumos[21:22:29] *** _ams_ has quit IRC[21:33:49] *** postwait has joined #illumos[21:34:14] *** sancho has quit IRC[21:42:03] *** _ams_ has joined #illumos[21:45:13] *** descipher has joined #illumos[21:54:10] *** Tenzer is now known as Tenzer|off[22:03:02] *** jamesd_laptop has joined #illumos[22:06:19] *** jamesd2 has quit IRC[22:09:07] *** glagasse has joined #illumos[22:15:55] *** cmbntr has joined #illumos[22:22:15] *** jamesd_laptop has quit IRC[22:22:43] *** jamesd_laptop has joined #illumos[22:31:41] *** lattera has quit IRC[22:33:27] *** gebi has quit IRC[22:35:43] *** n1c0b has joined #illumos[22:37:06] *** pothos_ has joined #illumos[22:38:44] *** pothos has quit IRC[22:39:23] *** pjfloyd has quit IRC[22:39:33] *** n1c0b has quit IRC[22:39:38] *** pothos_ has quit IRC[22:39:58] *** pothos has joined #illumos[22:46:29] *** HyperJohnGraham has joined #illumos[22:52:16] *** alhazred has quit IRC[23:14:26] *** Stellar has joined #illumos[23:26:37] *** Aram has quit IRC[23:48:27] *** mavhc has joined #illumos[23:57:24] *** dajhorn has quit IRC