[00:00:14] <oxygene> jengelh: huh?
[00:00:56] <jengelh> what part did not you understand
[00:01:11] *** stevel has quit IRC
[00:01:20] *** Tzoa has quit IRC
[00:01:52] *** stevel has joined #opensolaris
[00:01:53] *** ChanServ sets mode: +o stevel
[00:02:37] *** yippi has quit IRC
[00:04:16] <jengelh> oxygene: what part did not you understand
[00:04:36] <oxygene> jengelh: why are you uncomfortable with a simple build environment?
[00:04:47] <jengelh> seems error prone
[00:05:19] <jengelh> e.g. I don't see a vermagic string in solaris modules
[00:06:57] *** Jiko has joined #opensolaris
[00:10:08] *** bunker has quit IRC
[00:10:33] <jengelh> gdamore: ping
[00:12:25] *** gnu2it2 has joined #opensolaris
[00:13:10] 
[00:13:11] <asyd> oups
[00:14:02] <trs81> lovely: cc: Fatal error in /opt/SUNWspro/prod/bin/ipo : Segmentation Fault
[00:19:02] <jengelh> is bsd-style licensing compatible with cddl?
[00:20:32] *** gisburn has joined #opensolaris
[00:20:40] <gisburn> ugh
[00:20:51] <gisburn> WTF happened to sommerfeld&co ?
[00:20:56] *** gisburn is now known as nrubsig
[00:21:02] *** ChanServ sets mode: +o nrubsig
[00:22:05] * nrubsig curses the peson who stuffed 24000+ lines into /usr/pub/UTF-8
[00:24:16] <nrubsig> stevel: ping!
[00:24:28] *** sahafeez has joined #opensolaris
[00:26:36] *** sommerfeld has joined #opensolaris
[00:26:37] *** ChanServ sets mode: +o sommerfeld
[00:26:37] *** odla has quit IRC
[00:27:31] <jengelh> so much for that
[00:27:51] <stevel> nrubsig: pong
[00:28:26] <nrubsig> stevel: where are all the other opensolaris people today ?
[00:29:07] <stevel> there's 156 people in here...?
[00:29:10] <boyd> Monring, all
[00:29:13] <stevel> not enough for you?
[00:29:38] <alanc> most of the regulars seem to be here
[00:29:43] <boyd> Hi stevel
[00:29:48] <stevel> hey boyd
[00:29:59] * nrubsig misses a huge list of green dots in his users pannel, reflecting the /ops
[00:30:08] * elektronkind groggily raises his hand
[00:30:14] * nrubsig misses sommerfeld
[00:30:25] * stevel points at sommerfeld
[00:30:25] <Gman> there weren't many ops to begin with..
[00:30:30] <nrubsig> oh
[00:30:31] <boyd> Don't worry, nrubsig, I'm here :)
[00:30:33] *** coffman_ has joined #opensolaris
[00:30:42] <nrubsig> stevel: he wasn't here when I was looking first
[00:31:13] <nrubsig> s/looking fisrt/looking the first time for \/ops/
[00:31:27] <elektronkind>  /g
[00:31:37] <boyd> Anyone got a link handy for where to get current star source? I always find it a bit hard to find
[00:31:56] <boyd> Where's schily when you need him? :)
[00:32:58] <nrubsig> boyd: sleeing
[00:33:09] <boyd> Is that a new winter sport? :)
[00:33:42] <jengelh> when do you need scily?
[00:33:53] <jengelh> it's 12:33am over here
[00:33:53] <nrubsig> s/sleeing/sleeping/
[00:34:07] <boyd> I don't think that the star home page being a a URL with "old" and "private" in the path is very encouraging :)
[00:34:25] <boyd> jengelh: I demand that the entire universe run on my local timezone :)
[00:34:33] <jengelh> heh
[00:34:40] <jengelh> that would not be a problem
[00:35:04] <jengelh> just don't expect that they are awake at the same time
[00:35:13] <nrubsig> boyd: welcome to flatland where the universe has just two dimensions
[00:35:25] <boyd> It's only a problem if they selfishly want to sleep when it happens to be dark where they are :)
[00:35:30] <jengelh> there is no such thing as flatland
[00:35:35] <jengelh> it's mathematically impossible
[00:35:44] <nrubsig> jengelh: erm
[00:35:59] <nrubsig> jengelh: there is a story about "flatland". It's a story.
[00:36:04] <jengelh> nrubsig: Projecting an n dimensions onto (n-1) always incurs a loss of information
[00:36:08] <jengelh> (yes I know Flatland)
[00:36:27] <boyd> jengelh: You assume that the information lost is something we care about :)
[00:36:36] <jengelh> but since two things can't exist at the same dimensional coordinates, a projection is not possible :D
[00:36:43] <nrubsig> jengelh: if you prefer physics talk we can discuss the newest proposals and findings about string theory, including quantum noise etc.
[00:36:44] <jengelh> at least not in realLife
[00:36:49] <jengelh> lol!
[00:37:38] <boyd> Further to the above, the star home page has a link to "beta" versions, where the URL contains "alpha"
[00:37:41] <jengelh> Well let's assume we have (0,0,0) and (0,0,1) - if we now projected that onto flatland, we'd have two things at (0,0) and (0,0) at the same time, which can't be.
[00:37:50] <jengelh> boyd: illogical at best
[00:38:25] <boyd> jengelh: You are assuming a projection from 3->2. Why can it not originate in 2d?
[00:38:42] <jengelh> I don't live in Flatland right now
[00:39:07] <boyd> I'm not sure that anyone actually suggested that there is such a place.
[00:39:09] <jengelh> so I was assuming the origin is n>=3
[00:39:26] <jengelh> i'm just ruling out the possibility that Our World could become a Flatland.
[00:39:26] <nrubsig> jengelh: your assumption is wrong when you think about OUR universe. Nothing prevents our universe to have seperate 2D planes which have crosssections with our 3D/4D plane.
[00:39:42] <jengelh> note "become"
[00:40:17] <nrubsig> grraaaaaaahhhh
[00:40:21] <nrubsig> xxx@@!!!
[00:40:26] * nrubsig HATES waiting
[00:40:59] <boyd> What are you waiting for?
[00:41:29] <nrubsig> boyd: the next ast-ksh sources to get the "getconf" bulitin fixed.
[00:41:59] <nrubsig> boyd: the xx@@!!-"fcs quality all the time" rule prevents again any progress... ;-(
[00:42:21] <boyd> Apparantly you're right. That's why there's so little new stuff in solaris 10.
[00:42:46] <nrubsig> boyd: that was sarcasm, right ?
[00:42:51] <twincest> boyd++
[00:42:56] <boyd> I'm afraid so.
[00:43:23] <boyd> As my sister says "Sarcasm is the lowest form of wit...... and the funniest"
[00:43:43] <twincest> i would have to say toilet humour is lower
[00:43:58] <boyd> What about sarcastic toilet humor?
[00:44:35] <boyd> Actually, I think "funny voices" is the lowest.
[00:45:11] <dclarke> fyi : Sol10 article in eWeek : http://www.eweek.com/article2/0,1759,2035516,00.asp?kc=EWUCUEMNL102306EOAD
[00:45:19] *** kimc has joined #opensolaris
[00:45:23] <dclarke> Blastwave gets mentioned
[00:45:28] <dclarke> :-)
[00:45:52] <kimc> dclarke: are you still interested in the results of the tests from yesterday ?
[00:45:52] *** coffman has quit IRC
[00:46:05] <dclarke> yes yes
[00:46:14] <dclarke> I really wanted to see those results .. BUT
[00:46:25] <dclarke> your results are probably polluted
[00:46:36] <dclarke> because you tried to run du in the middle of the test
[00:46:39] <kimc> if they're not so good due to running 'du' I can try it again
[00:46:43] <dclarke> you would be better off to run it again
[00:46:48] <kimc> no problemo
[00:46:59] <dclarke> yeah .. and this time keep your fingers away from the keyboard :-)
[00:47:05] <dclarke> brb
[00:47:11] <kimc> no problem :)
[00:48:19] <kimc> ok test is running..
[00:49:18] *** coffman_ is now known as coffman
[00:50:31] *** yippi has joined #opensolaris
[00:54:37] *** sommerfeld has quit IRC
[00:55:02] *** ShadowHntr has joined #opensolaris
[00:59:17] *** koolniczka has left #opensolaris
[01:03:44] <kimc> test 2 running..
[01:04:31] <kimc> io_avg=42.132333 MB/sec
[01:05:30] *** piwi has joined #opensolaris
[01:06:30] <jamesd__> damm... what kind of box gives that kind of performance on dennis'  torchure test ;-)
[01:08:40] <kimc> right..
[01:08:53] *** sahafeez has quit IRC
[01:09:10] <kimc> test 3 running
[01:09:39] <kimc> io_avg=0.517229 MB/sec
[01:09:45] <jamesd__> but what hardware is it.. even my u20 with  250GB sata drive gets no where near that number.. like a factor of 500 slower..
[01:10:13] <jamesd__> well  50..
[01:11:01] <kimc> it has a an AMD am2 proc and 2 gigs ram
[01:11:12] <jamesd__> what disk hardware
[01:12:06] <kimc> its just Nvidia 590-level chipset, the drive is a seagate 250 GB sata
[01:12:46] *** dduvall has quit IRC
[01:12:53] <jamesd__> wow..  maybe i need to test again.. my u20 is the   2.6ghz single core,  2GB ram and the same harddisk
[01:13:02] <kimc> the mobo is an 'Asus m2n32-sli deluxe'
[01:13:32] <kimc> i'm in 'no touch' mode on the machine as this test is running :)
[01:14:29] <nrubsig> stevel: is the webserver scewiness solved ?
[01:14:59] *** stevel has quit IRC
[01:15:21] * nrubsig takes the "stevel has left freenode (Remote closed the connection)" as "NO"
[01:15:36] * nrubsig counts the number of /ops
[01:15:37] *** stevel has joined #opensolaris
[01:15:37] *** ChanServ sets mode: +o stevel
[01:15:44] * nrubsig counts the number of /ops again
[01:15:51] <nrubsig> stevel: is the webserver scewiness solved ?
[01:15:58] <stevel> no
[01:15:59] <kimc> jamesd: which opteron is in your u20 ?
[01:16:06] <jamesd__> the 152
[01:16:07] <nrubsig> stevel: what is the problem ?
[01:16:10] <kimc> ahh..
[01:16:44] <jamesd__> would be nice if sun decided to upgrade it but for now its a nice box..  some day i will get a dual core cpu for it
[01:18:52] *** stevel has quit IRC
[01:20:39] *** stevel has joined #opensolaris
[01:20:39] *** ChanServ sets mode: +o stevel
[01:21:29] *** stevel has quit IRC
[01:27:51] * elektronkind awaits his fios install
[01:30:46] *** Spawrq has quit IRC
[01:30:49] <jengelh> is it ok to call gettimeofday() from within kernel context?
[01:32:20] <nrubsig> jengelh: is the symbol available in the kernel ?
[01:32:44] <jengelh> basically yes but that does not mean it can blindly be called
[01:33:29] <nrubsig> jengelh: see http://cvs.opensolaris.org/source/search?q=gettimeofday&defs=&refs=&path=%2Fon%2Fusr%2Fsrc%2Futs&hist= for usage
[01:33:30] <jengelh> it may assume that 'struct timeval *tp' actually points to userspace
[01:34:15] <nrubsig> !summon kupfer
[01:34:20] <nrubsig> jengelh: ask mike kupfer
[01:34:22] <jengelh> linux otoh, has a "clear" distinction -- sys_gettimeofday() only accepts pointers to user mem while do_gettimeofday() takes pointers to kernel memory
[01:35:01] <jengelh> bsd otoh does not have such a distinction
[01:38:05] *** Spawrq has joined #opensolaris
[01:46:13] *** coffman has quit IRC
[01:46:18] *** lacaAFK is now known as laca
[01:46:37] *** Gadzooks has joined #opensolaris
[01:46:38] <dclarke> I return
[01:47:18] <jamesd__> dclarke, in most programming lanagues its    return  l;  ... i guess you are using reverse polish notation ;-p
[01:47:37] *** sahafeez has joined #opensolaris
[01:47:39] <dclarke> I have not done RPN in years
[01:47:44] <dclarke> well .. not true
[01:47:50] <dclarke> I still have a HP calculator
[01:48:20] <jamesd__> i know you have turned on your calculator so i knew you use rpn almost as often as  you use  windows ;-p
[01:49:18] <dclarke> the CBC people are running a really great bit about smoking
[01:49:31] <dclarke>  /optcsw/bin/xine --hide-video http://oggtrial.nm.cbc.ca:80/cbcr1-toronto.ogg
[01:49:36] <dclarke> funny as hell
[01:50:40] *** nwf has joined #opensolaris
[01:59:58] *** jamesd_ has joined #opensolaris
[02:00:42] *** maverickbna has joined #opensolaris
[02:00:52] <nrubsig> !summon kupfer
[02:00:59] *** ShadowHntr has quit IRC
[02:01:05] *** maverickbna is now known as ShadowHntr
[02:01:06] <jengelh> no spelling power left? :)
[02:01:06] <dclarke> okay .. wow .. the Enron guy goes to jail for 24 years
[02:01:15] <jengelh> i think I found it... uniqtime seems better
[02:02:05] *** Spawrq has quit IRC
[02:02:05] *** Odin- has quit IRC
[02:02:05] *** esproul has quit IRC
[02:02:07] *** Tpenta has quit IRC
[02:02:07] *** jamesd__ has quit IRC
[02:02:08] *** axisys has quit IRC
[02:02:08] *** Vanuatoo has quit IRC
[02:02:08] *** phips has quit IRC
[02:02:08] *** noyb has quit IRC
[02:02:08] *** the-decider has quit IRC
[02:02:08] *** milo has quit IRC
[02:02:08] *** Stric has quit IRC
[02:02:09] *** dwc- has quit IRC
[02:02:09] *** jacotton has quit IRC
[02:02:10] *** snyff has quit IRC
[02:02:10] *** richlowe has quit IRC
[02:02:10] *** xzi11a has quit IRC
[02:02:10] *** doownek has quit IRC
[02:02:11] *** TBCOOL has quit IRC
[02:02:11] *** fet-kviga has quit IRC
[02:02:11] *** kirma has quit IRC
[02:02:11] *** nightswim has quit IRC
[02:02:11] *** paulproteus has quit IRC
[02:02:28] *** Spawrq has joined #opensolaris
[02:02:28] *** Odin- has joined #opensolaris
[02:02:28] *** esproul has joined #opensolaris
[02:02:28] *** Tpenta has joined #opensolaris
[02:02:28] *** fet-kviga has joined #opensolaris
[02:02:28] *** axisys has joined #opensolaris
[02:02:28] *** richlowe has joined #opensolaris
[02:02:29] *** snyff has joined #opensolaris
[02:02:29] *** jacotton has joined #opensolaris
[02:02:30] *** dwc- has joined #opensolaris
[02:02:30] *** doownek has joined #opensolaris
[02:02:30] *** Vanuatoo has joined #opensolaris
[02:02:30] *** TBCOOL has joined #opensolaris
[02:02:30] *** xzi11a has joined #opensolaris
[02:02:30] *** phips has joined #opensolaris
[02:02:30] *** nightswim has joined #opensolaris
[02:02:30] *** paulproteus has joined #opensolaris
[02:02:30] *** noyb has joined #opensolaris
[02:02:30] *** kirma has joined #opensolaris
[02:02:30] *** the-decider has joined #opensolaris
[02:02:30] *** milo has joined #opensolaris
[02:02:30] *** Stric has joined #opensolaris
[02:02:44] *** Spawrq has quit IRC
[02:03:16] *** gnu2it2 has quit IRC
[02:03:30] *** uebayasi has joined #opensolaris
[02:04:48] <nrubsig> Welcome to the Innternet after the "great reset" ...
[02:05:01] <dclarke> has the porn been removed ?
[02:05:08] <dclarke> that really annoyed me
[02:05:13] <dclarke> oh .. and the newbies
[02:05:16] <nrubsig> dclarke: no.
[02:05:17] <dclarke> are they gone now ?
[02:05:23] <nrubsig> dclarke: no.
[02:05:28] <dclarke> well .. damn
[02:05:33] <dclarke> reboot it again !
[02:05:50] <gdamore> hey all.
[02:05:56] <nrubsig> dclarke: but http has been replaced with mspds - M$ Page Delivers Service
[02:06:43] <dclarke> oh .. the new MicroSoft internet ?
[02:06:48] <nrubsig> dclarke: and you have to pay $5.50/month to Microsoft for using their internet
[02:07:12] <dclarke> cool .. 1 billion users fork over 5.50 a month
[02:07:37] <Auralis> nah, more like 0.01 for each page hit you cause
[02:08:03] <nrubsig> dclarke: and the feds announced that Microsoft was granted the right to print $5 banknotes at will.
[02:08:20] <dclarke> sound like Canada
[02:08:26] <nrubsig> heh
[02:08:35] <dclarke> oh .. we need more money .. let's just print it !
[02:08:38] <dclarke> morons
[02:08:43] <nrubsig> dclarke: just curious: is "canducks" a offense in canada ?
[02:08:56] <dclarke> you mean "cnuck" ?
[02:09:01] <dclarke> canuck
[02:09:02] <nrubsig> dclarke: *shrug*
[02:09:07] <nrubsig> dclarke: yes
[02:09:10] <dclarke> canuck is fine
[02:09:12] <nrubsig> dclarke: canuck
[02:09:19] <dclarke> as in .. hey .. he's a canuck
[02:09:24] <dclarke> thats fine
[02:09:30] <nrubsig> dclarke: someone in #mozilla is calling all *ca people canucks
[02:09:43] <nrubsig> dclarke: what does it mean ?
[02:09:51] <dclarke> you get into trouble when you say "hey .. hes a real f#$khead"
[02:10:11] <dclarke> canuck == Canadian
[02:10:14] <nrubsig> ah
[02:10:19] <dclarke> it has no negative connotation
[02:10:38] <nrubsig> dclarke: even if I say Joseph Stalin had a real f#$khead ?
[02:10:47] <nrubsig> :-)
[02:10:52] <dclarke> unless you want to talk about our fag loving liberal dope smoking government
[02:11:11] <dclarke> then .. well .. then we have problems
[02:11:29] <dclarke> a breif flash of fascism would clean this country up
[02:12:10] <nrubsig> dclarke: my fear would be: would there be anything left after the facists are done with your country ?
[02:12:24] <dclarke> not much
[02:12:38] <nrubsig> dclarke: and the US would finnally have a reason to declare war on *.ca and make it the 57th state of the US
[02:12:41] <dclarke> then the USA could come up here and build better highways
[02:12:52] <dclarke> fine with me
[02:13:04] <nrubsig> heh
[02:13:18] <dclarke> I was an officer in the Canadian army, a combat unit, and I hold a commission from the Queen
[02:13:33] <nrubsig> dclarke: even if they throw some nukes first to lower the costs in the war against the icemen ?
[02:13:42] <dclarke> and I have no qualms about someone flushing the flag down the toilet and saying .. let's start over
[02:13:58] <dclarke> but hey .. I'm a radical nut
[02:14:14] <dclarke> the kind of guy that likes to support open source software .. thats like . .communism
[02:14:27] <jamesd_> politicians are like diapers... both should be changed frequently for the same reason
[02:14:40] <dclarke> ha ha
[02:14:47] <dclarke> but .. I do like the country
[02:14:56] <dclarke> the land .. the whole grand massive thing
[02:15:13] <dclarke> its really freaking big
[02:15:31] <nrubsig> jamesd_: I would agree with you about our guys in berlin... but I seriiously would disagree about the politicans on the village/town level.
[02:16:34] <jengelh> gdamore : this is what i needed
[02:16:35] <nrubsig> I think if I ever get god-like powers I would forbet men to create cities with more than 12000 inhabitants.
[02:16:38] *** solaris-user has joined #opensolaris
[02:16:59] <jengelh> gdamore: http://svn.sourceforge.net/viewvc/*checkout*/ttyrpld/trunk/kpatch/solaris-11_snv39.diff?revision=36
[02:17:01] <nrubsig> that way we could avoid the problems of larger cities
[02:17:06] <nrubsig> gdamore: ping!
[02:17:12] <gdamore> ack!
[02:17:58] <nrubsig> gdamore: remeber the secret e 4 5 0 project ?
[02:18:06] <gdamore> ?
[02:18:31] <gdamore> oh that.  no word yet.
[02:18:35] <nrubsig> gdamore: we talked about it and you said you have to ask u k for a p r o v a l...
[02:19:06] <gdamore> the people that have to approve it are travelling, which makes it harder...
[02:19:23] <dclarke> has anyone here installed the secure global desktop ?
[02:19:37] <dclarke> it says for Red Hat Enterprise Linux 3.0
[02:19:40] <gdamore> if it weren't for the noise/power factor, I'd just set up the machine here at home and give you a login.
[02:20:04] <dclarke> gdamore : SGD ?
[02:20:14] <dclarke> I am asking about SGD here
[02:20:27] <jamesd_> dclarke, yes i installed it. works fine
[02:20:28] <dclarke> I wonder if I can install it onto RHEL 4.0
[02:20:37] <gdamore> SGD?  No, i've not used it.
[02:20:40] <dclarke> on Sol x86 or what ?
[02:20:53] <jamesd_> but has a habbit of getting in a wierd situation with a thread locked up...
[02:21:00] <jamesd_> i used it on solaris
[02:21:07] <dclarke> oh
[02:21:14] <dclarke> I need to run VMware also
[02:21:32] <jamesd_> its a pig. you should put vmware on another box
[02:21:35] <dclarke> I just don't see a way to run VMWare and provide Windows XP sessions via Solarsi x86
[02:21:46] <dclarke> its a pig ?
[02:21:50] <dclarke> really ?
[02:21:57] <dclarke> what sort of resources required ?
[02:21:59] <jamesd_> yeah...
[02:22:07] <jamesd_> just a second booting its zone
[02:22:22] <dclarke> you have it in a zone .. cool
[02:22:26] * dclarke impressed
[02:22:48] <gdamore> Isn't SGD the tarantella acquisition tech?
[02:22:51] <jamesd_> of course... its easy to setup .. and i just halt its zone when not needed.
[02:22:54] <jamesd_> yes
[02:23:26] <dclarke> oh can you fill out this puppy ?  http://www.blastwave.org/survey/index.html
[02:23:36] <dclarke> just to get a test from someone else
[02:23:50] <gdamore> is that to me, or someone else?
[02:24:02] <dclarke> anyone
[02:24:44] *** lloy0076 has joined #opensolaris
[02:25:03] *** gnu2it2 has joined #opensolaris
[02:25:05] <jamesd_> Thank you for being counted .
[02:25:09] <dclarke> maybe I can install RHEL 4 and then VMWare GSX server and then Solaris 10 x86 into that
[02:25:10] *** nwf has quit IRC
[02:25:14] <gdamore> so, i'm seriously starting to think that it would be cool to productize a NAS device loaded with snv, zfs, and a via epia cpu.
[02:25:20] <dclarke> thanks .. I'll fix that dumb message
[02:25:53] <jamesd_> its a still booting but this explains its bloat
[02:25:54] <jamesd_> ZONEID    NPROC  SIZE   RSS MEMORY      TIME  CPU ZONE
[02:25:55] <jamesd_>      1       43  651M  355M    18%   0:04:34  74% sgd
[02:26:27] <twincest> gdamore: is that fast enough for raid-z?
[02:26:54] <gdamore> i don't see why not, but i don't know.
[02:27:03] <jamesd_>  22660 101       163M   93M sleep   59    0   0:01:54 0.1% java/112
[02:27:03] <jamesd_>   <--- the main sgd task
[02:27:19] <jamesd_> i haven't even logged into yet.
[02:27:26] <dclarke> okay .. tw oemails received
[02:27:38] <dclarke> I need to fix the multiple version choices
[02:27:47] <dclarke> and create a field for number of hosts
[02:28:49] <jamesd_> okay i logged into it now...
[02:28:50] <jamesd_> ZONEID    NPROC  SIZE   RSS MEMORY      TIME  CPU ZONE
[02:28:50] <jamesd_>      1       71  834M  508M    25%   0:06:00  51% sgd
[02:30:01] <gdamore> i was going to try setting it up on a USIIe running at 440MHz.
[02:30:09] <gdamore> pardon, USIIi
[02:30:38] <gdamore> the via chips run at around 1GHz.  if they're not doing anything except raidz and NFS, i don't see why it would be a problem
[02:31:25] *** lloy0076 has left #opensolaris
[02:32:01] *** jsubl2 has joined #opensolaris
[02:32:11] <Gadzooks> Try running the Centos 4.4 server environment as well. ;>
[02:33:28] * nrubsig throws a stone at gdamore's window in giscave
[02:34:04] <nrubsig> Gman: ping!
[02:34:11] <Gman> pong.
[02:34:21] <nrubsig> Gman: is kupfer around ?
[02:34:50] <Gman> no
[02:34:58] <nrubsig> Gman: ;-(
[02:35:02] <nrubsig> Gman: ;'-(
[02:35:03] *** sahafeez has quit IRC
[02:35:07] <nrubsig> Gman: ;'-((
[02:36:06] <Gman> nothing i can do.
[02:36:07] <Gman> get over it.
[02:36:51] * nrubsig gets depressed and seeks a huge cliff and a cabrio...
[02:37:34] * nrubsig can't remeber the movie's name which started with the scene where the cabrio jumped off the cliff...
[02:38:02] <Auralis> themla and luise?
[02:38:16] <Auralis> amv hell 3
[02:38:19] <nrubsig> Auralis: that was at the end.
[02:38:26] <nrubsig> no
[02:38:37] *** gnu2it2 has quit IRC
[02:38:46] <nrubsig> and it was not "santa's slay" either
[02:39:18] <nrubsig> nor "Jack Baht the massmurder"
[02:39:41] <nrubsig> "Gremlins 2" started with another scene, too
[02:40:00] <kimc> jmaesd: test run output at: http://www.pastebin.co.uk/4556
[02:40:22] <alanc> nrubsig: jurassic's down so a lot of the MPK engineers went home early (and it's about normal going-home time for many anyway, since it is 17:40)
[02:40:27] <kimc> err jamesd:
[02:41:13] <nrubsig> alanc: ah
[02:41:15] <nrubsig> alanc: ok
[02:41:32] <nrubsig> alanc: you know how commands like /kick work, right ?
[02:41:48] <alanc> vaguely
[02:42:13] <nrubsig> alanc: you're only saying that because you know my next question...
[02:42:16] <nrubsig> :-)
[02:42:25] *** yippi has quit IRC
[02:42:30] <Gman> hrm, looks like the opensolaris mailing lists are down too?
[02:42:58] *** Alias`b` has joined #opensolaris
[02:43:01] *** broadcast has joined #opensolaris
[02:43:05] <broadcast> hi
[02:43:08] <Alias`b`> hello
[02:43:12] <Gman> oh, seemingly they are
[02:43:16] <Gman> just my mail being slow..
[02:43:25] <alanc> if you're looking to make me an op, I'm about to leave - just need to finish typing in my e-mail about the 7.2rc1 packages and what I know is broken in the them
[02:43:26] *** nwf has joined #opensolaris
[02:43:45] *** Alias`b` has left #opensolaris
[02:44:07] <nrubsig> gdamore: ping!
[02:44:14] * nrubsig wonders if gdamore died.
[02:44:51] <nrubsig> alanc: why is jurassic down ?
[02:44:57] <gdamore> ack.
[02:45:08] <nrubsig> gdamore: see --> cave
[02:45:13] <gdamore> i was reviewing my daughter's homework
[02:45:15] <alanc> I haven't heard a reason yet
[02:46:13] * nrubsig suspects the wabanti of omicron thata 4.8712 behind the jurrassic outage as a piece in their great plan to destroy humankind.
[02:46:28] <Gman> bad hardware
[02:46:54] * gdamore suspects that nrubsig relaly suspects that komodo dragons are behind said outage
[02:46:55] <hile_> nrubsig is an op?
[02:46:59] <hile_> phear!
[02:47:21] <nrubsig> hile_: this happens only becase noone else wants to be an op here.
[02:47:47] *** ChanServ sets mode: +o Tpenta
[02:47:54] <hile_> ah
[02:47:55] <Tpenta> ok, so i forgot to identify
[02:48:11] <hile_> we'll forgive you just this once ;)
[02:49:34] <Gman> who cares about op status.
[02:49:42] <hile_> not i
[02:49:45] <hile_> i was just being a smartass
[02:49:59] <Tpenta> like i said teh otehr day, if we need an op in here, you can generally get one pretty quickyl
[02:50:02] <gdamore> how often does an op actually do anything here?
[02:50:14] <Gman> Tpenta, right
[02:50:31] <Tpenta> well i've been an op pretty much since it started and I havent had to do anything (tho it came close once)
[02:50:32] <gdamore> i mean obviously they aren't paying attention, because _I'm_ still here... :-)
[02:55:45] <nrubsig> steleman: ping!
[02:57:03] <nrubsig> Gman: just curious: which kind of machine is jurassic ? E25k ?
[02:57:12] <Gman> dunno
[02:57:21] <Gman> it's an mpk based machine, i'm in nz
[02:58:27] <hile_> just guessing wildly -- and i'd have no basis for making an educated guess, it's just a guess -- i'd probably reckon the box were a 6800 or 6900
[02:59:06] <gdamore> jurassic used to be an E10k IIRC, i suspect these days it is a E15K or 25K
[02:59:19] <hile_> ah
[02:59:24] <alanc> the web page for it says it's a 6800 now
[02:59:34] <Tpenta> it's answering pings
[03:00:32] <alanc> I know one of the former admins was suggesting replacing it with a cluster of thumpers, but I don't think that's happening
[03:00:38] <nrubsig> Tpenta: any idea why http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=5079668 has no description ?
[03:01:22] <nrubsig> alanc: since when is thumper a replacement for a e15k ?
[03:01:37] <Tpenta> because the engineer who logged it put them in the comments field?
[03:01:44] <alanc> when your primary duty is a file server, not a compute server
[03:01:56] <nrubsig> Tpenta: and the name of the engineer is ?
[03:02:11] <hile_> somehow I doubt he would give that out
[03:02:24] <Tpenta> it was logged by a japanese engineer more than 2 years ago
[03:02:30] <nrubsig> Tpenta: or better: is there a reason for hiding the issue ?
[03:02:31] <Tpenta> that's about all you are going to get
[03:02:52] <Tpenta> no, because 2 years ago, that was the "accepted practice"
[03:02:58] <alanc> oh funny - part of the fix for it was removing an "#ifdef apollo"
[03:03:01] *** sahafeez has joined #opensolaris
[03:03:02] <Tpenta> not everything has a valid conspiracy theory
[03:03:19] *** vmhobbes has quit IRC
[03:03:43] <alanc> we still have to fight with some engineers who were trained to do that and haven't reset their old habits yet
[03:03:57] <boyd> The invalid conspiracy theories are the most fun :)
[03:04:04] <Tpenta> and it is also teh case that there are so many of these that there is no way ewe are going to go back and "fix" every one of those bugs so that it has a valid description. That would be an utter waste of manpower
[03:04:05] <boyd> Hi Tpenta, alanc
[03:04:15] <Tpenta> hi boyd
[03:04:20] <hile_> howdy boyd
[03:04:42] <Tpenta> alanc: i was asked something yesterday that I'm sure you were discussing a few days ago, ... i was being asked about video on sunray
[03:04:43] <boyd> Hi hile_
[03:05:01] <nrubsig> Tpenta: was there any recent activity on the bug ?
[03:05:11] * hile_ needs another display and another sun ray
[03:05:22] <nrubsig> Tpenta: because it is listed as one of the first in my "ksh" query
[03:05:31] <Tpenta> yes, it had an engineer pick it up during the last week
[03:05:40] <nrubsig> Tpenta: why ?
[03:05:47] <dclarke> video on SunRay ..
[03:05:52] <dclarke> does that work now ?
[03:05:56] <Tpenta> roland: I see a modified date. I can't channel the person who picked it up
[03:06:00] <dclarke> 20 fps or more ?
[03:06:10] <alanc> what about video on Sun Ray?
[03:06:22] <alanc> it's always worked, just not very well
[03:06:25] <Tpenta> i vaguely recall you saying something on that topic in the last week or so
[03:06:26] <Gman> Tpenta, add roland to the interest list
[03:06:29] <dclarke> does video on sunray work well now ?
[03:06:33] <alanc> no
[03:06:40] <Gman> then he'll accidently see who makes the changes when someone updates the pages ;)
[03:06:42] <dclarke> oh
[03:06:43] <Tpenta> ok, must be imagining things
[03:07:02] <Gman> s/pages/bug/
[03:07:04] <alanc> Tpenta: the no was to dclarke
[03:07:12] <Tpenta> ahh
[03:07:12] <alanc> yes, I've talked to people about it here recently
[03:07:17] * nrubsig curses the fact that more tim gets wasted on the old Solaris /usr/bin/ksh monster.
[03:07:25] <Tpenta> was there anything being done with it?
[03:07:34] <broadcast> when i use format in install it sayme segmentation fault (core dump) http://www.pget.org/index.php?id=c5caf04b6a for see truss http://www.pget.org/index.php?id=076f5f31a1 :( i trying 2 days.
[03:07:35] *** bluejay has joined #opensolaris
[03:07:42] <alanc> Sun Ray is working on Xvideo support as part of their Xorg port
[03:08:02] <nrubsig> alanc: BTW: Xvideo for Xsun would be quite easy... :-)
[03:08:03] <alanc> but the hardware design simply will not allow full screen video at full frame rate no matter what we do
[03:08:07] <Tpenta> roland: reality check. We have customers running solaris 7,8,9,10. THese all have the old ksh on them. These customers pay us money. We fix things for them.
[03:08:10] <nrubsig> alanc: remeber my RFE :-)
[03:08:13] <gdamore> that would be good.  there is already YUV decode in sun ray
[03:08:22] <dclarke> in the mean time .. the new local login screen ( dtlogin ? ) has really choppy looking fonts
[03:08:22] <gdamore> alanc++
[03:08:30] <Tpenta> alanc: that's probably what I was trying to remember
[03:08:31] <Gman> Tpenta, though it doesn't look like that bug is an escalation
[03:08:32] <alanc> nrubsig: it would be damn near impossible actually
[03:08:41] <nrubsig> alanc: why ?
[03:08:42] <dclarke> the screen is crisp .. clean .. nice looking blue
[03:08:55] <dclarke> but the "Welcome to deimos" is blocky
[03:09:03] <Tpenta> true, but it was logged against 5.10, and could turn into one at some stage. Basically if a customer reports a bug, we log it.
[03:09:08] *** bluejay has left #opensolaris
[03:09:28] <alanc> because no one who can access the Xsun source code will do it, and no one who wants to do it has access to Xsun source code
[03:09:44] <dclarke> alanc : what ?
[03:09:51] <nrubsig> Tpenta: I am aware about the reality. but my comment was about the issue that it was reported against s10_63
[03:09:54] <dclarke> alanc : you mean the login screen fonts ?
[03:09:57] <Tpenta> alanc: so what you are saying is that once someone sees teh Xsun source code, they want nothing to do with it?
[03:09:59] <gdamore> and no one who has access to Xsun will let anyone else have access... :-)
[03:10:03] <alanc> AUUGH!
[03:10:11] <alanc> too many people talking at once!
[03:10:16] <Tpenta> :)
[03:10:16] <dclarke> alanc : sorry
[03:10:21] <gdamore> hehe.  i think alanc just exploded.
[03:10:28] * dclarke logs into his new build
[03:10:29] <Tpenta> that could be messy
[03:10:31] <alanc> dclarke: fonts are busted in a whole bunch of nv_4* builds
[03:10:40] <dclarke> alanc : thank you
[03:10:49] <dclarke> alanc : sorry for being pushy
[03:11:00] <gdamore> actually, i think i have Xsun source  -- but not recent, and certainly not the sun ray DDX module
[03:11:02] <alanc> dclarke: there's been a workaround posted a few times - real fix is in nv_50 I think
[03:11:04] <dclarke> today is an Official Blastwave birthday .. I'm in a good mood
[03:11:17] <dclarke> and playing with lots of new stuff is on my list today
[03:11:18] <Tpenta> happy bday blastwave
[03:11:27] <dclarke> but my new login screen .. looked a little choppy
[03:11:28] <gdamore> Tpenta++
[03:11:37] <nrubsig> gdamore: my proposal for Xvideo support is generic and does not require DDX access.
[03:11:42] <alanc> Tpenta: I'm saying that Sun employees with access are paid to work on Xorg instead and EOL'ing Xsun is on our goals, not improving it
[03:11:50] <dclarke> alanc : I guess the source is not open and there is no way to get sub-pixel hinting etc etc
[03:11:55] <nrubsig> gdamore: my idea was to add a Xvideo software driver available as fallback
[03:11:55] <gdamore> i didn't see your original proposal.
[03:12:11] <Tpenta> alanc: I was just trying to twist your words to reflect on an assumed code quality ;)
[03:12:12] <alanc> dclarke: the references to Xsun source were completely unrelated to the font thread
[03:12:19] <gdamore> what would a software driver give you?
[03:12:32] <alanc> dclarke: the Xsun source answers were in response to gisburn's questions, not yours
[03:12:48] <Tpenta> we need thread ids ;)
[03:13:05] <nrubsig> gdamore: https://bugs.freedesktop.org/show_bug.cgi?id=2635
[03:13:10] <alanc> dclarke: but yes, the Xsun font code is third-party licensed, and patent-encumbered, so there's no chance in hell of us opensourcing it
[03:13:24] <nrubsig> gdamore: it gives you server-side scaling which is one of the very heavywheight stuff in X11
[03:13:43] <dclarke> alanc : thanks
[03:13:44] <boyd> Tpenta: I agree... The screen could split vertically
[03:14:10] <alanc> dclarke: JDS font support uses the open source freetype library - if you want better font hinting in it, read the freetype docs about how to turn it on
[03:14:29] <nrubsig> gdamore: think about it: what is faster: copying 352x288 to the Xserver and let it scale up on the server side or scale on the client side and push 1600x1200 images to the Xserver ?
[03:14:29] * Gman snickers
[03:14:40] <gdamore> apart from bandwidth for a remote X server, i don't see much gain
[03:14:45] <alanc> dclarke: and by mentioning the evil "P" word I've probably gone way past what the lawyers are happy with me saying
[03:15:05] <dclarke> its okay
[03:15:06] <nrubsig> gdamore: it is a very real issue on local Xservers, too.
[03:15:10] <dclarke> there are no lawyers here
[03:15:11] *** piwi has quit IRC
[03:15:31] <alanc> Tpenta: if you'ld seen the Xsun source code, you'ld want nothing to do with it either
[03:15:34] <nrubsig> gdamore: and you get better load balancing on systems with more than one CPU when you move more load to the Xserver side.
[03:16:13] <alanc> now I'm going home so my head can stop spinning from the questions from all sides at once
[03:16:15] *** vmhobbes has joined #OpenSolaris
[03:16:18] <gdamore> i'm not sure i agree with that statement.  but, a software driver Xsun is very unlikely to happen.
[03:16:24] <dclarke> alanc : but we love you
[03:16:25] * gdamore wants Xsun to die, die, die.
[03:16:32] <dclarke> alanc : does that help
[03:16:37] * alanc is making Xsun die, die, die
[03:16:38] <dclarke> everyone hug alanc
[03:16:41] <gdamore> dclarke++ :-)
[03:16:42] <nrubsig> gdamore: even a Xvideo software driver for Xorg would make much sense.
[03:16:55] <nrubsig> gdamore: Xvideo hardware support is usually limited to one slot.
[03:16:55] <gdamore> roland: so write one. :-)
[03:17:04] *** nwf has quit IRC
[03:17:09] <nrubsig> gdamore: feel free to pay me for that... :-)
[03:17:24] <alanc> "Xsun EOL" is only a few cases down on my ARC case writing to-do list
[03:17:28] <gdamore> why would i?  _I_ am not the one complaining about it... :-)
[03:17:36] *** alanc is now known as alanc-away
[03:17:54] <gdamore> later alan, and as I said, if I can do anything at all to help with the Xsun EOL, please let me know.
[03:18:01] <nrubsig> gdamore: I've looked at the Xvideo stuff and I can only say that the authors must have enjoyed lots of dope while they wrote the extension
[03:18:03] <gdamore> i have mgmt buyin for helping you, i think.
[03:18:19] <gdamore> (last was to alan)
[03:18:42] <alanc-away> gdamore: thanks
[03:19:11] <alanc-away> gdamore: btw, SPARC graphics mgr came by today asking about your note...
[03:19:12] *** nwf has joined #opensolaris
[03:19:23] <gdamore> i'll just be glad when I can stop hacking away on pfb,m64 sources.
[03:19:25] <alanc-away> he may be responding soon
[03:19:28] <gdamore> tks.
[03:19:55] <nrubsig> gdamore: if you do m64 hacking please consider adding MPSS support to the driver.
[03:20:07] <nrubsig> gdamore: mapping the framebuffer with 64k or 512k pages would be a huge win
[03:20:42] <gdamore> i'm only doing the minimum hackng to support the LCDs on our laptops.  major changes to the framebuffer code is not something i want to get involved in.
[03:20:55] <nrubsig> gdamore: erm
[03:21:10] <nrubsig> gdamore: this is just one tiny call to make sure it doesn't use 8k mappings.
[03:21:40] <gdamore> i don't know anything about MPSS, is it sparc or x86 only?
[03:21:47] <nrubsig> gdamore: UltraSPARC-1/2 can cycle through all 64 MMU slots when drawing a diagonal line (which is sick)
[03:22:07] <nrubsig> gdamore: MPSS is generic but it only makes much sense on UltraSPARC right now.
[03:22:21] <gdamore> if you send me sample code for setting up a mapping, i can try to take a look at it.
[03:22:49] <gdamore> of course, it might not get back into any common sun code.  i don't own that code, only our private fork of the code for Tadpole laptops.
[03:22:52] <nrubsig> gdamore: x86 only supports 4k+4M or 4k+2m pages while ultrasparc usually supports at least 8k, 64k, 512k, 4m pages
[03:23:34] <gdamore> this would make sense for pfb as well, i think, if it doesn't already DTRT
[03:26:51] *** gnu2it2 has joined #opensolaris
[03:28:41] <dclarke> ssilly question : will 32-bit applications run on a T2000 ?
[03:28:58] <dclarke> my understanding is yes - full backward compatibility
[03:29:05] <dclarke> but I am not sure
[03:29:18] <jamesd_> yes they will but it requires solaris 10  a 64 bit OS.
[03:29:19] <nrubsig> dclarke: yes they will.
[03:29:44] <dclarke> well Solaris 10 or higher is all I would consider looking at
[03:29:46] <nrubsig> dclarke: but don't expect that FP-heavy stuff runs fast on these machines
[03:29:57] <dclarke> no .. I understand that fully
[03:30:00] <nrubsig> dclarke: and thinks like quake2 won't run faster as on a Blade100
[03:30:07] <dclarke> there is only one FPU on the internal bus
[03:30:07] <jengelh> nrubsig : as long as it keeps the pizza warm... it's good
[03:32:43] *** vmhobbes has quit IRC
[03:39:30] *** jsubl2 has quit IRC
[03:41:17] *** cylix has joined #opensolaris
[03:41:57] *** cylix has left #opensolaris
[03:42:24] *** gnu2it2 has quit IRC
[03:48:22] *** gnu2it2 has joined #opensolaris
[03:48:59] <elektronkind> yes, 32bit apps run fine on a solaris 10-running T1000
[03:49:30] <elektronkind> besides, only solaris 10 supports sun4v, so that's the only solaris you can run on a t1000
[03:51:22] <gdamore> i can't imagine that Sun would ever consider a processor that runs Solaris/SPARC and doesn't support 32-bit userland code.  it would break way, way too many ISVS.  nobody would buy it
[03:51:46] <gdamore> the kernel is totally different however.  all new Sun machines are 64-bit only.  (Sun/SPARC that is)
[03:51:55] *** mr_goose has joined #opensolaris
[03:52:01] <boyd> Many ON programs are 32 bit
[03:52:02] <nrubsig> grrrrr... where are the B51 sources ?
[03:52:33] <elektronkind> 98% of ON executables are 32bit...
[03:52:51] <gdamore> only 98%?  I would have thought it would be closer to like 99.5%.
[03:52:56] <jamesd_> nrubsig, a ksh script broke  so they can't realease it yet... if you beleve that I have a  bridge to sell you...
[03:53:00] <elektronkind> :)
[03:53:11] <boyd> Ok, so who's gonna run the "find" :)
[03:53:40] <gdamore> actually, i've seen sun products get delayed for "seemingly small" issues like this in the past.  it wouldn't surprise me if it were true.
[03:53:53] <gdamore> but, jamesd, i'm not in the market for any bridges, thank you.
[03:53:55] <nrubsig> gdamore: you soon will have to run a 64bit executable for your dialy work... :-)
[03:54:07] <nrubsig> gdamore: and you may not even notice it... :-)
[03:54:18] <gdamore> i already do.  its called "genunix". :-)
[03:54:24] <nrubsig> gdamore: userland
[03:54:49] <elektronkind> find /usr -type d -name sparcv9 -exec ls -l {} \;
[03:54:51] <nrubsig> gdamore: on sparc it's 64bit by default thanks to isaexec
[03:55:12] <nrubsig> gdamore: guess it...
[03:55:17] <gdamore> if some standard utilities were 64bit, that would be ok.  but if i _can't_ run 32-bit legacy apps, then I'd be pretty unhappy
[03:55:32] <gdamore> ksh93?
[03:55:37] <nrubsig> gdamore: yes
[03:55:50] <gdamore> but i'm a tcsh user! :-)
[03:55:55] <elektronkind> [daleg@ds1]~>find /usr/bin /usr/sbin /usr/openwin/bin -type d -name amd64 -exec ls -l {} \; | wc -l
[03:55:56] <elektronkind>       64
[03:56:01] <elektronkind> there you have it.
[03:56:02] <nrubsig> gdamore: basically we ship two flavors, a 32bit one and a 64bit one
[03:56:12] <elektronkind> how ironic. 64 64bit executables
[03:56:22] <hile_> gdamore: crazy ;-)
[03:56:25] <elektronkind> (for amd64)
[03:57:06] * boyd wonders what /usr/bin/java does, since it's 32bit and has a -d64 option
[03:57:23] *** jsubl2 has joined #opensolaris
[03:58:00] <gdamore> wtf? _ls_ is 64bit?
[03:59:12] <gdamore> and gtk-demo 64-bit is just nuts!
[04:00:18] <gdamore> roland: what is the point of a 64-bit ksh?  do we expect people to create humongous arrays in ksh scripts?
[04:01:01] <elektronkind> gdamore: well, is there a reason you should prevent them from doing so? :)
[04:01:05] <boyd> Ah... /usr/bin/java calls /usr/jdk/instances/jdk<something>/bin/{,sparcv9}/java
[04:01:06] <jamesd_> x64 has access to many more registers.. so runs faster whether or not it uses big arrays.
[04:01:31] <gdamore> ah, didnt know that
[04:01:33] *** gnu2it2 has quit IRC
[04:01:54] *** vmhobbes has joined #OpenSolaris
[04:02:24] <boyd> The trade of is that pointers are larger so it may be slower/use more memory in some ways. Registers win in most cases
[04:02:31] <boyd> s/trade of/trade off
[04:02:53] <gdamore> elektronkind: well, if it costs more to support two versions, then maybe it isn't worth doing.  see also boyds message about memory consumption
[04:03:16] <nrubsig> gdamore: the original reason was to allow applications use more than 2GB of data. ksh93 no longer has the artificial array size limitation of ksh88 and allows to processs a vast amount of data quite efficiently.
[04:03:17] <gdamore> i'm not sure about the trade off.  there is the issue of cache effects as well.
[04:03:40] <elektronkind> if you
[04:03:40] <mlh> I'm looking forward to having a HERE doc > 2gb
[04:03:41] <nrubsig> gdamore: another reason are binary plugins which may require 64bit code
[04:03:58] <elektronkind> if you're running huge shell scripts, cache effects are probably not a concern
[04:04:11] <gdamore> okay.  it sounds to me like ksh93 aspires to be perl, though.  i'm not convinced that is a good aspiration for a shell.
[04:04:21] <boyd> indeed.
[04:04:28] *** gnu2it2 has joined #opensolaris
[04:04:51] <gdamore> that said, i wrote a really cool application in dtksh once.
[04:04:53] <boyd> nrubsig: Another way to say that is that binary plugins *will* be required to be 64 bit code, since the shell is.
[04:04:54] <elektronkind> whatch out gdamore, that there is fighting words
[04:05:01] <nrubsig> gdamore: do you think there should be an artificial limitation for the a mount of data a shell script can process.
[04:05:15] <gdamore> i'm not sure the 2GB limitation is "artificial". :-)
[04:05:32] <nrubsig> gdamore: I've seen biosh processes with more than 8GB heap
[04:05:42] <nrubsig> (biosh is a hacked ksh93 for bioinformatics)
[04:06:01] <elektronkind> MPI-enabled ksh?
[04:06:03] <gdamore> and for portable scripting, you can't use it.  because 32-bit platforms won't be able to run it.
[04:06:04] * boyd wonders about a message from the shell that says "Hmm... you're using more than 2GB of memory for this shell script. Maybe you should be thinking about a different tool."
[04:06:15] <gdamore> boyd++
[04:06:25] <gdamore> can we have that for perl, too? :-)
[04:06:30] <boyd> heh
[04:06:33] <nrubsig> boyd: the same problem applies to perl and phyton
[04:06:50] <nrubsig> remember ksh93 uses compiled shell code internally.
[04:06:59] <gdamore> actually, i'd be happy, if it said "Hmm... you're using perl.  Maybe you should be thinking about a different tool." :-)
[04:07:06] <boyd> Of course, all these numbers would have to take into account the current date and inflate as hardware gets faster/bigger
[04:07:23] <gdamore> note the lack of numbers in my last.
[04:07:28] <boyd> :)
[04:07:33] <nrubsig> gdamore: and ksh93 is fast. the mandelbrot demo runs 1/3 of the speed of native, hand-optimized code.
[04:08:03] <nrubsig> which is IMO impressive.
[04:08:04] <gdamore> 1/3 is nothing for small data sets.  but if i'm going to process multiple gigabytes of data, i want something fast, i think.
[04:08:26] <nrubsig> gdamore: not if you only need to juggle around string data.
[04:08:45] <gdamore> nrubsig: btw, folks have said the same thing about perl.  i don't think it is a good tool for writing very big scripts in though.
[04:09:02] <boyd> I'm inclined to think that the attributes of the ksh *language* make it not the best tool for big jobs, not the implementation details and optimisation
[04:09:31] <gdamore> if the script is small, but needs to chew on large chunks of data, maybe its okay.
[04:09:41] * boyd nods
[04:10:10] <gdamore> i just hate qualifying two sets of tools, when 99% of use cases would never need the second qualification
[04:10:26] <boyd> But still... it feels to me a little like bolting objects onto Fortran
[04:10:37] <gdamore> yeah, well, it sort of is like that. :-)
[04:11:24] <gdamore> i also think that down this road leads: 64-bit grep, 64-bit awk, 64-bit sed, etc.. ad naseum.  and we still need the 32-bit versions...
[04:11:48] <gdamore> 64-bit code also runs slower on many platforms if you don't need the extra registers.
[04:12:26] <gdamore> (i don't think you get any more registers on sparc, but you can use 32-bit addresses with all other v9 features except 64-bit with sparcv8plus)
[04:12:41] <boyd> Indeed.. someone might want to create a 10G array in nawk
[04:13:26] <boyd> ... or process a 3GB line in sed :)
[04:14:09] <gdamore> heh.  i think sed/awk actually have non-32-bit related "artificial" limitations anyway.
[04:14:48] <boyd> Indeed. I'm glad the stupid terminal width limit has gone from vi
[04:15:11] <gdamore> personally, i'd be happy to see the 64-bit version of ksh93 put on the supplemental CD, and only ship a 32-bit version in the core OS.
[04:16:05] * gdamore wants to build a NAS appliance device from a minimized install, and doesn't want any bloat in SUNWc*r packages that isn't strictly necessary
[04:17:22] <boyd> What you really need is a sliced up SUNWcsr
[04:17:46] *** jsubl2 has left #opensolaris
[04:18:58] <nrubsig> gdamore: the 64bit ksh93 still makes sense. at least on AMD64 it is noticeable faster.
[04:19:13] <nrubsig> gdamore: and even on sparc it's a tiny bit faster
[04:19:39] <nrubsig> and if you use lots of math then 64bit ksh93 wins anyway.
[04:21:00] <nrubsig> and I disagree on the grep/sed thing.... these are stream filters.
[04:21:24] <nrubsig> awk may benefit from a 64bit version but I never saw an awk script requiring anything beyond 100MB
[04:21:27] <gdamore> what about nawk?
[04:21:43] <nrubsig> ksh93 can easily go beyond 8GB
[04:21:43] <gdamore> i can say the same about ksh. :-)
[04:22:04] <nrubsig> gdamore: ksh93 has associate arrays which can grow very very large.
[04:22:11] <gdamore> i'm still skeptical about the benefits.
[04:22:15] <boyd> can != should
[04:22:23] <nrubsig> you can even built complex trees and graphs in ksh93
[04:22:29] <boyd> can != should
[04:22:34] <nrubsig> gdamore: I am not sceptical.
[04:22:35] <gdamore> frankly if my *shell script* grows to > 1GB, then i'd rather have it crash, because it means something is wrong. :-)
[04:22:56] <nrubsig> gdamore: and 64bit is a selling poing for sun equipment.
[04:23:17] <gdamore> 64-bit binaries.  i don't think anyone is selling 64-bit shell scripts. :-)
[04:23:19] <nrubsig> gdamore: and ksh93 should be "better" than the old ksh in every imaginable point.
[04:24:04] <boyd> nrubsig: You have not established that it should be, or that it's better
[04:25:04] <Auralis> well, with the ever growing datasets and stuff people throw around, i think having a 64bit shell will be at one point needed anyway
[04:25:07] <gdamore> boyd: careful, or nrubsig will throw you into a pit full of reptiles. :-)
[04:25:37] <gdamore> i think there may be people who will want, for special purposes, a shell capable of groking 64-bits.
[04:25:42] <boyd> :)
[04:25:43] <nrubsig> gdamore: I can't do that as long as I am /op'ed. I can't threaten people with certain death.
[04:26:04] <boyd> Now *that's* an outcome I would not have predicted
[04:26:05] <gdamore> hahahaha.... you're fair game then! :-)
[04:26:58] <nrubsig> boyd: why ? As long as I have an official function in this channel I can't randomly attack people.
[04:27:15] <boyd> I only meant that from a rhetorical point of view the point was unsupported
[04:27:25] <nrubsig> boyd: for the same reason I can't kick&ban people without a reason.
[04:27:30] <gdamore> anyway, i think that those special cases can be handled separately from the "stock" ksh that is used for "sane" shell scripting and login shell.
[04:27:41] <boyd> nrubsig: Of course,I just didn't think it through
[04:28:20] <gdamore> i wouldn't have thought that a) the attack was random, b) was serious, or c) was likely to be taken as serious by anyone watching.
[04:28:27] <nrubsig> gdamore: what about the amd64 performance improvement.
[04:28:41] <gdamore> _that_ is the part i'm skeptical about
[04:29:08] <gdamore> for large data sets there may be a performance improvement, but the normal typical usage case i'm doubtful of.
[04:29:40] <elektronkind> give nrubsig a break and allow him to live (or rather, threaten) vicariously via komodo dragons
[04:29:41] <gdamore> and, having to call isaexec will *definitely* outweight any small gains.
[04:30:00] <nrubsig> gdamore: no, my comment is that I can't attack people while being an /op ... I need a reason (a good one) which I can quote in the public.
[04:30:36] <nrubsig> gdamore: and "I don't like his face" is not a suiteable reason.
[04:30:41] <gdamore> understood.  i'm just surprised there was a concern that it would not be seen as humorous and ignored. :-)
[04:30:47] <boyd> What's wrong with my face? :)
[04:31:00] <gdamore> now a direct insult, like "I don't like his face", _that_ is clearly a personal attack. :-)
[04:31:32] <gdamore> nothing.  what's wrong is nrubsig's personal preference for komodo dragons? :-)
[04:31:36] <boyd> Geez, will ya all stop picking on my face! :)
[04:31:44] <gdamore> ROTFL
[04:32:10] <nrubsig> gdamore: even if people attack me at this level I should not kick them. Only if they threaen the operation of the channel or seriously offend channel subscribers for a a long period then a kick may be considerable
[04:32:28] <boyd> hear hear, nrubsig. Well said
[04:32:38] <gdamore> ++boyd,nrubsig
[04:33:03] <boyd> Everything I say is in good faith
[04:33:29] <gdamore> going back to my earlier comment, I predict that the fork/exec cost implied by isaexec will definitely outweigh any small performance improvements for the vast majority of shell scripts
[04:33:50] <gdamore> as such, i think 64-bit execution of ksh should be _explicitly_ requested, and not implicit.
[04:34:28] <nrubsig> gdamore: I would agree with you on a 50MHz turbosparc. On a 200MHz UltraSPARC it does not matter anymore - and yes, we did benchmark this issue to death.
[04:34:51] <gdamore> wtf?  first you say the performance matters, now you don't?  well, which way is it?
[04:35:03] * boyd lunches
[04:35:03] <nrubsig> gdamore: and I even tweaked isaexec() to be faster.
[04:35:26] <nrubsig> gdamore: the overhead by isaexec(1) is just a exec(), not a fork()
[04:36:19] <gdamore> still, exec is not cheap.  for long lived shell scripts, its in the noise.  but for most short lived applications (which is true for most everything except interactive shells) it will probably cancel out any performance benefit.
[04:36:21] <nrubsig> gdamore: and the if the script runs longer than 2secs than the 1/50sec needed to run isaexec() is for /dev/null
[04:36:58] <gdamore> in my experience, ~all scripts that run longer than 2secs spend all their time waiting on child processes.
[04:37:22] <nrubsig>  $( ... ) in ksh93 no longer forks a child process
[04:37:46] <boyd> how on earth can you do that?
[04:37:50] <nrubsig> just a "subshell" which is just another Shell_t instance
[04:37:59] <gdamore> yes, and ksh93 is no doubt a big improvement.  but still, it has to fork/exec for things like sed,awk,grep, etc.
[04:38:08] <boyd> Only if you special case it for builtins.
[04:38:16] <nrubsig> boyd: it puts the current Shell_t on the stack and creates a new one.
[04:38:28] <boyd> you cant do it for the general case
[04:38:32] <nrubsig> boyd: ( ... ) & # will still require a fork()
[04:38:36] <twincest> boyd: it could dlopen() the program and call main()
[04:38:55] <boyd> myzone=$(zonename)
[04:38:56] <nrubsig> twincest: that's already done for builtin commands
[04:39:04] * gdamore hates to think of the debugging implications...
[04:39:04] <twincest> really?  how stupid
[04:39:22] <nrubsig> twincest: the symbol names are cached.
[04:39:26] <boyd> twincest: good grief! That's in evil thought
[04:39:31] <nrubsig> twincest: and there is a 2nd way
[04:39:40] <boyd> Of course if main() calls exit....
[04:39:47] <twincest> boyd: i was actually joking, sad to see i'm right :)
[04:39:48] <nrubsig> twincest: this is why there is a libdll with complex functionality :-)
[04:40:06] <gdamore> yikes.
[04:40:29] <nrubsig> boyd: this is why builtins need to include <shell.h> which #defines a different exit() function
[04:40:39] <nrubsig> which calls setjmp()
[04:40:51] <boyd> It's not builtins I'm talking about, it's the general case of $( )
[04:41:08] <nrubsig> boyd: $( ... ) # does not require a subprocess
[04:41:15] <gdamore> as in x = $(grep myname /etc/hosts)
[04:41:19] <nrubsig> boyd: only external applications need fork()
[04:41:20] <boyd> indeed
[04:41:28] <boyd> That is "the general case"
[04:41:35] <boyd> (or included in it)
[04:41:58] * gdamore thinks that nrubsig believes that he can replace /genunix with /bin/ksh93. :-)
[04:42:01] 
[04:42:14] <boyd> To say that $(..) does not require a fork for my scripts at least would be quite inaccurate in most cases
[04:42:19] *** B|nTaRa is now known as B|nTaRa-village
[04:42:38] <gdamore> why don't they just go all the way and incorporate full busybox into ksh93?
[04:42:41] * boyd lunches before he rants on the toolbox philisophy
[04:44:13] <gdamore> historically, shells (especially login shells) were used to execute lots of other helper programs.  what I'm hearing is that ksh93 is mostly abandoning that, and trying to incorporate as much of /usr/bin into itself as possible.  am i misunderstanding?
[04:44:42] <gdamore> (see also my earlier comments about ksh93 aspiring to become perl)
[04:45:23] *** rachel has quit IRC
[04:46:33] *** sahafeez has quit IRC
[04:48:49] <nrubsig> gdamore: yes and no. some common applications are "embedded" for performance or functionality reasons
[04:49:14] <nrubsig> for example printf(1) is builtin to make sure it can deal with the new datatypes (floating-point) correctly
[04:49:26] <nrubsig> or sleep(1) which supports sub-second timing.
[04:49:50] <nrubsig> find(1) should IMO be a builtin, too - that way you could call shell functions, too.
[04:50:09] <nrubsig> all other stuff is optional
[04:52:11] <jamesd_> nrubsig, why not just link in  schilly's  find library  then you can add in tar as well ;-)
[04:52:42] <nrubsig> james_: very funny.
[04:52:44] <gdamore> and i want a pony, too!
[04:53:01] *** jamesd_ is now known as jamesd
[04:53:04] <nrubsig> james_: tar(1) is not an application which is called often in a shell script
[04:53:08] *** ChanServ sets mode: +o jamesd
[04:53:21] <boyd> Actually, the find lib would be a good way to do it.
[04:53:32] <nrubsig> jamesd: things like printf(1) are called very often in a close loop.
[04:53:35] <boyd> It's not tar specific, and there is moves to integrate star
[04:53:46] <nrubsig> boyd: yes, I know.
[04:54:07] <nrubsig> there are... other... problems.
[04:54:10] <nrubsig> difficult ones.
[04:54:12] <gdamore> in the printf() case, it isn't that much, because you flood the console.  but i agree that its a good thing to have builtin
[04:54:16] <nrubsig> noone of them technical.
[04:54:24] <Auralis> vitualization, the thrid OS, solaris, emacs and now ksh93 :)
[04:54:28] <boyd> .. it's not written in ksh93? :)
[04:54:51] <nrubsig> gdamore: did you see my gnaw.ksh script ?
[04:54:56] <gdamore> ksh93 written in elisp, and an elisp implementation of ksh93. :-)
[04:55:04] <nrubsig> gdamore: it shows heavy usage of printf(1)
[04:55:15] <gdamore> i saw the screen shot.  but i don't think it is a typical script.
[04:55:28] <boyd> Little doubt there :)
[04:55:55] <gdamore> ksh93 is trying to step from being a shell scripting language to a full fledged development language.  i'm not sure that is a sensible move.
[04:56:23] <nrubsig> gdamore: I could cite other scripts. passing around results from shell functions usually requires to return them via echo(1), print(1) or printf(1)
[04:56:44] <gdamore> that said, there are some nice things.  i have used bitwise math in shell scripts, but that's probably a bit abnormal. (in my case it was to process netmasks/ip addresses)
[04:56:55] <broadcast> what's this shit? http://www.pget.org/index.php?id=c0d7b7d510
[04:57:25] <gdamore> okay, i give on printf() -- i've already agreed it should be a builtin.
[04:57:37] <gdamore> maybe ln -s /usr/bin/ksh93 /usr/bin/printf ?
[04:57:47] <nrubsig> gdamore: there is a patch for that.
[04:58:04] <nrubsig> gdamore: not via ln but via calling the b_printf code in libshell direcly.
[04:59:10] <boyd> broadcast: Is that a restricted networking install?
[04:59:13] <gdamore> i wouldn't mind having a "standard" builtin command, that used argv[0] to figure out what to do, and used libshell.  at least that would amortize the cost of keeping around two different versions of this stuff.
[04:59:20] <gdamore> (both 32 and 64 bit)
[05:00:01] <broadcast> no
[05:00:08] <nrubsig> gdamore: erm, you don't want that. the symbol lookup is very heavywheight.
[05:00:18] *** solaris-user has quit IRC
[05:00:19] <boyd> Hmm... because a minimal or restricted one failed for me like that last week
[05:00:26] <nrubsig> gdamore: and things like sleep(1), test(1) and printf(1) only need the 32bit libs
[05:00:46] <gdamore> sure.
[05:00:57] <nrubsig> and pwd(1)
[05:01:28] <gdamore> pwd is "weird", because it records shell history.  but i guess /bin/pwd doesn't.
[05:01:51] <icon> bah
[05:02:03] <icon> do any of you guys know how to figure out what the domain controller is for a win2k host?
[05:02:04] <nrubsig> gdamore: shell history ?!
[05:02:25] *** rachel has joined #opensolaris
[05:02:28] <nrubsig> IMO the peson who set PATH_MAX to 1024 in solaris should die a horrible fate.
[05:03:12] <boyd> icon: You might like to try somewhere where that question is slightly relevant
[05:03:30] <gdamore> nrubsig: if you have multiple ways to get to a path, pwd reports how you got there, but /bin/pwd reports by going back up .. etc.
[05:03:47] <icon> boyd: bleh, working on some ldap garbage
[05:04:17] <gdamore> icon: start powring of machines in the lab?  when the domain stops working, bam, that's the machine.
[05:04:23] <boyd> You're talking about pwd vs pwd -P
[05:04:29] <boyd> gdamore: hehe
[05:04:50] <icon> gdamore: haha that would be nice except chances are, my dn server is probably somewhere across the country
[05:04:51] <gdamore> of course, i wouldn't notice, because i don't have any windows clients.  but i don't have any windows servers, either. :-)
[05:05:04] <icon> meh, im at work, so its certainly by force
[05:05:34] <nrubsig> right now /bin/pwd has no -P or -L option
[05:05:43] <nrubsig> using b_pwd from libshell will solve that.
[05:06:04] <gdamore> i suspect the PATH_MAX stuff comes out of legacy from filesystems and tools that couldn't tolerate arbitrarily large path names.
[05:06:14] <boyd> I'm talking about the builtin (in ksh at least)
[05:06:17] <gdamore> i wouldn't be surprised if Linux has the same limits
[05:06:29] <nrubsig> gdamore: nope.
[05:06:38] <nrubsig> gdamore: 1024 is far too low.
[05:06:58] <gdamore> recall once upon a time Solaris supported SYSVFS.
[05:07:21] <nrubsig> gdamore: it bites back when you use multibyte characers - in some locales the maximum available characters (not bytes) gets down to 200 in the worst case (gb18030)
[05:07:39] <nrubsig> gdamore: PATH_MAX counts in bytes
[05:07:54] <gdamore> well, that limit _definitely_ predates using multibyte locales in Solaris. :-)
[05:08:03] <boyd> That's because pathnames are byte strings
[05:08:16] <gdamore> boyd++
[05:08:38] <nrubsig> gdamore: Linux uses  4096 for exactly that reason - that way they can safely have >= 1000 multibyte characters in most locales.
[05:08:42] <gdamore> using non-Latin1 names in path names is a really, really bad idea.  i suspect a lot of tools break that way.
[05:08:50] <nrubsig> gdamore: noipe.
[05:09:08] <nrubsig> gdamore: not unless the tools do special processing of the strings.
[05:09:15] <gdamore> (we just had this conversation wrt some odbc driver code, IIRC)
[05:09:28] <boyd> Actually, I didn't say that... I just mean that it's not the filesystem's job to concern itself with encodings
[05:10:05] <gdamore> boyd, true.  but you run into problems like alternative encodings of "/" or other special characters.
[05:10:47] <broadcast> boyd, minimal install
[05:10:47] <gdamore> the filesystem doesn't understand these encodings, so they aren't a problem for the filesystem.   but they might be a problem for the tools, and will _probably_ be a problem for the poor schmuck who has to use the tools.
[05:10:52] * boyd nods... but that's the apps problem
[05:10:58] <nrubsig> gdamore: alternative encodings do not count has ASCII '/'
[05:11:23] <boyd> broadcast: Yeah, It broke that way last week for me.. nobody had any idea when I asked
[05:11:46] <broadcast> boh
[05:11:46] <broadcast> xD
[05:12:16] <gdamore> no, but you have alternative encodings for the character "/", which are going to be confusing to the user _at least_.  and its common practice to "canonify" encodings.  if something gets canonicalized into "/", it could create problems.
[05:13:04] <nrubsig> gdamore: this is not valid for multibyte encodings in Unix
[05:13:24] <nrubsig> gdamore: any byte following the initial one will have the high bit set.
[05:13:35] <nrubsig> gdamore: that way you can't run into problems.
[05:13:51] <nrubsig> gdamore: and encodings have 1:1 mappings
[05:13:58] <nrubsig> (also a requirement)
[05:14:09] <gdamore> that's not what I'm saying.  what i'm saying is that some other multibyte character will have the same visual representation as "/".
[05:14:29] <nrubsig> it will still be a diffferent character
[05:14:30] <gdamore> and encodings definitely do not have 1:1 mappings.  at least not in UTF-8.
[05:15:23] <gdamore> e.g. you can express the spanish "n with a tilde" as either one accented character, or an "n" with a modifying character.
[05:15:58] <gdamore> going back, yes, it will be a different numerical value.  so no problem for the *software*, but a problem for the poor schmuck who has to use it.
[05:16:05] <jteo> morn *
[05:16:58] 
[05:17:15] <boyd> The userspace has to set canonicalisation rules
[05:17:20] <gdamore> another point, nrubsig, is that for _most_ of the locales with large multibyte characters (e.g. the chinese sets) you can express a _lot_ more in far fewer characters.  e.g. the word for "horse" in chinese takes 1 character, but in english it takes 5.   i don't think the limitation is too bad.
[05:18:03] *** Gadzooks has quit IRC
[05:18:20] <gdamore> boyd: yes, and some of the tools do things to avoid other problems.  like insertion of "/" alternate encodings, or alternate encodings for other "special" characters.
[05:21:22] <boyd> Indeed... it'd be sometimes nice to have a non-pathname-splitting slash. Maybe U2044 FRACTION SLASH but there's also U2571 BOX DRAWINGS LIGH DIAGONAL UPPER RIGHT TO LOWER LEFT and UFF0F FULLWIDTH SOLIDUS. Sheesh
[05:21:48] <gdamore> ew... going back to a sun type 5 keyboard mouse.  the keyboard is nice, but i miss the scroll wheel. :-)
[05:22:24] <gdamore> boyd must have a Unicode reference somewhere. :-)
[05:22:36] <boyd> From memory :)
[05:22:54] <boyd> Actually, the Mac lets me type a char and shows related unicode chars. Quite nice
[05:22:58] * nrubsig wants pacman symbols
[05:23:01] <nrubsig> and galaga
[05:23:05] <nrubsig> GNAW!!
[05:23:32] <nrubsig> If gatekeepers ever let gnaw pass the we can think about tetrtis, too :-)
[05:23:51] <gdamore> put it in /usr/demo
[05:24:01] <nrubsig> gdamore: already done.
[05:24:12] <boyd> U22DO DOUBLE SUBSET could pass for pacman ? (sorry about the UTF-8)
[05:24:20] <nrubsig> gdamore: you can find an early version via google code search
[05:24:52] <gdamore> that would be if i had more interest in this than i actually do. :-)
[05:25:43] <nrubsig> boyd: 22D0 has no glyth available in any solaris font.
[05:26:09] <boyd> :(
[05:26:10] <nrubsig> boyd: (e.g. "?")
[05:26:23] <boyd> Looks fine to me
[05:26:36] <boyd> (on the mac :) )
[05:26:42] <nrubsig> boyd: blame Ienup
[05:26:50] <boyd> who?
[05:27:03] <nrubsig> boyd: Ienup Sung
[05:27:22] <boyd> Ok, I'm not much more enlightened
[05:27:29] <nrubsig> boyd: The first person I usually call when I am in trouble with any i18n stuff
[05:41:36] *** Risky has quit IRC
[05:45:59] <gdamore> g'nite all.
[05:48:02] *** laca has quit IRC
[05:50:04] *** karrotx has joined #opensolaris
[05:50:17] <jamesd> hi karrotx
[05:50:50] <karrotx> hey
[06:05:11] *** Risky has joined #opensolaris
[06:08:03] *** sahafeez has joined #opensolaris
[06:10:47] <sahafeez> so rhythmbox work on sxcr 50 on sparc anyone?
[06:13:11] <jamesd> define sparc
[06:13:24] <jamesd> it will require a  usparcII cpu or better
[06:13:42] <nrubsig> jamesd: why ?
[06:13:43] <sahafeez> u80 w/4x450 and 4gb of ram
[06:13:56] <jamesd> should work
[06:14:12] <jamesd> nrubsig, just making sure he isn't trying to install on a  ss20 or something..
[06:14:18] <sahafeez> seems to break every other version and i need it so before i upgrade i want someone to say, yes works on my...bla just fine
[06:14:27] <jamesd> nrubsig, its allways required a  usparcII or greater
[06:14:28] <sahafeez> sold the SS20
[06:14:49] <nrubsig> jamesd: why does US-1 not work ?
[06:15:05] <sahafeez> this is funny, i need a pkg for openbsd but it only works on sparc, sparc64 and i386. no amd64.
[06:15:12] <jamesd> nrubsig, because sun got tired of supporting a buggy  cpu
[06:15:40] <jamesd> there is a patch to make it work on usparcI ... but its not supported by sun.
[06:15:41] <nrubsig> jamesd: umpf, you mean it's just the kernel which refuses to run on US1, right ?
[06:15:50] <delewis> jamesd: there are OpenSolaris distributions that already run on UltraSPARC-I
[06:16:11] <sahafeez> so does it work or not? i have been trying to get an answer for 2 days.
[06:16:12] <delewis> marTux being one that I know of
[06:16:13] <jamesd> nrubsig, i mean there are checks in the kernel code that disable support for usparcI
[06:16:14] <nrubsig> jamesd: note that we had the discussion recently - US-2 is a fixed US-2 and the US-1 bug could be worked-around quite easily.
[06:16:45] <sahafeez> it breaks so much i am going to put up a webpage of www.rhythmboxonsparcbrokenornot.com
[06:16:58] <jamesd> they can't work around the bug.. but most people don't care about it
[06:17:37] <nrubsig> jamesd: you can test for the matching code sequence in the elf loader module.
[06:17:54] <nrubsig> jamesd: that's damn easy.
[06:18:52] <nrubsig> (yes, i know... you can use mprotect() to mark pages executable etc. etc. but that is then a job for the same check to walk over new/changed executable pages. Linux does that, too.)
[06:19:51] <sahafeez> work or not????? ahahga!!!
[06:20:01] <nrubsig> and I guess Solaris x86 has already something like that to edal with the F00F pentium bug.
[06:20:11] <delewis> sahafeez: you could try sending a message to the sfwnv-discuss list
[06:20:12] <nrubsig> s/edal/deal/
[06:20:26] <delewis> nrubsig: does Solaris run on i586 hardware?
[06:20:33] <delewis> I suppose it would
[06:20:38] <LeftWing> I can't see why not.
[06:20:48] <nrubsig> delewis: dunno... I guess Pentium-1 is still supported.
[06:20:58] <LeftWing> I doubt anything prior to that would work...
[06:21:07] <delewis> that's the only place the F00F bug occurred, IIRC.
[06:21:12] <LeftWing> mmm
[06:21:34] <nrubsig> delewis: yes, but I guess the code is in OS/net.
[06:22:01] <LeftWing> leftwing@romulus ~ $ isalist -- amd64 pentium_pro+mmx pentium_pro pentium+mmx pentium i486 i386 i86
[06:22:10] <LeftWing> Perhaps we can use 386's =D
[06:22:20] *** fet-kviga has left #opensolaris
[06:22:21] <delewis> for some reason, I highly doubt that :-)
[06:22:26] <LeftWing> Yeah, me too.
[06:47:04] <nrubsig> bye
[06:47:12] *** nrubsig was kicked by nrubsig (bye)
[06:48:31] *** Kernel86 has quit IRC
[06:49:47] *** svoboda_ has joined #opensolaris
[06:55:34] *** svoboda__ has joined #opensolaris
[07:00:44] *** svoboda has quit IRC
[07:01:35] *** ShadowHntr has quit IRC
[07:03:07] *** svoboda has joined #opensolaris
[07:04:00] *** svoboda_ has quit IRC
[07:10:01] *** svoboda__ has quit IRC
[07:18:26] *** bougie has joined #opensolaris
[07:21:01] *** silly_girl22 has joined #opensolaris
[07:32:29] *** kimc has quit IRC
[07:48:12] *** gnu2it2 has quit IRC
[07:49:08] <sahafeez> anyone else having issues ftp.openbsd.org
[07:51:25] <Error_404> what a bizzare thing to aks
[07:56:06] *** bougie has quit IRC
[07:58:08] *** nwf has quit IRC
[07:58:11] <sahafeez> sorry. wrong window
[07:58:26] <sahafeez> both start with open..
[07:59:08] *** yongsun has joined #opensolaris
[07:59:11] *** nwf has joined #opensolaris
[07:59:30] <sahafeez> to many letters and i get confused
[08:03:40] <boyd> That's why I'm in #opensolaris and #closedwindows :)
[08:04:36] <g4lt-mordant> or #opensolaris and #freedos ;P
[08:05:53] <g4lt-mordant> and #netware FTW
[08:07:50] *** nwf has quit IRC
[08:12:02] <sahafeez> opensolaris, openbsd, asterisk..
[08:14:01] *** silly_girl22 has quit IRC
[08:14:20] <g4lt-mordant> #opensesame
[08:15:42] <boyd> #openopenwhereareyoupen?
[08:16:19] *** sahafeez has quit IRC
[08:19:32] *** axisys has quit IRC
[08:25:16] *** nwf has joined #opensolaris
[08:31:02] *** dunc has quit IRC
[08:48:25] *** Cyrille has joined #opensolaris
[08:48:27] *** Kernel86|Laptop has quit IRC
[08:49:10] *** yognix has joined #opensolaris
[08:49:11] *** Kernel86|Laptop has joined #OpenSolaris
[08:54:21] <boyd> :)
[08:55:51] * timeless could only use that line here
[08:57:46] *** sylvain_3 has joined #opensolaris
[08:59:16] *** sylvain_3 is now known as sylvain
[09:00:12] *** tsoome has joined #opensolaris
[09:05:21] *** nwf has quit IRC
[09:06:33] <asyd> \_o<
[09:07:55] *** Kernel86 has joined #OpenSolaris
[09:09:15] <yognix> how to get old source code for solaris 7 or 8 (i don't remember) that was published on sun.com some time ago?
[09:12:21] <noyb> dclarke: any chance of getting a songbird package for Solaris?  http://www.songbirdnest.com
[09:12:58] <noyb> >o_/       Quack back, asyd
[09:17:20] *** dunc has joined #opensolaris
[09:18:45] *** slowhog has left #opensolaris
[09:22:52] *** Gman has quit IRC
[09:22:53] *** Xh4 has joined #opensolaris
[09:24:49] *** pip` has joined #opensolaris
[09:26:47] *** alobbs has joined #opensolaris
[09:27:41] <tsoome> yognix: solaris/sunos code has been basically always available - for certain price and conditions, of course
[09:27:53] <tsoome> I have had access since 2.5.1, I think....
[09:28:00] <Xh4> tsoome, has it? Hm...
[09:28:01] <Xh4> Haven't seen it on the P2P networks ever. :p
[09:28:27] <tsoome> why it should be there?
[09:30:32] <tsoome> at some point of time it  was available for universities ($100) if bought with some server
[09:30:39] <yognix> as i know source code was available for free for all, but later it was closed
[09:30:44] <tsoome> obviously you had to sign some papers etc
[09:31:13] <tsoome> s8 base bits, yes
[09:31:16] <yognix> and yes, there is no source code in p2p networks :(
[09:31:45] <Error_404> probably not
[09:31:46] <delewis> yognix: given the audience, that seems logical.
[09:32:32] <delewis> I doubt the first thought a student sees at a university (say 5 years ago) that had Solaris 8 souce would be: "Woah, dudes I gotta get this on gnutella!"
[09:32:37] <yognix> my problem is living in xUSSR ;)
[09:32:43] <delewis> s/sees/thinks/
[09:33:18] <tsoome> hm, I'm from estonia - so it's the same background.
[09:33:31] <yognix> no, Estonia is more Europa
[09:33:37] <tsoome> true;)
[09:33:58] <yognix> Ukraine is a big hole as for companies in USA
[09:35:08] <Xh4> delewis, if I was a university student 5 years ago, trust me.. that would've been me. :P
[09:35:13] <yognix> what's the original name for sol8 source code archive?
[09:36:28] <tsoome> hm, why s8 afterall, go for s10;)
[09:37:02] <yognix> i'm interested in rootkit technology
[09:37:51] <delewis> yognix: Windows seems to he a popular area for that "field of study" -- not Solaris.
[09:38:25] <yognix> how many critical changes as for kernel were made in sol10? seems lot
[09:39:11] <delewis> yognix: lots -- zones, SMF, unified threads model, DTrace, and much more
[09:39:33] <yognix> delewis: so, you answer my question ;)
[09:40:06] <yognix> delewis: it's useful to use sol8 source code
[09:40:14] *** Gman has joined #opensolaris
[09:40:27] <yognix> delewis: not so critical, but useful
[09:40:32] <delewis> yognix: Solaris 8 source code is not publicly available.
[09:41:03] <yognix> delewis: it was for some time, AFAIK
[09:41:26] <delewis> yognix: no, it never was, unless you had a source license (hence, not "publicly" available)
[09:43:03] <yognix> delewis: ok, maybe for source license. but there is no as for now such proposition on sun.com. it was closed.
[09:43:56] <yognix> delewis: maybe you can give me the name for source code archive? ;)
[09:44:43] <delewis> yognix: not for Solaris 8, but here's the current source: http://cvs.opensolaris.org/source/
[09:45:27] <yognix> delewis: thanks, i have already
[09:49:34] *** bunker_work has joined #opensolaris
[09:51:00] <yognix> tsoome: what about s8 base bits? maybe i can find it in p2p?
[09:51:17] <tsoome> I doubt
[09:52:12] <tsoome> base - meaning that the published code was only originated from sun, all bits from others was removed
[09:52:23] <yognix> i understand
[09:52:44] <yognix> but it's more than nothing
[09:53:26] <tsoome> and it was for FCS version only, no updates
[09:54:56] <yognix> i don't care. i just wanna get something that's more close to "Solaris Internals Core Kernel Components" ;)
[10:00:42] *** solarisjon has joined #opensolaris
[10:02:35] <boyd> You want Solaris internals 2nd edition.
[10:06:25] <dwc-> there's some sol8 code in a file called sol_foundation-8-fcs-src-en.zip
[10:07:52] <Error_404> it's fantastic
[10:08:44] *** yongsun has quit IRC
[10:10:05] *** Gman has quit IRC
[10:10:06] *** solarisjon has quit IRC
[10:10:06] *** alanc_away has quit IRC
[10:10:45] *** solarisjon has joined #opensolaris
[10:11:02] *** alanc has joined #opensolaris
[10:12:43] *** yongsun has joined #opensolaris
[10:27:05] <raph_ael> hello
[10:29:26] *** cmihai has joined #OpenSolaris
[10:34:46] <trygvis> http://java.sun.com/community/opensource.jsp
[10:37:51] <yognix> sorry, was busy
[10:37:54] <yognix> thanks
[10:38:37] <yognix> dwc-: thanks
[10:48:54] <cmihai> http://in.sys-con.com/read/288823.htm
[10:49:48] <cmihai> Really funny how a 2500$ RHEL can reduce costs compared to a free Solaris with 360$ Premium tech support :)
[10:50:25] *** MattMan has joined #opensolaris
[10:51:01] <cmihai> Or how Itanics have better price / $ performance than Opterons...
[10:52:22] <oxygene> cmihai: for running EPIC apps, definitely..
[10:52:31] <oxygene> (otoh: where to find EPIC code?)
[10:53:30] <cmihai> Um... Oracle doesn't do that iirc.
[10:54:32] <Error_404> cmihai: linux voodoo... it's just that good
[10:54:55] <cmihai> Guess so.
[10:55:56] <Error_404> gotta spend money to save money and all
[10:56:51] <dwc-> yognix: 179216  sol_foundation-8-fcs-src-en.zip
[10:56:54] <dwc-> 180mb
[10:58:05] <yognix> dwc-: there is no such file on p2p
[10:58:32] <dwc-> yea, I didn't get it from p2p
[10:59:01] <dwc-> cmihai: lots of places switch to "rhel" and then install centos, except on the critical machines ....
[10:59:19] <yognix> dwc-: where I get it? ;)
[10:59:50] <cmihai> dwc-, yeah, well, it is the same thing only without the support and the logos.
[10:59:58] <dwc-> yognix: I don't know.
[11:00:03] *** solarisjon has quit IRC
[11:00:42] *** solarisjon has joined #opensolaris
[11:01:05] <dwc-> have had this one for awhile
[11:01:24] <yognix> dwc-: maybe you can upload it somewhere? or it's impossible?
[11:01:38] <dwc-> sorry, I don't remember having redist rights for it
[11:02:00] <yognix> dwc-: hm. ok.
[11:02:12] <dwc-> a fair number of the files I've had reason to look at
[11:02:23] <dwc-> are rather unchanged in opensolaris
[11:02:28] <dwc-> or have only minimal changes
[11:02:59] <dwc-> and this isn't full source code either -- there are binary bits
[11:04:26] *** mazon is now known as Mazon
[11:04:47] *** mikaeld_ has joined #opensolaris
[11:06:11] <yognix> dwc-: ok. thanks
[11:06:32] <yognix> i'm waiting for "solaris internals 2nd ed" in p2p ;)
[11:06:43] <asyd>  /kickban yognix !
[11:07:28] <yognix> asyd: wazzup?
[11:07:52] <asyd> buy it ?
[11:08:27] <dwc-> yognix: I could send you a list of files if you're really interested in what's in the zip, but I don't think I can send the contents
[11:08:42] <Berny> morning
[11:10:50] <yognix> dwc-: thanks. i'll try to use my docs, sol int. 1st ed., sol10 sources and kernel debugging to get needed info. if any troubles, i'll try to get info from you ;)
[11:11:33] <Berny> would anyone with some knowlegde of sun's C++ compiler enlighten me why the beast chokes on this line:     template <typename T, int N> char (&sizer(T (&)[N]))[N]; ?
[11:12:01] <trygvis> because that line is a beast? :)
[11:13:10] <Berny> ack!
[11:13:27] <Berny> ok, any volunteers to fix this? :-)
[11:13:29] <yognix> why asyd so worried about p2p networks and book? maybe he's an contributor to the book?
[11:13:35] <yognix> ;)
[11:13:44] <asyd> hey, was a joke for sure.
[11:14:08] <yognix> hey, sure i understand
[11:16:33] <lasseoe> so go buy the book
[11:16:52] <yognix> i can't
[11:17:31] <lasseoe> so if you need new hardware, do you then go steal it?
[11:17:54] <dwc-> I borrow books
[11:18:01] <jengelh> "Anything stolen is pure profit."
[11:18:02] <dwc-> and/or  tools
[11:19:14] <dwc-> looks like my city library only has the 1st rev now though
[11:19:15] <yognix> i don't need hardware, only information ;)
[11:19:20] <quasi> asyd: +1 on the kickban ;)
[11:19:29] <asyd> :)
[11:19:33] <quasi> yognix: read the source then
[11:19:38] <yognix> p2p has the 1st rev
[11:19:45] <asyd> quasi: btw have you ever tried hudson ?
[11:19:53] <trygvis> http://osnews.com/story.php?news_id=16268
[11:20:04] *** mikaeld has quit IRC
[11:21:07] <yognix> quasi: no problem with sources
[11:21:24] <quasi> asyd:  hudson?
[11:21:37] <yognix> quasi: only a lot of information
[11:21:44] <asyd> quasi: https://hudson.dev.java.net/
[11:24:17] <yognix> heh, book is available via safari. great
[11:24:33] <asyd> quasi: you can take a look in http://lei.asyd.net/hudson/ too
[11:26:18] <quasi> asyd: nope, I'd never seen that before - probably a bit too java-ish for me to notice on my own ;)
[11:26:25] <asyd> :)
[11:26:47] <asyd> well, since I work fulltime on a java project now..
[11:27:26] <trygvis> asyd: you're using it to do CI?
[11:27:59] <quasi> asyd: poor guy ;)
[11:28:16] <asyd> yeah...
[11:28:48] <asyd> after 6 years of sysadmin, I thought I'll be happy to leave production environment. but in fact, I miss it
[11:48:02] *** Ezechial_ has quit IRC
[11:54:39] <r3boot> asyd: hmm, I know that feeling
[11:57:24] *** cmihai has left #OpenSolaris
[11:58:59] *** yongsun has quit IRC
[12:09:53] *** axxl has joined #opensolaris
[12:11:17] *** yongsun has joined #opensolaris
[12:12:26] *** lplatypus has joined #opensolaris
[12:12:30] *** ka46 has joined #opensolaris
[12:13:23] *** yongsun has quit IRC
[12:13:38] *** ka46 has quit IRC
[12:34:26] *** pip` has quit IRC
[12:39:08] *** Joker^ has joined #opensolaris
[12:39:21] <Joker^> hi all
[12:39:55] <Joker^> what's the meaning of "defect list found" in format utility ?
[12:41:58] *** mikaeld has joined #opensolaris
[12:41:58] *** mikaeld_ has quit IRC
[12:44:54] *** Joker^ has quit IRC
[12:51:53] *** calMTG has joined #opensolaris
[12:52:19] *** kimc has joined #opensolaris
[12:55:55] *** bunker_work has quit IRC
[12:58:52] <kimc> got a racoon on the house
[12:59:01] <kimc> in the house actually..
[13:04:07] *** MattMan is now known as MattAFC
[13:11:31] *** Dr_Jekyll has joined #opensolaris
[13:18:36] *** LordKing has joined #opensolaris
[13:18:39] *** triplah has joined #opensolaris
[13:26:55] *** tsoome has quit IRC
[13:29:11] *** tsoome has joined #opensolaris
[13:32:37] *** coffman has joined #opensolaris
[13:32:45] *** tsoome has quit IRC
[13:33:28] <jteo> hello *
[13:33:43] *** tsoome has joined #opensolaris
[13:40:36] <kimc> hello jteo
[13:40:55] <jmcp> evening all
[13:41:33] <kimc> got a racoon in the house this morning..
[13:42:01] <kimc> came down the chimney
[13:42:08] <jmcp> not the natural environment for a racoon, I would have though
[13:42:16] <kimc> no..
[13:42:41] <kimc> now its in the laundry room, door closed, window open
[13:42:50] <kimc> won't leave
[13:43:08] <jmcp> you need something to entice it out
[13:43:11] *** Spawrq has joined #opensolaris
[13:43:34] <kimc> yes.. could maybe position the garbage can outside with the lid off
[13:44:27] <kimc> its just dawn here
[13:44:49] <kimc> was hoping the light in the open window would seem enticing
[13:50:24] <PerterB> as enticing as snuggling up for a doze in a pile of laundry? :)
[13:55:59] *** lplatypus has quit IRC
[14:02:18] *** salamanders has joined #opensolaris
[14:03:15] <kimc> thats pretty enticing for sure
[14:03:20] <jteo> mmm
[14:03:50] <kimc> the local animal control just opened.. lets see what their sotry is
[14:04:32] *** tsoome has quit IRC
[14:10:34] <jmcp> kimc: good luck with that
[14:10:40] <jmcp> much better to get the professionals in
[14:10:45] <Tpenta> nice blog on sgr james
[14:10:55] <kimc> thanks muchly..
[14:11:03] <jmcp> Tpenta: thankyou
[14:11:13] <jmcp> Tpenta: it's going to be interesting to see how that one pans out
[14:11:23] <Tpenta> :)
[14:11:28] <Tpenta> as long as it is not what they thought it was
[14:11:31] <jteo> sgr?
[14:11:37] <jmcp> sun global resolution
[14:11:39] <jmcp> aka SunResolve
[14:11:59] <jmcp> aka Sun's implementation of Kepner-Tregoe's Problem Solving and Decision Management methodology
[14:12:04] <jmcp> jteo: it's all in my blog :)
[14:12:14] <jmcp> Tpenta: I don't think I'm on the wrong track with this one
[14:12:49] <jmcp> Tpenta: I put my CPRE and PTS experience to good use, finally
[14:13:00] <Tpenta> :)
[14:14:35] <jmcp> you actually getting headcount for outside-of-India then?
[14:15:40] <jmcp> gah
[14:17:41] *** kimc has quit IRC
[14:23:43] *** LordKing has quit IRC
[14:23:56] *** chrisk has joined #opensolaris
[14:26:08] <trygvis> .win 22
[14:33:32] *** loke has joined #opensolaris
[14:33:39] *** chuck has joined #opensolaris
[14:39:40] <richlowe> blah
[14:40:23] *** bengtf has joined #opensolaris
[14:41:23] * jmcp starts the big migration
[14:41:26] <jmcp> back later folks
[14:41:38] *** jmcp has left #opensolaris
[14:44:31] *** vmhobbes has quit IRC
[14:50:44] *** bengtf has quit IRC
[15:03:34] *** laca has joined #opensolaris
[15:10:17] <asyd> so, I'm a bit surprise but... there is no vrrp daemon for solaris ?
[15:12:27] <quasi> so what would be a good setup for 20 disks on ZFS?
[15:12:59] <richlowe> asyd: there's IPMP, but that's not exactly the same thing.
[15:13:04] <asyd> yeah
[15:13:22] <richlowe> I don't know of any of the open VRRP-clones, work under Solaris, I know that VRRP is cisco-patented, though.
[15:13:35] <asyd> well, VRRP is RFCised
[15:13:36] <asyd> though..
[15:13:43] <richlowe> Isn't there a purely userland CARP (OpenBSD's non-patented vrrp-a-like) implementation somewhere?
[15:13:48] <asyd> (2338/3768)
[15:15:08] <richlowe> a quick google suggests that HSRP is patented, and there was concern that it basically overlapped VRRP.
[15:15:44] <asyd> yup
[15:16:10] <asyd> but vrrp is open at least
[15:16:29] *** loke has quit IRC
[15:16:41] *** bunker_work has joined #opensolaris
[15:18:22] <jteo> :)
[15:24:07] *** karrotx has quit IRC
[15:28:43] *** holzmann has joined #opensolaris
[15:29:22] *** loke has joined #opensolaris
[15:39:27] <richlowe> mornin' jteo.
[15:39:55] <jteo> richlowe: evening.
[15:42:16] <jteo> whee. refreshed MacBook Pro. more stuff i can't have. ;)
[15:42:51] <richlowe> I keep telling myself I'm waiting for a cheap laptop with a sensible configuration to exist.
[15:43:00] <richlowe> that way I can feel I'm being self righteous and stubborn, rather than cheap ;)
[15:45:52] *** alobbs has quit IRC
[15:47:41] <richlowe> Oh look, another message with implicit destination.
[15:47:55] <richlowe> I swear plocher sent out his last mail to osol-arc that way *purely* to re-inforce how annoying it is.
[15:48:32] <richlowe> hm, I can only guess psarc-ext@ is now attempting to forward to both?
[15:50:13] <jteo> i'm not in the loop regarding that debacle.
[15:50:22] <richlowe> You should be.
[15:52:46] *** alobbs has joined #opensolaris
[16:00:54] *** loke has quit IRC
[16:02:02] *** tsoome has joined #opensolaris
[16:03:42] *** laca is now known as lacaMtg
[16:15:01] <Doc> morning
[16:15:14] *** Kernel86 has quit IRC
[16:15:51] <jteo> Doc: moin
[16:16:47] <quasi> does anyone know of a shortcut for changing ips on a bunch of zones rather than removing and adding nets in zonecfg?
[16:17:31] <asyd> cd /etc/zones ; for file in *.xml ; do sed ... ; done ?
[16:18:12] <quasi> asyd: yeah, that was my thinking as well - I just wondered if I was missing something
[16:18:38] <Doc> script zonecfg
[16:19:55] *** calMTG is now known as calAFK
[16:20:44] <richlowe> all the stuff in /etc/zones is project private, I believe.
[16:21:07] *** tsoome has quit IRC
[16:21:25] <richlowe> but since zonecfg iirc will take a command as a command-line arg, you shouldn't need to muck around directly.
[16:21:28] <jteo> i've always regarded quasi as the adverturous sort.
[16:21:36] <asyd> hehe
[16:21:56] *** Dar is now known as Dar_HOME
[16:22:11] <asyd> few months ago, I even add a inherit-pkg-dir after the installation of a zone by editing the xml file
[16:22:14] <richlowe> for zone in <zones>; zonecfg -z $zone 'remove net ...; add net ...'
[16:22:15] <quasi> jteo: that must be the other quasi ;)
[16:22:20] <richlowe> pretty much what Doc said.
[16:22:56] <quasi> richlowe: yeah, thanks - that was about where I was
[16:23:12] <richlowe> s/remove/select/ s/add net/set/ :)
[16:23:25] <PerterB> If I'm not using HA NFS, can I safely remove the "exclude: lofs" that Sun Cluster adds to /etc/system?
[16:24:06] <quasi> richlowe: ah, that I hadn't thought of - I was just about to do the remove/add bits
[16:25:27] *** rodrickbrown has joined #opensolaris
[16:25:45] *** Kernel86 has joined #OpenSolaris
[16:26:28] <PerterB> ah, never mind... found the bit in the manual that says it's OK
[16:33:12] <jteo> i'm wondering why people use BSD jails instead of zones.
[16:33:13] <jteo> hmm.
[16:33:45] *** loke has joined #opensolaris
[16:34:07] <richlowe> jteo: Jail's are pre-existing.
[16:34:17] <quasi> jteo: because they run on bds rather than solaris?
[16:34:19] <richlowe> and if you're alreading running on FreeBSD, why change?
[16:34:32] <richlowe> also, they have differing abilities.
[16:34:45] *** MattAFC has quit IRC
[16:34:45] *** Cyrille has quit IRC
[16:34:57] <richlowe> with appropriate COMPAT_ options in the FreeBSD kernel, you can run the userland prior releases for quite a distance back within a jail.
[16:35:02] *** calAFK is now known as calumb
[16:35:05] <richlowe> 'the userland from'
[16:35:12] <quasi> and some of the linux vserver fans keep saying zones are useless because they are missing memory capping
[16:35:31] <rodrickbrown> useless?
[16:35:34] <rodrickbrown> hehe
[16:35:36] *** Cyrille has joined #opensolaris
[16:35:41] *** MattMan has joined #opensolaris
[16:35:43] *** kloczek has quit IRC
[16:35:52] <jteo> ah
[16:36:03] <quasi> rodrickbrown: quoting a guy who resells a fair number of vservers
[16:36:18] <richlowe> if you're reselling them as separate 'machines' I can understand the reasoning.
[16:36:29] <richlowe> ... on the other hand, I can't think why you wouldn't use some lower level virtualization to do that.
[16:36:42] <quasi> cost
[16:36:50] <quasi> seperation
[16:36:58] *** mrdeviant has joined #opensolaris
[16:37:06] <jteo> rcapd? (i have no experience)
[16:37:12] <hile_> hey derek
[16:37:47] <mrdeviant> hi hile_
[16:37:50] <rodrickbrown> quasi we use them extensily here
[16:37:55] *** mrdeviant has quit IRC
[16:38:05] *** mrdeviant has joined #opensolaris
[16:38:07] <quasi> jteo: except that needs to run inside the zone, so you'd need complete cooperation from the users
[16:38:08] <rodrickbrown> so i'm not sure why onewould consider them useless unless they have no idea how zones work
[16:38:11] *** mrdeviant has quit IRC
[16:38:27] <jteo> quasi: ah.
[16:38:35] *** mrdeviant has joined #opensolaris
[16:38:54] *** cajal has joined #opensolaris
[16:39:01] <quasi> rodrickbrown: well, the exact sentence was something like "unsuitable to resell" like he does with his vservers
[16:39:31] <rodrickbrown> zones are for application isolation
[16:39:36] <rodrickbrown> not really vservers
[16:39:40] <rodrickbrown> wait for xen integration
[16:39:55] <hile_> i agree there.
[16:40:09] <quasi> agreed - but the question was why people didn't switch to zones
[16:40:45] <jteo> is there upcoming functionality for memory capping?
[16:40:50] <quasi> yes
[16:41:30] <quasi> I think as an extension to the cpu capping stuff that went in a while back
[16:41:31] <rodrickbrown> even with memory caps its still a single kernel so if you want to sell them as vservers I would wait for xen
[16:42:51] <quasi> rodrickbrown: why? if that's the game you're in, you just need enough isolation to avoid one zone stealing resources from others
[16:44:05] <rodrickbrown> quasi theoritically its a single kernel so there will be a limitation on how much resources you can really cap/hide etc.. go with a true hypervisor
[16:44:25] <rodrickbrown> sells zones as vservers just seem hackish i'm pretty sure you will run into issues down the line
[16:44:46] <rodrickbrown> also what happens when X user needs to go to kernel patch XXXXX
[16:44:56] <rodrickbrown> it effects everyone :-)
[16:45:13] <richlowe> quasi: I don't believe cpu-caps has gone back yet.
[16:45:40] <quasi> richlowe: in navada?
[16:46:21] <richlowe> I know the bits are on opensolaris.org, but I don't think they're in Nevada.
[16:46:26] <richlowe> if they are, I must have not noticed.
[16:48:07] *** solarisjon has quit IRC
[16:51:59] *** hell` has quit IRC
[16:53:20] *** Spawrq has quit IRC
[16:53:53] *** nwf has joined #opensolaris
[16:53:55] *** mr_goose has quit IRC
[16:56:56] *** lacaMtg is now known as laca
[17:04:39] *** Kush- has joined #opensolaris
[17:05:50] *** Mazon is now known as mazon
[17:08:21] *** jcea has quit IRC
[17:17:11] *** hell` has joined #opensolaris
[17:17:37] *** mr_goose has joined #opensolaris
[17:19:41] *** james_ has quit IRC
[17:19:45] <quasi> yay, zoneadm: zone 'httpd': Failed to initialize privileges: No such file or directory
[17:19:48] <quasi> zoneadm: zone 'httpd': call to zoneadmd failed
[17:22:30] <quasi> seems like http://groups.google.com/group/comp.unix.solaris/browse_thread/thread/030a90e972279bad/75f4feeef46219cb?hl=fr#75f4feeef46219cb
[17:26:41] *** cylix has joined #opensolaris
[17:27:08] <gdamore> good morning *
[17:27:35] <jteo> gdamore: morn
[17:28:53] *** bunker_work has quit IRC
[17:35:29] *** axxl has quit IRC
[17:37:34] <gdamore> hmm.... has anyone ever considered building a storage appliance around niagra?  it would be a very nice chip for the purpose, i think
[17:37:56] <elektronkind> as long as it can XOR really well
[17:38:21] <elektronkind> storage is all integer anyway
[17:38:30] <gdamore> exactly.  lots of integer threads. :-)
[17:39:11] <mrdeviant> niagaraII would be interesting for zfs, since it'll have sha-256 in hardware on-die
[17:39:29] <gdamore> does zfs need sha-256?
[17:39:39] <elektronkind> gdamore: I'm working on a project at home right now to make a stripped-down nevada into a storage applicance OS
[17:39:48] *** broadcast has quit IRC
[17:39:54] <gdamore> really, i'm very interested in that.
[17:39:58] <mrdeviant> for some things, yes. i believe it's mandated for metadata. you can choose the checksum for the datablocks
[17:40:10] <elektronkind> gdamore: http://elektronkind.org/2006/07/making-opensolaris-into-a-storage-appliance
[17:40:11] <gdamore> i'm planning on building a NAS system around nevada myself.
[17:40:12] <mrdeviant> i believe they default to fletcher2
[17:40:25] <elektronkind> fletcher2 is the default, yes
[17:41:22] <gdamore> what is fletcher2?  a different checksum algo?
[17:42:37] <richlowe> does the zfs checksum stuff go through KCF?
[17:42:48] <mrdeviant> at the moment, no, but it would be a simple patch
[17:42:49] <gdamore> it _should_.
[17:42:58] <richlowe> gdamore: I agree, that's why I asked.
[17:43:04] <elektronkind> http://en.wikipedia.org/wiki/Fletcher's_checksum
[17:43:05] <gdamore> then you could add, e.g. an SCA 500, and get acceleration.
[17:43:06] <richlowe> It _should_, but I vaguely remember it not.
[17:43:14] <richlowe> I know darrenm's encryption stuff will.
[17:43:26] <mrdeviant> http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/fs/zfs/sha256.c
[17:44:53] <gdamore> i'm pretty familiar with kcf.  it took some inspiration from an earlier crypto framework for Solaris 8 that I designed for the SCA 500/1000.
[17:45:07] <mrdeviant> yea i'm getting familiar with it too
[17:45:23] <richlowe> I'm not, since the bits I'm more interested in can't be changed out here.
[17:45:31] <richlowe> curse the crypto signing key.
[17:45:41] <gdamore> heh.
[17:45:55] <mrdeviant> yea that is annoying. i was thinking of hacking the module verification daemon not to check for sigs
[17:46:02] <mrdeviant> just to make development easier.
[17:46:03] <richlowe> mrdeviant: good luck, that's closed too.
[17:46:06] <richlowe> they think faster than we do ;)
[17:46:10] <gdamore> i suspect it would be easy to work around that.  the code for kcf itself that does the signature check could be changed. :-)
[17:46:31] <mrdeviant> right. if you recompiled kcfd, i wonder if you could replace the closed one
[17:47:19] <richlowe> probably, but I'm not seeing kcfd sources.
[17:47:26] <gdamore> the signature check has a number of weaknesses.  it was never intended to be truly secure, just secure enough to meet export rules.
[17:47:45] <mrdeviant> richlowe, isn't http://cvs.opensolaris.org/source/xref/on/usr/src/uts/common/crypto/core/kcf.c#149 it ?
[17:47:50] <gdamore> i can't discuss the weaknesses here, but you can probably figure it out. :-)
[17:48:13] <richlowe> mrdeviant: that makes upcalls to kcfd.
[17:48:18] <richlowe> I thought kcfd did the actual verification.
[17:48:46] <gdamore> i don't think so.  i think the verify is in the kernel itself.
[17:49:35] <mrdeviant> according to the march 05 kcf whitepaper, it looks like the kernel does upcall to the mvd. but if you comment out that bit in the kernel, i wonder if it would just load the module and work
[17:49:36] <richlowe> the code mrdeviant linked to suggests it does.
[17:49:48] <gdamore> yeah, i see that now.
[17:49:48] <richlowe> that may work though, yeah.
[17:49:49] <richlowe> Hm.
[17:49:53] <gdamore> i think it would.
[17:50:50] <gdamore> there are special rules for "pluggable" crypto code wrt export, IIRC, and this was put in there to meet those rules (ITAR rules, IIRC)
[17:51:07] <richlowe> I thought it'd been said someone was working on figuring out if such things can be relaxed now.
[17:52:13] <mrdeviant> does the fact that solaris is now open-source change things? i thought oss crypto projects only needed to file a notification with the dept of commerce
[17:52:46] <richlowe> I think that's what they were trying to figure out.
[17:52:49] <mrdeviant> for instance, openssl doesn't have requirements that engine implementations are signed by them.
[17:52:50] <richlowe> I wish I could recall who mentioned it and where.
[17:53:00] <richlowe> I want to think it was darrenm somewhere.
[17:53:48] <gdamore> openssl is not developed in the US, IIRC
[17:54:29] <richlowe> and under solaris, only the pkcs11 engine and plain software are shipped.
[17:54:47] <quasi> gdamore: correct - although they do accept patches for it from .us people who remember to notify the right part of .gov
[17:55:05] *** dduvall has joined #opensolaris
[17:55:41] <richlowe> mornin' dduvall.
[17:55:48] <dduvall> Howdy.
[17:55:58] *** stevel has joined #opensolaris
[17:55:58] *** ChanServ sets mode: +o stevel
[17:56:50] <jteo> morn stevel
[17:56:53] <jengelh> Hm, I cp'ed my module (rpldev) into /platform/i86pc/kernel.WS/dev/, but modload -p still does not find it
[17:56:58] <stevel> morning jteo
[17:57:01] <jengelh>  /drv that is
[17:57:19] <stevel> quick poll: now that we have the Mercurial repository - can i stop delivering Mercurial bundles with the weekly drops?
[17:57:33] <richlowe> stevel: Dunno.
[17:57:49] <jteo> stevel: you'll still have to deliver the closed bins.
[17:57:50] <stevel> i should just posit the question to tools-discuss i suppose
[17:57:51] <richlowe> stevel: is loading a bundle then pull'ing to tip substantially faster than cloning onnv-gate?
[17:57:53] <gdamore> modload -p requires the module is on your kernel boot path.
[17:58:04] <stevel> jteo: yeah, closed-bins won't be going away anytime soon
[17:58:08] <richlowe> I think the incrementals can go though.
[17:58:10] <gdamore> sevel: yes, IMO.
[17:58:15] <elektronkind> if people can download mercurial bundles, I imaging they would have no problems syncing a mercurial tree directly
[17:58:17] <gdamore> stevel, i mean
[17:58:30] <richlowe> elektronkind: Yeah, I'm just wondering if the speed/load on os.o makes it worthwhile anyway.
[17:58:57] <gdamore> i have been meaning to try out Hg, myself.
[17:59:10] <stevel> richlowe: yeah i'm going to ditch the incrementals for sure.  i'm pretty sure i'll ditch the mercurial bundles too.  and honestly, i'd like to even ditch the source tarballs (for the 'nightly' deliveries anyway - i think it makes sense to keep the build-synchronised ones) - but i expect opposition to that last one
[17:59:28] <richlowe> I actually agree more with the last one than the bundles.
[17:59:28] <jengelh> gdamore : Is there anything special needed for exported variables, like EXPORT_SYMBOL? I got "undefined symbol rpl_open', even though the module containing the symbol is already loaded
[17:59:34] <richlowe> but only because I'm not in a position to compare speed right now.
[18:00:03] <stevel> well, i'll pose the question to tools-discuss and we'll see what people say.  i'd appreciate anyone's input as follow-up replies
[18:00:10] <richlowe> Actually, no, that's not the case.
[18:00:12] <Error_404> hmm... that's cute, this machine won't resolve things anymore
[18:00:17] <jteo> the build drops shouldn't go away.
[18:00:18] <richlowe> because stopping the tarball deliveries may well screw cyril and roland.
[18:00:27] *** nwf has quit IRC
[18:00:29] <richlowe> and the last think roland needs is more to complain about ;)
[18:00:38] <gdamore> jengelh: if the symbol is in a different module, it has to be declared extern, and you have to use -N<module> (e.g. drv/gld) to help the linker resolve it
[18:01:09] <stevel> richlowe: even for the nightly ones?
[18:01:21] *** sylvain has quit IRC
[18:01:28] <richlowe> stevel: I thought cyril synced their subversion stuff with every drop from the tarballs.
[18:01:41] <richlowe> though there's no real barrier to syncing from hg, it'd be worth asking them.
[18:01:44] <jengelh> gdamore : By gld, do you mean the gld lan driver or GNU ld?
[18:01:47] <stevel> richlowe: ah. hrm.
[18:03:02] <gdamore> gld lan driver, i was using it as an exmple.
[18:03:17] *** jmcp has joined #opensolaris
[18:03:23] <jteo> wb jmcp.
[18:03:26] <jmcp> hiya
[18:03:41] <jmcp> been futzing around with my you-beaut nvidia fx560
[18:03:56] <jmcp> with the monitors I have, it appears that I can't drive them with digital outputs
[18:04:11] <jmcp> or rather, I gave up on that for tonight in favour of getting some sleep
[18:04:18] <gdamore> anyone netinstalled b50 lately?  i notice a new warning from CDE in suninstall that DT messaging can't be started
[18:04:46] <jmcp> jteo: on a brighter note, I did manage to cutover my webserver, roller, postgres and dns to a zone
[18:05:16] <Doc> hrmmm
[18:05:46] <jteo> jmcp: mmm
[18:06:26] * hile_ lunches
[18:08:25] *** chrisk has quit IRC
[18:08:48] *** nwf has joined #opensolaris
[18:09:14] <jengelh> gdamore: What does it depend on?
[18:09:19] *** Cyrille has quit IRC
[18:09:29] * jmcp sleeps
[18:09:56] <jengelh> gdamore: uts/intel/gld/debug32/gld.o has "U gld_interpret_ether", gldutil.o in the same dir has "T gld_interpret_ether", but I can't figure out where the two are linked together
[18:10:17] <jengelh> since in /platform, there is only one 'gld'
[18:12:45] <jteo> jengelh: ld -dy -r -N"misc/gld"
[18:13:00] <jengelh> eh that's gotta be a hard time, I am using gnu ld :/
[18:13:15] <jteo> no idea jengelh.
[18:13:25] <jteo> why do you haveta use gnu kd?
[18:13:26] <jteo> *ld
[18:13:32] <jengelh> i don't really have to
[18:13:35] <jteo> i compiled with gcc, but used ld. works.
[18:13:38] <jengelh> it's just that the .c files are compiled with it
[18:13:43] <jengelh> ok, I'll try that!
[18:14:35] <gdamore> yes, you don't want to use gnu ld or gnu as, even if using gcc.  frankly, i'd recommend using studio 11 for all Solaris work, anyway.
[18:14:53] <jengelh> i'm still on sol11_snv39, requiring studio10 (all installed now)
[18:15:10] <jengelh> it's just that I worked with gcc before requiring studio for UTS
[18:15:38] *** Symmetria has joined #opensolaris
[18:15:39] <Symmetria> hey all
[18:15:48] <richlowe> studio shouldn't be *required*, iirc.
[18:15:54] <richlowe> if you're using studio, it must be the appropriate 11+patches.
[18:15:55] <Symmetria> quick question, and probably a stupid one, is there a way to mount a reiserfs file system under solaris?
[18:16:11] <richlowe> but I think wesolows work to make gcc work properly still... well, works.
[18:16:48] <jengelh> richlowe: Read: snv_39. studio10. installed. uts compiled. works. It's just that I had not needed it for the _module_ yet
[18:17:15] <jteo> Symmetria: no. do it the old fashioned way. Backup and restore. ;)
[18:17:35] <Symmetria> jteo errr yeah, just that its a 1.9 terabyte file system
[18:17:35] <Symmetria> :p
[18:17:38] <jengelh> this seems strange, a manpage for ld exists, but no ld program
[18:17:59] <richlowe> jengelh: /usr/ccs/bin/
[18:18:10] <jteo> Symmetria: ah.
[18:18:18] <jengelh> richlowe : Is that supposed to be in the default $PATH?
[18:18:26] <Symmetria> jteo heh not a problem, I'll transfer it to another SAN
[18:18:47] <richlowe> I don't think it's in the default path, no.
[18:19:24] *** bougie has joined #opensolaris
[18:20:05] <gdamore> default paths are always garbage anyway. :-)
[18:20:16] <jengelh> yeah but in makefiles...
[18:20:31] <jengelh> it's really helpful to not rely on paths (usually hardcoded then)
[18:20:42] <gdamore> on solaris, i always make sure /usr/bin:/usr/sbin:/usr/ccs/bin are in my path, and /usr/ucb is _not_ in my path. ;-)
[18:20:56] <asyd> you miss /usr/sfw :p
[18:21:32] <gdamore> i don't always want /usr/sfw in my path.  but I _always_ want the others.  (just like i don't always want /opt/SUNWspro in, but I _usually_ do.
[18:21:38] <asyd> ok
[18:21:47] <gdamore> stuff in /usr/sfw can collide with /opt/csw. :-)
[18:22:28] <asyd> oh well, I have most of path in my PATH, but with ~/bin at the begin. And sometimes I link in ~/bin to choose which version i want
[18:24:25] <Symmetria> shit, gotta board a flight, will work this out when Im back home again
[18:24:27] <Symmetria> *sigh*
[18:24:42] * Symmetria boards another flight for another 2 hours on the plane pushing the total to 10 hours for the day and 20 hours for the last 3 days :(
[18:26:37] *** Symmetria has quit IRC
[18:27:24] *** jcea has joined #opensolaris
[18:29:52] *** alobbs has quit IRC
[18:31:54] *** logic_ has joined #opensolaris
[18:32:56] *** MattMan has quit IRC
[18:34:33] *** loke has quit IRC
[18:35:56] *** logic__ has quit IRC
[18:39:52] *** solaris-user has joined #opensolaris
[18:43:26] *** mazon is now known as Mazon
[18:44:22] *** sahafeez has joined #opensolaris
[18:46:24] *** nwf has quit IRC
[18:46:42] *** cylix has quit IRC
[18:49:17] *** cylix has joined #opensolaris
[18:50:14] <cylix> I am trying to make a patch that i wrote compatable with solaris.  can someone give me a couple of lines sample of /proc/net/tcp?
[18:50:55] <cylix> I just want to see if the format needs adjusting compared to linux.
[18:51:03] <asyd> /proc/net ? it's only for liux
[18:51:04] *** nachox has joined #opensolaris
[18:51:32] <cylix> ak ok does solaris have a similar interface or way to map socket to user?
[18:51:51] <cylix> or rather uid
[18:55:10] <PerterB> not trivially.. you can use pfiles and search through each process' open files for the socket you're looking for, or install lsof and parse its output
[18:57:07] <cylix> PerterB, Thanks for the input.  lsof would be an easy solution but adds a depend.
[18:57:14] *** Kush- has quit IRC
[18:57:22] <asyd> what kind of patch you're writing ?
[18:57:44] <cylix> an authentication module for dansgardian.
[18:57:52] <asyd> ok
[18:58:16] <cylix> the only input i really have about the user is the socket...
[18:58:39] <asyd> "good luck" :)
[18:59:12] <cylix> The method I'm using on linux only works for local users of course but that doesn't matter since its mainly for thinclient use.
[19:00:58] <PerterB> if you have the socket, you could inist they install a suitable identd and do lookups via that (or crib the Solaris code from ident itself)
[19:01:12] <cylix> guess I need to get solaris up on the extra box again so I can hack on it for a bit.  http://www.dansguardian.org btw
[19:02:28] <PerterB> also if you're root and looking for a specific socket, grepping through the pfiles output is pretty straightforward (or better, crib the code from pfiles from opensolaris.org and implement the same functionality in your app)
[19:03:03] * crib the code!
[19:03:22] * PerterB laughs
[19:03:39] <crib> :D
[19:03:46] <cylix> right dansguardian has an ident module so that is true I was just trying to eliminate ident.    Now that you say ident though is there an opensource ident for solaris i could look at that would help a ton.  Thx for your help btw
[19:04:00] *** solaris-user has quit IRC
[19:04:15] *** uebayasi has quit IRC
[19:04:31] <cylix> and with dans gardian if i use ident other users could put a fake ident to bypass on the local network.....
[19:05:06] *** solarisjon has joined #opensolaris
[19:05:07] <cylix> to many holes to plug I say.
[19:05:20] <PerterB> http://www.lysator.liu.se/~pen/pidentd/
[19:05:30] <PerterB> yeah, identd is only trustworthy if you control the server
[19:06:27] <cylix> PerterB,  well even then it can be used to hack in if you have a packet sniffer...  thanks I look at that server.
[19:08:40] *** Kush- has joined #opensolaris
[19:09:07] *** sahafeez has quit IRC
[19:09:42] <g4lt-mordant> cylix, I'm not aware of any exploits for identd
[19:09:47] *** kloczek has joined #opensolaris
[19:10:21] <g4lt-mordant> if you know of one, you might want to post it to bugtraq
[19:11:25] <cylix> g4lt-mordant, I mean in the case of irc auth or dansgaurdian auth all you need is the username generaly which you can get with a packet sniffer as most idents are plain text.
[19:11:54] *** sahafeez has joined #opensolaris
[19:12:29] <g4lt-mordant> so how's that exploitable?
[19:15:48] <cylix> g4lt-mordant, ... well truth is I can't think of a way it could be used without setting up another fake ident server.  now that you put me to it.
[19:16:46] <PerterB> the connections to the ident server are TCP, so you need to be able to relaible inject a spoof packet into an existing TCP connection to spoof it
[19:16:54] <PerterB> *reliably
[19:17:31] <PerterB> which is raising the bar higher than most of the people trying to hack around a web gaurdian to download pr0n ;)
[19:18:18] <cylix> I'll say this Ident for solaris Is a ton more complex than most I have seen for linux.  Some of that is lack of /proc/net/tcp most I think is features.  very nice if you need one.
[19:18:31] <cylix> PerterB, Yeah :-) I think your right.
[19:19:18] <cylix> Well thanks for the help all I have to run.
[19:19:26] *** cylix has quit IRC
[19:19:27] <PerterB> Linux /proc is handy and evil at the same time... it's evil because of all the people writing code who just assume every Unix-like OS has a similar /proc
[19:19:41] <icon> heh
[19:20:26] <icon>  /proc isnt a horrible idea, its just that linprocfs is disgusting
[19:20:35] <icon> they try to mix sysctl into proc and its just... wrong
[19:20:39] <PerterB> agreed
[19:21:00] <icon>  /proc is just fine for vewing well... procs
[19:21:07] <icon> leave the tunables to sysctl et al
[19:21:32] <gdamore> right.  and of course, solaris has its own methods.  (kstat, ndd, etc.)
[19:21:53] <icon> yup
[19:22:03] <icon> havent played with ndd
[19:22:23] <gdamore> one thing i like about solaris (especially compared to e.g. NetBSD) is that you rarely really need or want to tweak a tunable.  most things auto-tune.
[19:22:34] <icon> yup
[19:22:46] <gdamore> ndd is "weird".  my afe driver supports it, but it is "undocumented" for most intents and purposes
[19:22:49] <icon> i like the kernel architecture quite a bit too
[19:22:54] <icon> solaris internals is a fantastic book
[19:22:55] <gdamore> well, duh. :-)
[19:23:29] <gdamore> i've yet to see any kernel architecture that even comes _close_ to the Solaris in quality, architecture, design, etc.
[19:23:38] * icon nods
[19:23:39] <gdamore> (of course, i've not looked at Windows internals.)
[19:23:56] <icon> 'kernel? we dont need no stinkin' kernel'
[19:23:59] <icon> 'we have gdi!'
[19:24:06] <gdamore> hehe.
[19:24:22] <icon> to be perfectly honest
[19:24:22] *** salamanders has quit IRC
[19:24:45] <gdamore> every now and then i get some headhunter who just can't understand why i'm not interested in working on windows systems. :-)
[19:24:46] <icon> the fact that i can boot straight into a 64bit kernel without having to tweak a damned thing is very nice
[19:24:53] <icon> ugh
[19:25:00] <icon> im still cursed to develop on windows boxes
[19:25:07] *** sahafeez has quit IRC
[19:25:32] <gdamore> not me.  Solaris, NetBSD, Sun Ray firmware, and other embedded bizarreness, sure.  but not Windows. :-)
[19:26:13] <gdamore> i try really hard to stay clear of loonix too, but i'm less successful at that than I am with windows.  (I wound up doing a bunch of hacking on JDS1/2/Suse for our Talin laptops)
[19:26:41] <icon> meh
[19:26:42] * gdamore cheered when JDS 3 abandoned Linux and only supported Solaris.
[19:27:01] <icon> i would *kill* for a good embedded job
[19:27:12] <gdamore> really, where are you located?
[19:27:15] <icon> st louis
[19:27:21] <gdamore> that's tough.
[19:27:26] <icon> yeah
[19:27:28] <icon> theres not much here
[19:27:51] <icon> well there is for j2ee, which i make a killing on (working as an arch now)
[19:27:53] <gdamore> we (Tadpole) are looking to replace some attrition, but we'd like to get developers in Cambridge UK, or the SF bay area.
[19:28:00] <icon> but im originally a systems programmer :/
[19:28:18] <icon> oh really?
[19:28:18] *** dunc has quit IRC
[19:28:25] <icon> do you relocate :)
[19:28:28] <gdamore> Solaris kernel/embedded/NetBSD hackers wanted. :-)
[19:28:31] <gdamore> maybe.
[19:28:35] <gdamore> for the right person.
[19:28:59] <icon> ill keep that in mind
[19:29:05] <icon> meh, time for yet another meeting
[19:29:06] * icon &
[19:29:33] <PerterB> do they need to be current? I haven't done any serious kernel work for a few years but I'm probably closer to Cambridge than most people in here
[19:31:32] <gdamore> well, it kind of depends, i suppose.
[19:31:52] <PerterB> it usually does :)
[19:31:57] <gdamore> heh .
[19:32:01] *** bondolo has joined #opensolaris
[19:32:33] <gdamore> you can e-mail me your resume: garrett_damore at tadpole dot com -- i can review and pass on to mgmt if it looks good. :-)
[19:32:47] <dclarke> tadpole ?
[19:32:52] <dclarke> I contacted them ...
[19:32:53] <gdamore> yes.
[19:32:58] <dclarke> they never replied to anything
[19:33:03] <dclarke> who is the CEO there ?
[19:33:20] <gdamore> CEO _was_ Mark Johnston.  but now we are owned by General Dynamics.
[19:33:31] <gdamore> who did you contact?
[19:33:35] <dclarke> so who is the man/guy in charge ?
[19:33:54] <PerterB> I just changed jobs a month or two back so I'm not looking just now... was more of a general enquiry for future reference
[19:33:54] <gdamore> Mark handles engineering at least.
[19:34:08] <dclarke> okay .. I'll get him on the phone
[19:34:17] <gdamore> what do you need?  maybe i can help?
[19:34:39] <dclarke> just looking for a sponsor
[19:34:42] <dclarke> the usual
[19:34:52] <dclarke> as I see it .. TadPole is a commercial user
[19:34:53] <gdamore> oh, yeah, call mark then. :-)
[19:35:08] <gdamore> i sent him e-mail suggesting we sponsor as well.  i will repeat.
[19:35:22] <dclarke> Blastwave hits its birthday a few days ago .. I still have not got the homepage updat edone
[19:35:48] <gdamore> btw, its Tadpole, not TadPole.
[19:36:27] <Triskelios> lol TadPole
[19:36:48] <gdamore> i don't know if Mark can approve donations, though.  he certainly has a budget, but i don't know how he decides spending.  it might be better to work a level down, at my level or Andy Giles (my manager) who can push it up.
[19:37:03] <jamesd> check out my latest web page:   solaris power search, it uses google but give preffence to  56 solaris related websites   http://unixconsult.org/solaris
[19:37:22] *** solaris-user has joined #opensolaris
[19:38:48] <gdamore> wow.  the website for www.tadpolecomputer.com hardly mentions tadpole anymore.  it is totally General Dynamics/Itronix branded now.
[19:39:58] <dclarke> well .. it could be worse .. it could be a big big company like .. say Sun .. that drags my server into the ground for weeks and then couldn't be bothered to return my phone calls AFTER I spoke with two execs about variosu things for two months
[19:40:05] <dclarke> geez .. what a weird world
[19:40:14] <dclarke> I was just on the phone with Red Hat also
[19:40:21] <gdamore> heh.
[19:40:24] *** sahafeez has joined #opensolaris
[19:40:25] <g4lt-mordant> gdamore, it's all your fault ;P
[19:40:57] *** coffman has quit IRC
[19:41:12] *** calumb has quit IRC
[19:41:13] <sahafeez> a question of traditional raids - not zfs - 4 disk, raid 0+1
[19:41:37] <gdamore> that is a running theme. ;-)
[19:41:54] <sahafeez> mirror A+B and C+D then stripe the mirrors or stripe A+B and C+D and mirror the stripes
[19:41:58] <sahafeez> which is better
[19:42:06] <hile_> stripe the mirrors
[19:42:28] <sahafeez> why
[19:42:33] <hile_> though SDS  RAID0+1 -> 1+0 conversion automagically.
[19:42:45] <hile_> you can tolerate more spindles failing
[19:42:46] <sahafeez> not sun, just theory
[19:43:10] <dclarke> bloody hell
[19:43:12] *** deather_ has joined #opensolaris
[19:43:18] <dclarke> is this still the topic of discussion ?
[19:43:25] <dclarke> RAID 1+0 etc ?
[19:43:37] <sahafeez> hum in both cases - 1 disk failure would not take anything down. as far as i can see..
[19:43:38] <gdamore> heh.
[19:44:02] <gdamore> no, but in one case you can tolerate 2 disk failures.
[19:44:08] <sahafeez> hum, but ok, 2 disk one in each mirror would
[19:44:09] <hile_> sahafeez - 1 disk failure in a 0+1 setup would take one entire mirror down
[19:44:30] <sahafeez> got it.
[19:44:37] <dclarke> http://www.blastwave.org/dclarke/stuff/mminteractive_00.txt
[19:44:38] *** Tzoa has joined #opensolaris
[19:44:41] <sahafeez> now speed wise...which is faster
[19:44:42] <dclarke> read that
[19:44:59] <sahafeez> reading..
[19:45:16] <sahafeez> there are times i find you quite helpful there dclarke
[19:45:23] <dclarke> it happens
[19:46:09] <gdamore> (shhh... don't tell _him_. :-)
[19:46:10] <dclarke> I just happens to wade into the topic on the OpenSolaris discuss list and
[19:46:25] <dclarke> I was outraged at the blazing ignorance I see tossed around
[19:46:32] <dclarke> brb
[19:50:07] <dclarke> in any case .. I need to drive away .. be back in .. 30 min or so
[19:52:35] <gdamore> how does blastwave get its bandwidth?
[19:54:07] <hile_> they manufacture it out of thin air ;)
[19:54:08] <jamesd> gdamore, it pays for it..
[19:54:29] <gdamore> i have to imagine that it uses a lot, and that it is pretty pricey
[19:54:42] <gdamore> but i don't know that for sure
[19:54:53] <jamesd> last i heard it was about $2000 a month for bandwidth and 20RU of space.
[19:55:22] <gdamore> that makes sense.  i wonder how dennis pays for it.  i mean does he have regular corporate sponsors?
[19:55:48] <jamesd> a few sponsors and and donations... there is more info on the site.
[19:56:00] <gdamore> i was looking for that info, but it wasn't clear.
[19:56:25] <jamesd> http://www.blastwave.org/sponsors/index.html
[19:56:38] <gdamore> yeah, but that doesn't indicate recurring sponsorship
[19:57:02] <gdamore> ah, its annual.  that makes sense.
[19:57:16] <jamesd> genessi is the only recuring sponsor ... besides a few individuals
[19:58:26] <gdamore> i'm drafting a message to my mgmt, encouraging us to sponsor as well.  frankly, having a high-level advert would probably be good for us.
[19:59:41] *** deather has quit IRC
[19:59:47] <gdamore> heh.  but genesi are not listed on the blastwave home page.  i think dclarke has to live up to his promises. :-)
[20:00:18] <jamesd> they were.. you probably have to talk to dennis for more details..
[20:00:57] <gdamore> hmmm.... i will.  neither dtu nor codersource.net are listed either, and they are gold sponsors.
[20:01:01] *** deather_ is now known as deather
[20:01:24] <jamesd> they are featured on http://polaris.blastwave.org/ which is probably more important since they make ppc based hardware
[20:01:42] <gdamore> sure, that makes sense.
[20:03:45] <icon> ugh
[20:03:46] <icon> that sucked horribly
[20:04:05] * icon pulls out the advil
[20:04:56] <dclarke> gdamore : whats the issue ?
[20:05:05] <icon> afternoon dennis
[20:05:17] <dclarke> gday
[20:05:26] <gdamore> i'm trying to draft a message to mgmt in support of blastwave sponsorship
[20:05:41] <dclarke> gdamore : genesi is no longer a sponsor of Blastwave but they have made great efforts in the past year
[20:05:42] <gdamore> but the other gold sponsors/diamond sponsors i see don't have their logos anywhere.
[20:05:54] <gdamore> what about the gold sponsors?
[20:06:01] <dclarke> which is why they are on the sponsors page and also on the polaris page and they will stay there
[20:06:15] <dclarke> but .. they don't contribute a nickle to the operating costs
[20:06:19] <dclarke> thats my job
[20:06:22] <gdamore> fair enough.
[20:06:25] <dclarke> make sense ?
[20:06:45] <gdamore> what about DTU and codersource?
[20:06:56] <dclarke> codersource made a donation ..
[20:07:05] <dclarke> DTU is working on one
[20:07:13] <gdamore> basically, it would be nice if mgmt could see an example of how this would help us with advertising.
[20:07:23] <dclarke> the DTU guys really just want to make sure that GTK and a few things are up to date
[20:07:24] *** sahafeez has quit IRC
[20:07:34] <dclarke> its a big university in Europe and it looks nice there
[20:07:48] <dclarke> nothing showed up as a donation .. yet
[20:07:51] <dclarke> they are working on it
[20:07:58] <gdamore> ok.
[20:08:07] <richlowe> gdamore: the problem with you guys and advertising is that so few people who see the material could afford the product :)
[20:08:14] <dclarke> same thing you are doing probably . .trying to convince some grey haired bean counter that they need the software
[20:08:29] <gdamore> but a lot of those people work for _companies_ who could afford the product. :-)
[20:08:36] <dclarke> the eWeek article released yesterday helps
[20:08:48] <dclarke> oh .. they can afford it fine
[20:08:51] <gdamore> hey, i have grey hair, you insensitive clod!
[20:08:54] <dclarke> they can afford to ignore it too
[20:09:05] <dclarke> I have a head of slowly vanishing grey hair
[20:09:14] <dclarke> but .. are you a grey haired bean counter ?
[20:09:18] <dclarke> not very likely
[20:09:19] <gdamore> no.
[20:09:22] <richlowe> dclarke: borrow some of Tpenta's ;)
[20:09:26] <dclarke> you don't need to read a manual to use a mouse
[20:09:46] <gdamore> but i know mgmt is going to ask to see an example of the advert that they would get.
[20:10:20] <dclarke> screw that
[20:10:25] <dclarke> that sa dumb approach
[20:10:30] <dclarke> all bloody wrong
[20:10:31] <gdamore> you don't want the money then?
[20:10:34] <dclarke> its the reverse
[20:10:43] <dclarke> they dictate the banne that they want
[20:10:50] <dclarke> thats how it works
[20:11:05] <gdamore> sure.  but they will want to see how you have treated other sponsors in the past, I mean.
[20:11:14] <dclarke> oh .. look to the sponsors page
[20:12:08] <gdamore> yes, i see that.  the problem is that it _appears_ that you have sponsors that have met a certain level, but that you don't give them the same treatment that is promised on the sponsorship page.   do you see my point?
[20:12:51] <dclarke> no
[20:12:52] <dclarke> try again
[20:12:56] <dclarke> what ?
[20:13:09] <dclarke> be real specific
[20:13:18] <gdamore> it _looks_ like you have 2006 sponsors, that don't have logos on the home page
[20:13:35] <dclarke> lets suppose that Genesi says that the can contribute some money .. I then say .. cool .. let's stick you on the home page of a few places
[20:13:47] <dclarke> no .. no logos on the home page
[20:14:03] <dclarke> probably because the bloody homepage and logos was a huge argument inside the Blastwave
[20:14:07] <dclarke> discuss mail list
[20:14:15] <dclarke> and I have a new homepage in the works
[20:14:20] <gdamore> to quote: The ?Diamond? level sponsor will have their company logo on the Blastwave homepage as well as on every webpage of significance
[20:14:20] <dclarke> but .. thats not the point
[20:14:27] <dclarke> exactly
[20:14:33] <dclarke> and that is where Genesi was
[20:14:41] <dclarke> on every page all over the place
[20:14:46] <dclarke> and on multiple sites
[20:15:00] <dclarke> Genesi actually shipped hardware all over the world for the PowerPC project
[20:15:05] <dclarke> they actually gace away hardware
[20:15:06] <gdamore> oh, now i see, i was confusing Gold and Platinum Sponsorship.
[20:15:23] <dclarke> and then .. to top it all off .. the dropped some cash into the Blastwave PayPal account
[20:15:34] <gdamore> right i recall that.  (recall i offered to work on Polaris if I had hardware?)
[20:15:40] <dclarke> those levels were dreamed up by James Dickens
[20:15:49] <dclarke> yep .. I recall
[20:16:37] <gdamore> anyway, to avoid confusion, you might want to indicate that Genesi is a "Previous Diamond Sponsor and Current Partner" or somesuch.
[20:16:50] <dclarke> or you can explain that in your email
[20:17:11] <gdamore> the longer this message is, the more likely that it will wind up in the round file.
[20:17:21] <dclarke> oh .. bloody hell
[20:17:27] <dclarke> okay .. so .. what am I changing
[20:17:33] <dclarke> the Genesi bit ?
[20:17:38] <dclarke> okay .. I'll change that
[20:17:42] <gdamore> i give up.  you run it the way you want.  i'll send my recommendation.  if you get the money or not, is up to you now.
[20:18:38] *** bank has joined #opensolaris
[20:18:41] <dclarke> it has never been up to me
[20:18:46] <dclarke> wtf ?
[20:19:28] <dclarke> look .. I appreciate the effort you are making but you need to know .. I'm not new at this and I have heard every sort of reply ( negative ) from everyone from IBM to Oracle
[20:19:33] *** bank has quit IRC
[20:20:01] <dclarke> in every single case they quote that its not open source .. it is Solaris and not Linux and that the project is run by Sun
[20:20:24] <gdamore> i've sent my recommendation, suggesting we sponsor at either platinum or diamond.  i would not be surprised if Mark or other higher mgmt asks to see an example of what they will get for their $$.   but then again they may just round-file my suggestion.
[20:20:31] <dclarke> in the case of Sun .. I have written absolute promises in my email from execs that never deliver _anything_
[20:20:56] <dclarke> you did your bit .. you see .. thats fine
[20:21:03] <dclarke> thats a damn good thing
[20:21:06] <dclarke> and it may help
[20:21:16] <dclarke> but I already know the way that management thinks
[20:21:25] <dclarke> its a bottom line issue and its worthless to them
[20:21:29] <dclarke> its "free" software
[20:21:32] <dclarke> so why pay ?
[20:21:36] <dclarke> its that simple
[20:21:43] <gdamore> who knows.  sometimes they listen to me.  i got both Tadple and ATI to agree to opensourcing some stuff that they would not normally have, for example
[20:21:59] <dclarke> oh .. in that case you haveinfluence
[20:22:04] <dclarke> or great charm
[20:22:07] <dclarke> I lack both
[20:22:11] <dclarke> :-)
[20:23:04] <dclarke> okay .. I'm really stressed here
[20:23:07] <gdamore> mgmt will never (almost) give something away for free.  instead you have to figure out how to present it as being worthwhile.
[20:23:08] <dclarke> its not your fault
[20:23:32] <gdamore> which is why i'm so keen on the advert line.  because _that_ is probably worth something to them.
[20:23:39] <dclarke> some jerk in the OpenSolaris world decided to plant a knife in my face today .. and I have no patience for that crap
[20:23:55] <gdamore> ???
[20:24:01] <dclarke> not you ..
[20:24:04] <dclarke> no one you know
[20:24:21] <dclarke> just a loud mouth out of control German that thinks I work for him as his personal slave
[20:24:56] <dclarke> I just had to kill off his account at Blastwave .. the second time I did that in four years
[20:25:03] <hile_> If it's the german I'm thinking of, most of us think he's a prick too, so you're not alone :)
[20:25:12] <dclarke> but I will not put up with slander and accusations under my own roof
[20:25:55] <dclarke> probably not the guy you are thinking of
[20:26:04] <dclarke> in any case .. I'm going to move on
[20:26:07] * gdamore suddenly is glad that he was heretofore unaware of the existing of "Blastwave Accounts" :-)
[20:26:20] <gdamore> s/existing/existence/
[20:26:30] <dclarke> its just user accounts
[20:26:37] <dclarke> for access to the build stack and all that jazz
[20:26:58] <gdamore> ah.
[20:27:13] <gdamore> doesn't everyone already have their own sparc and x86 hardware? :-)
[20:27:50] <gdamore> speaking of which, i know gisburn is trying to find faster sparc hardware (with on48 or newer, IIRC) to work with.
[20:28:15] <richlowe> I thought he already had access to the blastwave machines?
[20:28:16] <dclarke> four years ago when this thing started .. no one had both
[20:28:29] <richlowe> dclarke: many don't now, for what it's worth.
[20:28:31] <richlowe> at least at decent speeds.
[20:28:34] <dclarke> and .. no one had licenses to Studio gear .. or Forte
[20:28:41] <dclarke> so .. I provided all that
[20:28:53] <dclarke> then we grew and Sun got all open about things :-)
[20:29:14] <dclarke> so .. now we still have a stack for software building and integrated software work
[20:29:32] <gdamore> Getting Forte/Studio for free was a great thing. :-)
[20:29:39] <dclarke> but .. you need to have an account to at least create a package you want rolled out to the world
[20:29:44] <dclarke> brb
[20:30:04] <gdamore> heh.  I just roll out my own software outside of CSW.   :-)
[20:30:14] <dclarke> good man
[20:30:15] <icon> hrmm
[20:30:22] <dclarke> anyways ...
[20:30:30] <dclarke> I'll be back
[20:30:45] <icon> need to grab a couple of disks for home (throwing them in a zfs pool). anyone have any favorites?
[20:30:56] <icon> i have a sata 2 cab connected to the opteron
[20:36:09] *** sahafeez has joined #opensolaris
[20:36:34] <elektronkind> um, seagates
[20:36:37] <elektronkind> or hitachis
[20:36:46] <icon> ive always been a hitachi guy
[20:36:56] <icon> seagates... well *shrug*
[20:36:59] *** sahafeez has quit IRC
[20:37:16] *** sahafeez has joined #opensolaris
[20:37:23] <icon> scsi seagates are hard to beat, but every ide i have ever touched has gone to crap in a hurry
[20:38:14] <sahafeez> that is because they use the crappy parts to into the ide
[20:38:25] <icon> i havent messed with them for sata yet
[20:38:36] *** solaris-user has quit IRC
[20:38:38] <icon> saw a couple of 500G hit's that looked a little lonely
[20:48:08] * Kush- wishes there is a edubuntu like distribution using opensolaris.
[20:49:35] *** holzmann has quit IRC
[20:51:53] *** nwf has joined #opensolaris
[20:53:07] *** Kernel86 has quit IRC
[20:56:04] *** Cybernd has joined #opensolaris
[20:57:30] <gdamore> i must say, i'm reasonably pleased at how easy it was to setup a network install server for Solaris x86 b50.  its the first time I've loaded x86 since grub was added.
[21:00:12] *** myrkraverk has quit IRC
[21:00:45] <gdamore> i'm disappointed that it didn't detect my 1400x1050 laptop panel properly though.  (at least i could manually configure it)
[21:02:59] <oxygene> gdamore: some panels have broken ddc information - hard to detect correctly
[21:03:09] *** nwf has quit IRC
[21:04:17] *** zarathustra has quit IRC
[21:06:39] *** zarathustra has joined #opensolaris
[21:07:30] <gdamore> the radeon driver can get information for most of these panels from the BIOS however.
[21:07:49] <gdamore> trust me... i've written a radeon driver.  i know how to get that info.
[21:08:06] <gdamore> (you could also determine it by looking at the registers that BIOS configured)
[21:08:30] *** _william_ has joined #opensolaris
[21:08:30] <richlowe> so fix Xorg? :)
[21:08:35] <richlowe> (assuming it's Xorg that misbehaved?)
[21:08:45] <gdamore> i don't think it is Xorg, but kdmconfig that is busted.
[21:08:56] <gdamore> but i don't know how kdmconfig figures out its proposed settings
[21:08:59] <_william_> hi all
[21:09:00] <Auralis> kdmconfig is for Xsun
[21:09:36] * richlowe can't think of a single compelling reason to use Xsun on x86.
[21:09:43] * timeless is still kinda hoping to someday figure out how to get X to use this display @ native resolution [16x12] :(
[21:10:30] <gdamore> i have Xorg running on my desktop at 1600x1200 no problem.  actually, i'm using nvidia driver with twinview, one at 1920x1200 and the other at 1600x1200.  very nice.
[21:10:43] <gdamore> anyway, kdmconfig/Xsun is used during suninstall.
[21:10:46] <timeless> did you have to do any special configuring?
[21:10:57] <gdamore> probably, but it was a while ago.
[21:11:16] <gdamore> what graphics board do you have?
[21:11:33] <timeless> nvidia something
[21:11:46] <timeless> it has two dvi ports
[21:12:17] <gdamore> hmm... i'm using 1920x1200 at digital, and 1600x1200 with analog.  (only one dvi port)
[21:12:23] <gdamore> hang on a sec.
[21:14:20] <gdamore> first time trying pastebin.com... its slow...
[21:14:53] <richlowe> rafb.net tends to be less slow.
[21:14:54] <timeless> pastebin.mozilla.org
[21:14:55] <gdamore> i used nvidia-xconfig, it appears.
[21:14:56] <timeless> paste.lisp.org
[21:15:50] <timeless> that's some thing you download from nvidia?
[21:16:23] <gdamore> http://rafb.net/paste/results/cVvJVQ17.html
[21:16:46] *** Kernel86 has joined #OpenSolaris
[21:17:02] <gdamore> i think so, but i don't specifically recall.  it may have been one of the standard drivers included in my Ultra 20.
[21:17:13] <timeless> http://rafb.net/paste/results/cVvJVQ17.html
[21:17:14] <timeless> eep
[21:17:34] <timeless> http://www.nvidia.com/object/solaris_display_1.0-8776.html
[21:17:35] * timeless reads
[21:17:36] <gdamore> that file is /etc/xorg.conf
[21:17:45] <richlowe> /etc/X11/xorg.conf
[21:17:52] <gdamore> richlowe++
[21:18:24] <timeless> so i take it that this thing doesn't ship w/ snv?
[21:18:28] <timeless> (the driver)
[21:18:52] <gdamore> i don't think so.  its closed source.
[21:19:13] <timeless> in theory this shouldn't reqquire a reboot :)
[21:19:22] <timeless> hehh
[21:19:25] <timeless> step 3. reboot
[21:19:25] <richlowe> Hm, I thought integrating the nvidia driver properly was being worked on.
[21:20:03] <gdamore> nfi.
[21:20:04] <elektronkind> anyone know what the pcplusmp driver does?
[21:20:15] <timeless> how would i know?
[21:20:20] <richlowe> MP x86 nexus driver, I think.
[21:20:29] <gdamore> that sounds right.
[21:21:39] <gdamore> woot.  my dvd-rw (for laptop sparc) just arrived
[21:21:40] <timeless> oh, my driver right now would be nv :)
[21:23:11] *** myrkraverk has joined #opensolaris
[21:23:18] <gdamore> switch to the nvidia driver, I think.
[21:23:30] <gdamore> the "nv" driver is the stock Xorg driver, and isn't as capable.
[21:23:34] <timeless> that'll require downing the box
[21:23:48] <timeless> it's kinda busy, i'll schedule it for next sunday i suppose
[21:23:53] <gdamore> shouldn't -- just restart X?
[21:24:04] <timeless> the instructions really say to reboot
[21:24:13] <timeless> http://download.nvidia.com/solaris/1.0-8776/README/chapter-01.html
[21:24:21] <gdamore> really.  maybe there is a kernel helper.  i see modinfo shows "nvidia" module
[21:24:42] <timeless> the idea of rebooting for x11 is silly
[21:25:07] <gdamore> yeah, unless you need the kernel helper, i don't think you need to reboot.  check modinfo output
[21:25:43] <postwait> I'm trying to use dtrace to look at some I/O contention.
[21:26:00] <postwait> However, the io:::done stuff always has the fileinto_t bits as <none>
[21:26:13] <postwait> So, I can't see which file the I/O is happening on...
[21:26:24] <postwait> Is there an approach to this that will give me the filenames?
[21:26:46] <richlowe> Hm, can you use fds[]?
[21:27:45] <postwait> How would I do that?
[21:28:19] <timeless> swift% modinfo|grep -i nvi|wc -l
[21:28:20] <timeless>        0
[21:28:21] <postwait> It's system-wide, and processes are created and terminated, so mapping them with pfiles is a non-starter
[21:30:19] <richlowe> postwait: would depend on whether you had an idea which fd was involved.
[21:30:32] *** alanc-away is now known as alanc_work
[21:31:25] <alanc_work> you need to reboot to use the nvidia driver so the nvidia kernel driver can replace the vgatext driver as owning the video board, since you can't really unload the console driver while the system is running
[21:31:36] <postwait> tons... maybe 1000 files on three filesystems
[21:31:53] <gdamore> interesting.  you can't have solaris and and linux on the same disk.  is there a collision in partition types?
[21:31:54] <postwait> various processes have them open with different FDs.
[21:32:29] <quasi> gdamore: there used to be - I think it has been fixed
[21:32:59] <gdamore> well suninstall certainly doesn't know that.  it complained loudly about the presence of a linux partition on my disk.  (no big deal, it was an old install of JDS 2.)
[21:33:00] <alanc_work> the old Solaris fdisk and Linux swap partition ids were the same - S10 introduced a new one for Solaris to end the conflicts
[21:33:01] <quasi> gdamore: I think solaris used the same partition type as linux swap
[21:34:05] <richlowe> That changed, I believe.
[21:34:12] <gdamore> you can tell just how involved I am with x86 hardware... :-)  first time I've installed Solaris x86 since Solaris 8. ;-)  my U20 can with S10 x86 preinstalled.
[21:34:13] <richlowe> SOLARIS v. SOLARIS2 as far as the partition type goes.
[21:34:35] <richlowe> seems alanc beat me to saying that.
[21:34:54] * myrkraverk is hoping to install sol now (unfortunately, I probably don't have time to burn all images before going to work ;/ )
[21:35:15] <timeless> alanc: thanks, that sucks, it sounds like the nt4 problem
[21:35:48] <myrkraverk> btw, do I need all 6 discs of excr ?
[21:36:26] <richlowe> more than likely, yes.
[21:36:27] <richlowe> sadly.
[21:36:27] <timeless> alanc: if i had an ssh connection, shouldn't the kernel be clean enough to let the vga driver let go? or is the problem the generic "video drivers/cards don't know how to reinit to a clean slate"
[21:36:28] <gdamore> timeless: blame nvidia. :-)
[21:36:49] <alanc_work> timeless: you always have a console running though
[21:36:52] <gdamore> if you had the console on a serial port that would be true.  but the problem is that you still have a login process holding the driver.
[21:37:04] <gdamore> s/login/getty/
[21:37:08] <alanc_work> (and I just know what the kernel guys tell me - I don't know enough to explain better)
[21:37:15] <gdamore> (or whatever the current Solaris tool for getty is)
[21:38:24] <richlowe> sac
[21:38:31] <gdamore> there are a few drivers you cannot update with rebooting, no matter what.  console, root disk, cpu, etc.  these are things that it is impossible to remove the last reference for
[21:38:38] <richlowe> eh, ttymon, rather.
[21:38:39] <myrkraverk> is the sfw stuff included in excr, or do I need to get that somewhere else?
[21:38:51] <richlowe> SFW is in SX:CR, CCD is... well, on the companion CD.
[21:39:05] <myrkraverk> ccd?
[21:39:06] <gdamore> richlowe: which is why i didn't say.  solaris/sysv sac/ttymon/portmon is confusing to me. :-)
[21:39:30] <alanc_work> which sfw stuff?  /usr/sfw is in SXCR, /opt/sfw is in the companion CD
[21:39:55] <myrkraverk> alanc_work: oh, I didn't know there were 2 of them ;P
[21:40:09] <alanc_work> never assume unique or logical naming when it comes to Sun - you'll end up confused in the end
[21:40:13] <myrkraverk> alanc_work: I'll live with /usr/sfw - I think, for a bit
[21:42:24] <myrkraverk> oh, and one question I have, is it possible to put zfs on memory sticks? is it sane, if so?
[21:44:09] <myrkraverk> (my I infer from the silence that noone here has tried?)
[21:45:00] <Error_404> haven't tried, but i've read of people in the forums doing it
[21:45:12] <Error_404> good way to simulate a broken raidz i guess
[21:45:20] <timeless> heh
[21:45:33] <gdamore> wtf would you want to?  the data wouldn't port very well.
[21:45:42] <gdamore> i.e. only solaris has zfs.
[21:45:49] <myrkraverk> gdamore: does it need to port very well? ;P
[21:45:59] <timeless> gdamore: macosx or fbsd may get it
[21:46:04] <gdamore> eventually.
[21:46:19] <Auralis> dragonfly bsd wants to port zfs as well
[21:46:22] <gdamore> there was talk about doing a port in NetBSD.  but it was too "forward" thinking I think. :-)
[21:47:15] <timeless> >     I am tentatively scheduling the port for March.
[21:47:27] <icon> gdamore: NetBSD to the bone )
[21:47:29] <timeless> but that was dec 05
[21:47:42] <icon> as much as i like the nbsd guys, they really need to deal with the big issues first... threads, smp, etc.
[21:47:45] <icon> before YAFS
[21:47:53] <gdamore> yes.
[21:48:08] <icon> netbsd today is pretty much the same netbsd we used in college
[21:48:22] <gdamore> i think those were precisely my comments.  without good smp, zfs was pretty much useless to nbsd.
[21:48:25] <icon> even open has stagnated a bit. it seems like free is the only one doing anything different
[21:48:27] <myrkraverk> icon: if I'm feeling ... that I have too much time ... I may try to port zfs to zeta :P
[21:48:44] <icon> haha
[21:48:49] <timeless> ooh, that'd be interesting
[21:48:57] <timeless> at least zeta should have the smp support
[21:49:02] <icon> im about to throw together a zfs pool at home
[21:49:11] <gdamore> i've done some interesting work "porting" netbsd to new hardware, but apart from a few minor improvements (notably bluetooth), nbsd is largely stagnant.
[21:49:17] <icon> how well does raidz handle on 3 disks?
[21:49:26] <delewis> I thought even NetBSD had a better SMP implementation than OpenBSD (which was non-existent till a few years ago, and is still non-existent on most platforms) :-)
[21:49:37] <icon> can only have hardware failure on one disk, but still, it at least gets me 2/3 capacity
[21:49:51] <icon> delewis: depends on the arch
[21:49:57] <gdamore> netbsd's SMP implementation is not S.  They have MP, but barely.
[21:50:09] <icon> i used to boot my sparc 20 in netbsd and only have the use of one cpu
[21:50:22] <delewis> gdamore: which means only one processor has access to the kernel, right?
[21:50:25] <gdamore> nbsd has MP for sparc32,but not for sparc64.
[21:50:31] <gdamore> delewis: yes.
[21:50:36] <icon> MP for x86 too iirc
[21:50:39] <delewis> ouch
[21:50:41] <gdamore> right.
[21:50:44] <delewis> so much for processor hotswaps :-)
[21:50:45] <icon> also
[21:50:54] <icon> the netbsd kernel is still giant locked all to hell
[21:51:01] <icon> fbsd has been making some headway, but they still have a ways to go
[21:51:02] <gdamore> thereis a "BigLock" in nbsd.  Same technology as SunOS 4.x
[21:51:22] <icon> its a BSD thing :D
[21:51:29] <delewis> how does the virtual memory implementation in NetBSD fair?
[21:51:44] <icon> its been years for me, but it wasnt horrible
[21:51:46] <gdamore> there is a lot of work being done right now to address the locking situation in nbsd.  Andrew Doran and Yamato-san are working on it.
[21:51:50] <icon> much better than linux at the time
[21:51:51] *** Kernel86 has quit IRC
[21:52:00] <gdamore> the nbsd vm is pretty good, actually.
[21:52:17] <delewis> I assume it can't be that bad -- they've got to have a similar solution as SunOS that's highly compartmentalized for portability.
[21:52:19] <icon> BSD 4.4 based... gotta love it
[21:52:26] <icon> yup
[21:52:35] <icon> sunos 4.x was based off of bsd 4.3 wasnt it?
[21:52:38] <gdamore> yes.  and they've taken a page or two from Solaris as well.
[21:52:41] <icon> or was it 4.2?
[21:52:46] <delewis> the SunOS 4.x VM was completely re-vamped
[21:52:49] <delewis> icon: 4.2, IIRC.
[21:52:50] <icon> ahh
[21:52:56] <icon> i thought 4.2
[21:52:58] <delewis> it still had all the old VAX cruft in it
[21:53:01] <delewis> and it was beginning to show
[21:53:02] <icon> ahh
[21:53:13] <icon> im still pretty new to the solaris vm
[21:53:18] <gdamore> nbsd still has all the old VAX cruft in it, and it _definitely_ shows. :-)
[21:53:23] <delewis> (and at the time Sun was using M68k wide-spread, and SPARC was just coming along)
[21:53:23] <icon> im used to seeing 0 swap used from fbsd
[21:53:26] <delewis> gdamore: ew!
[21:53:55] <gdamore> people actually still run nbsd on vax, or so i'm told.  i can't for the life of me figure out why.
[21:54:09] <myrkraverk> I believe some ppl are still using some unix on vaxen - it may be nbsd
[21:54:12] <gdamore> but i've already decided that "N" no longer is for "Net", but for "Nostalgia" :-)
[21:54:24] <elektronkind> heh
[21:54:27] <Error_404> got a vax kicking around, need an OS for it'
[21:54:40] <gdamore> nbsd will work on it.
[21:55:04] <Error_404> gdamore: where "it" is an infinite set
[21:55:06] * myrkraverk is involved a bit in vms clubs - that's where I hear about vaxen and stuff
[21:55:15] <icon> wow
[21:55:16] <gdamore> seriously, there are people who consider sun2 "viable" paltforms for netbsd, and a number of folks still using sun3 hardware with nbsd in "production".
[21:55:23] <icon> i havent touched a vax in a very, very long time
[21:55:24] <delewis> according to the USENIX paper, "SunOS Virtual Memory Implementation" -- "In December 1985 our group at Sun Microsystems began a project to replace our 4.2BSD-based VM system with a VM system engineered for the future"
[21:55:32] <mrdeviant> it amazes me that anyone would use netbsd for production
[21:55:38] <Error_404> if someone gave me a vax, i'd use the opportunity to play with VMS
[21:55:46] <delewis> or BSD4.3 :-)
[21:55:47] <myrkraverk> ;)
[21:55:49] <Error_404> or an alpha, for that matter
[21:55:56] <icon> meh
[21:56:05] <gdamore> our nextgen comet (sunray mobile client) was originally going to be netbsd on au1550, which is why i became involved in that community
[21:56:14] <icon> i never did mess much with the sun3's. i think the sun4c and sun4m were the ones i used the most
[21:56:20] * myrkraverk would like an alpha - or if someone wants to give it to me - an itanium for vms ;P
[21:56:30] <icon> gdamore: what is it now?
[21:56:42] <gdamore> sun3 was before my time.  i became involved in Unix right after the switch to sun4.
[21:56:49] <gdamore> icon: cannae say.
[21:56:50] <icon> gdamore: same here
[21:57:04] <icon> man i loved my sparc 5 ;)
[21:57:08] <alanc_work> I learned C at berkeley on a cluster of diskless Sun 3/50's
[21:57:13] <delewis> sun4m was before my time, so I had to go a bit back in history to get a hold of one :-)
[21:57:21] <gdamore> actually, i still have some au1550 hardware kicking around here.  i want to port Solaris to it. :-)
[21:57:31] <alanc_work> amazing they could run X, emacs, and gcc with only 4Mb of RAM and swapping over the net
[21:57:38] <delewis> I started getting interested in computers roughly when UltraSPARC-II had come out
[21:57:43] <gdamore> delewis: youngster. :-)
[21:57:48] <icon> i think the sun4u's had just come out when i started getting into 'real' unices, so it wasnt too tough to go back a step ;)
[21:57:49] <delewis> gdamore: :-P
[21:58:05] <icon> i used to drool over the ultra 5's
[21:58:16] <gdamore> the ultra 2 was _much_ nicer.
[21:58:20] <delewis> icon: sick.
[21:58:24] <icon> haha
[21:58:28] <delewis> even an Ultra I would've been better to drool over :-)
[21:58:39] <icon> it was the case :P
[21:58:42] * myrkraverk is just a total noob here - with only linux (on x86 mostly) experience
[21:58:52] <icon> u10's werent too shabby
[21:59:10] <icon> for whatever disgusting reason, i just really liked the u5 cases
[21:59:13] <delewis> myrkraverk: hey, I started using Linux in 2000, myrkraverk, so I have some guilt to spread around, as well.
[21:59:26] <icon> delewis: 97/98 here
[21:59:31] <myrkraverk> delewis: erm, that's the same year as I ;P
[21:59:33] <gdamore> i recall having problems trying to Sun OS 4.1.x on a sparc 20 with ross hypersparcs.  the kernel was not robust with smp hypersparcs.  it was fast, but it would crash about once every two or three days.
[21:59:34] <icon> i think i hit unix proper around 99
[21:59:37] <delewis> and furthermore, I started using Solaris in the Fall of 2004, IIRC.
[21:59:52] <gdamore> o wow, you noob. :-)
[21:59:57] <delewis> yeah :-)
[21:59:58] <icon> hrmm
[22:00:04] <icon> i had a set of hsparc 115's
[22:00:07] <icon> man that 20 screamed
[22:00:18] <icon> the sun chips were actually quite a bit better than the ross chips
[22:00:23] <Auralis> ss20, best box of its days
[22:00:25] <icon> i never could get my hands on a dual 70
[22:00:28] <icon> yup
[22:00:29] <gdamore> they wanted us to install Solaris 2.  but Solaris 2 was crap back then.
[22:00:31] <icon> man i loved that box
[22:00:32] <delewis> so, I've got 6 years of UNIX/Linux experience under my belt (well, actually one 2 using a real UNIX of sorts)
[22:00:37] <gdamore> these were dual 70s.
[22:00:38] <delewis> quite a shame, actually.
[22:00:39] <myrkraverk> delewis: hmm, that's quite before me - I'm just getting into sol *now* ;)
[22:00:44] *** Tpenta has quit IRC
[22:00:51] <delewis> I used FreeBSD and NetBSD for about 2 years.
[22:00:53] <myrkraverk> (if these burns work, that is, I have my doubts)
[22:01:01] <delewis> (from 2002-2004 on SPARC on i386)
[22:01:20] <gdamore> i remember when the netbsd project was _just_ getting started.
[22:01:26] <delewis> gdamore: impressive. :-)
[22:01:27] <icon> in college i used my old ipc sparc as a doorstop ;)
[22:01:34] <gdamore> i wasn't involved in it back then.  but it was an interesting alternative to sunos 4.x
[22:01:38] * Auralis remembers installing netbsd 0.x on an amiga
[22:01:39] * delewis is still in college
[22:01:39] <icon> that thing weighed almost as much as my ss20
[22:01:40] <delewis> :-)
[22:01:56] * icon is a bout to go back and never leave
[22:02:05] <myrkraverk> delewis: well, at least I'm not *only* an x86 guy - I'm a lot more familiar with the ps2 ;)
[22:02:17] <icon> i was a mips/sparc fiend in college
[22:02:20] <gdamore> heh.  i though linux 0.99.7 was cool.  i played a lot of nettrek on it -- it was nice on my 33MHz 386. :-)
[22:02:30] <myrkraverk> . o O (too bad it won't make a sense to put sol on that)
[22:02:30] <icon> god, anything was better than 16bit asm on ix86
[22:02:48] <icon> gdamore: sx 33?
[22:02:56] <gdamore> right.
[22:02:59] <gdamore> with 8mb ram.
[22:03:04] <delewis> s*u*x :-)
[22:03:05] <icon> first box i ever built for myself from scratch... had 4mb of ram
[22:03:30] <icon> heh, but i threw windows on it so i could play myst :)
[22:03:35] <myrkraverk> icon: I avoid building boxes - I prefer to build software ;P
[22:03:36] <gdamore> oh crap, my first machine was a 12MHz 286. :-)  well unless you count the 8bit c64 i had before that.
[22:03:44] <icon> bah meeting again
[22:03:50] <icon> gdamore: my first box too ;)
[22:03:55] * icon &
[22:04:10] <myrkraverk> (that reminds me, I always meant to get myst for the ps1)
[22:04:26] <gdamore> hacking around on the 6510 was fun.  (6510 was a 6502 variant found in c64)
[22:05:05] <gdamore> so are there other mips weenies here that would be interested in helping out with a mips port of solaris?
[22:05:10] <myrkraverk> gdamore: oh - my only real exp points in assembly were z80 in '97 or so, and mips 5900 in 2002+
[22:05:34] <myrkraverk> gdamore: *maybe* - preferably without gcc ;P
[22:05:46] <gdamore> hard to do it without gcc unless sun opens up forte.
[22:05:52] <gdamore> s/forte/studio/
[22:06:00] <kleppari> not hard, just design a new compiler!
[22:06:10] <gdamore> see, "hard". :-)
[22:06:13] <myrkraverk> gdamore: hmm, true unless sgi opens mipspro ;P
[22:06:48] <gdamore> gcc does okay with MIPS.  i used it a bunch with the nbsd work i've done lately.  (au1550, ar5312, etc. are all mips variants)
[22:07:08] <delewis> gcc on SPARC still leaves a lot to be desired :-)
[22:07:21] <gdamore> yes, but i'm not talking about sparc.
[22:07:26] <delewis> register windows just seem to be a completely new concept to them
[22:07:34] <gdamore> i wonder how well gcc4 does, though.  gcc 4 got a lot better for MIPS.
[22:07:41] <delewis> gdamore: same as before
[22:07:44] <kleppari> well, GCC on SPARC is not at all very desirable..
[22:07:46] <Auralis> gcc4 is the same crap on sparc then 2.95
[22:07:51] <delewis> almost 0 improvement was done on SPARC code generation for gcc-4.x
[22:07:57] <hile_> afternoon auralis
[22:08:18] <Auralis> howeverm sun has a special gcc4 with sparc code generator for download, that thing is damn nice
[22:08:38] * delewis remembers a few months ago when Sun Studio developers posted on the gcc-devel list regarding their new back-end code generator for GCC
[22:08:41] <delewis> they were *not* happy.
[22:08:43] <twincest> Aur: will the gcc people integrate it?
[22:08:47] <gdamore> that's sad.  but i think Sun and the gcc team have had some "disagreements", such that I think the gcc people are not interested in sparc anymore
[22:08:50] <delewis> twincest: of course not
[22:08:57] <delewis> the gcc people went off the deep end when they heard about it
[22:09:10] <twincest> why?
[22:09:14] <delewis> it's in some mailing list archive somewhere -- google for it
[22:09:19] <richlowe> gcc have a history of refusing patches that may even *kind of* make that easier to do.
[22:09:23] <richlowe> even before someone went and actually did it.
[22:09:24] <twincest> hm, i don't care quite that much :)
[22:09:33] <myrkraverk> delewis: I've heard a lot of gripes about mips code generation improvements too - but I'm not really in those circles anymore
[22:09:46] <richlowe> the thought of being able to plug in a proprietary back-end is, for obvious reasons, anathema to them.
[22:09:47] <gdamore> yes, although i'd like to see studio 11 open sourced.
[22:09:59] <delewis> well, I think the biggest gripe the gcc-devel people had was the proprietary nature of the code generator
[22:10:04] <richlowe> (why yes, I did go through that just to use anathema in a sentence, it's one of those days)
[22:10:23] <myrkraverk> richlowe: thing is, at least in the past, it was wery hard to get anything not x86 into gcc
[22:10:24] <delewis> and they proceeded to give the "Well, it doesn't matter how good your code generator is. Our's is better in any way because it's `open'"
[22:10:56] <delewis> sorry, but the last thing I care about when I look at SPARC assembly is whether or not it was generated by something that was "open" or not.
[22:10:58] <libkeiser> gotta love the open license trumps good technology argument...
[22:11:15] <hile_> idiots :)
[22:11:18] <hile_> but what else is new.
[22:11:23] <twincest> so gcc proves open source is a failure because the open backend isn't being improved? :)
[22:11:26] <libkeiser> uh, the sky is blue?
[22:11:30] <Cybernd> delewis gpl evangilists got other priorities
[22:12:02] <myrkraverk> twincest: that would be a twist of the truth - but to by too much, I guess - it really depends on the ppl involved
[22:12:08] <gdamore> don't get me started on FSF bigotry. :-)
[22:12:47] <delewis> 3~n
[22:13:07] <delewis> hmm, stray key presses there.
[22:13:14] <myrkraverk> gdamore: ;) (that is one reason I'm starting to avoid linux)
[22:13:24] <delewis> myrkraverk: still go in #emacs very often? :-)
[22:13:31] <delewis> (you can tell I've dropped that channel)
[22:14:03] <delewis> and I've strongly considering dropping my role as the mplayer Solaris port maintainer -- it's hard to maintain software developed by a group of people that don't care about your platform, period.
[22:14:08] <myrkraverk> delewis: yes, I've been a regular for years (only, the channel spirit has been worsening lately)
[22:14:20] <gdamore> i still use gnu emacs.  its a nice program.  but i will never ever again release a piece of software under GPL if I can help it.
[22:14:21] <myrkraverk> delewis: ah, yes
[22:14:24] <timeless> has anyone considered openwatcom?
[22:14:36] <gdamore> isn't it x86 only?
[22:15:04] <myrkraverk> timeless: I have - but for me, its c++ support is rather crappy - I've even considered improving it - but so far, haven't had the time ;/
[22:15:30] <myrkraverk> gdamore: no, it's at least alpha too, but I don't know how many code generators it really has
[22:15:35] <timeless> it supports alpha at the least
[22:15:43] <timeless> and i thought it had at least one other generator
[22:15:52] *** Gman has joined #opensolaris
[22:15:55] <timeless> although trying to remember which it was isn't the easiest thing for my tiny little mind
[22:16:00] <Gman> morning dudes
[22:16:00] <twincest> what about tendra?  i thought that supported sparc
[22:16:01] <myrkraverk> last sol cd in the burner, I'll be running out any second now
[22:16:19] <twincest> (and some other odd target like m68k)
[22:16:30] <delewis> I still think Sun has plans to open source Forte
[22:16:43] <delewis> Jonathan has slipped at least once regarding it
[22:16:57] <delewis> (that or his definition of "open source" is muddled)
[22:17:06] * myrkraverk is out
[22:17:25] <richlowe> delewis: I'd bet the latter.
[22:17:34] <richlowe> or I'd bet he meant "netbeans" when he said it.
[22:17:36] <richlowe> rather than considering the rest.
[22:17:50] <gdamore> i would love to see studio open sourced.  then we could build cross compilers with it.  the one thing where netbsd really rocks is its cross compilation environment
[22:17:57] *** tek-ops has joined #opensolaris
[22:18:12] <delewis> richlowe: pity.
[22:18:17] <tek-ops> I'm attempting to migrate a home system to opensolaris, and going over the prep work now
[22:18:48] <tek-ops> since my storage drive is formatted ext3, i'd rather wait to convert it after i move the files to the opensolaris root drive
[22:18:59] <tek-ops> does opensolaris support ext3? does it support it by default?
[22:19:01] <delewis> and regardless (I forgot about this) -- a large part of Sun Studio is still closed
[22:19:04] <delewis> all the Rogue Wave C++ stuff
[22:19:06] <Auralis> tek-ops: it does not
[22:19:15] <tek-ops> can I build in support?
[22:19:16] <delewis> and probably much more we don't know about
[22:19:27] <Auralis> tek-ops: if you can write a driver, then yes
[22:19:48] <tek-ops> hmmm, alright, damn I really suck at filesystems...
[22:20:17] <twincest> ext3 can be downgraded to ext2, there is a third party ext2 driver
[22:20:22] <Auralis> i think there is a ext2 driver flying around on the net, no clue if it works, might be worth a look
[22:20:41] <tek-ops> hmmm, i probably should back up the drive first, hah
[22:20:42] <kleppari> I've used an  ext2 driver on Solaris..
[22:20:51] <kleppari> I can't remember where I got it though
[22:21:03] <tek-ops> kleppari, did it work for you?
[22:21:06] <kleppari> and IIRC it only had read-only support
[22:21:07] <kleppari> yeah
[22:21:34] <tek-ops> that's fine, i just need to pull the files off after install and before reformat
[22:22:03] <tek-ops> i'm curious, what is the most standardly supported filesystem
[22:22:23] <twincest> FAT
[22:22:23] <Auralis> ufs is default, zfs is the new baby
[22:22:27] <richlowe> hsfs, probably.
[22:22:28] <tek-ops> by say, OS X, windows, Solaris, Linux,a nd BSD
[22:22:39] <tek-ops> or as many of those as possible
[22:22:40] <richlowe> fat, or udf, if you mean writable.
[22:22:43] <Auralis> fat and hsfs
[22:22:49] <gdamore> probably everything except windows uses ufs.  if you want windows, you have to use fat
[22:23:00] <twincest> gda: i heard linux ufs is a bit fiddly
[22:23:04] <richlowe> gdamore: no no no!
[22:23:13] <delewis> UFS is also endian-dependent.
[22:23:15] <richlowe> gdamore: most support UFS, but they mostly support a slightly different UFS each :)
[22:23:16] <Error_404> twincest: linux ufs is read-only
[22:23:50] <gdamore> i've found that if you turn off ufs logging in solaris, it pretty much works across the board, but I don't use it with linux, only *bsd.
[22:23:55] <richlowe> (with added forcefullness, since I half recall mounting/fscking the 'wrong' UFS in the wrong place being especially fatal to the filesystem in general)
[22:24:12] <gdamore> well, that leaves FAT then. :-)
[22:25:06] <dwc-> iso9660
[22:25:19] <dwc-> but... with different extensions all around
[22:25:34] <gdamore> yeah, iso9660 is tricky.
[22:25:54] <gdamore> if you include network filesystems, NFS and CIFS.  (only CIFS for Windows. :-)
[22:26:09] <richlowe> I think there's NFS support in the services for unix thingy.
[22:26:21] <twincest> no cifs mounting in solaris
[22:26:24] <richlowe> I haven't made any effort to try and use it though.
[22:26:24] <twincest> (afaik)
[22:26:27] <Auralis> yes
[22:26:36] <gdamore> sure, but you can use samba client. :-)
[22:26:43] <dwc-> sneakernet+floppy
[22:26:52] <gdamore> *dd*.
[22:26:59] <timeless> twincest: erm?
[22:26:59] <dwc-> or I guess cdroms better than floppies
[22:27:05] <timeless> i use cifs here on opensolaris
[22:27:11] <Auralis> bah, line printer and ocr scanner :)
[22:27:12] <tek-ops> hmm, actually, i've been meaning to move storage to be networked
[22:27:17] <twincest> timeless: what is the fstype?
[22:27:20] <richlowe> timeless: there's smbclient, and similar.
[22:27:25] <richlowe> but there isn't kernel support (yet)
[22:27:36] <timeless> twincest: probably 'gnomevfs' :)
[22:27:39] <tek-ops> but all my experience with "directories" and networked filesystems has been NIS and NFS
[22:27:44] <twincest> timeless: that's not mounting, it's a userland shim
[22:27:49] <tek-ops> I was planning on doing that after moving to LDAP
[22:27:57] *** yippi has joined #opensolaris
[22:28:02] * timeless shrugs
[22:28:04] <timeless> i live in userland
[22:28:07] <timeless> i hate being root
[22:28:16] <twincest> userland as in non-kernel, not non-root
[22:28:26] *** nachox has quit IRC
[22:28:36] * timeless is used to mount being something reserved for root or root like
[22:28:50] <twincest> cifs mount on solaris is not reserved to anyone because it doesn't exist
[22:29:48] <timeless> anyway, i kinda like os'sthat do userland files (qnx?)
[22:30:44] <twincest> well if solaris supported it in userland that'd be fine too, but it doesn't do that either... only certain specific applications can
[22:30:47] *** mrdeviant has quit IRC
[22:31:11] <AbeFroman> automount might help there
[22:31:56] <tek-ops> well, i'm headed out, thanks for the help people
[22:32:01] <tek-ops> i'll be back again
[22:32:06] *** tek-ops has quit IRC
[22:32:39] <dwc-> hm, hp-ux has a mount_cifs
[22:34:17] <richlowe> http://opensolaris.org/os/project/smbfs/
[22:35:50] <gdamore> hmmm... pgx64 hardware, has problems with screen saver.  works fine until the screen saver comes on, then _boink_ no video.  video does not come back.
[22:35:58] <gdamore> this is on sparc hardware.
[22:37:58] <g4lt-mordant> gdamore, ?!  isn't the pgx64 on the SB100?
[22:38:45] <gdamore> i'm using it on a tadpole platform.
[22:38:52] <gdamore> i think the sb100 is m64 based.
[22:39:21] <g4lt-mordant> ahh
[22:39:30] <g4lt-mordant> ATI rage128?
[22:39:31] <alanc_work> the SB100 has a PGX-64 framebuffer, which uses the m64 driver
[22:39:51] <gdamore> huh.
[22:40:01] *** lolownia has joined #opensolaris
[22:40:50] <g4lt-mordant> if that's the case, my NV16 works fine with the screensaver
[22:41:09] <gdamore> no, i'm pretty sure this is pgx64 specific.  just a sec.
[22:41:09] <alanc_work> PGX-64 is the marketing name for an ATI Rage XL with 8Mb VRAM, as used on the SB100/SB150 motherboards and PCI addin boards
[22:41:32] <alanc_work> the m64 driver is used for the PGX-8, PGX-24 and PGX-64 models of ATI Rage-based fbs
[22:41:41] <gdamore> maybe i'm confused.  this is TSI gfxp driver.
[22:41:48] <alanc_work> that's the PGX-32
[22:41:59] <gdamore> ah.  well that's the one that is having problems.
[22:42:12] <alanc_work> which was the TechSource Raptor GFX with a Sun sticker on it
[22:42:13] <gdamore> alan: can you test?
[22:42:28] <alanc_work> I'd have to dig up a PGX-32 in our lab
[22:42:31] <gdamore> right.  we use the same part onboard in the Tadpole Voyager IIi.
[22:42:41] <gdamore> If you can't find one, let me know, and I can send you one.
[22:42:50] <gdamore> I don't have gfxp source, so I can't help debug.
[22:43:18] <alanc_work> we used to have a bunch - just haven't finished unpacking/organizing our lab since it moved to MPK from the being-closed-down NWK campus last month
[22:43:29] <alanc_work> I don't have gfxp source either
[22:43:40] <alanc_work> I don't know if Sun does or if we just call TechSource
[22:44:06] *** bougie has quit IRC
[22:44:18] <alanc_work> there have been some bugs/fixes in various Nevada builds in Xsun itself calling or not calling the FBIOSVIDEO ioctl
[22:46:33] <lolownia> Does opensolaris have AMD64 support ?
[22:46:39] <twincest> yes
[22:46:40] <alanc_work> yes
[22:47:01] <Error_404> lolownia: you better believe it does
[22:47:35] <lolownia> believe?what Do You mean? I have to keep my finger crossed during install?
[22:47:52] <lolownia> Or is it Plug'n'Pray :)
[22:49:40] *** proog has joined #opensolaris
[22:50:16] <Error_404> it'll work
[22:50:41] <Error_404> it was one of the major driving forces behind solaris10
[22:51:26] <proog> is there any way to install solaris without using cd/dvd? like booting a floppy and downloading the rest?
[22:51:40] *** wotdu has joined #opensolaris
[22:51:43] <wotdu> hey
[22:51:55] <wotdu> could you guys tell me where opensolaris/amd64 is? cos i can't find it
[22:52:22] <twincest> wotdu: assuming you want sx:cr amd64, it's part of the x86 version
[22:52:29] <alanc_work> it's the same ISO's as the x86 release - it has the binaries for both x86 & x64 in a single package
[22:52:31] <wotdu> sorry, sx:cr ?
[22:52:37] <twincest> solaris express community edition
[22:52:41] <wotdu> alanc, well, why?
[22:52:42] <richlowe> twincest: the same applies to 10, and SX, too.
[22:52:47] <twincest> rich: i know :)
[22:52:51] *** lolownia has quit IRC
[22:52:57] <alanc_work> wotdu: why make two copies when one works?
[22:52:58] <wotdu> what's the point in having both x86 and amd64 in the same
[22:53:00] <twincest> wotdu: solaris amd64 is not a separate platform like Linux
[22:53:10] <wotdu> alanc, cos its almost the double size
[22:53:14] <twincest> no it's not
[22:53:19] <delewis> wotdu: no!
[22:53:26] <alanc_work> amd64 version needs all the i386 libraries for running i386 applications
[22:53:27] <delewis> why does 'vi' need to be compiled for 64-bit?
[22:53:39] <wotdu> alanc, but i don't want to run i386 applications
[22:53:43] <richlowe> delewis: incase ip.c grows anymore? ;)
[22:53:48] <richlowe> delewis:.... oh, nm. :)
[22:53:52] <delewis> wotdu: you mis-understand 64-bit.
[22:54:05] <wotdu> delewis, enlighten me :)
[22:54:06] <delewis> or rather you've been contaminated by Linux.
[22:54:13] <alanc_work> you're not the typical customer then
[22:54:28] <delewis> wotdu: not everything in an operating system needs to be compiled for 64-bit
[22:54:31] <wotdu> customer *cof* *cof*
[22:54:47] <delewis> you should only very selectively compile applications for 64-bit on a case-by-case basis.
[22:54:50] <alanc_work> typical user then if you prefer
[22:54:57] <wotdu> delewis, delewis why?
[22:54:58] <delewis> "Does 'vi' need 64-bit? no"
[22:55:01] *** proog has quit IRC
[22:55:04] <wotdu> well it doesn't need
[22:55:05] <delewis> "Does my math library need to be compiled for 64-bit? yes"
[22:55:10] <wotdu> but what's the problem if it is?
[22:55:18] <delewis> wotdu: it's a waste
[22:55:22] <wotdu> saves me the trouble of maintaning 2 versions of libraries
[22:55:40] <rydis> delewis: But what if you want to edit your raw disk devices with vi? ;)
[22:55:44] <delewis> wotdu: sorry, that's life if you want binary compatibility across bit-ness.
[22:55:53] <wotdu> delewis, but i don't want :P
[22:55:57] <richlowe> you'd have to maintain two libraries *anyway*.
[22:55:58] <delewis> wotdu: go elsewhere.
[22:56:02] <richlowe> you just wouldn't have them installed at the same time.
[22:56:03] <twincest> wotdu: then don't maintain two versions of your library
[22:56:03] <wotdu> richlowe, why?
[22:56:04] <delewis> and that's true
[22:56:17] <alanc_work> you don't have to maintain two copies of the software you build yourself if you only ever want to use the 64-bit versions, but why take that choice away from everyone?
[22:56:30] <hile_> because we as a whole value compatibility a hell of a lot more than being "bleeding-edge"
[22:56:31] <wotdu> alanc, not taking the choice away
[22:56:40] <wotdu> i'm currently using gentoo
[22:56:48] <delewis> that explains it all.
[22:56:49] * Gman values bleeding edge ;)
[22:56:50] <wotdu> you have the option to install x86 compat libs
[22:56:57] <wotdu> but you're not forced into it
[22:57:08] <twincest> wotdu: what would you gain from using a solaris build with no 32-bit support?
[22:57:08] * hile_ :-Ps Glynn just on principle :)
[22:57:14] <delewis> wotdu: the majority of users of Solaris want to be able to use 32-bit and 64-bit applications without needing to take such steps.
[22:57:15] <wotdu> twincest, disk space
[22:57:19] <twincest> wotdu: very little
[22:57:20] <Error_404> solaris isn't brain dead, so there's no reason to bloat up programs with compiling them 64 bit
[22:57:28] <delewis> in 99% of cases, you're better of running 32-bit, anyway.
[22:57:36] <Error_404> unless there's a good reason to make them 64bit
[22:57:44] <Error_404> eg, bignum math operations
[22:58:08] <hile_> or things like perl where you might interact with a 64bit db2 or orrible instance
[22:58:56] <wotdu> what is solaris like btw? fancy configuration gui's and crap?
[22:58:59] <richlowe> mornin' Gman.
[22:59:16] <delewis> wotdu: it's a real UNIX.
[22:59:19] <Error_404> wotdu: there's smc, but it sucks so nobody really uses it
[22:59:24] * hile_ rolls eyes
[22:59:32] <wotdu> delewis, define real unix :P
[22:59:33] <alanc_work> heh - most users complain we don't have enough config guis in Solaris
[22:59:51] <Gman> hey richlowe
[22:59:52] <delewis> wotdu: one that's been certified throughout it's lifetime as being UNIX98 (workstation compliant as well) and UNIX2003.
[22:59:52] <Error_404> wotdu: code originating from bell labs unix
[23:00:04] <Error_404> or open group certified
[23:00:16] <delewis> Linux fails to meet either of these certification criterion, therefore, it should be considered a non-UNIX, by comparison.
[23:00:19] <delewis> or UNIX-like.
[23:00:25] *** alanc_work is now known as alanc-away
[23:00:31] <wotdu> and your point is?
[23:00:41] <hile_> that's not UNIX?
[23:00:50] <wotdu> i only use gnu/linux cos nvidia is taking shitloads of time to release fbsd/amd64 drivers
[23:01:53] <dwc-> GNU better not be unix ;)
[23:01:59] <delewis> dwc-: :-)
[23:02:02] <wotdu> hehe
[23:02:24] <wotdu> let's start a flamewar
[23:02:26] <delewis> wotdu: and again, FreeBSD is not a UNIX, by defintion (though, it does have far more UNIX-like roots than Linux does).
[23:02:27] <wotdu> gpl vs bsd
[23:02:31] <wotdu> :P
[23:02:46] <Error_404> wotdu: opensolaris is CDDL, not bsd
[23:02:47] <delewis> wotdu: in our case, it would be GPL v. CDDL, but thankfully, that's died down lately.
[23:02:57] <delewis> (please don't start it up again :-)
[23:03:15] * timeless chuckles
[23:03:32] <wotdu> where can i get a list of full UNIX OS's ?
[23:03:45] <Error_404> that'd be a long list
[23:03:46] <delewis> wotdu: opengroup.org and take a look at the UNIX98 and UNIX2003 "certified product" lists
[23:03:48] <jamesd> google.com
[23:03:49] <dwc-> opg
[23:03:53] <delewis> Error_404: it's actually quite short.
[23:04:07] <delewis> AIX, HP-UX, Solaris, and IRIX are UNIX98-certified (possibly Tru64)
[23:04:23] <delewis> the only operating systems that are UNIX2003 certified are Solaris 10 and AIX 5
[23:04:34] <Error_404> does it include every half-baked vendors' unix from the 80's? because that's the 'long list' i was referring to
[23:04:35] <wotdu> well, proprietary ....
[23:04:40] <delewis> Error_404: no
[23:04:45] <delewis> wotdu: point being?
[23:04:50] <dwc-> apparently unix didn't exist before '98 and '2003 ?
[23:04:56] <wotdu> delewis, i'm a free software zealot :P
[23:04:57] <delewis> dwc-: :-)
[23:05:07] <delewis> wotdu: oh, you're one of those.
[23:05:15] <Error_404> wotdu: most solaris types are "not-crap software" zealots
[23:05:41] <wotdu> Error_404, what are you calling crap to?
[23:05:44] *** sahafeez has left #opensolaris
[23:05:59] <delewis> wotdu: he was saying that the majority of 'free' software exhibits 'crap-like' features.
[23:06:16] <wotdu> i heard solaris uses gnome
[23:06:17] <delewis> if I could translate that anymore bluntly, I would.
[23:06:20] <wotdu> :P
[23:06:25] <Error_404> actually i was referring more to the "who cares if it's proprietary if it works best" attitude
[23:06:26] <dwc-> heck, a lot of non-free software exhibits 'crap-like' features too
[23:06:32] <delewis> wotdu: yes, a modified Gnome 2.x
[23:06:48] <wotdu> Error_404, yeah i know that attitude
[23:07:22] <wotdu> but if i can pick between two produtcs and the proprietary one isn't insanely better, i'll pick the free one
[23:08:22] <dwc-> I used to care more... but last weekend, I installed ndiswrapper and a proprietary windows driver on my linux machine
[23:08:32] <wotdu> dwc-, bah
[23:08:38] <dwc-> instead of rebuilding the kernel with the devicescape wireless stack
[23:08:39] <wotdu> just get a better card
[23:08:41] <delewis> wotdu: well, fortunately, OpenSolaris is "free," nowdays, so you not only have "insanely better" but "completely free"
[23:08:53] <dwc-> lazyness wins out
[23:08:54] <delewis> doesn't get much better than that :-)
[23:08:58] <wotdu> delewis, show me the benchmarks and i'll believe you
[23:09:03] <dclarke> wow .. Red Hat Enterprise Linux sucks ... http://www.blastwave.org/dclarke/up2date.png
[23:09:10] <delewis> wotdu: why are benchmarks the only thing that matters to you?
[23:09:11] <dclarke> on that note .. I have to step out
[23:09:12] <dwc-> aptitude install ndiswrapper && ndiswrapper -i foo.inf
[23:09:17] <wotdu> dclarke, really.. who gives a crap about Red Hat?
[23:09:26] <wotdu> they could die and i wouldn't give a shit
[23:09:29] <dwc-> wait wait
[23:09:29] <delewis> don't you want reliability, stability, and time-tested solutions to various operating systems problems (that most operating systems have only *begun* to address)?
[23:09:35] <dwc-> rhel sucks because of up2date?
[23:09:40] <dwc-> then solaris sucks because of smpatch
[23:09:47] <wotdu> delewis, not really
[23:09:50] <clee> rhel sucks for a lot of other reasons
[23:09:52] <gdamore> would like to click that link.  but my Xserver's mouse went out to lunch.  need to restart X ...
[23:09:53] <wotdu> not on the desktop
[23:09:56] <clee> up2date isn't *that* bad
[23:10:13] <wotdu> performance > extreme reliability on the desktop
[23:10:21] <delewis> wotdu: regardless, Sun has published a number of record-setting benchmarks on Solaris/AMD64
[23:10:21] <clee> rpm, on the other hand, is terrible
[23:10:26] <twincest> i prefer reliability on my desktop, actually
[23:10:54] <richlowe> performance is nice, losing my data is unforgivable.
[23:10:55] <wotdu> delewis, is there a closed Solaris anymore?
[23:10:56] <delewis> wotdu: most of us actually use are desktops to accomplish work, thus reliability and stability are just as important on my Blade 1000 as they are on my x86 fileserver.
[23:11:16] <Error_404> you can still download solaris9 iirc
[23:11:17] <delewis> wotdu: Solaris 10 uses a separate code-base from OpenSolaris, so you could consider it closed.
[23:11:30] <gdamore> kinda like staroffice vs. openoffice.
[23:11:31] <delewis> Solaris 11 will be based on the OpenSolaris bits, but will contain several other closed bits, as well, such as CDE, etc.
[23:11:43] <hile_> at least, we hope it still contains CDE
[23:11:53] <timeless> why?
[23:11:56] <gdamore> i'm not sure if i hope that or not.
[23:12:02] <delewis> currently, Solaris Express (and Solaris Express: Community Edition) are destined to evolve into Solaris 11, so they use the OpenSolaris bits, as well.
[23:12:13] <delewis> gdamore: I think it will make it in 11
[23:12:16] <hile_> well delewis and I do at least :p
[23:12:21] <delewis> whether or not it makes it into 12 will be more questionable.
[23:12:23] <wotdu> delewis, kinda like redhat and fedora thingie ?
[23:12:32] <twincest> not at all like fedora and fedora
[23:12:45] <delewis> wotdu: sort of, I suppose, though Solaris Express still has a lot more quality control around it than Fedora.
[23:12:58] <Error_404> like redhat & centos more i guess
[23:13:17] <Kush-> is there a reason sun keeps shipping with CDE? (if the lightweight'ed ness is the issue, why not XFCE)
[23:13:23] <gdamore> well, that goes all the way down to the kernel.  the quality processes associated with the solaris kernel are far more stringent than those associated with linux
[23:13:31] <delewis> Kush-: UNIX98-workstation compliance for one
[23:13:36] <twincest> kush: because you can't just drop a product that customers use with no warning
[23:13:37] <delewis> (CDE/Motif are required)
[23:13:46] <wotdu> gdamore, how do you know that?
[23:13:47] <delewis> and many, many customers are still using Motif-based applications.
[23:13:52] <gdamore> Kush: stay tuned to desktop-discuss.  i'm about to post a window manager comparision
[23:13:55] <jamesd> Kush-, its rock stable and people that consider a window manager a production ready app demand it...  gnome is not stable long term.
[23:14:05] <gdamore> wotdu: because I've read the friggin' source code.
[23:14:06] <wotdu> delewis, isnt motif dead and deeply burried ? :)
[23:14:22] <delewis> wotdu: to you perhaps, but to even 1% of Solaris customers, no.
[23:14:34] <gdamore> wotdu: no.  there are very large solaris installations that use motif/cde heavily
[23:14:54] <rydis> Motif is one of the few toolkits that doesn't suck, in that it actually works with X resources, editres, and all that.
[23:15:03] <wotdu> if solaris is so reliable and all
[23:15:18] <wotdu> why isn't used in really mission-critical stuff
[23:15:21] <gdamore> on sparc hardware it is.  x86 is getting there.
[23:15:27] <delewis> wotdu: it is
[23:15:30] <gdamore> and it _is_ used in mission critical stuff.  a lot.
[23:15:36] <delewis> far more than Linux, in most regards.
[23:15:44] <gdamore> i work with an army installation that uses Solaris to host air traffic control software.
[23:15:46] <wotdu> like spaceships
[23:15:48] <delewis> AIX and Solaris are still the two largest contenders for "mission-critical"
[23:15:49] <wotdu> an things
[23:15:56] <dwc-> the navy's s-3 flight sims for sure, use motif
[23:15:56] <delewis> wotdu: those require other operating systems
[23:15:57] <wotdu> *and
[23:16:00] <delewis> specifically, embedded ones
[23:16:09] <delewis> and real-time ones (though, Solaris has a real-time scheduler)
[23:16:14] <Kush-> the bad part is, i still see bug fixes for CDE.. must be costing to maintain it
[23:16:15] <jamesd> you think people buy  platinum level support contracts and have 100's of sun techs on site 24/7 ... aren't running mission critical apps
[23:16:18] <delewis> even Linux wouldn't be suitable for that.
[23:16:34] <gdamore> i've sysadmined sites were downtime was $1M/minute.
[23:16:45] <wotdu> gdamore, outch
[23:16:58] <gdamore> that's pretty mission critical.  they ran Solaris.  not unix, not HPUX.  Solaris.
[23:17:11] <dwc-> wait, so solaris isn't unix?
[23:17:14] <delewis> gdamore: and you certainly wouldn't purchase a $1k whitebox running a non-warranted operating system, like Linux, to run it.
[23:17:26] <richlowe> dwc-: it's pretty clear what the statement was intended to mean.
[23:17:32] <wotdu> poor me, i'm only 16, not really into bussiness
[23:17:43] <quasi> gdamore: I've done the same - *lots* of solaris
[23:17:55] <gdamore> hile_: if you are trying to reach me, i can't get to it, because i have no mouse ability right now.
[23:18:05] <hile_> ah well
[23:18:16] <Kush-> dwc: Unix is generally termed for HPUX/AIX/Solaris ...
[23:18:17] <hile_> wasn't important anyhow
[23:18:24] <gdamore> i never thought i was unique in the regard. just providing a concrete example
[23:19:04] <jamesd> wotdu, i have a friend she used to work for sun she made over $120k a year.. she was paid to sit on her a$$ at a client site to fix problems when they arose,  7.5 hours of an 8 hour shift was spent chatting and listening to mp3s and surfing... and the company was extremely happy to pay her salary...
[23:19:25] <gdamore> that's why you think great things about linux.  i thought linux was cool too, back in 1991.  i since started working on real kernel work, and have decided linux sucks. :-)
[23:19:28] <delewis> wotdu: regardless, you should ask (gdamore, for example) how stringent and slow the process is to commit code to OpenSolaris, specifically, the kernel.
[23:19:38] <twincest> i'd hate that job, the 0.5 hours you had to do something sound painful
[23:19:43] <richlowe> I'm entirely OS agnostic.
[23:19:46] <wotdu> gdamore, im not standing up for linux
[23:19:48] <richlowe> I guess that's rare these days.
[23:20:07] * dwc- trades his windows xp box for richlowe's solaris box
[23:20:09] <wotdu> but i do stand up for free software
[23:20:11] <gdamore> heh.  the one time you have an unplanned outage, folks come looking for your head. :-)
[23:20:14] <hile_> james: she made that much? i thought it was between 80-90, but I guess with the amount of OT she got...
[23:20:52] <jamesd> twincest, the .5 hours was usually doing something "proactive" to meet her bosses requiements that she did something productive... like  clean an air filter in a system, or make sure the networking cables are plugged in.
[23:20:54] <hile_> gdamore: worse - the one time you have a planned outage and the backup system decides to kernel panic and autoboot...
[23:21:06] <gdamore> haha.
[23:21:19] <gdamore> (i say that now...)
[23:21:34] <hile_> when somebody forgot to reset diag-level back to min on the 6900 that just decided to autoboot ;)
[23:21:43] <delewis> hile_: ugh.
[23:21:52] <delewis> talk about a long, irritable coffee-break.
[23:21:53] <wotdu> i think ill be downloading opensolaris and hope the configuration files aren't too obscure
[23:21:54] <gdamore> worse: try it on an E10K with full up domains.
[23:22:07] <Error_404> wotdu: docs.sun.com
[23:22:17] <Error_404> wotdu: it's fantastic resource
[23:22:20] <wotdu> Error_404, good configuration files don't need docs
[23:22:23] <delewis> wotdu: you'll be pleased to know that Solaris has some of *the* best documentation.
[23:22:26] <gdamore> if you can figure out linux perversity, you can figure out solaris. :-)
[23:22:38] <quasi> gdamore: we had a sun tech setting it to max on a 3800 - that took 18 hrs
[23:22:39] <delewis> docs.sun.com is a god-send.
[23:22:41] <wotdu> gdamore, what?
[23:22:53] *** nwf has joined #opensolaris
[23:22:57] <wotdu> the distros i use have very simple and clean configuration files
[23:23:05] <wotdu> and by very i mean very
[23:23:08] <wotdu> ^^
[23:23:11] <jamesd> a full diagnostic report on a  medium size sun box.
[23:23:15] <gdamore> try doing something "unusual" then.
[23:23:19] <delewis> wotdu: that's pretty naive.
[23:23:33] <wotdu> delewis, is it? tell me
[23:23:38] <delewis> I've told you.
[23:23:52] <Error_404> heh
[23:24:32] <hile_> wotdu: the sun box I had until June was only about 1/3 populated and it'd take a good half an hour to do a max POST
[23:24:52] <wotdu> sorry?
[23:24:53] <wotdu> POST ?
[23:25:01] <Stric> poweron self-test
[23:25:02] <twincest> power on self test
[23:25:11] <wotdu> yeah i know
[23:25:13] <twincest> "checking RAM" on i386, a bit more on sparc :)
[23:25:18] * delewis is thankful max-diag on workstations isn't as bad
[23:25:29] <wotdu> i though it was something different here
[23:25:38] <richlowe> same principle, a lot more done.
[23:25:39] <wotdu> since 30min is a huge time
[23:25:49] <wotdu> long ^^
[23:25:52] <delewis> on SPARC (at least for simple workstations) at the very least, it is checking all the caches on each of the processors, memory, etc.
[23:25:52] <richlowe> much more done, and in some cases much, much, much more to check.
[23:25:55] <gdamore> it was really, really painful on an E10K.  JTAG driven by an RPC implemented in Tcl.
[23:25:58] <wotdu> sorry, english is not my native lang :)
[23:26:18] <twincest> TCL?  that's kind of funny
[23:26:24] <gdamore> a full up E10K domain could take very many hours to POST.  so long that at full level it would break badly
[23:26:34] <wotdu> i hope opensolaris isnt a 4GB dvd or something
[23:26:45] <delewis> wotdu: Solaris Express is
[23:26:48] <gdamore> it is funny, and painful.  a tcl script execution for _each_ JTAG bit shift.  it took forever.
[23:26:50] <delewis> OpenSolaris is not an operating system.
[23:27:06] <delewis> it's basically a code dump of various bits (some optional, some required) to create an OpenSolaris-based distribution.
[23:27:07] <gdamore> this is why the E10K running a console in JTAG mode seems to be operating at like 100 baud.
[23:27:20] <wotdu> delewis, hmm
[23:27:34] <wotdu> is solaris a complete operating system?
[23:27:41] <delewis> you wouldn't call the Linux kernel (and possibly, let's say glibc) an operating system anymore than you would OpenSolaris.
[23:27:43] <wotdu> i mean, kernel + userland tools
[23:27:44] *** jsubl2 has joined #opensolaris
[23:27:50] <delewis> wotdu: absolutely.
[23:27:57] <richlowe> stevel: it's today's drop likely to happen?
[23:28:03] * delewis hopes so
[23:28:20] <wotdu> delewis, so where do i get the base system
[23:28:28] <delewis> I've been deciding whether I should wait for SXCR b51 or bother with SXCR b50 when a release is most likely emminent.
[23:29:13] <delewis> wotdu: https://sdlc2b.sun.com:443/ECom/EComActionServlet;jsessionid=99FE453DFD93EDE213DE3DD75BF24F17
[23:29:14] <gdamore> a release is almost always imminent.  i finally updated to b50
[23:29:36] <delewis> I'm still running b48.
[23:29:37] <wotdu> delewis, joke right? registration?
[23:29:42] <delewis> wotdu: no?
[23:29:58] <wotdu> To continue, please log in with your Sun Online Account  (Download Center,
[23:29:58] <wotdu> MySun, SunSolve, etc.). If you don't have an account, please Register.
[23:29:59] <Error_404> speaking of jsp, i managed to fit glassfish in 64M of ram
[23:30:00] <delewis> register. grab the DVD chunks. burn. boot. install.
[23:30:01] <Error_404> :D
[23:30:04] <delewis> wotdu: so register
[23:30:21] * wotdu is starting to hate sun :P
[23:30:34] <wotdu> register to grab a damn cd image ?
[23:30:35] <Error_404> wotdu: you could always download schillix/nexenta
[23:30:38] <delewis> wotdu: they're asking you to simply register to receive an enteprise-grade operating system for free.
[23:30:42] <delewis> big. fucking. deal.
[23:30:57] <wotdu> it's just VERY unsual
[23:31:11] <gdamore> not really.  to download _anything_ you almost always have to register.
[23:31:14] <delewis> wotdu: unusual in the aspect that you normally wouldn't get such things for free? absolutely.
[23:31:21] <richlowe> it's a click-through thing.
[23:31:23] <Error_404> delewis: but the CIA & Sun microsystems'll team up to rat him out to the RIAA for pirating free software
[23:31:30] <wotdu> you could count by hand the free software that requires registration to download
[23:31:31] <delewis> Error_404: oh, noes.
[23:31:44] <Error_404> heh
[23:31:50] <richlowe> if you want to avoid registering, download one of the non-sun distributions.
[23:31:51] <twincest> wotdu: what definition of free are you using?
[23:31:59] <wotdu> twincest, free as in freedom ofc
[23:31:59] <richlowe> nexenta ,or belenix, or schillix, or whatever.
[23:32:12] <twincest> well solaris express includes software which is not "free software"
[23:32:23] * gdamore thinks wotdu whines too much. :-)
[23:32:24] <Error_404> like CDE
[23:32:24] <twincest> (CDE for example)
[23:32:26] <delewis> most notably the CDE bits, SPARC graphics drivers, etc.
[23:32:27] <delewis> :-)
[23:32:42] *** coffman has joined #opensolaris
[23:32:47] * delewis throws wotdu out of the bathtub
[23:32:54] <gdamore> heh.
[23:32:58] * wotdu says bye to opensolaris , and advices them to remove the "open" prefix
[23:33:01] <twincest> (and you have to do a lot more than register to get CDE for linux, i.e. pay money :)
[23:33:07] <twincest> wotdu: ...?
[23:33:13] <delewis> wotdu: you mis-understand OpenSolaris
[23:33:13] <wotdu> twincest, who gives a shit about CDE?
[23:33:16] <wotdu> i though it was dead
[23:33:17] <delewis> as I sai, it is not an operating system.
[23:33:21] <gdamore> wotdu: if you want open software, use nexenta.  you get the crummy linux userland with an enterprise grade open kernel.
[23:33:22] <twincest> wotdu: if you don't want CDE, don't use a distribution that includes it
[23:33:30] <delewis> there are many, many OpenSolaris-based distributions that you are free to download
[23:33:35] <delewis> that do not require registration
[23:33:42] <wotdu> but you just told me
[23:33:45] <delewis> Solaris, itself, though, is one of the more notable ones.
[23:33:50] <twincest> told you what?
[23:33:53] <wotdu> that opensolaris was kernel + userland
[23:33:53] <delewis> and still necessary to build ON, I think.
[23:33:56] <wotdu> not only the kernel
[23:34:04] <delewis> wotdu: and what have I said to contradict that?
[23:34:15] <delewis> kernel+userland does not make it a usable, bootable operating system.
[23:34:20] <gdamore> opensolaris includes it all, but not all distros use it.  nexenta is notable in that it tries to use a GNU userland
[23:34:43] <wotdu> does solaris have sun-made base tools?
[23:34:48] <wotdu> like ls, cp, etc etc
[23:34:51] <wotdu> libc
[23:34:56] <gdamore> yes, and those are in opensolaris.
[23:34:59] <delewis> the "sun-made" base tools fall into the OpenSolaris ON bits.
[23:35:15] <wotdu> on bits?
[23:35:26] <delewis> wotdu: you might considering reading a bit about OpenSolaris
[23:35:27] <gdamore> think of ON as the "core" of the OS
[23:35:28] <twincest> isn't there a faq about this somewhere?
[23:35:33] <delewis> twincest: there is
[23:35:37] <delewis> but *of course*
[23:35:40] <gdamore> there has to be something about it on opensolaris.org
[23:35:41] <delewis> he hasn't bothered to read anything.
[23:35:51] <delewis> the Developer's Reference has a glossary that defines all of this
[23:35:55] <delewis> (and much more)
[23:36:02] <Error_404> delewis: reading things goes against the free software philosophy
[23:36:08] <gdamore> heh.
[23:36:14] <wotdu> i'm scared by all this enterprise company crap
[23:36:17] <wotdu> XD
[23:36:18] <Error_404> documentation is for dirty proprietary operating systems
[23:36:28] <delewis> wotdu: North Korea isn't too far away.
[23:36:55] <wotdu> delewis, 16-yeard olds dont read the newspapers :P
[23:36:58] <wotdu> what about it?
[23:37:03] <delewis> wotdu: *sigh*
[23:37:06] *** sahafeez has joined #opensolaris
[23:37:07] <delewis> I did when I was 16
[23:37:14] <delewis> (which actually wasn't too long ago)
[23:37:17] <wotdu> delewis, not general-purpose newspapers
[23:37:28] <gdamore> wotdu: my suggestion, please don't take this wrong, is to read up a bit.  don't act like a snotty 16-year old, and you will get taken a _lot_ more seriously
[23:37:44] <coffman> yeah, i came from windows/linux and i was scared, people told me that there was a documentation from the _vendor_ that was worth to read!
[23:37:59] <wotdu> coffman, windows.... all explained :P
[23:38:04] <gdamore> haha.
[23:38:10] <Gman> and you can be assured, almost everyone inside sun wants the registration to go away too
[23:38:13] <gdamore> this is true of linux, too.
[23:38:29] <Error_404> i read general purpose newspapers when i was 16
[23:38:36] <gdamore> i just figured the marketdroids created the registration page.  i have a spam filter, so i don't care too much.
[23:39:03] <gdamore> (have 16 year olds forgotten how to read?  what are we teaching our kids these days?)
[23:39:06] <coffman> hack, its worser on linux then on windows...
[23:39:09] <wotdu> gdamore, i do read
[23:39:25] <Gman> gdamore, yeah, it's mostly about marketing and legal
[23:39:26] <wotdu> but i don't give a shit about news like: "2 cars crashed and 3 people died"
[23:39:45] <gdamore> what about world news?  do you have any idea what is going on in the world today?
[23:40:00] * twincest doesn't read news and is a bit older than 16
[23:40:00] <gdamore> slashdot.org is _not_ a news site. :-)
[23:40:01] <wotdu> gdamore, a vague one
[23:40:08] <Error_404> i read (and still do) papers mostly for world events & politics
[23:40:08] <wotdu> gdamore, slashdot rocks :)
[23:40:16] <dwc-> slashdot sucks
[23:40:17] <gdamore> well, that says it all there.
[23:40:20] <Error_404> slashdot was abandoned ages ago
[23:40:31] <wotdu> was it?
[23:40:31] <dwc-> slashdot was decent about oh.... 6...7... years ago
[23:40:33] <gdamore> i still watch slashdot, but yeah, most of what is posted there is absolute crap
[23:40:38] <Error_404> everyone got tired of half-retarded 16 year olds & left
[23:40:42] <gdamore> hahaha
[23:40:44] <gdamore> rotfl
[23:40:47] *** sahafeez has quit IRC
[23:40:52] * wotdu headshots Error_404
[23:40:53] <stevel> richlowe: pushing now
[23:40:56] *** stevel changes topic to "ON Hg Mirror Live, please test : http://tinyurl.com/wgrd8 | SXCR: 50, ON build: 50, ON nightly: 20061023"
[23:41:07] <Gman> sweet
[23:41:21] <richlowe> stevel: cool.
[23:41:33] <wotdu> relogging to get into a group ^^
[23:41:33] <wotdu> brb
[23:41:37] <dwc-> hey steve, you should add a redirect servlet to www.o.org so you don't have to use tinyurl
[23:41:42] *** wotdu has quit IRC
[23:41:51] <gdamore> mostly, i think slashdot is full of kids that want to figure out how to pirate and distro their mp3s and dvds.  real tech stuff has moved elsewhere.
[23:42:05] <gdamore> (good riddance?)
[23:42:15] <dwc-> I dunno, the stories suck too
[23:42:20] <stevel> dwc-: that's not a bad idea
[23:42:23] <dwc-> so it's not just the comments
[23:42:31] <delewis> gdamore: how many times do you see articles regarding "<foo> circumvented" that show up on Slashdot?
[23:42:34] <dwc-> comment(er)s
[23:42:34] <stevel> i'll add it to my ridiculously long list of stuff to do
[23:42:36] <Gman> gdamore, to be honest, if people were doing that on solaris, i would be very very proud :)
[23:42:36] <gdamore> the same kids doing the poisting. :-)
[23:42:45] *** jsubl2 has quit IRC
[23:42:49] <twincest> delewis: isn't that valid news?
[23:43:08] <delewis> twincest: sure it is, but then that makes the press more than "ZFS released!".. I begin to get worried.
[23:43:17] *** wotdu has joined #opensolaris
[23:43:24] <delewis> as was the case with Slashdot -- it never even made the front-page.
[23:43:32] <delewis> but "Linux runs on Xbox!" did.
[23:43:49] *** Tpenta has joined #opensolaris
[23:43:53] <gdamore> i'm am afraid for our future.  one day these kids will actually enter the business world.  hopefully they will get some actual education in college first, though
[23:44:05] <wotdu> gdamore, im in portugal
[23:44:06] <delewis> gdamore: hey, that leaves more of a job market open for me :-)
[23:44:20] <wotdu> economical situation here is beyong control
[23:44:23] <wotdu> *beyond
[23:44:30] <Error_404> hmm... googlebot is my blog's biggest fan
[23:44:36] <wotdu> next year im going to the UK to the university
[23:44:39] <wotdu> and hopefully stay there
[23:44:48] * Gman often meets kids that are 10+ years younger than him and more technically competant
[23:45:03] <twincest> well technical competance isn't all you need in the real world
[23:45:06] <gdamore> there are a lot of technically competent kids out there, no doubt.
[23:45:08] <wotdu> Gman, what old are you? 50? :P
[23:45:09] <gdamore> twincest++
[23:45:17] <Gman> wotdu, 28
[23:45:23] <wotdu> i see
[23:45:26] <gdamore> pipsqueak youngster. ;-)
[23:45:27] <wotdu> gdamore, is it not?
[23:45:32] <delewis> gdamore: from my perspective (and keep in mind, that I'm 18) technical competence among youth has gone down the drain.
[23:45:45] <wotdu> so scamming and tricking and crap is also important i guess?
[23:46:04] <hile_> delewis: you're older than 18, I thought
[23:46:04] <wotdu> delewis, remember be about it...
[23:46:08] <gdamore> there's a reason they're called "script kiddies", and not "script parents" :-)
[23:46:12] <delewis> sure they know the "how to" behind things, but as far as theory is concerned - "Why does it work this way?", it's a complete non-concern for them.
[23:46:16] <hile_> i thought you were older than that, Gman
[23:46:22] <hile_> you're only a year older than I am
[23:46:23] <wotdu> i have to teach C in 2 months to a guy with no previous programming experience
[23:46:28] <dwc-> delewis: I could agree with your perspective
[23:46:42] <Gman> hile_, heh, you're not the only one who thought that
[23:46:47] <dwc-> just based on the quality of the students who have come through a couple of CS-related undergrad assocs at cal
[23:46:56] <Gman> plocher thought i had a grey beard when he met me
[23:47:03] <delewis> hile_: nope, why do you think I had such difficulty at the hospital?
[23:47:10] <gdamore> yes, a lot of kids are more concerned with high level GUI stuff, etc.  folks who know deep down bit fiddling are hard to find.  but i think that has always been true
[23:47:15] <Error_404> delewis: i disagree... there's the same amount of technical competence among youth, but it's become easier to fake for the rest of the dumb kids
[23:47:15] <delewis> (everyone thought I was out for their job, basically)
[23:47:38] <Error_404> just learn some basic linux usage & you're set to call yourself a "geek" these days
[23:47:41] <delewis> Error_404: well, also keep in mind my perspective comes from the South, so things may be different elsewhere.
[23:47:44] <gdamore> what is new is that now there is a place for people who don't want to know driver development, etc.  the former basic programmers can know get "real jobs" doing perl, java, python, etc.
[23:48:05] <wotdu> gdamore, that's not me :)
[23:48:07] <gdamore> Error_404++
[23:48:17] <wotdu> i hate high level programming
[23:48:19] <delewis> gdamore: indeed, hence why you see the distinction between "computer engineering" and "computer science/software development," nowadays
[23:48:36] *** ChanServ sets mode: +o Tpenta
[23:48:40] <gdamore> right.  i have a CS degree, but I am probably much closer to a CE in what I really do.
[23:48:43] <Error_404> delewis: my old Uni had "information technology" and "computer science"
[23:48:58] <gdamore> (but my CS degree is a BS, not BA)
[23:49:08] <Error_404> the former was using a computer, the latter was CS in the old derived from math style
[23:49:15] <delewis> well, all in all, I'm just decided to major in Mathematics, and stay away from the Computer Science/Information Technology programs.
[23:49:48] <delewis> besides, Computer Science is mostly Mathematics, anyway (at least in proper programs, I would think)
[23:50:01] <wotdu> delewis, Physics are important too
[23:50:21] <wotdu> i personally want to know a lot more about hardware
[23:50:36] <gdamore> i dunno.  there is a lot of low level stuff that i'm not sure is really math.  e.g. concurrency, interrupt handling, DMA, etc.  how do you cover that in "math"?
[23:51:01] <delewis> wotdu: you'll find that really a very small subset of Physics is actually used for hardware development, namely anything related to electromagnetism and solid-state physics.
[23:51:02] *** cilkay has joined #opensolaris
[23:51:33] <delewis> does a hardware developer need to know about the formal definition of a Lagrangian or Hamiltonian? (I doubt it).
[23:51:43] <gdamore> heh.  the only time i've ever, ever seen calculus in my entire career involved network theory.  one time.  ever.  never since.
[23:51:52] <wotdu> delewis, he doesnt, its mere curiosity in my case
[23:51:53] <dwc-> mostly mathematics? the algorithms/theory bits would be math related
[23:52:15] <dwc-> programming languages do to some extent ...
[23:52:16] <gdamore> maybe.  matrix math is quite useful.  and i suppose FSM qualifies as "math". :-)
[23:52:21] <delewis> gdamore: well, I was just referring to "computer science" as in "modern computer science", a.k.a "software development"
[23:52:24] <dwc-> not much of say... operating systems would be though
[23:52:33] <gdamore> s/FSM/finite state automata/
[23:52:37] <delewis> most proper CS programs (I would think) would place some focus on Discrete Mathematics and the theory of computation
[23:53:00] * gdamore wonders if maybe he was robbed of a proper education. :-)
[23:53:05] <twincest> delewis: i'm not sure how you define proper programs, but the project i'm working on at the moment (http proxy) doesn't involve much mathematics.  the only vague complex part is a radix tree
[23:53:06] <delewis> though, that's probably not the case
[23:53:23] <Tpenta> I managed to get a B.CompSC graduating in Maths colours at www.newcastle.edu.au
[23:53:33] <delewis> as most Java programmers that come out of most CS programs aren't even the slightest bit concerned with whatever algorithm they're using or rather or not "how expensive it is"
[23:53:50] <gdamore> delewis++
[23:54:00] <Tpenta> that's one of the failings of MHz covering a multitude of sins
[23:54:11] <dwc-> but... I can squeeze all sorts of performance out of a really really optimized shell sort!
[23:54:11] <Tpenta> but o(N^3) is still o(N^3) !!!
[23:54:21] <gdamore> MHz don't help with O(n^3) when n gets big enough.
[23:54:33] <dwc-> good thing intel's there to throw more MHz at it
[23:54:56] <Tpenta> didn't intel just admit (in the last few months) that they are running out of MHz?
[23:55:03] <delewis> Tpenta: yes
[23:55:06] <delewis> they're going to focus on adding more cores
[23:55:07] <gdamore> i still think a lot of my biggest contributions (especially to netbsd) are negative line counts and negative memory usage. :-)
[23:55:10] <dwc-> yep.  Better hope MS optimizes some of their code now
[23:55:13] <hile_> Tpenta: about 10 years ago now, when my alma mater's labs still had SS4s and SS5s in them, a distributed algorithms class had a sort 36M integers
[23:55:30] <hile_> one guy's was taking hours and hours to run when everybody else's took a few minutes
[23:55:47] <gdamore> hope he flunked. :-)
[23:55:50] <Tpenta> insertion sort?
[23:56:02] <dwc-> probably wasn't using qsort()
[23:56:13] <hile_> somebody asked this dumbass what he was using on each processor  for its sixth of the data; his response: "bubble sort"
[23:56:22] <gdamore> hahaha
[23:56:26] <delewis> ah, I was just going to say "bubblesort"
[23:56:30] <delewis> :-)
[23:56:33] <gdamore> me too, actually.
[23:57:22] <Tpenta> A colleague once wrote a perl script to do some work with mail aliasing and mail drop generation on our mail server which was at the time an ss5. he had a O(n^2) loop in there. It was supposed to run every 15 minutes. The problem was that it took 20 minutes to run
[23:57:26] <gdamore> so i suppose being able to understand O(n) and logarithms/exponents, and also work in alternate number bases (hex, octal) is useful.  that counts as math.
[23:57:34] <Tpenta> replaced the loop with hash lookups and made it O(N). ran in 20 seconds
[23:58:07] <gdamore> time for me to go now.  back tonight, maybe.  if not, tomorrow.
[23:58:08] <dwc-> why wouldn't you use hashes in the first place ... perl makes hashes so much easier
[23:58:24] <dwc-> sysadmin by training?
[23:58:27] <Tpenta> he didnt understand them (this was mid 90's)
[23:59:13] *** gdamore has quit IRC





top