[00:01:45] <bengtf> timeless: I have seen stupid build scripts that just adds stuff to path and/or environment which in the end when trying to execute something I would get 2 different errors like too long name and environement corrupted ..., back in the old days when path could be 1k long only ;)
[00:01:47] <gnu2it2> what is the easiest way to get a recient sol 10 install fully patched/updated? sorry, been a while for me
[00:03:13] *** lolownia has joined #opensolaris
[00:03:46] *** Trident has quit IRC
[00:04:05] <gtc> Hi all.  I'm under DHCP and the nameservers stopped working.  I can SSH in and get out using the direct IP.  How can I renew the DNS info?
[00:04:07] <ShadowHntr> gnu2it2: /usr/sbin/smpatch
[00:04:08] <ShadowHntr> :)
[00:04:23] <gnu2it2> thanks
[00:04:24] <lolownia> Hello, anybody here tried running OpenSolaris on Acer notebooks?
[00:05:08] <bengtf> will do so on my new one 5112 I think its names
[00:06:25] *** astinus has quit IRC
[00:07:20] <lolownia> I'm thinking aobut installing OSExpress or Nexe ta on aspire 1520, wonder about hw support
[00:07:42] *** Kush- has joined #opensolaris
[00:07:57] <bengtf> isnt that a slow one ?
[00:08:06] <lolownia> nvidia driver works i guess, at least if nvidia patches its holes... sound? no idea. wireless might work via ndis. acpi cpufreq with cool'n'quiet also is a big question
[00:08:18] <lolownia> slow one?
[00:08:30] <lolownia> ahh..its 1525 i quess. 1520 series..
[00:09:13] <bengtf> upgrade ;), the 5112 seems perfect ;)
[00:09:34] <lolownia> :P
[00:09:51] <lolownia> If I could upgrade hw every time the software doesn't want to speak to it :D
[00:10:27] <bengtf> the problem I had with old computer is that OpenS didnt have the gigabit card anyway ...
[00:10:29] <lolownia> what wifi card does 5112 have?
[00:10:47] <bengtf> dont know yet its not delivered yet
[00:11:18] <alanc> gdamore: your bug id is 6484159 (kb8042 should provide mappings for common keys on Tadpole hardware)
[00:11:32] <lolownia> bengtf, ouch.
[00:11:45] *** eboutilier has quit IRC
[00:11:54] <lolownia> anyway, i can see from os.com site that os community is very active
[00:12:13] <lolownia> and it seems os is making big big progress on bogus platforms
[00:12:19] <bengtf> I run os b43 on my x64, and the only thing not working is sound card
[00:12:44] <bengtf> but then I use usb headset , so that works nicely
[00:12:45] <gtc> Can I force OpenSolaris to renew its DHCP information?
[00:13:04] *** axisys has quit IRC
[00:13:18] <gtc> DNS stopped working.
[00:13:23] <trygvis> gtc: ifconfig stop; ifconfig dhcp
[00:13:31] <trygvis> ifconfig <if> ..
[00:13:42] <gtc> trygvis: Thanks -- I'll give it a go.
[00:14:09] <lolownia> ok. now a quite subjective question
[00:14:51] <gtc> trygvis: so, ifconfig e1000g0 stop; ifconfig e1000g0 dhcp, if e1000g0 is my interface?
[00:14:58] <bengtf> you may get an objective answer ;)
[00:15:35] <gdamore> hey all.
[00:15:51] <gdamore> alan: thanks.
[00:16:02] <lolownia> I have a bit of abckground in linux/unix. I'm using it as primary work station since 6 or 7 eayrs.. first Slackware and then switched to FreeBSD 4.5 tracking FreeBSD-STABLE from the on. Do You think I should go for OSExpress or start with (easier?) Nexenta to learn the solaris stuff?
[00:16:20] <Gman> http://www.opensolaris.org/os/
[00:16:23] <ShadowHntr> i'd suggest regular Solaris.
[00:16:23] <ShadowHntr> :)
[00:16:24] <Gman> oooop, new page!
[00:16:48] <lolownia> is it difficult to get? In general I have a preference for pure opensolaris, but the installation instruction is not quite welcoming
[00:17:02] <gtc> lolownia: I've tried both; IMHO straight solaris is the way to go.
[00:17:33] <lolownia> ok, great
[00:17:40] <bengtf> opensolaris that is ;)
[00:17:49] <ShadowHntr> lolownia: it's pretty simple. if you can handle either text or graphical install, it's not hard.
[00:17:54] <lolownia> Do Yu follow osexpress by doing builds from sorces or binary upgrades?
[00:18:19] <lolownia> I just want to install it on a partition and not hose the whole hard drive :)
[00:18:35] <lolownia> after that install some power management not to burn my over heating notebook
[00:18:38] <lolownia> after that i'm safe.
[00:18:41] <lolownia> :P
[00:18:42] <ShadowHntr> lolownia: it's already made for you - just like a beta of an os.
[00:19:22] <Gman> holy cow, that's sch without a tie! :)
[00:19:32] <lolownia> ok :)
[00:20:20] <ShadowHntr> http://www.sun.com/software/solaris/solaris-express/get.jsp
[00:22:08] <alanc> hmm...wonder if Sun could get product placement in the upcoming transformers movie...  http://blogs.sun.com/ThinGuy/date/20061017
[00:22:33] *** besonen_mobile has quit IRC
[00:22:36] <Gman> man, they're making a movie? ROCK! ;)
[00:22:51] <elektronkind> Boximus Prime!
[00:23:00] <gtc> trygvis: sudo ifconfig e1000g0 stop gives ifconfig: stop: bad address
[00:23:03] <gtc> trygvis: ???
[00:23:15] <elektronkind> heh "stop"
[00:23:32] <elektronkind> are you trying to down a interface?
[00:23:35] <richlowe> 'down' perhaps?
[00:23:37] <richlowe> 'unplumb' maybe?
[00:23:51] <ShadowHntr> ifdown doesn't work?
[00:23:55] <sommerfeld> down or unplumb
[00:24:00] <ShadowHntr> ahhhh.
[00:24:01] <gtc> elektronkind: My DNS isn't working under DHCP
[00:24:11] <twincest> ah, the well known solaris ifdown script :)
[00:24:13] <gtc> elektronkind: Just quit for some reason
[00:24:33] <elektronkind> gtc: because it's actually *my* DNS and I took it away from you
[00:24:46] <gtc> gtc: Unhand my DNS!
[00:24:57] <gtc> gtc: (or unplumb it)
[00:25:12] <elektronkind> gtc: did the contents of /etc/resolv.conf go blank when you noticed DNS failing?
[00:25:20] *** _william_ has quit IRC
[00:25:21] <gtc> elektronkind: yes
[00:25:30] <elektronkind> that's crazy
[00:26:12] <gtc> elektronkind: Wait, I lied, it's there
[00:26:25] <ShadowHntr> check /etc/nsswitch.conf
[00:26:28] <gtc> elektronkind: it's the dhcp stuff that's blank
[00:26:41] <gtc> elektronkind: any idea on how to figure out where the fault is?
[00:27:12] <elektronkind> that's even crazier. is your dns server on the same network? if it isn't perhaps your dhcp-origined default route disappeared.
[00:27:49] <elektronkind> i'd try a 'ifconfig e1000g0 dhcp'
[00:28:02] <gtc> elektronkind: My DNS is my local router.  I'm SSHd in to it right now (and typing from it here)
[00:28:20] <gtc> ifconfig: e1000g0: interface not in appropriate state for command
[00:28:26] <elektronkind> maybe there's some dhcp-specific options in ifconfig's man page you can use to get a better sense of what is/did happened
[00:29:17] <elektronkind> make sure dhcpagent is running, too.
[00:29:33] <elektronkind> ifconfig controls that daemon, which is the actual dhcp client
[00:30:15] <elektronkind> meanwhile, I need to go home. ttfn
[00:30:40] <gtc> elektronkind: thanks i'll keep hacking at it
[00:32:10] <sommerfeld> "ifconfig e1000g0 dhcp status"
[00:36:34] *** phretor has left #opensolaris
[00:37:43] *** gnu2it2 has quit IRC
[00:38:16] *** gisburn has joined #opensolaris
[00:38:23] <gisburn> ugh
[00:38:34] <gisburn> What happened to the list of ops ?
[00:38:36] *** gisburn is now known as nrubsig
[00:38:45] *** ChanServ sets mode: +o nrubsig
[00:38:54] <movement> somebody have today's onnv-notify email they could bounce to me, please?
[00:39:00] <richlowe> Yes.
[00:39:04] <richlowe> movement: Ptarget.c, btw.
[00:39:13] <nrubsig> !seen alanc
[00:39:14] <Drone> alanc is currently online in #opensolaris and last spoke on Thu 19 Oct 2006 22:23 GMT, saying 'hmm...wonder if Sun could get product placement in the upcoming transformers movie...  http://blogs.sun.com/ThinGuy/date/20061017'.
[00:39:19] *** besonen_mobile has joined #opensolaris
[00:39:27] <nrubsig> alanc: ping!
[00:39:32] <alanc> pong!
[00:39:49] <movement> richlowe: levon at movementarian dot org please
[00:39:52] <nrubsig> alanc: Did the internal triage queue list my "lint" bug yet ?
[00:39:57] <alanc> yes
[00:40:03] <nrubsig> alanc: CR' ?
[00:40:10] <alanc> I even fixed the typo in the title 8-)
[00:40:17] <nrubsig> s/'/#/
[00:40:27] <nrubsig> alanc: thanks! :-)
[00:40:46] <richlowe> movement: done.
[00:41:10] <alanc> 6484164: Sun Studio 11 "lint" does not recognize cc/CC's "-xstrconst" option
[00:41:48] *** gtc_ has joined #opensolaris
[00:42:07] <movement> richlowe: hmm, it has "from: ..." instead of "From:...", weird
[00:43:04] <gtc_> sommerfeld: Sorry -- killed the interace and knocked myself off the machine.
[00:43:16] <nrubsig> alanc: Do you still have an interest in my "solaris horror bug list" ?
[00:43:48] <lolownia> :>>>>
[00:43:55] <alanc> anything on it I need to fix?  or is it just a long list of things for you to feed to bugs.opensolaris.org?
[00:43:59] <richlowe> movement: aye.
[00:44:49] <alanc> you know what - I just remembered I have an old copy of it saved
[00:44:54] <nrubsig> alanc: mhhh, I don't know... I am filling the list right now...
[00:45:03] <nrubsig> alanc: It is a new list: http://svn.genunix.org/repos/on/branches/ksh93/gisburn/todo_list_gisburn.txt
[00:45:13] <richlowe> movement: what did that break?
[00:46:08] <alanc> the file timestamp on the one I have says 2004, but the last entry in it is 20010403
[00:46:54] <richlowe> nrubsig: gcc *does* include g77
[00:46:54] <nrubsig> alanc: I have at least four of these lists floating around somewhere.
[00:46:57] <richlowe> not ada though, I don't think
[00:47:04] <richlowe> > which g77
[00:47:04] <richlowe> /usr/sfw/bin/g77
[00:47:25] <nrubsig> richlowe: last time jenny checked g77 was missing in solaris and I got an angry mail.
[00:47:26] <alanc> the one I have starts out as "Future of the Common Desktop Environment in Solaris: 10th May 2000"
[00:47:34] <alanc> g77 is in Nevada but not S10
[00:47:50] <richlowe> nrubsig: it was added a few months back.
[00:48:24] <alanc> 6223618: need to ship gnu autoconf, f77, and gdb in Solaris
[00:48:33] <alanc> Integrated in Build: snv_22
[00:49:56] <alanc> good, I don't see anything I can fix on there 8-)
[00:50:07] * nrubsig wonders whether fortran code would be allowed in OS/Net... =:-)
[00:51:20] <alanc> you want me to mail you this old one so you can add all those to your bugs-to-file list too?   it's mostly CDE stuff though
[00:51:37] <alanc> dtksh: # BUG: Is not POSIX/Unix98 conformant (no examples requires - simply read the docs: dtksh is a "hacked" ksh93 shell...)
[00:51:38] <alanc> heh
[00:52:11] *** gtc has quit IRC
[00:52:57] <alanc> I think this one can be ignored now: BUG: solregis doesn't work properly if HotJAVA hasn't been installed on the machine.
[00:53:10] *** delewis has joined #opensolaris
[00:53:44] <alanc> and this will only have taken 7 years to be fixed: RFE: Support for multiple X servers at the same time. - Linux supports multiple X servers running on multiple "virtual" consoles (swapping between servers is done via a special key and a function key).
[00:55:38] <richlowe> nrubsig: since alanc is complaining he has nothing to do on there, feel free to add "Solaris should ship setxkbmap" ;)
[00:55:40] <movement> richlowe: CIA parser
[00:56:04] <alanc> you would like that, wouldn't you?
[00:56:46] <alanc> I'm this close --> <-- to deleting all the XKB config files from the X consolidation - that will make me so much happier...
[00:57:19] <Tpenta> oooo he's down to a single space
[00:57:46] <alanc> just waiting for test results to say it's okay
[00:58:26] <Tpenta> http://blogs.sun.com/ThinGuy/entry/more_than_meets_the_eye   ROFLMAO
[00:58:45] <alanc> sadly, we'll still be shipping them, they'll just be in G11n instead, since they make all the changes for the localized keyboard layouts, so it will reduce the coordination effort between them and us
[00:59:33] <alanc> Tpenta: the next entry on there suggests the even better title of "Opteron Prime"
[00:59:47] <Tpenta> o gawd, the lauighing hurts
[00:59:50] *** lolownia has quit IRC
[00:59:59] *** sommerfeld has quit IRC
[01:00:10] <Tpenta> I like the suggestion for product placement in the movie
[01:00:27] <nrubsig> OK, emergency alert: we're own to one /op in in this channel...
[01:00:41] * oxygene applies and waits to be kicked
[01:00:44] <nrubsig> alanc: Do you know how to kick&ban people...
[01:00:48] <nrubsig> ?
[01:00:52] <oxygene> (it's not as if being op is a full-time job here)
[01:00:53] <alanc> no
[01:01:10] * nrubsig looks for eboutillier
[01:01:10] <alanc> but I know ops can be found if needed
[01:01:27] * nrubsig feels lonely and frightened
[01:01:36] <oxygene> yay, p7zip builds with sunpro again *runs testsuite*
[01:01:37] *** ChanServ sets mode: +o Tpenta
[01:01:56] <alanc> see, there's one now
[01:02:09] <nrubsig> Tpenta: Thanks! :-)
[01:02:24] <Tpenta> why do you want to kick ban someone?
[01:02:44] <Tpenta> i just hadnt identified
[01:03:13] <delewis> interesting, the banlist in here is larger than I thoght.
[01:04:01] <nrubsig> Tpenta: I do not want to kick or ban someone, except the person gives a matching reason for such an action.
[01:04:10] <Tpenta> ok :-D
[01:05:16] <nrubsig> Tpenta: usually I warn people multiple times before I even think about kicking them. and when I kick them I don't ban them immediately. Only if they rejoin and continue ofensive actions I kick them again. If they rejoin again and continue offensive behaviour THEN in consider a ban.
[01:06:08] <Tpenta> yea, the ban is th epenultimate resort
[01:06:30] <Tpenta> actually, the pen-penultimate (I can think of two worse things)
[01:06:33] <nrubsig> Tpenta: well, I ban people usually for 10mins and/or remove the ban before I leave.
[01:07:02] <nrubsig> Tpenta: if the behaviour continues over multiple days then I notify lilo to take action.
[01:07:12] <Tpenta> the other two things take an irc op with an O line
[01:07:35] <alanc> umm, you do know about lilo not being here anymore, right?
[01:07:41] <Tpenta> I used to have one of those on efnet
[01:07:45] <Tpenta> yes
[01:07:54] <nrubsig> alanc: I know that... very well...
[01:07:58] <Tpenta> that was incredibley sad
[01:07:59] <nrubsig> and it hurts.
[01:08:22] *** ShadowHntr has quit IRC
[01:08:31] <alanc> okay, just saw you speaking about lilo in the present tense and wasn't sure if it was denial or not having heard
[01:08:48] <richlowe> It really does not matter in the slightest whether anyone is currently opped or not.
[01:09:16] <Tpenta> if needs be we can have an op in here pretty quickly
[01:09:52] <Tpenta> I think I only ever had to G-line someone once (and I can't remember why that was)
[01:10:24] <delewis> and efnet, there's plenty of reasons :-)
[01:10:31] <delewis> s/and/on/
[01:10:49] <nrubsig> alanc: well, I sometimes forget it when writing some comments here. It needs some time until the fingers realised that something is... gone.
[01:10:50] <Tpenta> only had an O line as a ran a server
[01:12:16] <nrubsig> steleman: ping!
[01:14:44] *** nwf has quit IRC
[01:15:32] *** nwf has joined #opensolaris
[01:19:24] <nrubsig> Is stevel around ?
[01:22:39] <gdamore> cool.  got my first bullfrog I.  its _so_ much quieter than the two processor unit.
[01:23:34] <nrubsig> gdamore: did your company ever considered liquid cooling for your units ?
[01:23:57] <gdamore> no.
[01:24:08] <nrubsig> gdamore: the casings of your machines are large enougth to make one side a radiator... :-)
[01:24:17] <nrubsig> gdamore: why ?
[01:24:24] <gdamore> some are.  but we have some smaller systems.
[01:24:34] <gdamore> water is heavy.
[01:24:48] <nrubsig> gdamore: it would remove the sensivity of the fan vs. sand and dust...
[01:24:57] *** sahafeez has quit IRC
[01:25:00] *** Mazon is now known as mazon
[01:25:12] <gdamore> the hardened units have special filters for that. :-)
[01:25:28] <nrubsig> gdamore: erm... we're talking about less than 200ml in modern systems. I didn't thought about to use the Cray2 solution... :-)
[01:25:59] <gdamore> with so little water, can you effectively dissipate heat enough?
[01:26:05] <nrubsig> gdamore: yes
[01:26:19] <gdamore> well, i'm a software guy.  so i dunno.
[01:26:32] <nrubsig> gdamore: since the fluid is only the heat transport medium.
[01:26:35] <oxygene> hm.. the shuttle pc stuff has some vapor<->liquid cycle in pretty small pipes
[01:26:54] <nrubsig> gdamore: and I don't suuggest water... flourent is better (as used in the Cray2)
[01:27:28] <nrubsig> Is Linda Bernal offline ?
[01:27:48] <gdamore> ah, if you get a phase transition involved, then yes, the thermals make sense to me.
[01:35:57] <Gman> nrubsig, don't think she irc's
[01:37:17] <richlowe> nrubsig: the delay when having CRs forwarded to you is often fairly random.
[01:37:38] <richlowe> I filed a CR to have b.o.o automatically add the submitter to the interest list, and pass out initial notification that way.
[01:37:46] <richlowe> which should totally remove humans from the process, if it's done.
[01:39:18] *** piwi has joined #opensolaris
[01:40:55] * nrubsig files the next bug...
[01:42:40] <alanc> they've only recently made it possible to have non-sun.com e-mail addresses on the bug interest list
[01:44:24] <nrubsig> Does anyone got a subscription notification from the on-nv-commit list yet ?
[01:44:27] <nrubsig> !summon stevel
[01:46:16] *** Kush- has quit IRC
[01:46:25] <richlowe> alanc: yes, I filed the RFE after that was done.
[01:46:38] <richlowe> alanc: though frankly, the information handed out isn't at all appropriate for the initial notification.
[01:46:41] <richlowe> (or in general, to be frank)
[01:50:12] <Gman> nrubsig, i successfully subscribed to it, and got a mail through just this morning
[01:51:41] <richlowe> nrubsig: do you mean notification you're subscribed, or an actual putback notification?
[01:52:08] <nrubsig> richlowe: notification that I have subscribed.
[01:52:49] <richlowe> ah.
[01:53:53] <nrubsig> Gman: did you had time yet to check what's wrong with gnome-terminal vs. GNAW! ?
[01:54:17] <Gman> no
[01:54:25] <Gman> did you file a bug? :)
[01:55:21] *** Seawolf99 has quit IRC
[01:56:41] <nrubsig> Gman: erm... I doubt it is usefull to reveal the existence of GNAW! to bugs.opensolaris.org before the putback...
[01:57:51] <Gman> then it probably won't get fixed
[01:57:56] <nrubsig> Gman: as long as I label it as "technology demo" it should pass the review and OS/Net guards... but I doubt it is usefull to make them aware what GNAW! really is...
[01:58:16] <Gman> mmm
[01:58:34] <Gman> i assume you've provided a pkg manifest in the case?
[01:58:46] <nrubsig> Gman: yes
[01:58:57] <Gman> ok, so they know stuff goes into /usr/demo?
[01:59:11] <nrubsig> Gman: usr/demo/ksh/fun/ to be exactly.
[01:59:15] <Gman> then i doubt they're worried too much
[01:59:36] <Gman> [apart from being a waste of space..]
[01:59:46] <nrubsig> Gman: so far noone was told was this technology demo really does... :-)
[01:59:56] <nrubsig> Gman: it's 16k right now.
[02:00:06] <nrubsig> Gman: that are two UFS blocks
[02:00:31] <nrubsig> Gman: Gnome has background images 500times larger than that.
[02:00:32] <Gman> i'll blame it on you if we're 16k over and creating a 7th cd ;)
[02:01:29] <nrubsig> actually it's now 22k
[02:01:34] <nrubsig> I added two new levels.
[02:02:01] <nrubsig> and an EXIT trap to make sure the teminal state gets restored.
[02:02:11] <oxygene> komodo nethack?
[02:02:20] <nrubsig> oxygene: ?!
[02:02:44] <alanc> maybe the next McNealy magazine interview will have him showing gnaw off the way he was talking about the GNOME games in the last one 8-)
[02:02:44] <oxygene> just some wild guess at gnaw's function
[02:02:57] *** nachox has joined #opensolaris
[02:03:08] <nrubsig> oxygene: nope.
[02:03:11] *** jacotton has joined #opensolaris
[02:03:57] <nrubsig> oxygene: it's something "forbidden" because solaris is such a dry^H^H^H"professional" operating system.
[02:04:28] <nachox> hey all
[02:04:58] * Gman has a fun zenity easter egg in there :)
[02:05:13] * Gman was mostly just winding nrubsig up
[02:05:23] <richlowe> we're all humourless old men, really.
[02:05:30] <gdamore> what is gnaw?
[02:05:34] <richlowe> think of those two guys from the muppet show, but more so.
[02:05:36] <richlowe> ;)
[02:05:39] <Gman> heh
[02:07:10] <nrubsig> Gman: my point is that konsole gets this right while gnome-terminal sucks at this point.
[02:07:15] <nrubsig> er
[02:07:18] <nrubsig> bad englsh
[02:07:22] <nrubsig> umpf
[02:07:25] <nrubsig> bad english
[02:07:26] <alanc> I noticed /usr/share/pixmaps/poetry/ is missing in Nevada
[02:07:32] <nrubsig> heh
[02:07:46] <nrubsig> alanc: please file a bug.
[02:08:10] <alanc> Bug: /usr/share/pixmaps/poetry/Glynn.aha no longer contains Gman's ugly mug
[02:08:21] *** coffman has quit IRC
[02:08:26] <boyd> hehe
[02:08:29] <boyd> Morning, all
[02:08:30] <Gman> alanc, man, do we still install that?
[02:08:47] <alanc> it's in 10, but I didn't see it on machines with vermillion
[02:08:55] <Gman> yeah, we removed the poetry patch
[02:09:04] <Gman> i need to hook up the easter egg to something else
[02:09:10] <Gman> because it's the best easter egg *ever*
[02:09:18] <alanc> heh
[02:09:35] <nrubsig> Gman: doom3 ?
[02:09:42] <Gman> nautilus gestures on the background spelling out 'sun'
[02:09:46] <Gman> how fucking cool is that? :)
[02:09:47] <alanc> closest I've got to any easter eggs was sprinkling a few shoutouts into the random word list in the xscreensaver barcode hack
[02:09:52] <boyd> I never noticed it... how do I trigger the easter egg?
[02:10:19] <Gman> middle mouse button
[02:10:22] <Gman> takes practice
[02:11:44] <Gman> i'll replace it with something about opensolaris when i have time
[02:14:05] <alanc> have to put that on the list of features that would be removed if we put JDS4 into S10 8-)
[02:14:15] <Gman> heh
[02:14:50] <gdamore> hehe.  gnaw.  hey, with the new ksh we can bring back all those funky old C64 games that just used the graphics chars, right? :-)
[02:14:51] *** sickness has quit IRC
[02:14:52] * Gman shakes head, drops it, and sighs.
[02:16:52] <nrubsig> gdamore: yes
[02:17:34] <nrubsig> gdamore: as long as you get the JDS people (Hi Gman!) to fix gnome-terminal
[02:18:36] <gdamore> i hate gnome-terminal.
[02:18:47] <gdamore> give me xterm, or dtterm, or give me death!
[02:18:52] <boyd> It's better in 2.14
[02:19:04] <boyd> It's still slow, though
[02:19:11] <gdamore> unless they fixed the meta keybinding, so I can use meta key bindings, i don't care.
[02:19:55] <boyd> I can use meta fine if I change a pref from the default
[02:19:55] * nachox wonders if opensolaris will have mono integrared into jds
[02:21:03] <gdamore> the problem is that i need gnome-terminal to convert Meta-X to Esc-X or the shell won't grok it.
[02:21:32] <gdamore> (Perhaps the problem should be considered a bug in tcsh/ksh?)
[02:23:19] <boyd> Hmm... I think it works ... or maybe I'm using alt as meta
[02:35:30] *** gtc_ has quit IRC
[02:44:10] *** sommerfeld has joined #opensolaris
[02:44:10] *** ChanServ sets mode: +o sommerfeld
[02:46:55] *** lloy0076 has joined #opensolaris
[02:50:24] <twincest> so i'm trying to profile this application with the Studio 11 express performance analyzer, but the vast majority of runtime is spent in '<Unknown>'.  is there any way t get a more useful result?
[02:53:04] *** sommerfeld has quit IRC
[02:55:41] <nrubsig> Will Sun Studio 11 be shipped with Solaris 11 FCS ?
[02:56:07] <alanc> probably not
[02:56:13] *** nachox_ has joined #opensolaris
[02:56:16] <nrubsig> alanc: why ?
[02:56:22] <richlowe> I'd think it depends on what you mean by "With"
[02:56:27] <richlowe> on the same media? I'd assume not.
[02:56:30] <alanc> because a later version of Studio is more likely to be current then
[02:56:32] <richlowe> in the media kit? who knows.
[02:56:37] <richlowe> "available", well, it's available *now*
[02:56:58] <alanc> I think Studio 12 will be out before Solaris 11, not that I know when either one will actually ship
[02:57:13] <Gman> one hopes that studio will be a part of solaris someday...
[02:57:13] <richlowe> previous discussions about that mentioned the Studio folks treating every release as a major release, and various other things that make integrating it painful.
[02:57:17] <richlowe> most of it smelled like politics.
[02:58:06] <alanc> Studio 11 should be shipping in the Solaris 10 media kits already I thought
[02:58:26] <alanc> as just another DVD in the package, not integrated into the OS
[03:00:46] *** nachox_ has quit IRC
[03:01:28] *** nachox has quit IRC
[03:03:00] <richlowe> alanc: I thought so too, but wasn't sure.
[03:03:39] <richlowe> shipping it on the same physical media would add around 900M (two CDs), and cause whatever politics.
[03:04:24] <richlowe> if it could be pared down to the core tools, it would be good to see though.
[03:04:32] <richlowe> (cc, dmake, lint, cscope, lock_lint, etc)
[03:05:36] <delewis> no Netbeans, XEmacs, or gvim (source included) :-)
[03:06:03] <delewis> I'd be just as satisfied with shipping it as a separate CD in the media kit (it may now, but the initial FCS media kits, of course, didn't have it, as Sun Studio wasn't "free" at the time)
[03:06:17] *** cypromis has quit IRC
[03:06:23] <delewis> $ cd /opt/SUNWspro
[03:06:24] <delewis> $ du -sh contrib
[03:06:24] <delewis>  298M contrib
[03:07:03] <alanc> I thought the original media kits had the 60-day-free-trial/Try-and-Buy studio CD's in
[03:07:08] <delewis> alanc: nope
[03:07:16] <delewis> just Solaris SPARC/x86 media, and JES stuff.
[03:07:27] <richlowe> heh, I spent a long long time hopping from try-and-buy to beta to try-and-buy, and on and on. :)
[03:07:33] <Gman> someone does need to put some thought into how they deliver..
[03:07:34] <delewis> richlowe: yeah :-)
[03:07:37] <alanc> I remember older solaris releases did - maybe that got dropped at some point
[03:07:42] <delewis> every 60 days... re-install :-)
[03:07:49] <Gman>  /usr/sfw/bin/gcc isn't exactly a good message
[03:07:58] <delewis> and also, 'contrib' *really* needs to die
[03:08:09] <delewis> even sfwnv doesn't ship source for every piece of freeware.
[03:08:20] <alanc> we did do an LSARC review for having the Studio compilers install links into /usr/bin so we could finally have /usr/bin/cc again, but I don't know if they've shipped with that yet
[03:08:23] <delewis> the gcc patches are just thrown onto some publicly accessible site
[03:08:30] <delewis> the same should be down for SUNWspro's 'contrib'
[03:08:38] <Gman> yeah
[03:09:01] <delewis> that's a *huge* hunk of the install footprint, as I pasted earlier.
[03:09:12] <alanc> sadly, it's much easier to comply with GPL terms by bundling the source than by trying to make sure you have a stable long-term spot on a Sun download site somewhere
[03:09:23] <delewis> alanc: true
[03:09:50] <delewis> still doesn't make it the best solution, though :-)
[03:09:51] <alanc> the GNOME solution of putting stuff on an ancient cobalt.com CVS server always struck me as even more funny
[03:10:11] <delewis> hehe, interesting.
[03:10:27] *** cypromis has joined #opensolaris
[03:10:50] <richlowe> delewis: sfwnv doesn't?
[03:10:59] <alanc> http://www.sun.com/software/star/gnome/get/index.xml still has a link to it
[03:11:07] <delewis> richlowe: not on media
[03:11:20] <alanc> (the URL amuses me too, since of course, everyone knows GNOME is part of StarOffice)
[03:11:31] <delewis> :-)
[03:12:14] <delewis> I'm not sure sfwnv changes any source, though
[03:12:27] <richlowe> there's a few patches, I believe.
[03:12:29] <delewis> and I would guess that SUNWspro doesn't either, which makes it even more pointless.
[03:12:40] <delewis> the Sun Workshop stuff is integrated into XEmacs
[03:12:57] <alanc> oh that's funny - search for "GNOME" on Sun.com, and the first link is:
[03:12:57] <alanc> User Desktop Environments
[03:12:57] <alanc> Download the latest desktop environment software from Sun!
[03:12:57] <alanc> Java Desktop System | GNOME | Sun Secure Global Desktop Software | Common Desktop Environment (CDE)
[03:13:08] <alanc> I'd hardly call CDE " the latest desktop environment"...
[03:13:29] <delewis> alanc: nonsense!
[03:13:30] <delewis> :-)
[03:13:48] <delewis> actually, I'm proud to say that I've finally transitioned to JDS.
[03:13:56] <alanc> and the CDE link doesn't even point to a Sun page - it's a link to the OpenGroup page!
[03:14:07] <delewis> for awhile, I was moving back and fourth between 'em -- I'd get tired of one and start using the other.
[03:15:53] <delewis> why is that all of my Solaris systems print to the printer fine, but print from a Linux system and you get pure Postscript *sigh*.
[03:15:56] <delewis> I hate CUPS.
[03:18:01] *** bengtf has quit IRC
[03:18:06] *** sparc-kly has joined #opensolaris
[03:18:07] *** ChanServ sets mode: +o sparc-kly
[03:18:32] <Gman> whoah, the new airbuses all have powerports..
[03:19:23] <nrubsig> Gman: powerports ?
[03:19:33] <Gman> http://seatguru.com/airlines/Singapore_Air/Singapore_Air_Airbus_A345.php
[03:19:46] <Gman> in economy, there's power supplies!
[03:21:34] <boyd> Yay :)
[03:21:45] <boyd> Now as long as the wireless broadband is there...
[03:21:48] <boyd> oh, wait
[03:21:55] <Gman> actually boeing have taken out that service
[03:21:59] <Gman> wasn't making them any money :(
[03:22:06] <boyd> That's what I mean... :(
[03:22:08] <Gman> so it's free up until end dec
[03:22:22] <Gman> it's like, so close, but yet so far :)
[03:22:23] <boyd> It wasn't making money so they made it free?
[03:22:25] *** yjsong has left #opensolaris
[03:22:36] <Gman> i think they had to because they were pulling out
[03:22:59] <boyd> I wasn't 100% that that was the service SQ was using. I guess you're saying it was
[03:23:12] <Gman> yeah, unfortunately
[03:23:21] <Gman> i'll get free wifi on the way back to dublin in december
[03:23:30] <Gman> but none on the way back to nz in jan :(
[03:23:39] *** janus_man has quit IRC
[03:24:03] <nrubsig> Gman: horrible. No Internet. No drugs. Horrible... =:-)
[03:24:13] <boyd> I always thought that having wifi on the plane made a mockery of the "no radio transmitters" regulation
[03:24:15] *** sahafeez has joined #opensolaris
[03:24:27] <alanc> must be weird to get on the plane in the middle of winter and get off in the middle of summer
[03:24:28] * nrubsig wonders how long the internet-addicted people here can surive without their laptop ?
[03:24:31] <nrubsig> 5mins ?
[03:24:36] <Gman> boyd, all that stuff is getting relaxed
[03:24:36] *** janus_man has joined #opensolaris
[03:24:47] <Gman> nrubsig, dude, when you're on the plane for 12 or 13 hours, having internet really helps
[03:24:56] <boyd> Gman: Maybe because everyone knows it's ridiculous
[03:25:09] <Gman> nod
[03:25:10] <boyd> Gman: 12 hours? That's a short flight! :)
[03:25:38] <Gman> alanc, by then, i'll be stir crazy having lived with my parents for 4 weeks
[03:25:39] <nrubsig> Gman: I usually prefer to sleep during that time. Or read a good book. Something to clear my mind.
[03:25:46] <boyd> Gman: when you're on the plane for 12 hours, power really helps :)
[03:25:55] <Gman> boyd, yeah
[03:26:01] <Gman> don't think i'll be on an airbus however
[03:26:15] <boyd> nrubsig: For Aus/NZ to Europe it's like 20-24 hrs. You can't sleep that long
[03:26:22] <boyd> well... *I* can't
[03:26:26] <alanc> the pilot down the hall swears the cell phone ban is not because it messes with the plane instruments, but because it screws up the cell phone networks when you move between towers that fast
[03:26:39] <boyd> alanc: I'd buy that.
[03:26:41] <nrubsig> Gman: the last flight to boston was spend playing with some children because they were starting to rip the plane into pieces and noone was caring about them (and the stewardeses fed them with coca-cola, making it even much more worse... ;-/ ).
[03:26:56] <boyd> It they're so dangerous, why the hell are we allowed to have them on the plane?
[03:27:23] <Gman> heh, jeff just mentioned that having power was good for more people with explosive laptops :)
[03:27:37] <nrubsig> laptop terrorism!
[03:28:20] <boyd> Speaking of ripping the plane to pieces, I like Ze Frank's comment about why do they make such a point about tampering with the smoke detectors in the bathrooms.... it's the freakin' *wings* I don't wan't people tampering with :)
[03:30:56] *** yongsun has joined #opensolaris
[03:32:45] *** nwf has quit IRC
[03:33:02] <richlowe> alanc: I could fully believe (and support) it being because cellphones are irritating as all hell :)
[03:33:43] *** nwf has joined #opensolaris
[03:33:58] <alanc> of course, the conspiracy theory that the airlines & cell phone companies support the ban so they can make more money off the in-flight calls wouldn't be that surprising either
[03:34:23] <Tpenta> i can agree with that, spending 3hours a day on trains and having people shout into trheir cellphones (why does no-one use a normal talking voice on the cell phone?)
[03:34:44] <boyd> But that only applies when they *offer* in flight calls. Which no airline here does.
[03:35:20] <boyd> Tpenta: I try to use a normal voice. It's because they don't understand automatic gain control
[03:35:23] <sahafeez> 1. there is no technical reason now. 2. the airlines will let cell phones when they get setup for them (need a relay on the plane because of speed) becuase they will take a slice of the charges per min for the call. its a done deal.
[03:35:59] *** sp0rq has joined #opensolaris
[03:36:09] *** sp0rq is now known as Spawrq
[03:36:15] <boyd> Tpenta: I find that most people in East Asia are very quiet on their phone
[03:36:19] <Spawrq> Hola.
[03:36:38] <Spawrq> he kernel occupies the top virtual memory addresses, and the process occupies the lower memory addresses. This means that part of the virtual address space available to the process is consumed by the kernel, limiting the size of usable process virtual memory to between 3.5 and 3.75 Gbytes, depending on the size of the kernel's virtual address space.
[03:36:39] <Tpenta> most people who take the northern line trains on a weekday are not
[03:37:04] <Spawrq> so is that saying a single process can consume roughly 3.5gb of physical ram?
[03:37:27] <Tpenta> sprawq, the kernel has memory in a comletely different context to user space. your assumtion is invalid
[03:37:42] <LeftWing_> Tpenta: Is there a lot of "No.. NO.. NO I'M ON A TRAIN. NO A *TRAIN*. TRAIN."
[03:37:42] <boyd> Also, all those numbers are for 32 bit
[03:38:02] <Tpenta> no, but some confused looking people when the calls drop out in tunnels
[03:38:03] <boyd> Also Virtual address space size != Physical memory usage
[03:38:04] <Spawrq> tpenta- right, this is for sparc v7
[03:38:08] <LeftWing_> Tpenta: haha.
[03:38:33] *** Disorganized has joined #OpenSolaris
[03:39:15] <boyd> Sparc v7! I think the answer for them depends on how much coal is heating the boiler
[03:39:26] <Tpenta> if you want to have a read about issues with user address space, tim uglow has just written some interesting stuff on his blog at http://blogs.sun.com/timatworkhomeandinbetween
[03:39:43] <Tpenta> the fight is between allocatable memory and stack (why does everyone forget about stack)
[03:40:04] <Tpenta> if we had user processes able to access kernel memory, we would have a real security problem
[03:40:48] <Tpenta> what consumes the top of the address space is the user stack
[03:41:18] <Tpenta> if you set your rlimits for available stack too high, they eat into how much memory you can allocate to the process address space
[03:41:31] <Spawrq> perhaps this o'reilly book should be peer reviewed.
[03:41:42] <Tpenta> this is in an orielly book?
[03:41:45] <Tpenta> which one?
[03:41:47] <Spawrq> yes.
[03:41:56] <boyd> Tpenta: But surely this is practically only a problem for 32 bit AS's
[03:42:06] <Spawrq> Solaris Internals: Solaris 10 and OS Kernel Arch, 2nd
[03:42:11] <Tpenta> boyd: absolutely
[03:42:16] <Spawrq> OS = opensolaris
[03:42:17] <Tpenta> hmmmmm
[03:42:24] <boyd> Umm... I don't think that's an O'Reilly book
[03:42:30] <Tpenta> i dont have my copy here, but can you send me a reference?
[03:42:35] <Tpenta> it's prentice hall
[03:42:36] <boyd> At least my copy isn't
[03:42:41] <Spawrq> The SPARC V7 systems use a shared address space between the kernel and process and use the processor's privilege levels to prevent user processes from accessing the kernel's address space. The kernel occupies the top virtual memory addresses, and the process occupies the lower memory addresses.
[03:42:52] <Spawrq> This means that part of the virtual address space available to the process is consumed by the kernel, limiting the size of usable process virtual memory to between 3.5 and 3.75 Gbytes, depending on the size of the kernel's virtual address space.
[03:43:06] <Spawrq> This also means that the kernel has a limited size, ranging between 128 and 512 Mbytes.
[03:43:25] <boyd> Spawrq: Is there a reason you are so interested in Sparcv7?
[03:43:26] <Tpenta> given who wrote that, I am now less confident in my statement
[03:43:47] <Spawrq> boyd- nope.  it's just in the book.
[03:43:57] <boyd> I'm pretty sure it's changed in sparcv9
[03:44:03] <Spawrq> I figure if it were important enough to include, it's important enough to read.
[03:44:05] <boyd> so that Tpenta is correct for v9
[03:44:20] <Spawrq> yes, it goes on to say that v9 is significantly different.
[03:44:38] <Spawrq> and I'm just simply trying to fully understand.
[03:44:41] <boyd> I have no idea why they mention v7 at all. except that they couldn't bring themslef to remove it :)
[03:44:42] <Tpenta> it's quite possible they are right; I don't deal with a lot of sparcv7 anymore
[03:45:23] <boyd> The description of v7 rings a bell to me... I think maybe Rich Teer has some mention of it too.
[03:45:26] <Spawrq> safari.oreilly.com ++
[03:45:37] <boyd> Oh, "Safari"!
[03:45:49] <boyd> Not all the books on there are published by O'Reilly
[03:46:02] <Spawrq> right
[03:46:09] <Spawrq> o
[03:46:31] <Spawrq> this is a prentice hall book
[03:46:38] <Spawrq> holy crap, its new too
[03:46:45] <boyd> SI2 is very well respected
[03:46:55] <Spawrq> i can get rid of my old crappy solaris internals.
[03:47:08] <boyd> It's the second edition of the same book
[03:47:12] <Tpenta> you want to buy the second volume too. One of the authors hangs out in here
[03:47:15] <Spawrq> the parrot chewed up the cover anyway.
[03:47:48] <boyd> There are some errors in SI2... I noticed a few in the UFS material particularly
[03:49:26] <Tpenta> don't say that, I helped review ufs
[03:49:30] <Tpenta> :(
[03:49:33] <Spawrq> ho ho
[03:49:51] <boyd> :) Sorry... but it's true
[03:50:18] <boyd> of the top there were a few at the start in the discussion of the vtoc, boot block and superblock
[03:50:23] <boyd> s/of/off
[03:53:32] *** sparc-kly_ has joined #opensolaris
[03:53:41] <Spawrq> ok
[03:53:54] <Spawrq> so v9 seperates the kernel and user space all together
[03:54:05] <Tpenta> yup
[03:54:27] <boyd> That alternate address space in v9 is pretty nifty
[03:54:30] <Spawrq> so how large can a kernel be on a v9?
[03:54:33] <delewis> Spawrq: yes, in v8, there was a kernel context space at the top of every memory space.
[03:54:40] <delewis> Spawrq: as large as the memory space, itself.
[03:54:43] <Spawrq> o
[03:54:49] <delewis> in v8, that was limited to 256MB, IIRC.
[03:54:54] <Tpenta> it's limited by how much memory you can physically put in th ebox
[03:54:55] <boyd> 2^64 is... large :)
[03:55:00] <Tpenta> you're not likely to run out of bits
[03:55:10] <Spawrq> does the same apply for x64?
[03:55:28] <Tpenta> i believe so
[03:55:31] <delewis> in v8, you don't want a huge kernel, otherwise, the memory space for your program won't be very large (as there is a kernel context at the top)
[03:55:34] <delewis> Spawrq: no
[03:55:39] <boyd> I think there is a hole in the middle of the x64 address space
[03:55:42] <delewis> x64 has a very v9-like memory space.
[03:55:53] <delewis> no kernel context
[03:55:57] <delewis> and the stack is in the right place :-)
[03:56:03] <boyd> hehe
[03:56:15] <Spawrq> sun4u doens't have a kernel context either?
[03:56:22] <delewis> sun4u == v9
[03:56:29] <Spawrq> ok got it
[03:57:04] <boyd> Anyone remember the hole? I think x64 sign extends addresses by about 5 bits from meme
[03:57:11] <boyd> s/meme/memory
[03:57:21] <Spawrq> the only hole i see is in regards to US1 and 2 processors
[03:57:40] <delewis> boyd: not sure
[03:57:48] <delewis> but there was a hole in all US1 and US2 processors
[03:58:01] * boyd doesn't have his amd64 arch guid handy
[03:58:10] <Spawrq> he Ultra-SPARC-I and -II implementations, however, support only 44 bits of the address space, which means that there is a virtual address space hole in the middle of the address space.
[03:58:17] <delewis> boyd: I'm fairly sure there isn't a hole
[03:58:25] <delewis> x64 can address a full 64-bit memory space
[03:58:26] <Spawrq> I don't see the hole in x64
[03:58:29] <delewis> US1 and US2 couldn't
[03:58:39] * boyd shrugs... I may have a memory parity error
[03:58:47] <boyd> (or a hole :) )
[03:58:56] <delewis> only 16TBytes on US1 and US2 were addressable :-)
[03:59:00] <delewis> vs. 16 Ebytes
[03:59:24] <twincest> drat, i was about to build a USII system with 32TB
[03:59:30] <delewis> hehe
[03:59:31] <boyd> haha
[03:59:34] <delewis> shop elsewhere :-)
[04:00:29] <delewis> I do have some complaints about the layout of SI2 -- I really wish they would've introduced locking primitives earlier than they did.
[04:00:34] <jamesd_> time and 64 bit cpus is the only thing that can make  16 dimm sockets look  inadequate ;-)
[04:00:37] <delewis> in SI1 it was within the first 100 or 200 pages.
[04:00:45] <delewis> in SI2, it's at the *very* end.
[04:01:20] <nrubsig> Doesn't Niagara1 have the same MMU ("spitfire") as US-1 and US-2 and therefore the giant hole in the middle of the address space ?
[04:02:16] <jamesd_> i doubt it... and why would it matter its still has 16TB of address... who will notice a whole in the middle.
[04:02:21] <delewis> not sure, I just know that VIS is non-existing on Niagra1
[04:02:29] <delewis> :-)
[04:02:58] <nrubsig> delewis: it would be a huge problem for mmap()'ing a sparse file larger then 1TB
[04:03:10] <nrubsig> which is quite common for some HPC applications
[04:03:13] <Spawrq> o
[04:03:14] <Spawrq> southpark time
[04:03:15] <delewis> nrubsig: how so?
[04:03:41] <delewis> because the hole is in the middle, which is in the heap segment?
[04:03:56] <jamesd_> isn't the whole in physical memory not  in  virtual memory?
[04:04:13] <delewis> jamesd_: it's in both, because it's in virtual memory.
[04:04:32] <delewis> well, actually the hole is in virtual memory only, which is why physical memory is only limited to 16Ebytes.
[04:04:34] <jamesd_> can't the os  hide the whole?
[04:05:01] <nrubsig> delewis: I can think about seven or eight large simulation packages which will not run on niagara1 then (well, it doesn't matter because niagara1 has non FP power)
[04:05:15] <delewis> nrubsig: yes :-)
[04:05:20] <delewis> though, VIS emulation would've been nice
[04:05:31] <nrubsig> delewis: niagara1 has VIS emulation
[04:05:33] <delewis> actually, I think it's possible
[04:05:34] <nrubsig> er
[04:05:39] <delewis> as a VIS instruction just generates a trap
[04:05:42] <nrubsig> delewis: solaris has VIS emulation
[04:05:43] <delewis> forgot about that.
[04:05:44] <delewis> right
[04:07:17] *** axisys has joined #opensolaris
[04:07:22] <nrubsig> jamesd_: how should the OS be able to hide the hole ?
[04:07:39] <nrubsig> jamesd_: we are talking about a continues > 1TB file mapping.
[04:07:43] <delewis> nrubsig: it's an MMU problem, which is the OS has no control over.
[04:07:56] <nrubsig> delewis: yes, I know.
[04:08:01] <delewis> oh
[04:08:02] <axisys> anyone knew of this url http://sun.com/solaris/teachme which turns into a refrence url most of u may saw
[04:08:06] <delewis> wrong tab complete -- I meant jamesd
[04:08:10] <jamesd_> trap on accesses to the hole.. and  fill it with good memory.
[04:08:11] <nrubsig> I wrote one of the simulation packages myself.
[04:08:19] <delewis> nrubsig: the whole is above the heap, so I'm not sure why mmap()'ing a file would be a problem
[04:08:20] *** boyd has left #opensolaris
[04:08:22] <axisys> i wonder if there is any more hidden urls like this
[04:08:28] *** boyd has joined #opensolaris
[04:08:33] <delewis> jamesd_: that's exactly what happens.
[04:08:41] <delewis> a trap is generated whenever an access to that hole happens
[04:09:05] <nrubsig> delewis: if the mmu has the hole in the virtual address space then you can't use virtual addresses in the hole area
[04:09:10] <delewis> it's hidden to the program, but it still doesn't mean you can address more than 44 bites (16Tbytes)
[04:09:15] <axisys> i think there is one called freetrain or freetraining.. but cant find it
[04:09:17] <boyd> Ok, I worked out what I was thinking of with the hole. I read some of the amd arch docs and UA2005 and UST1 at the same time. I think I was thinking of niagara
[04:09:35] <boyd> 44 *bits*
[04:09:41] <jamesd_> also couldn't the os not use the hole...  if it asks for a block of ram always give well beyond the hole
[04:09:42] <nrubsig> yes
[04:10:08] <Tpenta>  12:09pm  up 9 day(s), 14:15,  4 users,  load average: 126.77, 53.39, 25.89
[04:10:09] <Tpenta> wow
[04:10:12] *** slowhog has joined #opensolaris
[04:10:19] <boyd> Yes... all the OS can do to hide the hole is to make it unusable. Which is why it's a "hole" :)
[04:10:21] <delewis> jamesd_: that would work for allocation only
[04:10:29] <Tpenta>  12:10pm  up 9 day(s), 14:15,  4 users,  load average: 169.47, 67.66, 31.35
[04:10:30] <Tpenta> double wow
[04:10:30] <delewis> a trap still has to be generated for every access, in general.
[04:10:33] <nrubsig> jamesd_: it will not work when the requested mapping is larger than the remaining address apce
[04:10:44] <nrubsig> s/apce/space/
[04:10:48] <jamesd_> Tpenta,  remember you will go blind if you watch more than 100 pr0n movies at once.
[04:11:05] <Tpenta> looks to have peaked at about 195
[04:11:13] <boyd> Haha! but 99 is still ok, right?
[04:11:14] <delewis> I've gotten my p640 up to 400 before with a forkbomb :-)
[04:11:21] <richlowe> Tpenta: having fun?
[04:11:22] <delewis> and of course, it recovered :-)
[04:11:29] <Tpenta> this is a 4800 doing real work (on build)
[04:11:30] <nrubsig> jamesd_: remeber I said "mmap()'ing a sparse file" - which can be as large as 2^63 byte
[04:11:36] <delewis> Tpenta: *impressive*
[04:11:43] <jamesd_> boyd, of course, but only prefessionals should go above 50.
[04:12:03] <Tpenta> I've only got 12 1050MHz cpus in this box
[04:12:12] <nrubsig> "only" ?
[04:12:13] <boyd> hehe... /me pictures a professional watcher, then realises that was a mistake :)
[04:12:18] <jamesd_> nrubsig, okay... i thought a 32 terabytes would be enough...
[04:12:23] <nrubsig> Tpenta: give me access to that box! NOW!
[04:12:24] <Tpenta> boyd :-D
[04:12:34] <Tpenta> did you get the raincoat in the image boyd?
[04:12:47] * boyd shudders
[04:12:56] <jamesd_> Tpenta, only 12? they make you share a  t2000?
[04:13:04] <boyd> It was draped over a chair
[04:13:07] <Tpenta> nope, it's a 4800
[04:13:27] <Tpenta> with a photon attached for teh zfs pool holding my workspace
[04:13:38] * gdamore drools.
[04:13:39] <nrubsig> jamesd_: no, because the idea of the mmap()'ing of the sparse file is to get an "infinite" address space for the simulation application and let the filesystem to the real allocations. This saves a GIGANTIC amount of complexity and hides it in the OS.
[04:13:46] *** sparc-kly has quit IRC
[04:13:51] <nrubsig> s/to the/do the/
[04:14:20] <jamesd_> nrubsig, okay.. so i guess you will have to confine those programs to nice 4800's  or bigger
[04:14:40] <gdamore> sorry for stepping in the middle.  what is this for, the next version of firefox?
[04:14:44] <nrubsig> jamesd_: SF12K
[04:15:07] * Tpenta was wondering why his builds were taking 3 times longer than he thought the yshould. well after making the domain have 3 boards in it instead of 1, my expectations were about right
[04:15:10] <nrubsig> gdamore: version 666, sponsored by sa.tan.inc
[04:15:41] <nrubsig> Tpenta: BTW: was the bug fixed tha OS/Net did not compile successfully when used in tmpfs ?
[04:15:55] <Tpenta> no idea
[04:16:08] <gdamore> who has a tmpfs big enough to compile OS/Net in it?
[04:16:18] * Tpenta grins
[04:16:30] <gdamore> *smack!*
[04:16:34] <nrubsig> gdamore: get a box with 4GB memory and you're happy... :-)
[04:16:39] <Tpenta> $ df -h /tmp
[04:16:39] <Tpenta> Filesystem             size   used  avail capacity  Mounted on
[04:16:39] <Tpenta> swap                    25G   541M    24G     3%    /tmp
[04:16:46] * boyd has access to a 2900 with 16G
[04:16:49] <boyd> is that enough
[04:16:50] <delewis> OS/Net only takes up about 2.65G on SPARC
[04:16:50] <boyd> >
[04:16:51] <boyd> ?
[04:16:57] <gdamore> is that all you need? well, sheeeit, i have one one of those right here...
[04:17:00] <nrubsig> Tpenta: why don't you build your stuff in tmpfs ?
[04:17:02] <delewis> gdamore: yeah.
[04:17:23] <Tpenta> because this is a workspace with a bugfix in it that I'd rather not have to redo if /tmp goes away
[04:17:30] * delewis is still pissed about one of his DIMMs frying in the SB1000
[04:17:35] <delewis> going from 2GB to 1GB *sucks*.
[04:17:37] <nrubsig> Tpenta: what about using subversion ? :-)
[04:17:38] <gdamore> actually, the BF2 I have coming back to me has 8GB.
[04:17:43] * jamesd_ is happy with his 2  2GB boxes...   though i would love to have another 2GB of ram in my blade 1500 or u20
[04:17:45] <delewis> gdamore: nice :-)
[04:17:50] <nrubsig> delewis: my laptop has 2GB memory
[04:17:52] <delewis> 8GB in a portable must rock
[04:17:59] <gdamore> its coming back because i toasted the OBP prom on it. :-)
[04:18:00] <delewis> nrubsig: my Ferrari 5k will have 2GB :-)
[04:18:05] <delewis> gdamore: lol
[04:18:17] <nrubsig> delewis: yes, but my laptop is already two years old
[04:18:29] <gdamore> well, it rocks, but only if you haven't properly strapped it down -- all the fans and disks have some interesting resonance... :-)
[04:18:31] <axisys> i found another hidden url http://www.sun.com/solaris/freetraining
[04:18:32] <delewis> normally, if you had told me a BF2's OBP prom got toasted I would be a bit worried, but I guess working at Tadpole has it's perks :-)
[04:18:35] <delewis> "No big deal"
[04:18:51] <gdamore> i was testing a new prom, with new Radeon 9250 support.
[04:19:08] <axisys> anyone know of any other hidden url that actually expands to a real url on solaris site?
[04:19:29] <nrubsig> gdamore: see invites
[04:19:49] <Tpenta> nrubsig, the gate is still teamware, if I'm going to put back to the gate, ...
[04:20:36] <richlowe> and the gate *will* be mercurial.
[04:20:38] <richlowe> so svn would still screw you.
[04:21:22] <nrubsig> Tpenta: yes, but you could setup a temporary svn server for this purpose. at least mozilla takes around 38% less time to compile on tmpfs - and this is IMO *impressive*
[04:21:28] <Tpenta> absolutely
[04:21:47] <gdamore> unfortunately, the person who burned the image for me sent me the wrong file.
[04:21:58] <gdamore> so, how is the conversion to Hg going?
[04:22:07] <richlowe> also, svn wouldn't even vaguely solve the problem of your stuff vanishing if your machine vanishes.
[04:22:12] <richlowe> tmpfs is tmpfs.
[04:22:13] <richlowe> gdamore: check the /topic
[04:22:25] *** eboutilier has joined #opensolaris
[04:22:26] <gdamore> doh.
[04:22:51] <boyd> Well.. you could rsync your on-disk workspace to /tmp for the build :)
[04:23:14] <Gman> hey eric
[04:23:15] <Tpenta> I could, but this build is going to take an hour, I'm really not going to gain anything significant
[04:24:04] <Spawrq> so ...
[04:24:09] <eboutilier> hi glynn
[04:24:19] <Tpenta> hi eric
[04:24:24] <jamesd_> boyd, i did that a while ago, and it didn't improve performance... on a 4GB box...
[04:24:26] <eboutilier> hi alan
[04:24:37] <Spawrq> if you're running a sparc v9 in 64 bit mode, and you're using a 32 bit program, the program will only be able to address 3.9gb ram?
[04:25:04] <Tpenta> 6475492 finally annoyed me enough to take it and fix it
[04:25:21] <richlowe> and if you don't have a large chunk of ram, lint is going to hate you for having stuff in tmpfs.
[04:25:26] <richlowe> since lint chews up so much of it itself.
[04:25:30] <boyd> Spawrq: Closer to 4GB I'd think.
[04:25:39] <Spawrq> k
[04:25:43] <Spawrq> that's what i was thinking.
[04:25:48] <gdamore> so we need to get Hg in blastwave, so i can install it trivally. :-)  where's dclarke?
[04:26:09] <richlowe> gdamore: it's in Nevada
[04:26:15] <richlowe> since around snv_46, I think.
[04:26:22] * boyd lunches
[04:26:52] <richlowe> There are also packages in the tools community.
[04:27:37] <boyd> They require python 2.4. I don't know if there is a build that has 2.4 and not hg
[04:27:45] * boyd *really* lunches
[04:27:52] *** aijin has joined #opensolaris
[04:28:01] * gdamore looks around in the tree....
[04:28:58] <gdamore> hmm... my local tree might be downrev.  where in Nevada is it?
[04:29:48] <richlowe> SFW, I think.
[04:29:53] <richlowe> and /usr/bin/hg
[04:29:55] <richlowe> SUNWmercurial
[04:32:07] <gdamore> thanks.
[04:32:26] <Gman> is it possible [those with leet opensolaris.org web power] to create links on the sidebar that are just links, and don't have a page attached to them?
[04:32:54] <nrubsig> umpf
[04:33:01] <nrubsig> you ship HG but not SVN ?
[04:33:14] <Gman> we ship svn too
[04:33:21] <nrubsig> Gman: where ?
[04:33:22] <Gman> it's in the final stages of integration
[04:33:35] <Gman> arc case is approved
[04:33:48] <nrubsig> Gman: ah
[04:33:56] <nrubsig> Gman: putback in... 2008 ?
[04:34:02] <richlowe> svn is a whole lot larger than hg.
[04:34:09] <nrubsig> (sorry for the sarcasm)
[04:35:03] <Gman> it's more complicated
[04:35:07] <Gman> bigger deps
[04:35:20] <nrubsig> Gman: more or less complex tha ksh93 ?
[04:35:22] <nrubsig> =:-)
[04:35:26] <Gman> i half think whether bzr would have been a good solution
[04:35:33] <Gman> just not ready in time i guess
[04:35:34] <Gman> anyway
[04:35:39] <Gman> nrubsig, you have it easy dude
[04:35:42] <richlowe> bzr was very, very, very slow at the time of the evaluation.
[04:35:47] <Gman> you have most of the case figured out, you're well on your way
[04:35:50] <Gman> richlowe, nod
[04:35:54] <richlowe> I only once managed to fully import an ON tree.
[04:35:55] <Error_404> richlowe: it still is, in my experience
[04:36:00] <richlowe> I recall it took the better part of a day.
[04:36:09] <richlowe> (I didn't have the patience for operations after that)
[04:36:25] <Gman> ooof
[04:39:08] *** piwi has quit IRC
[04:39:44] <richlowe> ah looking back at the mail, I had it backwards.  There was a period of time where initial import was fairly fast, but a bringover took the better part of a day.
[04:41:19] <Error_404> bzr push takes the better part of an afternoon
[04:42:39] *** mercykiller has joined #opensolaris
[04:44:33] <Gman> mmm, just noticed a bug with svn - it's not syncing to opengrok
[04:47:00] <richlowe> I don't believe that's a bug.
[04:47:09] <richlowe> I'd bet it was just not considered.
[04:47:15] <Gman> boo!
[04:51:41] *** lloy0076 has left #opensolaris
[04:58:23] <Tpenta>  12:58pm  up 9 day(s), 15:03,  4 users,  load average: 241.64, 165.79, 102.14
[05:00:03] <Spawrq> hmm
[05:01:22] <nrubsig> Tpenta: you get your own private swimming pool with extra hot boiling oil in hell when you reach load average: 666.666, 666.666, 666.666
[05:01:31] <Tpenta> :)
[05:01:52] <Tpenta> i had to fix a bug in the load average code because starcats were displaying negative numbers for large loads
[05:02:05] <nrubsig> heh
[05:02:15] <nrubsig> Tpenta: was the code using |short| ?
[05:02:21] <Auralis> this will be sorta hard to do, since load only uses two decimal values
[05:02:34] <Tpenta> no it was a little more complex than that in the degrading code
[05:02:42] <nrubsig> Tpenta: why ?
[05:03:01] <Tpenta> It's a 32 bit number, BUT
[05:03:20] <nrubsig> Tpenta: fixed16bit.fixed16bit ?
[05:03:26] <Tpenta> its only really a 24 bit number with 16 bits in front of the decimal point and 8 bits behind
[05:03:35] <nrubsig> ouch
[05:03:46] <nrubsig> Tpenta: I hope you extended it to 64bit...
[05:03:55] <Tpenta> couldnt
[05:03:58] <elektronkind> wait, so you managed to get a load average to fill up 24bits ?
[05:04:00] <Tpenta> that would be an interface change
[05:04:03] <nrubsig> Tpenta: the solaris is in trouble
[05:04:04] <Tpenta> in a released system
[05:04:10] <Tpenta> and it's a very well defined interface
[05:04:26] <Tpenta> the trick was actually being a little more clever in the sequence of operations
[05:04:28] <nrubsig> Tpenta: I expect that high-end rock machines may in the rage of >= 1024 cores
[05:05:04] <Tpenta> the real fix is to take it to 64 bit, but that's a rather large interface change
[05:05:04] <nrubsig> Tpenta: which will likely lead to 8192 strands per machine
[05:05:51] <nrubsig> Tpenta: I suggest to fix it now - before later rock-based machine run into that problem.
[05:06:19] <nrubsig> Tpenta: which API/function is affected ?
[05:06:31] <Tpenta> this was about 18 months ago roland
[05:06:45] <Tpenta> i actually don't recall
[05:07:37] <elektronkind> hmm
[05:07:38] <elektronkind> int getloadavg(double loadavg[], int nelem);
[05:07:47] <Tpenta> that's the one
[05:07:59] <elektronkind> that's from solaris 8's man page
[05:08:04] <gdamore> just define a new API for loadaverages, and let the old ones max out.
[05:08:17] <gdamore> getloadavg64().
[05:08:24] <elektronkind> cumbersome
[05:08:45] <gdamore> you could have new code get the new API by way of some macro magic.
[05:09:14] <nrubsig> Tpenta: was the syscall the problem or the libc function ? |int getloadavg(double loadavg[], int nelem);| looks like that it won't run into problems...
[05:09:15] <Tpenta> i have a vague recollection that my change makes use of 23 or 24 bits
[05:10:54] <gdamore> above a certain number, load averages are not likey to be meaningful.  e.g. after 64,000, you _know_ your machine is horked, regardless of whether you have 1 core or 1000 cores.
[05:11:10] <Tpenta> :)
[05:11:39] <richlowe> it's not particularly meaningful either way.
[05:11:58] * richlowe pictures a dashboard warning light that just says "Something could be wrong, maybe"
[05:12:10] <elektronkind> yeah I agree
[05:12:17] <gdamore> heh
[05:12:19] <elektronkind> load average is like a "check engine" light
[05:13:05] <elektronkind> like, the o2 sensor in your FC  HBA needs to be replaced after 80000 CPU minutes
[05:14:51] <gdamore> to say nothing of your disks
[05:17:46] *** alanc is now known as alanc-away
[05:18:56] <elektronkind> sigh. it's raining outside and that means my DSL has gone from 3Mb/s to 130Kb/s
[05:19:15] <elektronkind> what a wretched technology
[05:21:26] <sahafeez> what the hell? what a bad telco. no reason at all for that cept the telco is cheap ass and does not built out well
[05:21:35] <gdamore> why is your DSL dependant on the weather?
[05:21:41] <nrubsig> elektronkind: since when is DSL whether-sensitive ?
[05:21:50] <Tpenta> bad phone line would be my guess
[05:21:53] <elektronkind> well, it's verizon and I live in a 60 year old neighborhood
[05:21:55] <nrubsig> yep
[05:22:20] <elektronkind> the tech is coming out tomorrow (3rd time!) to take a look at it. I'm going to demand to be put on new pairs.
[05:22:20] <gdamore> i have verizon, and its new, and i can't get DSL yet.  i've been waiting hoping my neighborhood would get FIOS.
[05:22:40] <gdamore> the neighborhood is about 1.5 years old.
[05:22:42] <elektronkind> mm, FIOS. I have that available to me. I'm thinking of replacing my dsl with it.
[05:22:52] <gdamore> uh, duh.  no brainer.
[05:23:05] <elektronkind> I was astounded to learn that's it's available in this crummy 'hood I live in
[05:23:16] <gdamore> heh.  lucky you.
[05:23:17] <elektronkind> gdamore: yeah, but I'm addicted to my static IP
[05:23:18] <sahafeez> well i am 15k from the co, which should be fine for 768/768 which is want i want but they will not do it. no dsl beyond 14k now for sbc..
[05:23:46] <gdamore> i asked timewarner (my cable ISP) for a static IP.  they want $40/month for it.  insane.
[05:23:55] <elektronkind> that's crazy talk
[05:24:17] <gdamore> i am already paying ~$60/month for "extreme high speed"  (16Mdown, 2Mup)
[05:24:18] <elektronkind> I guess it's priced so no one would buy it, but they can still say they offer it
[05:24:23] <sahafeez> i have cox right now. ip changes about once every 3 months or so when their network goes blam
[05:24:46] <gdamore> i've not had my IP change in over year.  even thru outages.
[05:24:53] <sahafeez> hell i am paying $55 for 6d/512k up and i only get 4 down
[05:25:16] <elektronkind> that's still a great price compared to dsl
[05:25:19] *** jsubl2 has joined #opensolaris
[05:25:35] <sahafeez> ah, sbc dsl for 1.5 mb is $13
[05:25:40] <sahafeez> if you can get it
[05:26:03] <sahafeez> of course you are sharing your packets with the nsa but..
[05:26:31] <gdamore> what was hilarious was before i went to cable, i was testing an 802.11 service.  they only offered about 2M/384 for ~$80.  and, the amazing thing is, 802.11 is _symmetric_. they put in special rules just to kill upload speed.
[05:26:55] <elektronkind> wow.. that's asinine
[05:27:12] <Auralis> welcome to consumer ISPs
[05:27:17] <gdamore> and, after the 30 day trial, it was a 2 year commitment.  needless to say, i didn't stick with them.
[05:27:18] <elektronkind> obviously crippled products don't fly well with tech-aware people
[05:27:25] *** nwf has quit IRC
[05:27:51] <sahafeez> hum, wifi is half-duplex. that is funny
[05:28:27] <elektronkind> I wonder if they included a free packet loss option
[05:28:34] <gdamore> they didn't kill it at the mac layer, they throttled it  at the IP layer.
[05:28:54] <twincest> saha: so is ethernet, it just uses two cables, i think?
[05:29:51] <sahafeez> yah, i did a startup in CA. 10000 aps in apartment complexes. i but in freebsd routers running on a settop box style on flash card. it shaped via ip with ipfw and dummynet on a per ip
[05:29:52] <gdamore> uh, that's the only way get full duplex.  you need separate conductors.  kinda basic physics, right?
[05:29:57] <LeftWing_> Ethernet has a TX and an RX pair, it depends on the method of aggregation (switched or shared) as to whether you can get Full duplex.
[05:30:26] <sahafeez> thus the reason wifi is half duplex. transmit or listen
[05:30:39] * LeftWing_ nods.
[05:30:43] *** LeftWing_ is now known as LeftWing
[05:30:58] <gdamore> right, though they could have done two antennas and slightly offset frequencies to get full duplex.
[05:31:15] <timeless> hrm
[05:31:18] <gdamore> but it would have been a waste of bandwidth.  most network traffic is highly assymetric.
[05:31:22] <timeless> i think i "lost" init :)
[05:31:33] <sahafeez> bah! they still have not got the 1 antenna bit right.
[05:31:57] <sahafeez> wifi sucks
[05:32:10] <gdamore> heh.
[05:32:19] * gdamore has hacked on a lot of wifi code over the years.
[05:32:20] <sahafeez> hope the multi-path multi-antenna in N will fix it
[05:33:02] <gdamore> 11N chipsets are still too power hungry to use anywere you want good battery life.
[05:33:09] <sahafeez> http://www.smartbridges.com/about/articles.asp?id=112 3rd down. EdgeFoucs
[05:33:26] <gdamore> which means we'll probably be getting them in our laptops as soon as mgmt can figure out how to get them.
[05:33:36] <sahafeez> put out alot of APs. they all suck.
[05:33:52] <Auralis> gdamore: hehe
[05:34:10] <gdamore> fyi, check out www.meraki.net.   I have a port of NetBSD to this thing.
[05:34:35] <sahafeez> not really deisgned for high density deployment. stomp on each other. N will fix it (crossing fingers)
[05:35:09] <sahafeez> mesh cna work done right.
[05:35:21] <sahafeez> s/cna/can
[05:35:48] <gdamore> david young is trying to get some improvements to do mesh better on these units.  the linux code is closed-source.
[05:35:49] *** nbkk6fo has joined #opensolaris
[05:35:52] *** nbkk6fo is now known as rodrickbrown
[05:35:55] *** yongsun has quit IRC
[05:36:02] <sahafeez> I think N will fix it. the tech is smarter and Apple is betting alot on it. The release date of the iTV is the same as the final spec approval for N.
[05:36:27] <sahafeez> did not know you could close source linux code or is is code that runs on linux i guess
[05:36:33] <gdamore> heh.  i'm still skeptical that 802.11n will live up to the promise.
[05:36:48] <gdamore> the code is a loadable kernel module. it is closed.
[05:36:53] <sahafeez> me, i have learned not to bet against apple.
[05:37:02] <gdamore> the basic core OS has source.
[05:37:03] *** slowhog has quit IRC
[05:37:21] <gdamore> yah, i remember thinking the Airport (the original) was the greatest thing since sliced bread.
[05:38:22] <sahafeez> it is. hell, i just went to a friends and setup to airport express to do music for him to talk to his linksys via wep. it was simple.
[05:38:45] <elektronkind> I <3 my ap express
[05:39:02] <elektronkind> I have two of them
[05:39:13] <elektronkind> both primarily for their music ports
[05:39:28] * Auralis hugs her cat7 cables :)
[05:39:32] <sahafeez> i have an extreme and 2 express in my house. heck planning music right now on the computer, living room and bedroom.
[05:39:51] <sahafeez> playing even. shit cannot type tonight.
[05:40:07] <elektronkind> and in reasonably good sync! (if youre using itunes of course)
[05:41:02] <sahafeez> if you want to use something else get hum. there is an app that will hijack any sound and put it out to and AE
[05:41:05] <elektronkind> I like it. I'm a audio guy and going from room to room like you have yours and there's no off-time clash in the sound due to lag and buffers is excellent. It's like I have speaker cable going around the whole house
[05:41:20] <sahafeez> trying to remember the name. windows and mac version
[05:41:21] <elektronkind> sahafeez: yeah, by rogue amoeba
[05:41:33] <elektronkind> Airfoil
[05:41:34] <sahafeez> no that is not the one i am thinking of
[05:41:40] <sahafeez> airfoil
[05:41:45] <sahafeez> thanks
[05:42:10] <gdamore> so now that ON is going to Hg, wtf is the use of having an svn repo for jds?  am i just being dense?
[05:42:25] <elektronkind> gdamore: I was wondering that, too
[05:42:54] <richlowe> gdamore: different requirements.
[05:43:16] <richlowe> gdamore: JDS (and I guess the CCD, though I'm not sure I agree there), don't have the really long-term projects that make distributed stuff useful.
[05:43:19] <richlowe> they say.
[05:43:34] <gdamore> sometimes i think i live in SCM hell -- I have to interface with CVS, SVN, now Hg, Teamware, and even Bitkeeper for some stuff.  I want fewer SCMs, not more.
[05:43:57] <richlowe> gdamore: so convince your employer (I assume) to ditch TW :)
[05:43:57] <nrubsig> Auralis: you're female ?
[05:43:58] <richlowe> that's one less.
[05:44:07] <gdamore> the TW stuff comes from Sun.
[05:44:18] <sahafeez> well as soon as my evil plan to take over the world comes thru i will mandate one SCM
[05:44:20] <gdamore> (e.g. Sun Ray firmware sources)
[05:44:23] <richlowe> I know it's sun's product, but I assumed you were using it for tadpole-ish things.
[05:44:27] <richlowe> Ah.
[05:44:44] <gdamore> we use CVS for most of our internal stuff.
[05:44:52] <richlowe> I still miss bitkeeper having the free version.
[05:44:57] <richlowe> whatever else it may be, it was damn nice.
[05:45:17] <movement> shurely only because of the merge thingy
[05:45:28] <gdamore> bitkeeper was the best damn SCM on the planet.  I tried _really_ really_ hard to get Sun to buy it out and get it into Forte.
[05:46:07] <gdamore> (that was right after Sun announced internally the EOF of Teamware.)
[05:46:08] <richlowe> movement: merge thingy, speed, at the time the only other thing with even close to the same distributed model and freely available was arch...
[05:46:29] <timeless> EOF?
[05:46:35] <richlowe> timeless: End Of Feature.
[05:46:38] <gdamore> end-of-feature.  when they canned it.
[05:46:39] <nrubsig> EndOfFeature
[05:47:05] <gdamore> i was one of probably hundreds of Sun engineers who thought at the time: WTF?!?
[05:47:12] <nrubsig> This is typical for sun: First they cann it and then they want it back.
[05:47:16] <richlowe> movement: oh, and though I hate to say it, revtool (I think it was) also presented a really nice view of things.
[05:47:21] <richlowe> movement: though hgk does that well enough.
[05:47:22] <gdamore> and in regards to that particular decision, I still think: WTF?!?!
[05:47:27] <richlowe> (or would, if the Tk in nevada was new enough)
[05:48:12] <richlowe> movement: 'from:' thing looks to be fixed, by the way.
[05:48:22] <gdamore> for the first time i tried out our synaptic track pad configuration tool written in Tk, and I was mightily impressed.  it was a lot snazzier than i expected.
[05:48:24] <sahafeez> http://en.wikipedia.org/wiki/Lighthouse_Design
[05:48:48] <sahafeez> some get Jonathan to release the source for this. i want them for my mac
[05:49:46] *** laca has quit IRC
[05:50:37] *** mercykiller has quit IRC
[05:50:48] *** sunyong has joined #opensolaris
[05:50:54] <sahafeez> everyone sign this pls http://www.petitiononline.com/laafs/petition.html
[05:51:17] <timeless> hrm
[05:51:21] *** sunyong has quit IRC
[05:51:30] <timeless> is it bad that i have /var/crash/cores/vold.*
[05:51:41] *** jsubl2 has left #opensolaris
[05:52:00] <nrubsig> sahafeez: what is "lighthouse" ?
[05:52:09] *** yongsun has joined #opensolaris
[05:52:47] <sahafeez> nrubsig: http://en.wikipedia.org/wiki/Lighthouse_Design
[05:53:01] <sahafeez> some of the best application ever done - buy JS no less
[05:53:24] *** yongsun has joined #opensolaris
[05:54:50] *** yongsun has quit IRC
[05:55:05] <sahafeez> but then again i want steve jobs to kill the top menu and bring back the next side menu
[05:56:37] <timeless> oh wow
[05:56:51] <timeless> glimpseindex crashed and ate 2.5g each time it crashed
[05:56:57] <timeless> heh :)
[05:58:48] <timeless> hrm
[05:59:08] <timeless> i just crashed dbx, and it said "dbx's coredump will appear in /tmp
[05:59:21] <Gman> known bug
[05:59:28] <timeless> but i can't find its coredump there, i do see two copies of its coredump in /var/crash/cores
[06:00:07] * timeless sighs
[06:00:18] <timeless> yeah, it's the stupid ps_pbrandname thing
[06:00:32] <timeless> i thought i applied a patch that was supposed to fix that
[06:01:34] *** yongsun has joined #opensolaris
[06:01:38] <timeless> oh
[06:01:55] <timeless> so, it looks like i get two copies of each coredump
[06:02:16] <timeless> is there a way to tell the system to only dump something as global if it doesn't dump it as user?
[06:02:41] <timeless> ooh /var/crash/cores/init.1.global* :)
[06:04:42] *** Tpenta has quit IRC
[06:09:22] <nrubsig> ouch.
[06:09:23] *** Tpenta has joined #opensolaris
[06:09:28] * nrubsig is alone ias /op again
[06:09:36] <nrubsig> s/ias/as/
[06:09:43] *** nrubsig sets mode: +o Tpenta
[06:10:12] <Tpenta> my vpn dropped out
[06:10:24] <nrubsig> bad VPM
[06:12:18] <nrubsig> YIIPPIIEEE
[06:12:21] <nrubsig> YIPPIIIEEE
[06:12:23] <nrubsig> hurray
[06:12:28] * nrubsig hugs anyone
[06:12:38] <nrubsig> New ast-ksh93 source tarball out...
[06:12:51] * nrubsig implodes
[06:14:56] <timeless> http://www.webwizardry.net/~timeless/solaris/init.1.global.0.1161281387.swift.stack
[06:29:46] <dclarke> nrubsig: I needed a hug :-)
[06:30:22] *** yongsun has quit IRC
[06:30:47] * rodrickbrown gives dclarke a hug
[06:30:56] *** sahafeez has quit IRC
[06:31:27] <dclarke> er .. its getting a little too friendly in here
[06:31:34] * dclarke runs away
[06:31:42] *** ShadowHntr has joined #opensolaris
[06:36:08] <dclarke> damn ..
[06:36:25] <dclarke> can't figure out where the Studio 11 downloads are located
[06:36:35] * nrubsig gives dclarke a hug, tooooooooo
[06:36:39] <timeless> sun.com/studio iirc
[06:36:53] <dclarke> http://developers.sun.com/sunstudio/downloads/
[06:37:11] * nrubsig waits for Tpenta giving dclarke a hug, too.
[06:37:11] <dclarke> http://developers.sun.com/sunstudio/downloads/index.jsp
[06:37:22] <nrubsig> Now... everyone gives dclarke a hug, Ok ?
[06:37:37] * dclarke blush
[06:37:45] <Tpenta> good evening dennis
[06:37:46] * dclarke runs away .. fast
[06:37:54] <dclarke> g'day
[06:38:35] *** sahafeez has joined #opensolaris
[06:38:53] <dclarke> damn .. my browser just spins and spins on the download for Studio 11 ... nothing ever happens
[06:39:23] *** karrotx has joined #opensolaris
[06:39:35] <nrubsig> dclarke: maybe ir wants a hug ?
[06:39:46] <nrubsig> s/ir/it/
[06:39:49] <dclarke> or a kick
[06:40:12] <nrubsig> dclarke: hug it first!
[06:41:21] <dclarke> geez .. all it wanted was a correct password
[06:41:24] <dclarke> go figure
[06:41:31] <nrubsig> everyone gives dclarke a hug! 1... 2... 3... huuuuugggg
[06:41:34] <dclarke> sheesh .. it should just "know"
[06:41:39] <dclarke> stop that !
[06:41:56] <nrubsig> rotfl
[06:43:01] <dclarke> here .. do something useful and identify this : http://maps.google.com/?ie=UTF8&om=1&z=15&ll=40.884026,-72.875962&spn=0.019922,0.036564&t=k
[06:43:18] <dclarke> hint .. its not an impact crater
[06:44:03] <nrubsig> looks like a ring accelerator
[06:44:14] <dclarke> my thoughts exactly
[06:44:37] <dclarke> its waay too perfect a circle
[06:44:44] <dclarke> in the middle of nowhere
[06:45:02] <nrubsig> dclarke: and it is underground and has some labs around
[06:45:07] <nrubsig> and some injectors
[06:45:40] <Tpenta> that url crashed my firefox :(
[06:45:59] <dclarke> sorry
[06:46:08] <dclarke> Mozilla works well
[06:46:21] <dclarke> it was the accelerator that did it
[06:46:46] <dclarke> I seem to spend my time during long compiles looking at the earths surface for odd features
[06:46:56] <dclarke> things that don't look "right"
[06:47:20] <dclarke> and that definately looks out of place
[06:47:58] <timeless> if i try to move 28G from one zfs partition to another zfs partition in the same pool
[06:48:03] <timeless> and i only have 6G free
[06:48:09] <timeless> it won't work properly, right?
[06:48:10] <dclarke> my mozilla can not seem to download Sun Studio
[06:48:28] <dclarke> timeless : unless compression is really working well
[06:48:34] <timeless> dclarke: mozilla1.7 (not firefox)
[06:48:39] <timeless> ?
[06:48:52] <timeless> (firefox1.5.0.6 ~ mozilla1.8)
[06:48:53] <dclarke> compression ..
[06:49:16] <dclarke> if the destination ZFS filesystem has compression enabled
[06:49:19] <timeless> i know i can ask, but i suspect that the compression algorithm isn't as smart as i want it to be
[06:49:20] <dclarke> it may work fine
[06:49:27] <timeless> they all have it enabled
[06:49:32] <dclarke> its just a block algorithm
[06:49:38] <timeless> most files in this thing are duplicates of others in the same thing
[06:49:40] <dclarke> LZW I think
[06:49:54] <dclarke> it works real well
[06:49:57] <timeless> i want something that recognizes duplicate files and only wastes space once
[06:50:10] <timeless> but that's zfs / app hacking that i need to do later
[06:50:17] <dclarke> I think you want a hierarchal file system
[06:50:27] <timeless> HFS--?
[06:50:38] <dclarke> HFS was an IBM construct back in the days of mainframes
[06:50:48] <nrubsig> samfs
[06:50:48] <timeless> sounds like what i want
[06:51:10] <dclarke> something I used with ESA Systems on IBM 3090 mainframes
[06:51:11] <timeless> anyway, i have a zfs volume foopy and in it is barby
[06:51:14] <dclarke> its big money
[06:51:24] <timeless> can i move barby to be a parent of foopy and move foopy to be a child of barby?
[06:51:34] <dclarke> say what ?
[06:51:49] <timeless> root_pool/this/that
[06:51:52] <timeless> i want:
[06:51:55] <timeless> root_pool/that/this
[06:52:08] <dclarke> hrmmm ...
[06:52:13] <dclarke> tough to say
[06:52:19] <dclarke> I don't think you can do that
[06:52:28] <timeless> :(
[06:52:35] <dclarke> once the hierarchy is set and inheritance has taken place
[06:52:49] <dclarke> no .. you would need to backup this and that
[06:52:55] <dclarke> foopy and barpy
[06:53:05] <timeless> i don't have 30G to backup
[06:53:08] <timeless> i have 6G!
[06:53:10] <dclarke> then destroy and recreate with a new hieararchy
[06:53:29] <dclarke> so create a new filesystem .. do a backup
[06:53:46] <timeless> i don't have backup space. just a lousy 80g drive :(
[06:54:10] <timeless> oh well, the problems of an engineer w/ a small disk drive
[06:54:29] <dclarke> got tape ?
[06:55:04] <timeless> heh, no :)
[06:55:28] *** jerome__ has joined #opensolaris
[06:55:37] <timeless> so basically the one thing i really can't do is change my hierarchy at all?
[06:56:02] <dclarke> its tough to carry a bucket of water without a bucket
[06:56:15] * libkeiser mutters something about afs getting that right 20+ years ago...
[06:56:24] * libkeiser ducks and runs
[06:56:28] <dclarke> you need a third register to swap the contents of two registers without loss
[06:56:45] <dclarke> regardless of what the opcode says
[06:56:57] <jerome__> Quick question, I run build 50 on a blade 150 with a usb keyboard, and my keyboard has the wrong keymap, the keys are not where they're supposed to be, where can I change this? It's supposed to be a US keyboard (happy hacker keyboard) and it puts the signs at the wrong place, I don't even have ':' for instance.
[06:57:02] <timeless> i'm ok with incremental
[06:57:28] <timeless> try the shift-. key? :(
[06:57:35] <jerome__> snot working...
[06:57:41] <jerome__> go figure
[06:57:48] <jerome__> it display 'dell usb keyboard' when it boots
[06:57:58] * timeless has a dell usb keyboard
[06:58:01] <jerome__> but the keymap is totally wrong
[06:58:04] <timeless> unfortunately it's a scandinavian layout
[06:58:11] <jerome__> this keyboard works great on my linux box tho
[06:59:04] <rydis> Don't HHKBs have a dip-switch for Sun/PC/Mac, or was that the older ones?
[06:59:16] <jerome__> that was the older ones I think
[06:59:21] *** qbit has joined #opensolaris
[06:59:23] <jerome__> I've looked
[06:59:26] <jerome__> no dip
[06:59:27] <jerome__> for that
[06:59:53] <jerome__> OH
[06:59:56] <dclarke> I have to drive early in the morning
[06:59:58] <jerome__> bash-3.00# kbd -l
[06:59:58] <jerome__> type=6
[06:59:58] <jerome__> layout=33 (0x21)
[06:59:58] <jerome__> delay(ms)=500
[06:59:58] <jerome__> rate(ms)=40
[07:00:04] <dclarke> so .. thanks for the hugs !
[07:00:05] <jerome__> 33 is spanish???
[07:00:06] <jerome__> wtf
[07:00:07] <dclarke> gotta go
[07:04:53] *** jerome__ has quit IRC
[07:08:19] *** eboutilier has quit IRC
[07:10:11] *** jerome__ has joined #opensolaris
[07:10:26] <jerome__> hm there's a bug or something
[07:10:34] <jerome__> # kbd -s US-English
[07:10:34] <jerome__> # kbd -l
[07:10:34] <jerome__> type=6
[07:10:34] <jerome__> layout=33 (0x21)
[07:10:34] <jerome__> delay(ms)=500
[07:10:35] <jerome__> rate(ms)=40
[07:10:45] <jerome__> layout should show 41 (US-English)
[07:14:06] <jerome__> any clues?
[07:15:58] <gdamore> hey.  33 is US.
[07:16:26] <nrubsig> good night
[07:16:36] <jerome__> 33. Spanish
[07:16:43] <gdamore> hex or decimal?
[07:16:44] <jerome__> 41. US-English
[07:16:51] *** nrubsig was kicked by nrubsig (looser)
[07:16:56] <jerome__> layout=33 (0x21)
[07:17:08] <jerome__> 0x21 is the hex value
[07:17:29] <gdamore> these are USB country codes.  i'm 99% sure 33 is US, but hang on while i check the sources.
[07:18:01] <gdamore> yep.  its US.  as i thought.
[07:18:16] <gdamore> i actually have been in that code for a lot of today. :-)
[07:18:22] <jerome__> hm
[07:18:34] <jerome__> then I dunno why my keyboard has the wrong map
[07:18:35] *** bougie has joined #opensolaris
[07:18:36] <gdamore> you want spanish.  that is 25.
[07:18:40] <jerome__> no thanks
[07:19:10] <gdamore> did someone do KIOCSLAYOUT on your keyboard?
[07:19:18] <jerome__> nope
[07:19:26] <jerome__> I just installed b50 from scratch
[07:20:18] <gdamore> i've not played with b50 yet, i only just installed b49.  but b49 works fine with a type 6 keyboard.
[07:20:41] <rydis> Thing is he doesn't have a Type6, but a HHKB, which seems to be misidentified.
[07:20:48] <gdamore> is the problem just in the X server?  or also at the console?
[07:21:01] <gdamore> wtf is HHKB?
[07:21:02] <jerome__> hm
[07:21:07] <jerome__> I didn't check
[07:21:11] <rydis> "Happy Hacking Keyboard"
[07:21:13] <jerome__> Im away from my blade now
[07:21:52] <gdamore> hm... well there are a _lot_ of cheap keyboards that either don't identify themselves or misidentify themselves.
[07:22:07] <gdamore> the code 33 is used as a "default" in Solaris for this case.
[07:22:22] <gdamore> do you _think_ it has a spanish layout?
[07:22:30] *** DataStream has joined #opensolaris
[07:23:28] <jerome__> dunno
[07:23:33] <jerome__> I can check tomorrow...
[07:23:48] <jerome__> can I force it to be type US-English anyway?
[07:23:53] <gdamore> you can "force" a keyboard layout using KIOCSLAYOUT.
[07:24:06] <gdamore> (which you can probably do with kbd(1M))
[07:24:35] <delewis> kbd -s I believe
[07:24:58] *** bank has joined #opensolaris
[07:25:04] <bank> cd..
[07:25:28] *** sickness has joined #opensolaris
[07:25:50] <bank> Hello, I configure vhost in apache2 but I can't find /www
[07:26:04] <bank> Where does the default apache page located?
[07:26:20] <gdamore>  /var/apache/htdocs?
[07:26:41] <delewis>  /var/apache2 in the case of apache2
[07:26:45] <gdamore> it might also be /var/apache2/htdocs.
[07:26:59] <bank> Oh yes. why apache2 is separate in several location in solaris?
[07:27:17] <Error_404> anyone know if the determinate of a matrix is restricted to R, or are values in C fair game?
[07:27:17] <Triskelios> the other one is apache1
[07:27:20] <bank> I ever install from pkg.gz by myself. but everything seem to be in one location.
[07:27:49] <delewis> bank: yes
[07:27:50] <gdamore> jerome: you don't want to force it to be english, you want it to match the keystrokes that you're typing.  KIOCSLAYOUT doesn't reprogram the keyboard -- it just tells the usb keyboard module how to interpret the USB keycodes.
[07:28:05] *** sparc-kly__ has joined #opensolaris
[07:28:07] <delewis> Solaris still ships two Apaches (and rightly so, for users that still want to use Apache 1.3.x)
[07:28:19] <delewis> therefore separate htdocs and config dirs must be maintained
[07:28:22] <gdamore> i'm assuming this is a USB keyboard.  All bets are off for PS/2 keyboards.
[07:28:37] <Triskelios> it's confusing that apache1 automatically starts when a config file is created for it
[07:28:46] <gdamore> i think Sun should ditch Apache 1.3.x with SNV.  I can't see there is much point in keeping apache 1 around anymore.
[07:28:49] <delewis> Triskelios: that's how most daemons work in Solaris
[07:28:54] <delewis> named, smbd, etc.
[07:29:04] <delewis> at least in Solaris 9 where smf wasn't around :-)
[07:29:12] <delewis> and legacy services, nowadays.
[07:29:16] *** yongsun has joined #opensolaris
[07:29:33] <bank> a little more please..  If I mapping virtual host to some location such as /www/test  where should the www located?
[07:29:46] <bank> I remember it is in apache folder somewhere...
[07:29:55] <delewis> gdamore: probably -- vendors that ship applications that depend on Apache 1.3.x are probably shipping their own bundled Apache, anyway.
[07:29:58] <delewis> but you know Sun :-)
[07:30:05] <gdamore> yes.  and yes.
[07:30:06] <delewis> Apache 1.3.x will still be around in Solaris 22 :-)
[07:30:12] * Triskelios shudders
[07:30:23] <bank> Solaris 22. .
[07:30:40] <delewis> bank: yes, complete with quantum computing support
[07:30:47] <delewis> :-)
[07:30:51] <gdamore> i had to deal with apache/openssl madness when I was a part of Sun (for Sun Crypto Accelerator 1000).  years ago I couldn't want for Apache 1.3.x to get nuked.
[07:30:54] <Tpenta> do you really think we can go that many releases without a name change?
[07:31:26] <delewis> Tpenta: considering how much confusion the SunOS->Solaris change caused, I think so :-)
[07:31:27] <bank> umm... do you know where is /www located?
[07:31:34] <delewis> and that confusion (obviously) still lives on today
[07:31:52] <Tpenta> i'm thinking more 2.3 2.4 2.5 2.5.1 2.6 7 8 9 10 11
[07:31:59] <delewis> oh, I see.
[07:32:00] <gdamore> heh.
[07:32:09] <delewis> we'll hit roman numerals eventually.
[07:32:21] <Tpenta> thats exactly what i was thinking
[07:32:36] <delewis> why can't marketing come up with something cool like converging towards pi..
[07:32:45] <delewis> like LaTeX :-)
[07:32:56] <gdamore> there is code in the E10K controller that I modified to use SunOS 5.x numbering instead of Solaris 2.x/7 numbering.  I converted it after the Solaris 7 fiasco.  I never wanted the code to be at the mercy of marketing again. :-)
[07:32:57] <Tpenta> i had a friend who had e in his phone number
[07:33:05] <delewis> Tpenta: sweet :-)
[07:33:14] <rydis> That's one long phone number... ;)
[07:33:19] <bank> oh i see, we just map to any location in the Harddisk ... right?
[07:33:32] <Tpenta> well the first 8 digits
[07:33:33] <delewis> rydis: :-)
[07:33:47] <delewis> nonsense, 22/7 *is* pi, I swear.
[07:33:48] * delewis ducks
[07:33:49] <gdamore> actually, i suspect for any 7 or 10 digits, you can find them in sequence in digits of either pi or e, you just have to look far enough.
[07:34:02] *** lloy0076 has joined #opensolaris
[07:34:07] *** tsoome has quit IRC
[07:34:12] <Tpenta> someone put up a website for that
[07:34:13] <lloy0076> I'm trying to buld PHP with mysql support on Solaris Community Edition 10. In /usr/sfw/include/mysql/ there are what appears to be mysql header files. However, ./configure ... --with-mysql="/usr/sfw/include" doesn't find them, nor does --with-mysql="/usr/sfw/include/mysql". I've got the PHPv 5.1.6 from the mirrors...
[07:34:29] <Tpenta> delewis: didnt one state once legislate pi = 3 ?
[07:34:32] <gdamore> Tpenta: figures.  another not so great idea of mine, that someone else beat me to it.
[07:34:35] <delewis> gdamore: the digits of pi have failed the randomness test
[07:34:36] <gdamore> indiana.
[07:34:41] <delewis> Tpenta: yeah.
[07:34:44] <gdamore> actually, i think they legislated it to 4.
[07:34:54] <delewis> the egyptians had the same approximation 3,000 years ago :-)
[07:35:02] <delewis> (at the same time, the Chinese knew pi up to 8 digits)
[07:35:40] <bank> mysql?
[07:35:51] <lloy0076> Yes, mysql. The mysql extension.
[07:36:11] <gdamore> how do you know the digits of pi have failed randomness?  you obviously just haven't taken a large enough sample... :-)
[07:36:13] <lloy0076> I have a Mysql 4.0ish server running, and it does appear to have its library headers in /usr/sfw/include/mysql/
[07:36:16] <delewis> lloy0076: look at the contents of config.log and see what exactly is going on.
[07:36:48] <delewis> gdamore: it can be shown that a certain size sample is sufficient to prove {non}-randomness
[07:37:44] *** sparc-kly_ has quit IRC
[07:38:01] <gdamore> really?  i'm skeptical.  there are statistical tests that try to predict randomness/quality, but the thing is with random numbers, well, they're random.
[07:38:38] <lloy0076> delewis: There's no error at the bottom. All it says is: configure:58244: checking for MySQL support but doesn't say where or what it's trying to do.
[07:38:41] <delewis> gdamore: that may only be the case with a finite set, or perhaps the randomness test only used the set of known pi digits and a sample from that.
[07:38:45] <gdamore> e.g. if you search long enough, you could get a string of 1 million digits that are all "1".  it would fail randomness, but doesn't mean that your random source is bad.
[07:38:47] <delewis> I never bothered to read the analysis
[07:39:04] <delewis> and keep in mind, I've only had one statistics course in my life as a Mathematics major :-)
[07:39:54] <gdamore> heh.  I washed out of chemical engineering at UCB.   thermodynamics killed me.  went to a state school and got honors as a CompSci major. :-)
[07:40:13] * lloy0076 CONFOUND IT
[07:40:28] <delewis> ah, thermodynamics.
[07:40:28] <lloy0076> It wanted: --with-mysql="/usr/sfw"
[07:40:32] <yongsun> twincest, are you around?
[07:40:34] <delewis> helps to know a bit of statistics in that case :-)
[07:40:38] <twincest> hi
[07:40:39] <Error_404> so... back to my simple question... det(A) = elementOf( R ), or is C okay too?
[07:40:43] <timeless> has anyone here used gnome system monitor?
[07:40:46] <Error_404> for some matrix A
[07:41:15] *** Xh4 has joined #opensolaris
[07:41:31] <gdamore> actually, I can attribute my failure directly to Gibbs' Free Energy.  :-)
[07:41:38] *** tsoome has joined #opensolaris
[07:41:40] <yongsun> twincest, remember that you mentioned CDDL lib should not be linked with GPL code in linux, do you mean both Kernel and User lands?
[07:41:42] <delewis> Error_404: C would be acceptable if the matrix was complex (only in certain cases, though)
[07:41:43] <bank> My svcs -l apache2   go maintenance .  /  inspect on tail error_log there is nothing useful ...
[07:41:52] <Error_404> *nod*
[07:42:03] <delewis> you certainly can't say det(A), where A is a complex matrix is always in C (ex. Hermitian)
[07:42:32] <bank> oh i got /var/svc/log/network-http:apache2.log
[07:43:26] <delewis> ugh. 'make clobber' takes forever.
[07:43:56] <bank> can we spilt the httpd.conf? .... it very long to reach the <virtualhost> at the bottom of the file...
[07:43:59] <Error_404> the question in my homework was just conditions for n such that an nxn matrix A where A^2+I=0 holds
[07:44:04] <delewis> bank: yes
[07:44:08] <delewis> there's an include directive
[07:44:15] <delewis> though, I forget the exact name of it
[07:44:16] <jamesd_> delewis, that is why you take a snapshot before you build.
[07:44:23] <gdamore> woot!  woot!  my flash translation driver for SPI connected flash on NetBSD is working!
[07:44:29] <Error_404> if n is odd, det(-I) = -1, so det(A) = i
[07:44:34] <delewis> jamesd_: good idea
[07:45:04] *** Gman has quit IRC
[07:46:02] <gdamore> just need to add a simple redboot partition parser (to read a disklabel), and I'll commit the bad boy.
[07:46:30] <Error_404> it's probably a decent assumption that he didn't mean to bring C in to the mix, but i'll put a caveat on the question anyways
[07:47:35] <delewis> it's a shame that I've forgotten a lot of Linear Algebra, but the thing about Linear Algebra is *everything* is connected, so I figure when I take a look at it again, it'll all come back quickly.
[07:47:54] * Tpenta wishes that Acer didn't force the TI sd card reader in the fr4000 into the proprietry mode. It supports the standards, but it's a hard wiring thing to set it that way
[07:48:33] <bank> ah include directive!
[07:48:34] <delewis> Tpenta: isn't the ACPI implementation fairly proprietry, too?
[07:48:50] <Tpenta> not sure, all i know is that stuff works
[07:48:56] <delewis> oh, cool :-)
[07:49:06] <delewis> I'm considering getting a Ferrari 5000 next week
[07:49:13] <delewis> so, I've been doing a bit of research.
[07:49:17] <Tpenta> i have no idea abt the 5k
[07:49:50] <delewis> I think the 5k is identical to the 4k in most aspects (except the processor, default physical memory size, disk, and graphics card)
[07:50:08] <Tpenta> and a few extra gadgets
[07:50:11] <Tpenta> like built in camera
[07:50:22] <delewis> it'll be interesting to see, anyway, and if it turns out I'm wrong about that, I'll have something to contribute :-)
[07:51:08] *** bougie has quit IRC
[07:53:00] *** tsoome has quit IRC
[07:55:41] <bank> err...
[07:55:54] <bank> I already set the virtualhost but it still map to default apache page.
[08:00:16] *** sahafeez has quit IRC
[08:07:41] <bank> sorry. I will try more.
[08:09:20] *** tsoome has joined #opensolaris
[08:12:45] <delewis> boyd: *ping* you were right -- current AMD64 implementations only use the first 48-bits of a virtual address are actually used, so there is a hole.
[08:16:51] * delewis doubts anyone is going to whine in the near future about not being able to address more than 256TB of memory :-)
[08:23:21] <bank> delevis
[08:23:27] <bank> delewis
[08:23:51] <ShadowHntr> we'll be there in about 10 years. ;)
[08:24:25] <bank> If I have BIND run on global zone, and APACHE on non-global zone call site-zone. I have to configure on both zone in one time.
[08:24:43] <bank> If I would like to write the web application to automate this process.
[08:25:13] <bank> I should place this app on the app-server in the global zone to access and modify those configuration of apache in non-global zone ?
[08:25:37] <bank> for ex.   site-zone , ftp-zone , mail-zone live inside the global-zone that have BIND running.
[08:26:33] *** uebayasi has joined #opensolaris
[08:26:36] <gdamore> Tpenta: wrt SD card readers.  If you can get a USB controller that exposes them as mass storage great.  but otherwise you're screwed, cause the only way to get SDcard otherwise is to use very NDA information.
[08:26:58] * gdamore wrote the SDcard reader driver/Winbond nexus on SPARCLE.
[08:28:02] <Tpenta> and i thnk really bad manners on acers part to elect to use the closed proto rather than the standard
[08:28:38] *** tsoome has quit IRC
[08:30:04] *** lloy0076 has quit IRC
[08:31:23] *** bank has quit IRC
[08:32:58] <Triskelios> gdamore: someone did manage to RE winbond and sdhci before, though...
[08:35:13] <boyd> Anyone seen that the SUNWCrnet seems to be broken in b49
[08:35:47] *** tsoome has joined #opensolaris
[08:38:04] *** ShadowHntr has quit IRC
[08:40:18] *** Jiko_ is now known as Jiko
[08:42:10] * timeless sighs
[08:42:19] <timeless> each gnome status applet really takes 10-12mb?
[08:43:44] <gdamore> hmm... RE of the winbond and sdhci would be "challenging".  there are stuff in there that would require some good guess work.  e.g. there is stuff laid out on SDcard filesystems (partitioning) that is unlike what you'd find elsewhere.   the sdcard driver or usb mass storage adapter hides that from you.
[08:44:03] <gdamore> timeless: probably.  modern GUIs are very, very piggy.
[08:45:44] <gdamore> gnome-about is 12M on my system.
[08:45:51] <gdamore> (12M resident)
[08:46:48] <gdamore> i don't know if top counts shared pages properly though.  i sort of don't think so.
[08:47:31] <gdamore> speaking of which, it looks like i should think about restarting X soon.  its approaching 500M.
[08:51:03] *** DataStream has quit IRC
[08:51:08] * timeless sighs
[08:51:34] <timeless> is there any way for a user application to create files owned by 1000:1000 that point off into devices?
[08:51:48] *** DataStream has joined #opensolaris
[08:52:19] <gdamore> sure.  mknod/chown
[08:52:34] <gdamore> of course, the code has to run as root to do that.
[08:53:08] <timeless> user application=application running as user != root
[08:53:39] <gdamore> well, there is symlinks.  but to do mknod, you have to have root.
[08:55:15] <timeless> ok, dbx insists on ccrashing, i should be able to use mdb...
[08:58:47] <asyd> IE7 actually seems to support JavaScript much better which probably is partially the result of the warming relationship (sort of) between Sun and Microsoft.
[08:58:49] <asyd> \o/
[08:58:51] <asyd> http://news.digitaltrends.com/talkback152.html
[08:59:25] <timeless> asyd: eh?
[08:59:30] <timeless> how does javascript relate to sun?
[08:59:36] <timeless> ignoring the fact that sun owns the name
[08:59:46] <asyd> you should ask to the "Expert" who wrote the article :)
[08:59:51] <timeless> which is almost entirely irrelevant
[09:00:16] <timeless> there we go, " Sid on Oct 19th, 2006 at 7:19 PM"
[09:03:39] *** Jiko_ has joined #opensolaris
[09:03:41] *** snyff has joined #opensolaris
[09:03:59] <snyff> hi
[09:04:43] <asyd> hi snyff
[09:07:49] *** mazon is now known as Mazon
[09:11:19] *** Cyrille has joined #opensolaris
[09:17:09] *** Jiko has quit IRC
[09:18:39] *** tsoome has quit IRC
[09:19:11] *** mikaeld has quit IRC
[09:21:07] *** yongsun has quit IRC
[09:21:21] *** mikaeld_ is now known as mikaeld
[09:21:53] *** yongsun has joined #opensolaris
[09:32:55] *** Disorganized has quit IRC
[09:37:23] *** bengtf__ has joined #opensolaris
[09:37:27] *** bengtf__ is now known as bengtf
[09:51:18] *** MattMan has joined #opensolaris
[09:51:19] *** jcea has quit IRC
[09:52:52] *** yongsun has quit IRC
[09:59:09] *** jmcp has quit IRC
[09:59:23] *** jmcp has joined #opensolaris
[10:10:05] *** alanc_away has quit IRC
[10:10:25] *** yongsun has joined #opensolaris
[10:10:56] *** alanc_away has joined #opensolaris
[10:31:53] *** Gman has joined #opensolaris
[10:31:55] *** yongsun has quit IRC
[10:32:02] *** yongsun has joined #opensolaris
[10:41:33] *** tsoome has joined #opensolaris
[10:48:42] *** janus_man has quit IRC
[10:49:03] *** yongsun has quit IRC
[10:55:42] *** yongsun has joined #opensolaris
[11:01:13] *** peteh has joined #opensolaris
[11:01:53] *** dunc has quit IRC
[11:02:41] <PerterB> morning
[11:02:51] <asyd> hi PerterB
[11:04:11] <PerterB> oh...you know we were talking about free space with soft partitions? turns out metarecover will give you a nice listing of where the partitions and free space are
[11:04:36] <asyd> oh
[11:04:51] <quasi> so you just have to make a disk fail first? ;)
[11:05:04] <PerterB> metarecover, not metareplace :)
[11:05:49] <PerterB> eg http://pastebin.ca/211388
[11:06:18] <asyd> nice
[11:07:17] <PerterB> yeah
[11:07:46] <PerterB> now all I have to do is figure out how to remove the broken metaset I managed to get in my cluster config due to idiocy
[11:09:31] *** raph_ael has joined #opensolaris
[11:09:34] <raph_ael> hello
[11:09:44] <asyd> hi raph_ael
[11:09:51] <PerterB> salut
[11:09:56] <raph_ael> :)
[11:09:58] <raph_ael> salut asyd
[11:10:17] *** PosixCompliant has joined #opensolaris
[11:17:14] <raph_ael> I'm trying to use driver update floppies when installing the last solaris express, but it never finds the floppy, or doesn't recognize the DU
[11:17:32] <raph_ael> is there incompatibilites between DU floppies and solaris express ?
[11:20:09] <oxygene> sure that this isn't some oldboot driver?
[11:21:42] <raph_ael> well I'm quite new to solaris world
[11:21:59] <raph_ael> I got the floppy for solaris 10 on sun website
[11:22:54] <Cyrille> do we release Solaris 10 on floppies?!? I guess that's what the BlackBox must contain, all the floppies for the OS... ;-)
[11:23:24] <raph_ael> :)
[11:23:44] <raph_ael> it's a library of floppies the blackbox maybe
[11:24:14] <raph_ael> the floppy is just a driver, and i'm installing solaris express with cd's of course :)
[11:27:28] *** uebayasi has quit IRC
[11:27:38] *** yongsun has quit IRC
[11:28:02] *** yongsun has joined #opensolaris
[11:35:37] *** dunc has joined #opensolaris
[11:41:41] *** yongsun has quit IRC
[11:50:38] <PosixCompliant> I have an access to a solaris machine on which there is a solaris express; my sysadmin is not sure wheter it is Solaris Express Community Release or Solaris express ; is there a way to know it ? BTW: cat /etc/release shows:
[11:50:40] <PosixCompliant>                             Solaris Nevada snv_44 X86
[11:50:40] <PosixCompliant>            Copyright 2006 Sun Microsystems, Inc.  All Rights Reserved.
[11:50:40] <PosixCompliant>                         Use is subject to license terms.
[11:50:40] <PosixCompliant>                              Assembled 17 July 2006
[11:50:59] <PosixCompliant> is it SXCR 44 or solaris express 44  ?
[11:52:14] *** yongsun has joined #opensolaris
[11:59:21] *** sparc-kly_ has joined #opensolaris
[11:59:26] <PerterB> Well, that's what SXCR looks like, I would expect Solaris Express to have a slightly different /etc/release, but you never know
[11:59:35] *** chain-lightning has joined #opensolaris
[12:01:03] *** chain-lightning has left #opensolaris
[12:12:58] *** calumb has joined #opensolaris
[12:16:08] *** Dar_HOME is now known as Dar
[12:16:13] *** aijin has quit IRC
[12:21:43] *** sparc-kly__ has quit IRC
[12:23:38] <raph_ael> no idea on how to add drivers at install time ?
[12:24:25] <asyd> raph_ael: what kind of driver you need ? can't you load them from usb card after initial install ?
[12:28:43] *** janus_man has joined #opensolaris
[12:28:44] *** bougie has joined #opensolaris
[12:38:16] <jteo> hello all.
[12:39:21] *** Gr|ffous has joined #opensolaris
[12:41:22] *** Tzoa has joined #opensolaris
[12:44:19] <oxygene> hmm.. any idea how I can get lazy expanded variables in shell, similar to foo=bar assignments in make?
[12:47:20] <twincest> b=stuff; a='$b'; ...; eval c=$a
[12:48:39] <oxygene> hmm.. hardly transparent, but I guess that's the only reasonable way (apart from building my own parser for the config files I use here).. thanks
[12:49:10] <oxygene> hmm.. or - maybe I could just use make?
[12:55:41] *** Gman is now known as GmanAFK
[12:56:40] <raph_ael> asud lsi megaraid drivers, maybe i should try from a usb key
[12:57:07] *** PosixCompliant has quit IRC
[12:57:48] *** bruci has joined #opensolaris
[13:03:40] <raph_ael> asyd the porblem is I can't install as the disks are not viewed
[13:04:38] <jmcp> raph_ael: istr that there's an option early in the boot process which allows you to install drivers which are necessary for your machine
[13:09:52] <raph_ael> yes there is
[13:10:00] <raph_ael> but it doesn't see the floppy
[13:10:09] <raph_ael> on two different machines, which is weird :)
[13:20:36] *** calumb is now known as calLNCH
[13:20:48] <jmcp> bummer
[13:20:53] <jmcp> does it see the usb disk?
[13:21:04] <jmcp> s/disk/key|stick/
[13:22:03] *** calLNCH has quit IRC
[13:27:03] <timeless> hrm
[13:28:23] <timeless> so if i want the web server to have a higher priority than some 'mv' processes, and some x11 to be responsive to user input, but not steal cpu from mv while i'm not actively giving input, [how] can i do that?
[13:28:26] <raph_ael> jmcp: well I tried on two machines, but didn't knew how to investigate more
[13:28:52] <jmcp> raph_ael: um, hang on a bit
[13:29:05] <raph_ael> thanks btw
[13:37:33] <timeless> jmcp: is it possible that gnome-system-monitor's hard disk monitor applet view doesn't work w/ zfs?
[13:37:44] <jmcp> timeless: yes, it's possible
[13:37:47] <jmcp> I haven't tried it
[13:44:09] <jteo> mail.opensolaris.org is a much nicer interface than Jive.
[13:46:55] *** yongsun has quit IRC
[13:49:55] <timeless> mailman gets annoying fairly quickly for a certain class of user
[13:50:00] <timeless> but i can't remember which class that is
[13:50:11] <timeless> it might be the mmpr or something
[13:53:50] <jmcp> timeless: the clueful class of user?
[13:54:54] <timeless> i can't remember
[13:55:17] <timeless> one of the mail systems annoys admins
[13:55:20] <timeless> another annoys users
[14:01:53] <sickness> hi all
[14:02:31] * jmcp starts installing snv_50 on the lappy
[14:03:08] * sickness is about to liveupgrade on the zfs pr0n fileserver :P
[14:04:14] * tsoome just upgraded to snv_50 and now I need to build 64bit nss_winbind.so ....
[14:04:15] <jmcp> raph_ael: I just got to the {choose 1..6} screen after grub, chose "5 apply driver updates" and was prompted for cd/dvd | non-usb floppy | all other removable mass storage |all of the above
[14:04:48] <raph_ael> I got those too jmcp
[14:04:59] <raph_ael> well except the non-usb floppy
[14:05:00] <jmcp> and which option did you choose?
[14:05:08] <raph_ael> "floppy"
[14:05:15] <raph_ael> as it was a floppy :)
[14:05:31] <jmcp> which shows up fine under other OSes?
[14:05:39] <raph_ael> maybe I should try to use a usb key, mount it and load it manually if it's possible
[14:05:48] <jmcp> yes, I'd try that
[14:06:13] <raph_ael> I'll have to rtfm a little bit to find things :)
[14:06:18] <raph_ael> thanks a lot for your help
[14:06:23] <jmcp> you're welcome
[14:06:59] <jmcp> raph_ael: as long as your usb key is fat32-formatted then the installer ought to be able to find it
[14:07:13] <raph_ael> I'll try this way then :)
[14:07:29] *** qbit has left #opensolaris
[14:09:27] <icon> *yawn*
[14:09:28] <icon> morning all
[14:09:31] <jmcp> gday
[14:17:19] *** lplatypus has joined #opensolaris
[14:20:46] *** coffman has joined #opensolaris
[14:30:00] *** piwi has joined #opensolaris
[14:31:45] *** piwi has quit IRC
[14:36:46] <oxygene> where can I find information about the interaction of zones with packages (for package builders)?
[14:38:20] *** JohnDoe8 has joined #opensolaris
[14:38:25] <jmcp> oxygene: in the packaging part of the admin guide
[14:38:39] *** JohnDoe8 has quit IRC
[14:39:19] <oxygene> "Application Packaging Developer's Guide" is where I'd look for it.. admin guide for zones?
[14:39:30] <jmcp> Solaris Containers guide
[14:39:51] <oxygene> thanks
[14:41:30] *** calumb has joined #opensolaris
[14:41:40] <rodrickbrown> oxygene, nothing really changed
[14:41:47] <rodrickbrown> other than the new -g option i believe
[14:41:52] <rodrickbrown> to make packages zone aware
[14:42:01] *** laca has joined #opensolaris
[14:43:04] <quasi> there's a bit about package flags as well to show if they are global only or installable in local zones as well
[14:43:41] *** lplatypus has quit IRC
[14:44:33] <oxygene> rodrickbrown: I need to know about stuff like where to put configuration (/opt/*/etc is obviously a bad place, so I chose /etc/*/ for now, but that's a standard violation again, yada yada). interaction of class action scripts and postinstall etc. is interesting, too (what if they try to modify files on an inherited filesystem, which is a) unnecessary, b) impossible. how to stop them?)
[14:44:52] <oxygene> rodrickbrown: the structure didn't change, but there are several things that must be taken into account now that weren't important before
[14:45:33] <oxygene> (as for modifying files, the 4front oss package is a good example of zones-unawareness; or any of my pmpkg files that handle gnu info stuff)
[14:45:44] <jmcp> oxygene: doesn't the guide mention /opt/[stockticker]package/etc  ?
[14:46:15] <oxygene> jmcp: inherit-pkg-dir -> read-only, and global configuration (instead of per-zone).. bad idea
[14:46:28] * jmcp starts paying attention
[14:46:29] <jmcp> good point
[14:46:51] <rodrickbrown> oxygene it depends on the package
[14:47:02] <rodrickbrown> inherit-pkg-dir is a bad idea in general I think
[14:47:03] <rodrickbrown> disk is cheap
[14:47:26] <rodrickbrown> everyone zone should really be self contained
[14:47:42] <oxygene> it's a maintenance issue, too - one update vs. 10^x
[14:48:01] <quasi> rodrickbrown: if you don't have zfs cloning, then no inherit-pkg-dir on /opt will mean copying bunches of files
[14:48:04] <oxygene> oh, and 1000 full zones make disks look less cheap, too ;)
[14:48:22] <quasi> get a thumper ;)
[14:48:35] <rodrickbrown> if your running 1000 zones on one box you got much bigger problems than disk :-)
[14:49:09] <oxygene> not to forget that this translates to ram multiplication, too: inherit-pkg-dir should allow the kernel to share read-only sections of apps between zones
[14:49:33] <oxygene> and even with 20 zones, it's a difference between 20 full instances of apache in memory or 1 + 20 times data stuff
[14:50:29] <rodrickbrown> yeah but then what happens when 1 person needs to go to PHP6/7 which requires apache 2.X.Bleeding
[14:50:53] <rodrickbrown> we tested it for a while and found it, it just became too much work later down the line to make everyone happy
[14:51:02] <oxygene> install a full zone for them (conversion spare->full would be a nice feature for that)
[14:51:02] <rodrickbrown> so we just keep everyone isolated
[14:51:24] <rodrickbrown> then in the end you end up with 18 full zones :-)
[14:52:01] <oxygene> depends on what you do.. currently, our setup is one apache with vhosts, even spare zones are a huge improvement and our clients don't care about php versions usually
[14:52:52] <oxygene> anyway, the feature (spare zones) is there, and I want my packages to work with them - so, saying "you shouldn't do it" is nice and all, but I still think that the standards should handle those scenarios
[14:58:21] <rodrickbrown> seems like unneccessary overhead to seperate web servers in general
[14:58:40] <rodrickbrown> each customer gets their own zone? and apache instance ?
[15:00:30] <quasi> rodrickbrown: that's actually quite a nice feature - making sure nobody steals other customers data
[15:00:49] <rodrickbrown> proper unix file permissions can do that too :-)
[15:00:56] <quasi> rodrickbrown: no
[15:01:01] <oxygene> yes.. or one of those many "backdoor-by-default" php scripts affects only the customer who installed it
[15:01:23] <oxygene> had our share of embarrassing (but harmless) incidents with those already
[15:01:48] <quasi> rodrickbrown: take the instance where a script is saving submitted data - someone else could just write another script to read it
[15:02:00] <oxygene> defacement of several dozens sites because some tool did a system($_GET["whatever"]);
[15:02:54] *** Xh4 has quit IRC
[15:03:33] <quasi> oxygene: but at least they only get their own sites defaced when they're locked in a zone
[15:03:38] <oxygene> yes
[15:03:48] <oxygene> that's the idea :)
[15:03:56] *** calumb has quit IRC
[15:03:59] <quasi> "containing stupidity" ;)
[15:06:04] <quasi> oxygene: and I've been fiddling with letting those zones only have non-routable ip and using apache to proxy - that makes hacks even harder
[15:06:23] <oxygene> that's my plan, too
[15:07:52] <icon> hrmm
[15:08:12] <icon> quasi: mess with zones much?
[15:08:30] <oxygene> even if they manage to plant some eggdrop (or whatever ircbot - yep, that syste() issue was really "fun"), it's utterly useless
[15:12:51] <icon> alright... has anyone messed with zones much :)
[15:13:55] *** MattMan is now known as MattAFC
[15:15:48] <quasi> icon: a fair bit - http://soulfood.dk/archives/2006/03/08/T16_17_05/index.html ;)
[15:16:01] <icon> had some questions for you if you dont mind :)
[15:16:25] <icon> ive only run stuff in the global zone to date, but im wanting to consolodate my main workstation and server
[15:16:31] <icon> and just pickup a ray
[15:16:54] <icon> thing is, i dont want them trouncing each other. ive done a bit with bsd jails, im assuming zones is similiar
[15:17:19] <quasi> zones have somewhat better isolation
[15:17:27] <icon> so, would it make sense to setup a box with a minimal global zone with two local zones, one for the server, one for the workstation?
[15:17:48] <quasi> sure - that's certainly doable
[15:18:02] <icon> k
[15:18:13] <icon> how nasty of a hit to performance is it?
[15:18:21] <trygvis> not much
[15:18:22] <icon> its more disk cost, but is there any other overhead?
[15:18:33] <icon> i wouldnt think there would be
[15:18:34] <quasi> almost no overhead
[15:18:40] <icon> okay
[15:19:00] <icon> so are the filesystems kept separate from each other in the zones?
[15:19:15] <quasi> depending on which dirs you inherit, it costs somewhere from a few hundred megs to a few gigs
[15:19:19] <quasi> yes
[15:19:25] <icon> k
[15:19:40] <icon> how nasty would it be to just do a minimal global, and a fully decked out global for dev?
[15:20:01] <quasi> if you have inherited dirs, they will be mostly read-only mounts of the global zone
[15:20:09] <quasi> you can't have two global zones
[15:20:16] <icon> right
[15:20:23] <icon> sorry i meant local, not global :)
[15:20:28] <icon> havent had my coffee yet :)
[15:20:30] <quasi> there's only one and it isn't really a zone
[15:20:56] <icon> what makes sense to inherit?
[15:21:01] <quasi> speaking of coffee - I've got to run
[15:22:46] <oxygene> icon: if your global zone has a full install, you could inherit /usr into both server and desktop (if appropriate for your needs)
[15:22:56] <icon> hmm
[15:23:06] <icon> im almost tempted to do no inheritence (i know its a bitch for patch management)
[15:24:04] <icon> question:
[15:24:25] <icon> what happens if i inherit /bin and i install a package on a localzone that adds a new file into /bin
[15:25:55] <oxygene> you can't
[15:26:00] <icon> ahh
[15:26:26] <oxygene> which might be a good argument for full zones for your scenario
[15:26:42] <icon> hm
[15:27:18] <icon> im wondering if it makes more sense to have the server as the global and a workstation as a local
[15:27:23] <icon> or just having a single global with two locals
[15:27:33] <oxygene> in this case I'd go for two locals
[15:27:38] <icon> why?
[15:27:57] <oxygene> simply because the global zone, if hacked, is capable of affecting the local one
[15:28:10] <icon> good point
[15:28:52] <oxygene> some kernel hack will end any zones isolation as well, but that's a smaller attack vector than "any service with root permissions"
[15:29:12] <icon> hrm
[15:29:19] <icon> i suppose i could start messing with zfs too
[15:29:24] <icon> thats been on my list for a while
[15:29:49] <icon> i wonder if it is beneficial to share /home
[15:30:04] <oxygene> for /home, I'd use loopback nfs mounts
[15:30:14] <trygvis> why not lookup lofs mounts?
[15:30:33] <oxygene> with 10u2+latest patches or 10u3, you can use zfs datasets - ie. delegate control over subtrees of your zfs hierarchy to zones
[15:30:34] <trygvis> I'm using loopback nfs myself now, but want to try to use lofs instead
[15:30:35] <oxygene> quite nice
[15:30:48] <icon> im using loopback nfs right now
[15:31:23] <oxygene> won't help with sharing /home, though (but with organizing your zone filesystems)
[15:31:56] <icon> so loopback nfs can be used on any zone then?
[15:34:27] *** axisys has quit IRC
[15:38:34] *** rodrickbrown has quit IRC
[15:43:10] *** axisys has joined #opensolaris
[15:45:06] *** mage2 has joined #opensolaris
[15:45:15] <mage2> Hellloooo :) I am back
[15:48:38] <mage2> Anyone awake yet?
[15:49:14] * PerterB yawns and stretches
[15:49:26] <PerterB> I was just about to go for my Friday afternoon nap
[15:49:31] <PerterB> (if only)
[15:49:39] *** eboutilier has joined #opensolaris
[15:49:48] <mage2> haha
[15:49:57] <mage2> so i have yet another problem
[15:50:07] <oxygene> hello eboutilier
[15:50:22] <mage2> does dladm  support bcme drivers?
[15:50:35] <eboutilier> hey oxygene
[15:50:46] <mage2> meaning i am using bcme (broadcom drivers) and need to create vlans
[15:50:51] <mage2> is that possible?
[15:50:58] <jteo> eboutilier: +1 for the "ports knowledge database thingy"
[15:51:14] <PerterB> "ports knowledge database thingy"?
[15:51:19] <eboutilier> jteo: Thanks. :)
[15:51:40] <icon> ports knowledge base?
[15:51:42] <oxygene> eboutilier: I'm just splitting pmpkg in nice little units, esp. the various "magic" data gathering parts. just in case you want to reuse some of them
[15:51:47] <eboutilier> http://blogs.sun.com/eric_boutilier/entry/an_opensolaris_ports_freeware_idea
[15:52:53] <eboutilier> oxygene: Cool. So people can more easily use it as a learning tool too?
[15:53:15] <eboutilier> oxygene: Or a resource for re-using code?
[15:53:17] <icon> ahh
[15:53:22] <icon> porting 'code' :)
[15:53:29] <icon> i misread that too the first time around
[15:53:35] <eboutilier> Hi icon
[15:53:38] <oxygene> eboutilier: should be easier to understand, too
[15:53:39] <icon> morning
[15:54:03] <icon> it would be really nice to have a porting handbook so to speak
[15:54:10] <eboutilier> icon: Yeah, the word ports carries a lot of pre-conceived meaning these days. My bad.
[15:54:17] <icon> ie: welcome to the neverending fun of not having ldconfig :)
[15:55:14] <eboutilier> oxygene: Excellent. BTW, when I first started playing w/ pmpkg, I learned,e.g.  how to use ld to check for dependencies by reading your code.
[15:55:21] <oxygene> eboutilier: in fact, it became hard for me to keep in mind what happens when, so now it's one file/directory-of-files with script snippets, each with a single purpose (eg. fetch ELF dependencies, inherit permissions from live filesystem, ..)
[15:55:57] <quasi> "Sun Solaris containers - is that project blackbox? (quoting a former co-worker ;)
[15:55:57] <eboutilier> icon: Online porting "handbook". I like that...
[15:56:15] <icon> eboutilier: not mine, its freebsd's, but it applies more towards porting the os to other archs
[15:56:40] <icon> integration handbook?
[15:56:42] <icon> *shrug*
[15:57:19] <jteo> even just a short snippet showing the magic configure flags would be a great help.
[15:57:24] <axisys> what is a way to clone a OS that has ZFS partition?
[15:57:26] <mage2> Any one have to work with bcme drivers and vlans?
[15:57:31] <jteo> as opposed to googling mailing lists.
[15:57:51] <axisys> jumpstart w/ flar wont take care of ZFS i know
[15:57:59] <icon> eboutilier: we could definitely use something like that for the ports collection
[15:58:00] <eboutilier> oxygene: The thing I like best about pmpkg is that if a source package builds transparently, one can add it to pmpkg in about 2 minutes.
[15:58:42] <icon> afk
[15:59:36] <eboutilier> icon (steve S., correct?): Sounds good. An ideal knowledgebase would be useful to _all_ Solaris porting/repo/DIY/etc. interests -- is the way I'm looking at it.
[16:00:01] <oxygene> eboutilier: otherwise it takes 2 minutes + the time needed to add a patch to files/ to make it run :) that was one of the goals (I did debian packages for a while, and back then, the tools just sucked)
[16:00:40] <eboutilier> icon: But I should qualify that by saying, with a bias towards SunOS 5.11 (Nevada)
[16:01:27] *** solaris-user has joined #opensolaris
[16:02:20] <eboutilier> oxygene: Yeah, there's nothing quite like it. And your design goal of minimal non-SUNW package dependencies and focusing on SX also makes it stand out, IMO.
[16:05:43] *** loke has joined #opensolaris
[16:05:44] <oxygene> right now I focus on solaris10. php5 didn't properly enter the repo until recently because libxml2 is too old there - coffman added both libxml2 and php5 to remedy this. of course, php5 builds with a single line pkg.mk change without our libxml2 on SX :)
[16:06:02] <asyd> php5 works good with pkgsrc :P
[16:10:20] <eboutilier> oxygene: Interesting. well, my using it exclusively on Nevada w/out noticing that it's developed for 10 is just another good testimonial...
[16:10:35] <eboutilier> Or a sign that I'm not very astute? :-)
[16:10:39] <oxygene> heh
[16:10:45] <jteo> :)
[16:11:02] *** Griffous has joined #opensolaris
[16:11:03] *** Gr|ffous has quit IRC
[16:11:17] *** stefanjo has quit IRC
[16:11:35] *** Trinisan_ has joined #opensolaris
[16:11:46] *** Trinisan has quit IRC
[16:12:07] <mage2> Anyone have experiance with bcme drivers and vlans?
[16:16:13] <axisys> how to clone? disk 1 = slice0 ufs, slice1 swap, slice6 ufs and slice7 zfs; disk 2 sitting on the same X2100
[16:17:08] *** Mazon is now known as mazon
[16:17:46] *** loke has quit IRC
[16:18:28] <trygvis> axisys: dd if=/dev/rdsk/c0t0d0s2 of=/dev/rdsk/c1t0d0s2 should work, at least for the ufs partitions
[16:18:43] <trygvis> but you need to make sure they're not in use which might be hard
[16:18:56] <trygvis> usfdump and restore might be a better option
[16:19:03] *** bruci has quit IRC
[16:19:06] <trygvis> dunno if the zfs pool will like it
[16:19:20] <trygvis> the process is documented pretty will in on the of the admin guides
[16:19:26] *** qufk has joined #opensolaris
[16:19:46] <axisys> how about zpool mirror slice7 of disk2
[16:19:47] *** kFuQ has quit IRC
[16:19:55] <axisys> wil it copy whole slice?
[16:20:12] <axisys> and disksuite mirror the other slices?
[16:20:41] *** qufk is now known as kFuQ
[16:21:27] *** Belgar has joined #opensolaris
[16:22:08] <axisys> man we really need either zfs option during reboot OR a dd like tool take copy both zfs and ufs
[16:22:14] *** Disorganized has joined #OpenSolaris
[16:22:19] <axisys> s/rebbot/install/
[16:22:32] <axisys> s/reboot/install/
[16:22:50] <axisys> i have to build 10 of those second disks
[16:23:08] <Belgar> can you dd a zfs fs?
[16:23:16] <trygvis> dunno
[16:23:19] <axisys> Belgar: nawp
[16:23:36] <eboutilier> whoops, just need to insert a correction: Earlier I referred to pmpkg using ld to check for dependencies... wrong, it uses ldd.
[16:24:13] <Belgar> anyone managed to compile  emacs on Sol 10?
[16:24:47] <axisys> pkg-get install emacs :-)
[16:24:58] <Belgar> i wanted to modit and need it as a tarball
[16:24:59] <Belgar> :)
[16:25:33] <PerterB> at the C level? most emacs mods are just in lisp, no?
[16:25:36] <eboutilier> Belgar: emacs is also in CSW's (blastwave) GAR system...
[16:25:51] <PerterB> and on sunfreeware, but has about 10 dependant packages
[16:26:00] <axisys> trygvis: i can do it in single user mode, the dd that is for ufs slices.. that should catch all I think
[16:26:12] <eboutilier> And althought its still at beta stages, that means that all the build parameters being used to create the CSW emacs package are "open".
[16:26:14] <axisys> trygvis: the master system is not in production yet
[16:26:33] <Belgar> PerterB: so it should be possible to compile it ;)
[16:27:01] <PerterB> ah ok
[16:34:20] <eboutilier> IOW, connecting the question "anybody know how to build emacs..." to the discussion earlier about the need for a knowledgebase: Although it's in machine rather than human readable form, the following files contain all the raw material for a knowledgebase entry (i.e. per my blog post) for emacs (albeit for Solaris 8):
[16:34:25] <eboutilier> http://svn.blastwave.org:5957/csw/trunk/utils/emacs/
[16:37:08] <Belgar> hmm
[16:37:14] <mage2> hmm
[16:37:15] <mage2> ?
[16:37:20] <coffman> eboutilier: yeah i like to have that to
[16:39:14] <coffman> there is no point on dinging how to fix a app to build under solaris if some one else only did the work (only cause i dont like the packet/port system or so)
[16:40:03] *** karrotx has quit IRC
[16:40:20] *** MattAFC is now known as MattMan
[16:40:47] *** karrotx has joined #opensolaris
[16:40:54] *** karrotx has quit IRC
[16:40:57] *** karrotx has joined #opensolaris
[16:42:39] *** karrotx has quit IRC
[16:42:42] *** karrotx has joined #opensolaris
[16:42:49] *** Trinisan_ has quit IRC
[16:42:55] *** Trinisan has joined #opensolaris
[16:43:36] <mage2> Anyone have experiance with bcme drivers and vlans?
[16:48:00] <axisys> Belgar: i was wrong... dd src dest will work - Darren Moffat from SUN just explained..
[16:48:10] *** solaris-user has left #opensolaris
[16:48:19] <Belgar> it should ;)
[16:48:34] <axisys> Belgar: look for the zfs-discuss email w/ subject Cloning a disk w/ ZFS
[16:48:37] <Belgar> i just haven't got around to try it a lot
[16:48:46] <axisys> Belgar: sorry for the confusion
[16:49:44] <Belgar> shift happens
[16:49:53] <eboutilier> coffman: Totally agree. Having open source is only half the battle. Good (and easily findable/searchable) documentation is what makes things truly open. (Attribution: I stole that observation from a post I read by Casper a long time ago... it was an online debate -- including why Solaris is actually more open than most (all?) open-source OSs, because of Solaris' relative superiority in technical documentation.)
[16:50:49] <Error_404> what's funny is i pronounce "ZFS" as "zed-eff-ess", but brandz as "brand-zee"
[16:51:13] <coffman> my move to solaris was less technical - more about quality
[16:51:35] *** loke has joined #opensolaris
[16:51:51] <Belgar> loke :)
[16:52:02] <Belgar> loke; what are you doing in the land of the lah?
[16:52:03] <loke> hej
[16:52:19] <loke> working hard, without any thought of personal gain. Like any good worker should
[16:52:26] <Belgar> hehe
[16:52:30] <loke> you?
[16:52:35] <oxygene> eboutilier: and that you can reuse stuff from one version to the next (see you using pmpkg on SX) helps because the interfaces are stable helps keeping things open
[16:52:37] <Belgar> stuck in sweden
[16:53:24] <loke> Belgar: that must suck
[16:53:31] <Belgar> loke; indeed. In particular this time of year
[16:53:34] <loke> Belgar: didnt you have a wife from here?
[16:54:02] <eboutilier> coffman: Cool. And I'd say it's not accident either that Solaris is known both for quality and comprehensive technical documentation... one begets the other and vice versa?
[16:54:03] <Belgar> Malaysia
[16:54:12] <loke> Belgar: ah
[16:54:42] <Belgar> .. a bit to many muslims there though..
[16:54:43] <eboutilier> oxygene: Very true too.
[16:55:04] <coffman> yeah
[16:56:22] *** kgrimm has joined #opensolaris
[16:57:57] *** Trinisan_ has joined #opensolaris
[16:58:26] <eboutilier> Recently on opensolaris-discuss someone said "I'm new to Solaris, coming from a Windows admin background, where should I start documentation-wise?" And you had to love ux-admin's reply, which was basically: Simple, everything you need to know is at docs.sun.com. Read it, love it, learn it.
[16:58:39] <eboutilier> Although I'm not sure that's where I'd send a total newbie...
[16:58:40] <loke> Belgar: they are less strict here :-)
[16:59:11] <Belgar> loke: indeed :)
[16:59:30] *** janus_man has quit IRC
[16:59:43] <Error_404> makes it easy to get in to contributing too...
[16:59:54] <Error_404> i'm thinking "solaris internals" specifically
[16:59:55] <loke> Belgar: I like the cultural diversity. And people are very tolerant
[17:00:05] <eboutilier> The fact is, docs.sun.com, a resource that Sun and Solaris engineering takes extremely seriously, has _incredible_ depth and breadth.
[17:00:22] <Belgar> loke: yup, and its about the total opposite of .my
[17:00:32] <coffman> while coming  from windows/linux it was pretty shocking to have documentation from the vendor it self :)
[17:00:43] <eboutilier> LOL!
[17:00:48] *** mazon is now known as Mazon
[17:01:23] *** Trinisan has quit IRC
[17:02:14] <eboutilier> (geez, I'm staring to sound like a plant sent here from marketing. :) )
[17:02:23] <icon> ouch
[17:02:30] <eboutilier> :)
[17:02:50] <icon> docs.sun.com is great, but the format is a little rough on the eyes at 2am :)
[17:02:57] <icon> but in all , its a fantastic resource
[17:03:09] <oxygene> and it's way to much to send it to the printer
[17:03:16] <icon> nah
[17:03:19] <loke> Belgar: your wife is chinese?
[17:03:20] <icon> print it out at work :D
[17:03:34] <Belgar> loke; indian
[17:03:39] <loke> ah
[17:04:05] *** stefanjo has joined #opensolaris
[17:04:20] <coffman> eboutilier: no worrys, oxygene is always telling me i should do marketing rather then being a admin :)
[17:05:01] <oxygene> coffman: additionally to being an admin.. it's just amazing how quick you get into marketing mode :)
[17:05:33] <eboutilier> Back in about 1990, I worked at a place where we bought all the  Sun docs, and had them organized in binders on a waist-high shelf designed for that purpose (reference). It was at least 10 feet long, no joke.
[17:06:22] <icon> eboutilier: i do something similar here with mks crap
[17:06:30] <icon> its maybe a foot or so, but its still a ton of docage
[17:07:15] <eboutilier> coffman: I was a product manager in Solaris marketing for three years. Turns out, there are many extremely smart technical Solaris people there, e.g. Larry Wake just for starters.
[17:07:38] <Belgar> Hmm.. When did you quit eboutilier?
[17:08:52] <coffman> hehe
[17:08:53] <eboutilier> icon: Yeah, I'd still opt for having physical docs as much as possible, if it weren't so impractical relatively speaking these days.
[17:09:24] <icon> eboutilier: true enough. i usually only print out stuff that requires a lot of reference. im working on writing a new driver for mks, so im constantly in that damnedable cli doc
[17:09:44] <icon> the first mks developer i come across i am going to strangle
[17:09:46] <coffman> im not smart :) but i have to sell the shit that i do directly to my bosses - i work in a realy small company
[17:10:08] <eboutilier> Belgar: I was a product manager (which falls under marketing at Sun) from about 2000 - 2003.
[17:10:26] <Belgar> eboutilier: ok, i worked at sun service 00-05
[17:11:05] <eboutilier> icon: :-) Let me know us know if you need an alibi.
[17:11:13] <icon> haha will do
[17:11:46] <icon> eboutilier: got a quick ports question for you
[17:12:25] <eboutilier> Belgar: Cool, I was also an SE from 1997 - 2000, so I worked a lot with the sun service folks here in Chicago
[17:12:32] <icon> ive talked with danek off and on, and he brough up a good point that several packages wont live in SUNW space, would it make sense to publish a list of aceptable package prefixes?
[17:13:22] <icon> ie: GNUautomake
[17:13:40] <icon> i think laca converted the pkgbuild repo to use SFE now as well
[17:13:47] <eboutilier> icon: Hmm, I think I know what you're talking about... e.g. GNU tar can't live in /usr/bin as /usr/bin/tar, etc., correct?
[17:14:13] <jengelh> no. On BSDs it lives as /usr/bin/gtar
[17:14:16] <jengelh> (and solarises)
[17:14:18] <icon> well, say there is a gnu tar port. the ports collection will actually build a solaris package to install
[17:14:24] <icon> bsd's they are /usr/bin/tar
[17:14:31] <eboutilier> Oh, you're talking about the sanctity of using the SUNW prefix for packages.
[17:14:34] <icon> so each port will need an actual package prefix
[17:14:36] <icon> exactly
[17:14:47] <icon> which btw, /usr/gnu really needs to happen
[17:15:08] <gdamore> woot!  my first opensolaris code contribution is committed.  (kinda trivial, PSARC 2006/510)
[17:15:22] <icon> gdamore: grats ;)
[17:15:48] <Belgar> whats wrong with /usr/sfw/ though?
[17:16:19] <gdamore> now if only i can get the bigger stuff i want to commit done.  mostly waiting for confirmation that my company has submitted an SCA.
[17:16:21] <icon> because gnu is an active replacement for several 'core' commands
[17:16:30] <icon> just like xpg4 or xpg6
[17:16:41] <icon> not all gnu stuff, just things like tar etc.
[17:16:42] *** schily__ has joined #opensolaris
[17:17:19] <eboutilier> icon: Yes, anything but using SUNW. That's only for packages that get into Solaris (past the ARC, etc.). Even Sun's Companion CD doesn't use the SUNW prefix.
[17:17:30] <icon> ahh
[17:17:33] <icon> what does the scd use?
[17:17:57] *** tsoome has quit IRC
[17:18:33] <gdamore> i thought the packaging standard was that if the packaging was done by sun, that it should be SUNW.  i.e. you use the stock ticker of the company doing the packaging.
[17:18:41] <jengelh> solaris gets confused when woken up from vmware suspend
[17:18:46] <icon> eboutilier: its really on the honor system for the most part. if jds is moved into ports, obviously theyll need to support using SUNW package names.
[17:18:49] <eboutilier> The trick to Sun's plan for /usr/gnu is it's not really a place to put FOSS/GNU packages, it only exists (will exist) as a place to relegate FOSS/GNU files when there's a name conflict in /usr/bin, and that's the only time it's to be used (I think).
[17:18:54] <Belgar> i never managed to get my laptop to wake up after being suspended
[17:18:56] <jengelh> WARNING: Time of Day clock error: reason [Jumped by 0x48c3c]. -- Stopped tracking Time Of Day clock.
[17:19:07] <icon> eboutilier: right
[17:19:16] <icon> eboutilier: at least thats what i understood from the project page
[17:19:29] <eboutilier> icon: Not sure what the scd is?
[17:19:43] <eboutilier> Oh, companion CD.
[17:19:43] <jengelh> scsi cd?
[17:19:48] <icon> sorry, shorthand :)
[17:19:58] <eboutilier> It uses SFW (which was not a good idea, in retrospect.)
[17:20:03] <icon> ahh okay
[17:20:28] <icon> well, i think it would make sense to organize prefixes, but there would still need to be a root 'generic' prefix for a port
[17:20:47] <icon> ie: gnu under GNU, sun under SUNW, and so on
[17:21:15] <icon> otherwise youll have everyone and their mother pimping their own prefix, and it will really pollute the tree
[17:21:32] <delewis> a script to change the links in /usr/bin to /usr/xpg4, /usr/sunw, or /usr/gnu would be cool :-)
[17:21:44] <icon> delewis: why not just change your PATH?
[17:22:08] <icon> set path = (/usr/xpg4 $path) ?
[17:22:11] <delewis> icon: that's exactly what I do.
[17:22:28] <icon> delewis: why mod the links then? (just curious)
[17:22:45] <eboutilier> icon: I think what you're saying depends on whether the ports project aims to be a conduit for full integration of its packages into Solaris. That's a lofty goal to say the least...
[17:22:54] *** mwaterl has joined #opensolaris
[17:23:14] *** bengtf has quit IRC
[17:23:27] <icon> eboutilier: its a primary goal
[17:23:33] <icon> and actually not that tough
[17:23:46] <icon> the pkg utils deal with most of the pain
[17:24:28] <eboutilier> Sounds good. I'm game.
[17:24:35] <jengelh> let me be your prey :D
[17:24:40] <icon> so for examlple, say you issue 'port install tar', a GNUtar package is built and installed
[17:24:45] <Belgar> laters folks!
[17:24:46] *** Belgar has left #opensolaris
[17:25:01] <icon> its silly to ignore solaris packaging when sun has invested so much in it
[17:25:21] *** schily_ has quit IRC
[17:25:40] <icon> so with additional repository support, sun could even have their own officially supported packages with patchsets
[17:25:47] <icon> or just distrubute the built packages
[17:25:56] <icon> and everyone lives in the same happy world ;)
[17:26:04] <eboutilier> Here are the prefixes in use for Solaris freeware that I'm aware of: SUNW, SFW, SFE, CSW, SMC, PMP(?) (oxygene... help)
[17:26:26] <icon> im familiar with the first 4
[17:26:31] <icon> what is SMC and PMP used with?
[17:26:35] <delewis> there's also DEL -- my personal repository :-)
[17:26:36] <jengelh> I'll declare SOW for me!
[17:26:39] <delewis> SMC is Sun Freeware
[17:26:40] <icon> haha
[17:26:44] <icon> ahh thats right
[17:26:51] <icon> its been years since ive used them
[17:27:13] <icon> well, id like to see just a few that we use in the repo
[17:27:23] <delewis> $ pkginfo | grep DEL | wc -l
[17:27:24] <delewis>       13
[17:28:03] <icon> not bad
[17:28:17] <icon> well once this gets going, everyone can move their packages into ports proper
[17:28:18] <eboutilier> Patrick's (oxygene's) pmpkg system uses something like PM...
[17:28:25] <icon> ahh okay
[17:28:32] <icon> DEL, PMP, CSW ;)
[17:28:43] <delewis> icon: not really intrested. I prefer to build my own packages. Pure, plain, and simple.
[17:28:57] <icon> delewis: ahh
[17:29:07] <icon> one thing thats planned is 3rd party repos
[17:29:19] <icon> been talking with laca, and it looks like we might end up goign with pkgbuild for the build system
[17:29:29] <jengelh> PMP my CSW
[17:29:38] <icon> so you can maintain recipies and packages outside of the ports system, but still take advantage of the utils
[17:29:52] <delewis> most of my stuff is multimedia in nature and installs in /opt/local
[17:30:01] <icon> thats most of what i build at home too
[17:30:03] <delewis> compiled for SPARCv8 (in some cases SPARCv9 in the case of ffmpeg and mplayer)
[17:30:11] <icon> ahh
[17:30:30] <eboutilier> icon: I agree with that choice, FWIW. One reason is the JDS consolidation is huge and committed to pkgbuild, you gain a lot of leverage right off the bat.
[17:30:47] <icon> eboutilier: yup. its a good system too. it really gets us about 80% of the way there
[17:31:18] <eboutilier> icon: Agreed.
[17:31:20] <icon> eboutilier: any suggestions on a generic package prefix?
[17:31:38] *** Trinisan has joined #opensolaris
[17:32:16] <eboutilier> icon: Have you ruled out SFE, the prefix laca uses for the unsupported pkgbuild packages (spec-files-extra)?
[17:32:23] <icon> nope
[17:32:38] <icon> what does it stand for?
[17:32:41] <eboutilier> I guess that's what I'd suggest then.
[17:32:46] <eboutilier> spec-files-extra
[17:32:53] <icon> hmm
[17:33:04] <icon> do we have a generic package name for opensolaris?
[17:33:09] <eboutilier> So that's a downside, it would sort of be a mis-nomer for what you want to do.
[17:33:12] <icon> yeah
[17:33:30] <icon> OSOL pops out, but i dont particularly care for it
[17:34:10] *** sommerfeld has joined #opensolaris
[17:34:10] *** ChanServ sets mode: +o sommerfeld
[17:34:31] <icon> it would be nice to have something the entire community can use to differentiate themselves from SUNW if needed
[17:34:52] <icon> we need our own SUNW :)
[17:34:54] *** Trinisan_ has quit IRC
[17:34:58] <andersmo> Is there a "the entire community"? =)
[17:35:01] <eboutilier> icon: A generic package name for opensolaris. that's takes us down an interesting path. OpenSolaris is a code base and its community; technically there is no concept of a package at the code-base level.
[17:35:10] <andersmo> There's CSW for "Community Software" from blastwave.
[17:35:40] <icon> well, most package prefixes indicate a topic rather than an entity
[17:35:55] <icon> SUNW comes from sun, we should have one that indicates a package from the opensol community
[17:36:24] <icon> eboutilier: I completely agree, but there are projects that will more than likely never be merged into solaris proper.
[17:36:29] <icon> im betting ports will probably be one of those
[17:36:50] <icon> which i dont have a problem with, because it creates a logistical issue with solaris users that sun supports
[17:37:37] <eboutilier> icon: OK, I'm with you there. So a package prefix that conotates (is that the right word) the opensolaris community
[17:37:44] <delewis> it would more likely be an optional bit (similar to JDS) that a distribution can either include or decline to use (in the case of Sun)
[17:38:17] <icon> eboutilier: exactly
[17:38:28] <icon> i agree with delewis
[17:38:46] <eboutilier> icon: OK, ports isn't aimed at being a conduit into Solaris then. We got our wires crossed on that  earlier I think.
[17:38:53] <icon> well
[17:38:57] <icon> i would love to see solaris use it
[17:39:16] <icon> and i want to do everything possible to see that happen
[17:39:17] <eboutilier> Gotcha.
[17:39:27] <icon> i guess what i meant was, im carefully optimistic :)
[17:39:54] <icon> but if it doesnt, and it stays in opensol only, im okay with that too. whats important to me is improving solaris for everyday use
[17:40:55] *** mwaterl has quit IRC
[17:42:05] <eboutilier> Understood: Just as Companion CD, SFE, and PM packages are designed (perhaps coincidentally) to be easy to pick up and get ARC'd into Nevada (aka "Solaris 11"), the ports model would be designed with that in mind too...?
[17:43:25] <eboutilier> icon: We should keep in close contact w/ Moinak and co. Their goals for Belenix and yours sound quite similar.
[17:43:40] <eboutilier> icon: Moinak was one of the earliest people to +1 your proposal
[17:45:02] <icon> eboutilier: yes it can
[17:45:08] <icon> very easily actually
[17:45:31] <icon> all of the overhead of creating packages by hand is removed by tooling. you create your spec file, and your package is generated
[17:45:40] <icon> targeted for whatever platform etc. you need
[17:45:58] <icon> ports just closes the gap by installing and maintaining those packages
[17:46:35] <eboutilier> Cool.
[17:46:45] <icon> once its done, it will be pretty slick :)
[17:47:22] <icon> it will certainly reduce the number of smf scripts i have to write each week
[17:48:22] <eboutilier> :)
[17:48:25] *** besonen_mobile has quit IRC
[17:50:53] *** tsoome has joined #opensolaris
[17:51:34] <icon> eboutilier: i think the ultimate test will be to create the port package as a port
[17:51:39] <icon> that would make selfupdates pretty easy
[17:52:14] <jengelh>  /msg #aurora Not anymore, no :)
[17:52:19] <jengelh> blah
[17:52:38] <jengelh>  /usr/ucb, /usr/sfw and /opt/csw are all names to me. But what on Earth is /usr/dt?
[17:52:50] <eboutilier> Hmm.
[17:52:51] <icon> ?
[17:52:53] *** sopor_aeternus is now known as zarathustra
[17:52:56] <icon> not a cde user i take it :D
[17:52:59] <trygvis> jengelh: X stuff
[17:53:04] <trygvis> yah, cde stuff
[17:53:08] <jengelh> well I've seen dt before, but it is the abbreviation for what?
[17:53:16] <icon> was it desktop?
[17:53:20] * icon digs
[17:53:25] <jengelh> possibly, but that seemed to easy
[17:53:27] <jengelh> :p
[17:53:57] <icon> yeah
[17:53:58] <icon> desktop
[17:55:38] <eboutilier> CDE, one of the great detente event of the early 90's. Let's see, Sun, HP, and who-else? agree on the desktop standard to end all desktop standards.
[17:55:41] <eboutilier> Hee hee.
[17:56:25] <icon> haha
[17:56:40] <icon> i think xfce is all that really survived, and its more of just a lookalike
[17:57:06] <icon> ahh those days of firing up a fresh sparc and being greeted by dtlogin ;)
[17:57:13] <icon> speaking of which, when is gdm going to be used?
[17:57:37] <eboutilier> xfce is CDE progeny... wow, I didnt' know that.
[17:58:24] <icon> let me dig up something on it
[17:58:57] <jteo> xfce is more of a clone in spirit
[17:59:05] <icon> yeah
[17:59:13] <eboutilier> gdm plans... Best to defer to one of the JDS guys for that answer...
[17:59:16] * gdamore prefers xfce
[17:59:26] <icon> from xfce.org:
[17:59:28] <icon> Xfce is a lightweight desktop environment for UNIX platforms. It is similar to the commercial CDE, and is now based on the GTK+ toolkit.
[17:59:34] <jteo> gdamore: you never answered my question about NetBSD.
[17:59:40] <icon> its actually not half bad
[17:59:45] <icon> very simple, and quick
[18:00:03] <gdamore> on the phone now, i'll tell you in a few minutes when the concall is done.
[18:00:10] <jteo> gdamore: no worries. ;)
[18:00:17] <jengelh> hakuna matata!
[18:00:58] *** solarisjon has joined #opensolaris
[18:01:05] <icon> hrmm back on my sunray kick... has anyone here used local devices on a ray 2 versus a 2fs?
[18:01:33] <jengelh> local as in usb sticks and such?
[18:02:06] <gdamore> sunray 2 and sunray 2fs have the same usb hardware.
[18:02:07] <icon> yeah
[18:02:10] <icon> oh?
[18:02:14] <icon> i thought the 2 had 1.1 and the 2fs was 2.0
[18:02:19] <eboutilier> Re: xfce: It'd definitely be my default desktop if it weren't so important (for me anyway) to be JDS-aware.
[18:02:31] <gdamore> the 2fs has a USB 2.0 hub.  but it uses the 1.1 host controller on the Au1550
[18:02:46] <icon> gdamore: could you elaborate?
[18:02:56] * Stric uses xfce on a lowend laptop just because "it works" and it doesn't use much resources
[18:02:59] <icon> eboutilier: same here. xfce with nimbus would be dead sexy :)
[18:03:04] <jteo> nimbus?
[18:03:05] <gdamore> you can run devices at full speed, but not high speed.  ie. there is no point.
[18:03:07] * jteo googles
[18:03:11] <icon> xfce 4.4 actually has thunar in it... which is looking *very* nice
[18:03:21] <icon> gdamore: ahh okay
[18:03:37] <sickness> hey what's that post-install thing that I see after a liveupgrade reboot to snv50?
[18:03:42] <sickness> anyone knows?
[18:03:48] <icon> so theres no reason for me to shell out an extra 250 for a 2fs if i dont dual head and i only use a single usb port
[18:03:58] <gdamore> usb 2.0 was added to satisfy a SunFed marketing tick, but it doesn't have any actual technical merit since they only have 1.1 at the CPU.
[18:04:01] * jengelh tackles src again
[18:04:08] <gdamore> correct.
[18:04:13] <icon> gotcha
[18:04:36] <gdamore> (its possible that they would prefer to have that bit of detail not commonly known)
[18:04:41] <icon> the comet is tempting, but i really dont need a portable
[18:04:48] <icon> yeah... i kept coming across conflicting info
[18:04:51] <icon> some said 2.0, some said 1.1
[18:04:52] *** Dar is now known as Dar_HOME
[18:04:53] <gdamore> comet is a lot more $$.
[18:04:57] <icon> how much?
[18:05:07] <gdamore> i think they are ~$999
[18:05:10] <icon> ouch
[18:05:19] *** mega has joined #opensolaris
[18:05:25] <gdamore> but you get full wireless, and a beautiful 15" screen for that.
[18:05:29] <icon> true
[18:05:37] <icon> but i have a gorgeous 19" lcd ready for the ray
[18:05:55] <icon> actually i was watching csi last night, and they had a 1g on the show
[18:05:56] <gdamore> if you don't need mobility, and have your own monitor, then a sunray 2 is the right choice.
[18:05:58] <jteo> comet's are nice.
[18:06:13] <icon> id love to work in an environment that would just pass out comet's
[18:06:14] <jteo> i remember using one as a glorified irc terminal. :)
[18:06:14] <jengelh> Besides comets and rays, do they also sell stars?
[18:06:22] <icon> stars?
[18:06:25] <jengelh> SCNR
[18:06:31] <gdamore> heh.
[18:06:38] <gdamore> ask shily if you want a star.
[18:06:43] <gdamore> schily i mean
[18:06:50] <jengelh> no please don't do that to me!
[18:07:03] <icon> alright, im playing the patsy here, but what is a star?
[18:07:06] <jengelh> or rather do they sell quasars?
[18:07:19] <jengelh> icon: Think of star as in astrology.
[18:07:28] <icon> crap
[18:07:31] <gdamore> icon: we have talked about customers using comets in environments like that.  imagine a class room, where the teacher just handed them out.  or on a plane.
[18:07:32] * icon feels like a moron now
[18:07:38] <jengelh> http://en.wikipedia.org/wiki/Star
[18:07:49] <icon> gdamore: yup, they would be perfect for coders too
[18:08:33] <icon> picking one up for home because ill be decking out the opteron box
[18:08:42] <icon> so the noise level is going to go from loud to retarded
[18:08:54] <icon> need to hide it back in the hvac again
[18:08:55] <jteo> icon: earplugs are cheaper.
[18:09:04] <icon> jteo: but not as cool
[18:09:04] <jengelh> buy silent fans
[18:09:16] <jengelh> use the powernow fluff that comes with amd
[18:09:23] <icon> silent fans are great, but when you have a 5 disk array in there with a 500w psu, it gets a bit loud
[18:09:31] <icon> cool n quiet doesnt do much when under load
[18:09:37] <jengelh> you need a bigger fan
[18:09:42] <gdamore> if you want builtin VPN support, then you need a comet.  sun ray 2 will get it later.
[18:09:53] <jengelh> http://www.ahajokes.com/cartoon/bigfan.jpg
[18:09:58] <eboutilier> gotta run. bbl
[18:09:59] *** eboutilier has quit IRC
[18:10:00] <icon> gdamore: with a firmware update, or as a new product?
[18:10:13] <gdamore> i *think* with a firmware update
[18:10:16] <icon> haha
[18:10:32] <icon> hmm
[18:10:37] <gdamore> i don't work for the group though, so i only know what engineers were planning as of a few months ago
[18:11:14] <icon> well
[18:11:18] <gdamore> the biggest problem was getting thru the ARC with persistent data store (required to store VPN keys) and the UI.  (They are using the UI I wrote for Comet, btw.)
[18:11:25] <icon> i know the Au has some builtin vpn stuff
[18:11:50] <gdamore> the au1550 has a very nice crypto accelerator on it.  last i heard they weren't using it yet, and werent' sure if they were going too
[18:11:51] <icon> you know, i actually thought tadpole was extinct until you told me about the comets the other day
[18:12:08] <icon> hrmm
[18:12:11] <gdamore> we were acquired by General Dynamics.  not extinct.
[18:12:14] <icon> the au is a mips32 isnt it?
[18:12:24] <gdamore> yes.  they are running it at 400MHz, IIRC.
[18:12:31] <icon> thats suprising
[18:12:47] <icon> i figured they would have dropped the clock on the rays
[18:12:48] <gdamore> its a _lot_ zippier than what is in the Comet or Sun Ray 1g.  you don't notice unless you do VPN though.
[18:13:14] <gdamore> no.  even at 400MHz the power consumption and heat are negligible.
[18:13:29] <icon> specs say 8w... is that true?
[18:13:32] * Stric replaced 100MHz linux indys running X -query  over 10Mbit -> SunRay1 .. "Whoa."
[18:13:34] <gdamore> yes.
[18:13:37] <icon> wow
[18:13:41] <icon> 12vdc?
[18:13:44] *** dunc has quit IRC
[18:13:58] <gdamore> i guess.  i've not seen an actual Sun Ray 2 yet.
[18:14:08] <icon> ahh
[18:14:11] <icon> hrmm
[18:14:12] <LeftWing> I have.
[18:14:15] <gdamore> the au1550 itself runs much lower voltage.  but you'd expect that.
[18:14:17] <icon> LeftWing: what do you think?
[18:14:18] <LeftWing> They're pretty spanky.
[18:14:40] <LeftWing> The card slot lights up. =P
[18:14:43] <icon> LeftWing: have you tried any throughput tests on it yet?
[18:14:45] <gdamore> the au1550 would be one of my prime targets if ever I got started on a MIPS port of Solaris.
[18:14:46] <icon> oh cool
[18:14:57] <LeftWing> icon: Nah, it's sitting on the desk of a guy who doesn't ... really use it.
[18:15:00] <icon> gdamore: let me know if you need help, im an old mips junkie ;)
[18:15:21] <icon> in fact, i have something funny to show you
[18:15:30] <gdamore> i am still trying to get interest in a MIPS Solaris port.
[18:16:11] <icon> way back in college i ported spim over to sparc solaris 7
[18:16:28] <icon> i was digging around in archive.org last week and actually found an old screenshot i posted
[18:16:32] <icon> http://web.archive.org/web/20010627025109/www.cs.smsu.edu/~stallion/spim-ss.jpeg
[18:16:37] <jteo> is there affordable MIPS hardware i can actually *buy*?
[18:16:48] <icon> jteo: all over the place
[18:16:48] <gdamore> heh.  when i was in college "Solaris" didn't exist.  we had  SunOS 4.1 though.
[18:17:08] <gdamore> trivial.  most wireless APs have MIPS in them.
[18:17:26] <icon> yup, mips has been doing very well in embedded for *years*
[18:17:29] <gdamore> of course, they also have ~32MB or less RAM, and usually only 8MB or less flash.
[18:17:34] <icon> hell, i think the snes was a mips variant wasnt it?
[18:17:45] <gdamore> i think so.
[18:17:48] <gdamore> the ps2 is MIPS
[18:17:52] <icon> yup
[18:17:57] <jteo> i was thinking workstation class.
[18:17:58] <jteo> ;)
[18:18:01] <icon> well
[18:18:10] <icon> you could go dig up some old sgi boxes
[18:18:15] <icon> should be going for cheap now :D
[18:18:20] <gdamore> ebay.  for sgi or even a cobalt qube
[18:18:31] <icon> ahh thats right cobalts were mips too
[18:18:37] <gdamore> there are other options, pmax, etc.
[18:18:40] <jteo> i gathered as much.
[18:19:00] <sickness> back from work
[18:19:23] <icon> here you go
[18:19:24] <icon> http://cgi.ebay.com/Cobalt-Qube-3-Server-No-Reserve-Refurbished_W0QQitemZ200037275304QQihZ010QQcategoryZ51236QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[18:23:07] *** peteh has quit IRC
[18:23:20] <delewis> n64 was MIPS
[18:23:26] <delewis> I highly doubt the SNES was for some reason
[18:24:05] <delewis> for some reason I want to say it was a Saturn or something
[18:26:59] *** slowhog has joined #opensolaris
[18:29:58] <icon> ahh
[18:30:05] <icon> it was a 16bit extension to the mos6502
[18:30:52] <icon> the 6502 was alot of fun
[18:32:09] *** loke has quit IRC
[18:32:48] *** solarisjon has left #opensolaris
[18:33:36] *** solarisjon has joined #opensolaris
[18:33:46] *** solarisjon has left #opensolaris
[18:33:59] <jengelh> 07:46 sa0:~ > man -f /opt/onbld/man/man1/bldenv.1
[18:34:01] <jengelh>  /usr/share/man/windex: No such file or directory
[18:34:03] <jengelh> wtf
[18:35:24] *** kgrimm has quit IRC
[18:36:22] <coffman> cobalt qube
[18:36:34] <coffman> ill wonder why the put this on ice
[18:38:16] *** mega has quit IRC
[18:42:24] <delewis> jengelh: catman is your friend.
[18:42:34] <delewis> catman -w, specifically.
[18:43:38] *** Cyrille has quit IRC
[18:48:04] <gdamore> the url for the qube is a qube3.  its not mips, but x86.  you need an original qube or (better) a qube2.
[18:48:11] <gdamore> if you want mips, that is
[18:48:28] <icon> ahh okay
[18:48:38] <icon> i thought they kept mips in the 3
[18:48:49] <g4lt-mordant> yeah, but the MIPS I WANT is an R14400 ;P
[18:48:52] <gdamore> and it will be less $$.  (dont pay more than $50)
[18:49:15] <gdamore> i want one of those new 16-way cavium 1GHz mips chips.
[18:49:24] <gdamore> but not just the chip, but put it ina system for me.
[18:49:27] <gdamore> please.
[18:49:31] <gdamore> and a pony too.
[18:50:05] <g4lt-mordant> gdamore, they make new MIPS?
[18:50:23] <g4lt-mordant> I thoiught they died when SGI dumped them
[18:50:48] <jerome__> Hi
[18:51:03] <jerome__> I was wondering how I could detect what sunpci card I have in my machine
[18:51:10] <jerome__> I know I have one, I can see the back
[18:51:16] <jerome__> but I don't know which one it is
[18:51:20] <jerome__> is there a trick?
[18:51:27] <g4lt-mordant> describe the back, I might be able to help
[18:51:43] <jerome__> it has video ethernet and an orange light
[18:51:44] <icon> g4lt-mordant: amd's alchemy chips are mips32 based
[18:51:47] <jerome__> and some stuff for sound
[18:51:54] <jerome__> and usb
[18:52:00] <jerome__> I have a picture
[18:52:12] <g4lt-mordant> cool, pastebin it
[18:52:17] <jerome__> w8
[18:52:43] <icon> g4lt-mordant: one sec
[18:53:31] <jerome__> http://i12.ebayimg.com/02/i/08/7c/89/14_12.JPG
[18:53:34] <jerome__> that's it
[18:53:50] <g4lt-mordant> looking....
[18:54:39] <jerome__> any clues
[18:54:50] <jerome__> I got it for free
[18:54:57] <jerome__> no clue what it is
[18:55:01] <jerome__> or if it even works
[18:55:21] <g4lt-mordant> looks like a spci2
[18:55:39] <jerome__> can I use the drivers 3.2.2? (latest)
[18:55:45] <jerome__> on build 50
[18:55:58] <g4lt-mordant> and it looks like that's a configuration I know works, because I've used a spci2 in a sb100 before myself ;)
[18:56:08] <g4lt-mordant> not drivers 3.X, use 2.X
[18:56:13] <jerome__> ok
[18:56:18] <jerome__> let me try
[18:56:27] <jengelh> catman takes quite some time
[18:56:49] <jerome__> sure does
[18:56:52] <jengelh> ugh
[18:57:25] <jengelh> i killed that now
[18:57:44] <jengelh> and ran catman -w instead, now man bldenv still does not find a thing
[18:57:57] <jengelh> man -f /opt/onbld/man/man1/bldenv.1
[18:58:02] <jengelh> returns nothing, finishes instantaneously
[18:58:08] <g4lt-mordant> catman doesn't add manpath, catman adds apropos entries
[18:58:31] <jengelh> so what's the easiest way to display an arbitrary manpage file without having it indexed?
[18:59:24] <g4lt-mordant> and man -f uses windex, which is what catman builds
[18:59:34] <jerome__> this blade 150 is slow
[18:59:42] <g4lt-mordant> jerome__, DUH
[18:59:44] <jengelh> man man says -f uses file
[18:59:46] <jerome__> I can't believe they still sell them
[19:00:11] <g4lt-mordant>      -f file ...     man attempts to locate manual pages  related
[19:00:11] <g4lt-mordant>                      to  any  of  the  given files. It strips the
[19:00:11] <g4lt-mordant>                      leading path name components from each file,
[19:00:11] <g4lt-mordant>                      and  then prints one-line summaries contain-
[19:00:11] <g4lt-mordant>                      ing the resulting basename  or  names.  This
[19:00:12] <g4lt-mordant>                      option also uses the windex database.
[19:00:29] <g4lt-mordant> hmm, last line pretty much says that it uses windex
[19:01:02] <jengelh> so what's the option to just display a file
[19:01:15] <g4lt-mordant> wait, display?  that's man -m /opt/onbld/man bldenv
[19:01:31] <jengelh> does not look like it's able at all
[19:01:38] <mage2> any networking guru's awaken?
[19:02:01] <jengelh> linux knows man -l /fullpath/to/here.1
[19:02:10] <jengelh> man -M /fullpath/to here
[19:02:14] <jengelh> big shrug for solaris
[19:02:24] <jengelh> what if the file does not end in .[1-9]?
[19:03:23] <jerome__> I wonder if sun still really sells the blade 150, how many orders they get a year
[19:03:27] <jerome__> like 5
[19:03:59] *** yippi has joined #opensolaris
[19:06:48] <jerome__> okay let's try this sunpci
[19:06:55] <jerome__> tjat c
[19:06:59] <jerome__> oops
[19:08:19] *** MattMan has quit IRC
[19:08:19] <jengelh> http://rafb.net/paste/results/09IwNg86.html
[19:08:48] <jengelh> nm
[19:08:57] <jerome__> devfsadm: driver failed to attach: sunpci2drv
[19:08:58] <jerome__> Warning: Driver (sunpci2drv) successfully added to system but failed to attach
[19:09:04] <jerome__> can't be good
[19:09:33] *** alanc_away is now known as alanc
[19:09:38] <jengelh> ah I hate this makefilesystem
[19:10:16] <alanc> Blade 150 should be off the price list now, along with all other old equipment that doesn't meet current European RoHS requirements to be able to sell it in Europe
[19:10:19] *** alanc is now known as alanc_away
[19:10:42] <jengelh> http://rafb.net/paste/results/MieHsS66.html
[19:13:07] *** Trinisan_ has joined #opensolaris
[19:13:15] <g4lt-mordant> what, europe has a legal requirement for non-underpowered machines?  how do they sell any X86 in europe then? ;P
[19:13:32] <asyd> ahah
[19:13:46] <jerome__> the requirement for europe is to have computers running on coal
[19:14:42] <g4lt-mordant> in that case, may I never use a computer in europe.  mine all run on mostly hydroelectric around here ;
[19:14:47] <g4lt-mordant> ;) even
[19:16:37] <jengelh> where is the differencet between SUNWwgetr and SUNWwgetu?
[19:16:38] *** Trinisan has quit IRC
[19:17:08] <g4lt-mordant> user binaries versus root binaries
[19:17:11] <g4lt-mordant> you need both
[19:17:20] <jengelh> difference?
[19:17:23] <jengelh> er
[19:17:36] <jengelh>  /usr/bin/wget can be both used by user and root...
[19:18:21] <jengelh> unless I am missing something
[19:19:19] <sommerfeld> packages are split between "/usr" vs root *filesystem*
[19:19:35] *** nwf has joined #opensolaris
[19:19:44] <sommerfeld> SUNWwgetr installs /etc/wgetrc
[19:19:49] <g4lt-mordant> okay, so I mistyepd and put in a useless E
[19:20:08] <sommerfeld> SUNWwgetu installs /usr/sfw/bin/wget
[19:20:29] <sommerfeld> the reason for the split is to allow for diskless installs where /usr is shared
[19:20:42] *** nwf has quit IRC
[19:21:29] *** nwf has joined #opensolaris
[19:21:58] <sommerfeld> it looks a little silly when the package is this fine-grained.
[19:22:07] *** hell` has joined #opensolaris
[19:22:27] <sommerfeld> jengelh: did that answer your question?
[19:23:41] *** bank__ has joined #opensolaris
[19:23:44] <bank__> hi
[19:23:52] <jengelh> yes
[19:23:55] <jengelh> thank you
[19:24:05] <bank__> I am installing sugarcrm. but i got this http://lh4.google.com/vichai.ar/RTjyLrbRABI/AAAAAAAAAAs/NmcaIjI1jLE/sugarcrm.JPG
[19:24:14] <bank__> I am not sure it involve with solaris permission
[19:24:21] <jengelh> just a little frustrated that I have to install so much additional software :(
[19:24:28] *** bengtf__ has joined #opensolaris
[19:24:32] *** bengtf__ is now known as bengtf
[19:24:41] <bank__> also MB Strings...
[19:25:27] <bank__> I install php , libxml2 on apache2 that reflect on all non-global zone. php is working.
[19:25:37] <bank__> then I have mysql on another non-global zone.
[19:26:02] <bank__> I installing sugarcrm. and got several unexpected results .... ( I expect to see error only for mysql)
[19:35:56] <coffman> bank__: u need the files writeable
[19:36:04] <coffman> and it didnt find ur mysql
[19:36:44] <coffman> and u will need curl and zlib
[19:36:58] <coffman> u need to point sugar on it
[19:37:21] <coffman> u
[19:37:35] <coffman> this reminds me that i need to install sugar
[19:37:36] <coffman> shit
[19:41:22] *** sahafeez has joined #opensolaris
[19:41:55] <jengelh> gosh why are all package names (not only SUNW but also CSW) so crumbled? Like CSWoldaprt... could not that be called CSWopenldap_rt to make things more clear
[19:42:33] *** deather_ has joined #opensolaris
[19:42:37] <jengelh> like we're living in ISO9660 Level1 days..
[19:43:42] <bank__> hello coffman
[19:43:48] <bank__> still overthere??
[19:44:33] <bank__> What is files writeable ?
[19:44:36] *** Triskelio- has joined #opensolaris
[19:44:37] <bank__> chmod something ?
[19:44:38] *** Triskelios has quit IRC
[19:44:46] <bank__> curl and zlib?
[19:47:03] <sickness> little question, when a system is using liveupgrade profiles, where is the grub config file stored? I can't find it but grub is correctly configured and boots ok :)
[19:47:18] <sickness> there's no menu.lst in /boot/grub
[19:49:50] <sickness> uhm, nevermind, found it in /etc/lu/GRUB_backup_menu :)
[19:49:58] *** _Tzoa has joined #opensolaris
[19:50:46] *** Tzoa has quit IRC
[19:50:47] *** _Tzoa is now known as Tzoa
[19:51:24] <gdamore> woo hoo..  i'm told to expect a signed SCA so that i can start contributing code early next week.
[19:51:44] <gdamore> that said, anyone want a driver for synaptics ps2 touch pads?
[19:53:37] <quasi> doesn't X have one of those?
[19:53:55] *** bunker has joined #opensolaris
[19:53:56] <gdamore> dunno.  i haven't checked Xorg.
[19:54:18] <gdamore> we have a driver that works with standard X (a vuid mouse module)
[19:54:27] <gdamore> standard Xsun i meant
[19:54:28] <quasi> it does on lunix
[19:54:55] <gdamore> does it offer the full feature set (scrolling, tapping, palm detection, etc.?)
[19:56:08] <bank__> coffeeman...
[19:56:11] <gdamore> it looks like the Xorg driver is Linux only
[19:56:16] <quasi> I couldn't say - I hate the darn thing ;)
[19:56:47] <gdamore> our solaris module is quite nice.  we also have a very nice tcl/tk (hate to say it) configuration UI
[19:57:08] <gdamore> i basically got approval to give this to Solaris today
[19:59:01] <gdamore> we also have a mroe advanced driver the alps usb touchpad, which we use in some of our units.
[19:59:44] <gdamore> (in the case of the alps driver, it is changes to the standard usbms.c driver.)
[19:59:46] *** deather has quit IRC
[19:59:55] <sommerfeld> what I want is a  "cat detection" mode :-)
[19:59:56] <quasi> cool (if you like touchy-feely pads)
[20:00:47] <gdamore> i actually _much_ prefer an external mouse.  but with a dozen or so different laptops, its inconvenient moving a mouse around, so I use the track pads quite a bit.  admittedly my usage case is a bit unconventional.
[20:01:03] <coffman> bank__: u need to make a couple of folders and files write able
[20:01:12] <AbeFroman> i used to hate them, but the one on the macbook pro is nice
[20:01:19] <AbeFroman> though of course, the keyboard is always faster
[20:01:20] <coffman> gnar
[20:01:32] <coffman> didnt get that primepower thing
[20:01:37] <coffman> 118 euros
[20:01:39] <coffman> cheap shit
[20:02:00] <gdamore> the synaptics device is very, very nice.  but you want to get a synaptics driver so you can use its full feature set.  in normal "compatibility" mode the thing is horrid
[20:02:35] <gdamore> (mostly you really want palm detection, because the position is such that i brush up on it with a palm, and it changes the mouse focus)
[20:02:58] <quasi> gdamore: oh, it does that too?
[20:03:03] <gdamore> yes.
[20:03:14] <quasi> nifty
[20:03:35] <gdamore> i think i'm gonna file a bug/rfe for it.
[20:04:02] <sommerfeld> the palm detection causes it to ignore touches which hit too big an area?
[20:04:08] <gdamore> yes.
[20:04:39] <gdamore> sommerfeld, are you willing to sponsor an arc case for this?  (and i also need a sponsor for the kb8042 changes to add support for the extra Sun buttons on SPARCLE)
[20:05:05] * gdamore seems to recall sommerfeld has a SPARCLE of his own
[20:05:49] <sommerfeld> sure.
[20:06:06] <gdamore> thanks.
[20:06:54] <bank__> gdamore
[20:07:00] <gdamore> yes?
[20:07:48] <bank__> I am not sure you advice me about dragon kodomo
[20:08:01] <bank__> let me search.
[20:08:04] <gdamore> ?
[20:08:45] <bank__> no no sorry that is gisburn
[20:09:00] <gdamore> yeah, that sounds more likely.  he's got a komodo dragon fetish.
[20:09:23] <sommerfeld> I was going to say "unhealthy obsession with komodo dragons."  I'm not sure which is worse
[20:09:31] <gdamore> heh.
[20:09:45] <bank__> <gisburn> bank: and contrary to my constant claims it does not eat humans.
[20:10:39] <AbeFroman> no but they will bite, and they have lots of bacteria in their mouths
[20:10:44] <AbeFroman> bad bacteria
[20:10:57] <gdamore> i think i'd still rather not be thrown into a pit filled with them.
[20:11:15] <bank__> :O
[20:11:16] <gdamore> alligators don't eat humans either, but that doesn't mean they are nice to play with.
[20:11:49] <Error_404> they'd probably kill you, but not because they're hungry
[20:12:19] <Error_404> they're like rich fratboys with too much free time... killing hookers just for the thrill
[20:12:47] <bank__> it's my fault to open dragon kodomo issue :P
[20:16:31] <Error_404> yum, gyoza... they're japan-tastic
[20:17:41] <jerome__> hm
[20:17:46] *** Fish- has joined #opensolaris
[20:18:12] *** alanc-away is now known as alanc
[20:18:22] <Fish-> hello
[20:18:35] <jerome__> server: bash-3.00# share -F nfs -o rw -d "zfs" /zfs
[20:18:41] <jerome__> client: amanda:/mnt/nfs3# touch a
[20:18:41] <jerome__> touch: cannot touch `a': Permission denied
[20:18:52] <jerome__> any clues?
[20:19:15] <quasi> you're root on that client?
[20:19:20] <jerome__> yes
[20:19:25] <jerome__> see the '#'
[20:19:36] <alanc> gdamore: a VUID driver would be nice - currently we've been using Casper's hacks for synaptics in the Xorg mouse module, but that doesn't help Xsun users
[20:19:39] <gdamore> root is "nobody" over NFS usuually
[20:19:51] <alanc> and hasn't actually been integrated anywhere other than frkit
[20:20:09] <jerome__> gdamore thanks
[20:20:13] <jerome__> gdamore that was my problem
[20:20:16] <gdamore> well, i can probably submit code as early as next week.  i do need to test it on x86 though
[20:20:34] <alanc> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6220332
[20:20:35] <gdamore> np.  quasi was already on to it, i think.
[20:20:36] <quasi> jerome__: but there's an option to allow it if you wish
[20:20:57] <quasi> gdamore: yeah, thanks for saving me the typing ;)
[20:21:06] <gdamore> heh.
[20:21:19] <alanc> Casper had done some work, so maybe you can get him to sponsor you for that bug id
[20:21:28] <gdamore> alan: i think vuidsyn would be better than a mouse_drv.so.  but that's just my opinion.
[20:21:48] <gdamore> does Xorg use the vuid modules normally?
[20:21:50] <alanc> whoops - yeah I pasted the wrong bug id
[20:21:52] *** Fish- is now known as Fish
[20:22:00] <alanc> 6223897 vuid3ps2 module should be extended to support synaptics touchpad.
[20:22:07] <alanc> but that's not on bugs.os.o
[20:22:18] <gdamore> bummer.
[20:22:21] <alanc> though vuidsyn would make more sense if SPARC can use it too
[20:22:28] <gdamore> well, you can associate mine with it.
[20:22:46] <alanc> yes - Xorg uses VUID unless you manually configure otherwise
[20:22:54] <gdamore> the problem is that a vuidsyn really wants to run in "absolute mode", and is hard to get all that functionality in a compatible way with normal ps2 support
[20:23:14] <gdamore> so its better to do it as a seperate vuid module.
[20:23:17] <alanc> and with virtual mouse, the kernel picks the right VUID module now, instead of the X server
[20:23:36] <alanc> and can have different VUID modules for different mice coming into the same stream
[20:24:02] *** Netwolf has joined #opensolaris
[20:25:08] <gdamore> right.
[20:25:30] <gdamore> i'm using vuidsyn on my SPARCLE right now.  it works really nice.  (This is in OpenSolaris).
[20:25:58] <gdamore> The only thing I have to do is add code to mouse8042 to detect a synaptics pad and configure the the right minor node for dacf.
[20:26:11] <alanc> cool - I'd love to see it, but you'll need an ON sponsor, which would have to be someone other than me
[20:26:16] <gdamore> (right now we just unconditionally change the dacf rule on our hardware)
[20:26:40] <gdamore> i think sommerfeld will sponsor.
[20:26:42] <alanc> but I can help with ARC if needed, or any changes on the X side
[20:26:55] <gdamore> no changes needed for X.  their just vuid events.
[20:27:00] <alanc> good
[20:27:27] <gdamore> our configuration UI is a different story, that will probably be contentious.  Its a Tcl/Tk beast, and it doesn't look like gnome.
[20:27:49] <alanc> does it send VUID scroll wheel events for the scroll areas?
[20:27:52] <gdamore> yes.
[20:28:09] <gdamore> (you have to enable them.  there is the GUI for that, or you can use a CLI app)
[20:28:31] <gdamore> likewise, we support configuration of the scroll buttons as either a middle mouse button or as scrollers
[20:30:18] <sommerfeld> i've seen that UI in action -- the live plot of what's being touched and how hard is particularly cool
[20:30:34] <jerome__> can I prevent root from being mapped as nobody?
[20:30:45] <jerome__> yes I understand the security implications
[20:30:47] <sommerfeld> jerome__: yes.
[20:30:51] <jerome__> but I doubt my benchmark thing would care
[20:30:56] <Error_404> no-root-squash
[20:31:33] <sommerfeld> I think you mean "anon=0"
[20:31:59] <sommerfeld> see share_nfs(1M)
[20:32:02] <jerome__> thanks
[20:32:15] <twincest> no_root_squash is the Linux option
[20:33:03] <sommerfeld> jumpstart requires -o "ro,anon=0"
[20:34:57] *** dunc has joined #opensolaris
[20:35:45] *** Netwolf has quit IRC
[20:39:04] *** mage2 has quit IRC
[20:59:11] <delewis> anon=0 isn't necessary, but ro is
[21:01:08] <delewis> well, nevermind, I suppose it would be if weren't using root=
[21:02:20] *** astinus has joined #opensolaris
[21:02:44] * astinus cheers because he's got Solaris running (first time!)  :D
[21:02:58] *** Triskelio- is now known as Triskelios
[21:03:19] <delewis> astinus: congratulations.
[21:03:36] <astinus> delewis: thanks, I have a few kinks to iron out, but BigAdmin seems great
[21:03:56] <delewis> docs.sun.com is even better :-)
[21:04:03] <astinus> Failed to get my Highpoint RocketRAID working
[21:04:20] <astinus> It's a Marvell 88sx chipset; but the revision has 'issues'
[21:05:12] <delewis> heh, sweet. the Ferrari 5k has an Athlon X2 == dual core :-)
[21:05:43] <astinus> almost bought one of those (well, a Ferrari 1k)
[21:06:04] <astinus> ended up with a Samsung Q35 - which is really neat
[21:06:18] <quasi> delewis: sounds like the way to get 5 minute battery life ;)
[21:06:19] <delewis> I'm in need of a portable solution to run Solaris on, and SPARC portable solutions are a bit pricey, so a Ferrari seems like the next best bet.
[21:06:25] <delewis> quasi: :-)
[21:06:44] <delewis> after dealing with my Powerbook's 2 hour battery life for the last 4 years, I can cope with anything
[21:07:11] <astinus> delewis: clive king came to give a guest talk recently at my university, he had Solaris running very nicely on a Ferrari - it wasn't a 5k, one of the older Turion ones that preceded the stuff with Turion X2s
[21:07:20] * quasi pats the stinkpad with 5hr battery life
[21:07:33] <delewis> astinus: yeah, everything before the 5k had a Turion (including the 4k series)
[21:07:39] <delewis> which I was going to go with, until I noticed the 5k :-)
[21:07:51] <delewis> almost all the specs are doubled, compared to the 4k
[21:08:00] <astinus> yikes, pricetag?
[21:08:12] <delewis> (2GB of memory, 160GB disk, Radeon x1600 (instead of the x800), etc.)
[21:08:18] <delewis> astinus: about the same as the 4k when it was new
[21:08:26] <delewis> $2k-$2.5k
[21:08:37] 
[21:08:44] <delewis> and the 4k is roughly $1.k, nowadays.
[21:08:45] <delewis> er
[21:08:48] * astinus is a scummy Brit ;P
[21:08:49] <delewis> $1.8k
[21:09:04] <delewis> so a couple hundred more isn't that bad when the specs are almost doubled in every aspect :-)
[21:09:33] <delewis> regardless, I was willing to drop $3.5k on a SPARC portable a few weeks ago, so the Ferrari 5k isn't all that pricey.
[21:10:24] <astinus> Total newb question, but how do I list all the disks recognised in the system?  I have a Maxtor Atlas 15K3 which I installed onto, with nothing else connected during install, I just plugged in a Marvell 88sx controller and a 120G SATA for testing..
[21:10:35] <quasi> delewis: and you're getting an awful lot more for you money (even if it is lacking the cool factor of portable sparc ;)
[21:10:44] <sickness> astinus: format?
[21:10:54] <delewis> quasi: my thinking, precisely.
[21:10:57] <sommerfeld> astinus: "format" will start with a menu of disks it sees
[21:11:05] <astinus> Cool :D  Thanks!
[21:11:10] <delewis> my mindset in buying the SPARC portable was, though, it wouldn't be fastest thing in the world, it was certainly more elegant.
[21:11:14] <quasi> you may need to run devfsadm first
[21:11:14] <sickness> astinus: cfgadm should list the marvell 88sx connections too, since it's supported in the sata framework...
[21:11:15] <delewis> then I woke up :-)
[21:11:38] * gdamore is not listening to the current discussion.  la la la la....
[21:11:43] <quasi> delewis: the insane pricetag tends to do that to you ;)
[21:11:44] <delewis> gdamore: hehe :-)
[21:11:51] <astinus> sickness, showed up right away .. cfgadm lists all disks connected to my 88sx
[21:11:58] 
[21:12:09] <delewis> now on the other hand, if a Meso 999 or Bullfrog was the $3.5k..
[21:12:11] <sickness> yeah, I wish I could find one controller like that too :'
[21:12:20] <delewis> I would've bought one without a thought
[21:12:25] <sickness> I can't find one in a shop that ships to europe, only usa :(
[21:12:29] <gdamore> Bullfrog is _not_ portable.
[21:12:38] 
[21:12:41] <gdamore> (well, it is portable, but you don't want it on your lap)
[21:12:46] <delewis> gdamore: hehe
[21:12:46] <sickness> astinus: really?!?
[21:12:48] <quasi> sickness: I see it in .nl and I think they were in .uk as well
[21:12:50] <sickness> astinus: lemme check :)))
[21:12:57] <gdamore> a Viper (15" US3i) would be nice however.
[21:13:02] <sickness> mmm
[21:13:05] <delewis> the Meso999 is Naturetech's high-end portable offering -- SB1500 clone with 17" display
[21:13:08] 
[21:13:17] <delewis> gdamore: that sounds more like the Meso999, yeah.
[21:13:27] <astinus> sickness: i guarantee its there ;P I ordered one on Wednesday, it arrived today, its in my chassis now :P
[21:13:39] <gdamore> i've not seen the meso 999.  i've not heard great things about nature's h/w quality though.
[21:13:40] <delewis> which retails for roughly $10k
[21:14:07] <gdamore> i don't _think_ the viper is anything close to $10k, but i could be mistaken
[21:14:41] <astinus> sickness:  sata0/1::dsk/c3t1d0     disk     connected     configured     ok
[21:14:53] <gdamore> viper2 in a hardened case might be that much.
[21:15:18] <delewis> http://flickr.com/photos/mrbill/sets/489821/
[21:16:17] <delewis> http://www.sunhelp.org/review-hardware-naturetech-meso999/
[21:16:21] <gdamore> do you have one?  i'm curious to know if they are using a PS/2 controller for their intenral keyboards
[21:16:28] <astinus> Next question now I know the controller/disk is working .. can Solaris mount any kinda Linux partition?  Specifically interested in ext2 :P
[21:16:39] <delewis> gdamore: I wish!
[21:16:58] <astinus> Tried to find some info about this, but didn't see much easily visible on BigAdmin or docs.sun.com :/
[21:17:39] <gdamore> heh.  its basically a mobile 2500.  they didn't even bother to change uname.
[21:17:45] <quasi> sickness: http://www.scan.co.uk/Products/ProductInfo.asp?WebProductID=258223
[21:17:51] <delewis> gdamore: yeah, just saw that
[21:17:56] <delewis> for some reason I was thinking it was a 1500 clone
[21:18:09] <astinus> quasi, That's the one :D
[21:18:42] <sickness> astinus: cool :)
[21:18:54] <gdamore> in his review he's saying the lesser equipped model is $3400.  the lesser models on sun store are _not_ meso 999's, but tadpole sparcles.
[21:19:04] <delewis> gdamore: yeah.
[21:19:10] <sickness> quasi: tnx :)
[21:19:21] <delewis> but of course, the Ultra 3 is now dead, so it doesn't *really* matter.
[21:19:29] <gdamore> true.
[21:19:38] 
[21:19:39] <quasi> astinus: but looking at the image - what kind of slot does it fit?
[21:19:54] <astinus> quasi: PCI-X
[21:19:58] <delewis> I remember the Slashdot/osnews/uninformed commentary when the Ultra 3 premiered -- "OMG, SUN IS IN THE NOTEBOOK BUZINESS"
[21:19:59] <gdamore> basically, the difference between natures platform and ours is that nature is just repackaging sun hardware, with no additional software or drivers, etc.
[21:20:09] <delewis> gdamore: interesting
[21:20:14] <gdamore> we have signficant value add in some of our software.
[21:20:18] <quasi> astinus: ah, cool - they looked a bit longer than that
[21:20:19] <gdamore> which now i'm trying to get into OpenSolaris
[21:20:23] <astinus> quasi: Its also good for PCI usage, fwiw .. I tried it in both types on my server.
[21:20:24] <delewis> gdamore: sweet
[21:20:37] <icon> gdamore: do you think there would ever be an open firmware for the rays?
[21:20:41] <astinus> quasi: the unit comes with both brackets, for 2U (half height) and full height deployment.
[21:20:44] <gdamore> icon: no
[21:20:44] <quasi> astinus: excellent - I see the text now ;)
[21:20:48] * astinus nods
[21:20:52] <icon> gdamore: oh well ;)
[21:21:03] <gdamore> you wouldn't want it, anyway, trust me.
[21:21:07] <icon> bad?
[21:21:10] <astinus> i think its a bloody bargain :D
[21:21:24] <gdamore> sun ray 1g has no mmu, only 4MB ram, and only 100MHz.  and only 1MB flash.
[21:21:31] <astinus> can i 'mount -F ext2 /dev/dsk/c3t1d0p0 /mnt/linux' ?
[21:21:44] <gdamore> actually, i think the 1g has 512K flash.  Comets have bigger 1 or 2M flash.
[21:21:54] <quasi> astinus: no, no ext2 in solaris
[21:22:14] <astinus> quasi: Any kinda common filesystem support between Linux <-> Solaris?
[21:22:18] <delewis> astinus: pcfs
[21:22:19] <sickness> astinus: heh, I gone the ultracheap way for my home zfs server, I have a simple sil3114 pci controller, but I'll buy one of these supermicro 88sx things for my next upgrade ;)
[21:22:20] <delewis> NFS
[21:22:22] <delewis> CIFS
[21:22:31] * astinus nods
[21:22:32] <delewis> but that's about it.
[21:22:34] <quasi> astinus: fat32 ;)
[21:22:39] <gdamore> it also needs custom compilers.  gcc dropped support for the needed -mflat option a while ago
[21:22:48] <astinus> delewis, Thanks - I'll mount these drives in a Linux server, and NFS the data over via Gig-E
[21:22:50] <delewis> Linux filesystems would have a hard time making it past Solaris' quality-control mandate.
[21:23:05] <gdamore> s/filesystems//
[21:23:08] <delewis> :-)
[21:23:23] <delewis> I was trying not to be *too* general, you know.
[21:23:23] <icon> yup
[21:23:32] <gdamore> i call 'em as i see 'em
[21:23:33] <icon> ext2fs was a bane
[21:23:43] <delewis> well, even filesystems like XFS and ReiserFS
[21:23:49] <delewis> *especially* ReiserFS
[21:23:52] <icon> *nods*
[21:23:58] <astinus> sickness:  sounds like you're in my shoes 2-3 years ago .. I bought 4 * 120GB drives, onboard controller, and deployed Linux.  Now I need a lot more space, I've got a Dual 3GHz Xeon with 2GB RAM in a 4U rackmount chassis .. and I'm using a 73GB scsi for the OS, 8 * 250GB and 4 * 200GB for storage, and just getting Solaris + ZFS installed today
[21:23:58] <delewis> XFS and JFS would have a chance, I think
[21:24:02] <icon> reiser required far too much tweaking
[21:24:05] <gdamore> hmmm.... i wonder, can i do setup_install_server on a netbsd server?
[21:24:06] <delewis> but why bother when ZFS is around :-)
[21:24:12] <icon> delewis++ ;)
[21:24:22] <delewis> gdamore: you should be able to -- I know people have used it on OS X/FreeBSD.
[21:24:29] <delewis> AIX's NIM is a beast, though.
[21:24:30] <astinus> delewis, I'm really watching the whole ZFS-on-Linux via FUSE development. That's going to rock.
[21:24:34] *** mjf has joined #opensolaris
[21:24:37] <delewis> I've been *trying* for months to get NIM on Solaris working
[21:24:38] <gdamore> i want to use my cobalt raq2 as an NFS server for my various ISOs
[21:24:46] <mjf> can the log area for journalling in UFS be on a seperate device?
[21:24:57] <gdamore> i don't think so
[21:25:00] <icon> delewis: will zfs as a root device be supported in nevada?
[21:25:07] <icon> err root fs
[21:25:10] <delewis> icon: sure
[21:25:12] * icon grabs more coffee
[21:25:15] <delewis> well
[21:25:18] <delewis> I take that back
[21:25:22] <icon> without using zones :)
[21:25:25] <astinus> I just loaded b48 of Nevada, and didn't manage a ZFS root
[21:25:28] <delewis> *probably*
[21:25:33] <icon> hmm
[21:25:34] <delewis> but will it get back-ported to 10, doubtful.
[21:25:40] <icon> *nod*
[21:25:45] <icon> well
[21:25:49] <icon> hmm
[21:25:57] <astinus> icon: I saw a guide describing how to setup a ZFS root the other day
[21:26:03] <icon> im really torn on how to divide up a consolodated box at home
[21:26:14] <astinus> icon: it involved putting the stuff required for boot into a small UFS partition ....
[21:26:17] *** gnu2it2 has joined #opensolaris
[21:26:18] <quasi> icon: still?
[21:26:18] <delewis> some were saying it would be in the Solaris 10 release (or the one after that), but I still very much doubt that, as it hasn't even been integrated in Nevada, yet.
[21:26:19] <icon> two local zones, wondering if i should use zfs for both of those
[21:26:24] <sommerfeld> the missing pieces are zfs boot & install to zfs
[21:26:24] <icon> quasi: still
[21:26:35] <delewis> and the installer is a dinosaur
[21:26:42] <quasi> icon: I'd definetely use zfs for zones
[21:26:45] *** mjf has left #opensolaris
[21:26:46] * astinus nods
[21:26:49] <delewis> so whenever someone will get around to adding ZFS support to it, is well, TBD.
[21:26:51] <icon> i have always used ufs - ive read about zfs quite a bit for  awhile, but never got around to actually using it
[21:26:54] <gnu2it2> how stable is opensolaris compared to solaris 10 ?
[21:27:01] <postwait> Anyone have a large ZFS filesystem that they can run a quick (non destructive) performance test on?
[21:27:08] <delewis> gnu2it2: OpenSolaris is not an operating system.
[21:27:09] <quasi> icon: http://prefetch.net/blog/index.php/2006/10/19/speeding-up-solaris-zone-creation-with-cloning/
[21:27:12] <delewis> your comparison is invalid.
[21:27:13] <icon> reading...
[21:27:30] <astinus> postwait, Give me 2-3 hours and I'll have about 2TB in a storage pool I can do tests on
[21:27:36] <astinus> postwait, right now, nope :(
[21:27:50] <jengelh> add a bunch of 950 G virtual disks to vmware
[21:27:57] <jengelh> makes you have 2 TB in a glimpse
[21:28:01] <quasi> postwait: once I get this thumper install done I can
[21:28:15] * astinus wants a thumper :/
[21:28:26] <astinus> heheheh
[21:28:35] <astinus> quasi, 12 or 24TB model?
[21:28:38] <icon> hmm
[21:28:45] <gdamore> gnu2it: solaris nevada should be reasonably stable, but it will depend on what hardware you have.
[21:28:54] * quasi mumbles some displeasure at sun giving out a loaner thumper with some hacked up linux version that kernel panics
[21:29:01] <gnu2it2> ultra 5
[21:29:17] <quasi> astinus: it would have been 12TB if they had given me all 48 drives
[21:29:21] <gdamore> the SX (not community releases) will be pretty well tested.  the SXCR less so.  and with ON nightlies still less so
[21:29:33] <gnu2it2> 277 mhz 128 mb.. kinda lightweight
[21:29:38] <slowhog> is the svn repo allow http browse?
[21:29:39] <astinus> quasi: cool :D
[21:29:47] <icon> i tend to start with SX builds, and then update the bits that i work off of
[21:29:50] <gdamore> you need more RAM for sure, and a faster processor wouldn't hurt
[21:29:51] <icon> stays pretty stable
[21:30:01] <icon> jds will probably be your biggest painpoint
[21:30:15] <gdamore> just run xfce and be much happier.
[21:30:31] <gnu2it2> sad thing is pay more for freight then parts on ebay for this puppy
[21:30:31] <quasi> astinus: so there's hardly any space on those 22 disks
[21:30:36] <gdamore> until you start firefox. :-)
[21:30:58] <icon> quasi: would i be far off by saying its unneccessary to run hardware raid arrays with zfs pool?
[21:31:05] <icon> err a pool
[21:31:40] <astinus> quasi:  I'm a student, gonna be on industrial placement before long .. looking to get a bigger case, two of those lovely 88sx controllers, and move from 250GB to 400 or 500GB drives - should be almost thumper-size :P
[21:31:44] <quasi> icon: it isn't recommended to run hardware raid under zfs
[21:31:51] <Stric> postwait: I've got a 1.4TB raidz2
[21:31:54] <astinus> that is until they release a 36TB model of Thumper using 750GB drives
[21:32:33] <icon> quasi: *nod* makes sense
[21:32:38] <astinus> Stric: your name looks mighty familiar .. you a sponsor / staffer? or am I going loopy?
[21:32:38] <quasi> astinus: I have seen a thumper clone that was a 36TB model for roughly the price of a 12TB thumper
[21:32:49] <Stric> astinus: leguin?
[21:32:51] * gdamore remembers when a Terabyte was something incredibly special.
[21:32:53] <icon> in practice, how stable is it compared to a ufs system on raid5?
[21:32:54] * astinus nods
[21:32:57] <astinus> Stric, thought so :)
[21:33:19] <icon> reading about it, it makes you wonder why you dropped the cash for hardware raid...
[21:33:24] *** Triskelios has quit IRC
[21:33:27] <astinus> quasi: not in 4U though?
[21:33:33] <quasi> astinus: in 4U
[21:33:37] <astinus> owch
[21:33:38] <astinus> cool
[21:33:45] <quasi> astinus: using 750G drives
[21:33:51] <astinus> quasi, link?
[21:33:56] <gdamore> i've not played with zfs yet.  but hardware raid _historically_ has been necessary to get decent write perf with raid5.
[21:34:03] <Auralis> those thumber clones miss one critical aspect, the multiple buses to handle the bandwidth
[21:34:05] *** mage2 has joined #opensolaris
[21:34:09] <mage2> Hellloo all
[21:34:23] <Stric> astinus: www.sun.com, x4500
[21:34:32] <quasi> astinus: unfortunately not - I saw it at the .nl storage expo last week, but haven't found it on the web
[21:34:39] <quasi> Stric: not for the clone
[21:34:44] <Stric> .. clone?
[21:34:49] <astinus> Stric, whats leguin? runs Solaris, right?
[21:34:58] <Stric> astinus: AIX :P
[21:34:59] <astinus> Stric, clone = cheap imitation
[21:35:18] <astinus> Stric: its been a while, I remember SSH'ing to it months ago, and noticing it wasn't Linux
[21:35:32] * astinus scratches his head
[21:35:47] <astinus> I was sure one of the boxes runs Solaris, and now I'm trying to remember which one....
[21:35:55] <Stric> astinus: err. no.. wait.. it's linux.. our other irc server is aix though :)
[21:36:07] * astinus sighs
[21:36:12] <astinus> now I'm confused :P
[21:36:16] * Stric too
[21:36:16] <astinus> ah well :)
[21:36:36] <astinus> Stric:  just one of those "am I going mental" moments ;)
[21:36:55] <astinus> Stric: I'm no longer staff, so its not easy to just check, heheheh
[21:37:23] <Stric> our new backend for cdimage.debian.org, ftp.se.debian.org, se.ubuntu.releases.com, ftp.gnome.org etc is running a 7x300G raidz2 on SXCR.. seems to be getting performance then out of the netapp fas270
[21:37:44] <Stric> s/then/than/
[21:37:48] <astinus> nice, what controller?
[21:37:55] <quasi> Stric: not bad
[21:38:08] <Stric> well.. some HP and some IBM scsi controllers :)
[21:38:19] <Stric> on a 2x750 blade1000
[21:38:28] <astinus> ah, 300G scsi drives - nice :)
[21:38:47] <gdamore> Stric: that's awesome.
[21:38:54] <Stric> the "writing over nfs" performance issue is biting hard though
[21:38:55] <jengelh> and by no means cheap
[21:39:03] <astinus> indeed
[21:39:06] <jengelh> 300G in scsi == big bucks
[21:39:06] <Stric> donations from canonical etc :)
[21:39:23] <astinus> jengelh, 7 * 300G in SCSI == mega bucks
[21:39:33] <jengelh> == 25 * 500G in SATA
[21:39:38] <jengelh> ("slightly" rounded)
[21:39:39] <Stric> http://stric.se/tmp/fs.html <- performance
[21:39:54] <Stric> write //   means two writes in parallel
[21:40:05] <Stric> ufs does -not- look good there :P
[21:40:20] <Stric> and raidz2 is chewing on those poor cpus..
[21:44:39] *** AbeFroman has quit IRC
[21:47:19] *** AbeFroman has joined #opensolaris
[21:47:21] <astinus> Okay, I have two graphics cards in this computer .. they're both nVidia 6200's on PCIe .. during setup I only had one monitor connected, but I'd like all three to work now everything is installed. Is there any 'easy' way to reconfigure X.org to make magic happen in Solaris, or should I just monkey with the config file manually?
[21:48:19] <alanc> install the nvidia driver and use the /usr/bin/nvidia-xconfig tool to update your xorg.conf for you
[21:48:50] <astinus> alanc++ :)  thanks!
[21:51:49] *** _william_ has joined #opensolaris
[21:52:19] <_william_> hi all
[21:52:43] <astinus> hiya :)
[21:55:50] <gnu2it2> anyone ever clean a type 5 keyboard? my neighbor dumped a beer in it last weekend. i dried it out best i could, but still fails to function
[21:56:43] *** nwf has quit IRC
[21:58:43] <Stric> gnu2it2: put it in the bath tub. add water. wait. hang out to dry. if still not working, it probably fried while pouring beer.
[21:59:38] <astinus> gnu2it2, if you have an airing cupboard, put it in there to dry, leave for /at least/ 24 hours
[22:02:30] *** Netwolf has joined #opensolaris
[22:02:35] <gnu2it2> afraid to soak on tub.. has a circuit  board
[22:02:46] <mage2> Anyone ever try to put a bcme interface on a vlan?
[22:03:09] <gnu2it2> been shaken out and drying since last saturday night
[22:03:17] <Stric> gnu2it2: chips doesn't care about water while not connected
[22:04:02] <Stric> anyway, what do you have to lose if it's not working now?
[22:04:24] <gdamore> the problem is drying it out before you test.  water gets _everywhere_.  it is hard to ensure that it is fully dry afterwards.
[22:04:43] <Error_404> i ran a model M through the dishwasher... it ran fine
[22:04:45] <dwc-> water > beer
[22:05:21] * gdamore has a stockpile of type 6 keyboards just to cover accidents
[22:06:11] <gdamore> i only have one or two type 5 keyboards, though, and if those die i'm hosed.
[22:06:20] <gdamore> (i think i have a type 4 somewhere though.)
[22:07:23] <astinus> gnu2it2, seriously - if its not working now.. you have very little to lose, and thats why I said to leave it in a warm room for >24 hours (preferably 2 days) *after* seriously shaking the sh*t out of it to remove loose water
[22:07:39] <alanc> I have a type 2 on my shelf that I found cleaning out a closet in an old building we were moving out of
[22:07:52] <astinus> aren't they up to type7 now?
[22:08:00] <alanc> talk about a solid keyboard - I think the case is steel
[22:08:14] <alanc> astinus: yes - I'm typing on a Type 7 rright now
[22:09:01] <astinus> alanc, how do you rate it?  i'm a picky person about keyboards... does it give good tactile response?
[22:09:10] <Stric> astinus: 7 is usb, 6 is usb and sun, 5- is sun
[22:09:15] <gdamore> believe it or not, my favorite keyboard of all time is still the type 5 clone keyboard found on the UltraBook IIi.
[22:09:20] <Error_404> model M... the only keyboard
[22:09:31] <alanc> it's okay to me - been using it almost a year now - but I'm not that picky - I never understood why people hated the type 6 so much
[22:09:44] <gdamore> i like the type 6, and i am picky
[22:09:54] <Error_404> alanc: because the type6 is like typing on bread
[22:09:58] <Error_404> it's so... mushy
[22:09:58] *** yippi has quit IRC
[22:10:01] <gdamore> but its a let down compared the old type 5
[22:10:08] 
[22:10:23] <gdamore> didn't they just release type 7?  i've never even seen one
[22:10:35] <alanc> they've only been on sale since last December, so probably not many on eBay yet
[22:10:38] * Stric likes Logitech Ultra-X (older versions) and Microsoft Natural Ergonomic 4000 (not the earlier naturals)
[22:10:42] <gdamore> i seem to recall sun having a "fire sale" to get rid of their type 6s
[22:11:08] <gdamore> i don't want any keyboard that doesn't put control and esc in the right places.
[22:11:35] <alanc> had to get rid of the Type 6's before RoHS made it illegal to sell them in Europe, since Sun doesn't like having separate European vs. rest-of-world price lists
[22:11:47] <gdamore> ah, that makes sense.
[22:11:49] *** Disorganized has quit IRC
[22:12:04] <spike723> I bet a lot of stuff will come onto ebay soon cause of rohs
[22:12:05] <gdamore> i think we picked up a bunch.
[22:12:34] <gdamore> not so much, i think.  rohs has been known for some time, so everyone has already converted or sold off inventory
[22:12:49] <gdamore> doesn't california have rohs laws too?
[22:12:51] <astinus> why can't I find just a keyboard on the bloody store
[22:13:06] <alanc> that was the motiviation for a hell of a lot of product refreshes in the last year - the new Ultra 25/45 SPARC workstations, the Sun Ray 2 series, the Type 7 keyboards... all had to get the lead out of all their components, and most suppliers would only certify they'd done that for new stuff, not the components used in the older models
[22:13:12] <gdamore> you have to get the country kit, which includes a mouse and power cable, probably
[22:13:35] <alanc> it's even the main reason Sun mice finally have a scroll wheel 8-)
[22:13:45] <AbeFroman> thank you RoHS
[22:13:46] <gdamore> heh.  infineon refused to do a rohs copernicus (microsparc variant)
[22:14:16] <astinus> gdamore, found it :) 30 quid, lol!
[22:14:43] <gdamore> when i bought my U20 last year, it came with a scroll mouse, and it wasn't clearly sun branded.
[22:14:50] <alanc> I think California doesn't ban the sale of non-RoHS components, they just charge extra disposal fees for them
[22:15:01] <gdamore> apparently type 6 mice didn't work properly with the U20.
[22:15:16] <alanc> that was because the Type 6 mouse didn't work with the U20 USB controller and the Type 7 wasn't out yet
[22:15:48] <gdamore> there is another problem with some sun type 6 mice being out of USB spec.  there are tales of "dark tailed" and "light tailed" type 6 mice (i have some of each)
[22:16:22] <gdamore> it impacted compatibility on some of our hardware, and it took us a while to figure out the sun mouse was actually not fully usb compliant
[22:18:47] <astinus> i think a mouse is even more personal than keyboards ..
[22:18:55] <astinus> if its not your "use most of the time" computer - fine
[22:19:51] * astinus shrugs
[22:20:13] <astinus> guess it stems from having a bad wrist, got thrown off a pony as a kid .. some kb/mouse aggravate it more than others
[22:21:23] <sickness> astinus: yeah, I always did linux+swraid or openbsd+swraid, but now I wanted to try zfs, and I'm really happy :) bought 3 320gb sata2 disks at a local computer fair and a cheap asus barbone (vintage) with a 64bit sempron, works like a charm with opensolaris :)))
[22:21:24] * astinus buys a country kit to try out type 7 kb/mouse :)
[22:21:49] <sickness> http://www.sickness.it/zfshomeserver.txt  http://www.sickness.it/zfshomeserver.jpg  http://www.sickness.it/zfshomeserver.png
[22:21:52] <sickness> :)))
[22:22:25] <astinus> sickness: cool - I'm a gentoo developer (hence the gentoo/developer/astinus cloak) .. but I'm sick and tired of linux filesystems being unreliable, the age old concept of partitioning and LVM2 is piss poor for big arrays
[22:22:35] <astinus> ZFS looked like the silver bullet :P
[22:22:46] <sickness> astinus: yeah, I know
[22:23:23] * gdamore wishes he had some hardware where it made sense to run zfs.
[22:23:47] <oxygene> zfs helps me a lot on my single-disk desktop and laptop already
[22:23:58] <gdamore> really? how so?
[22:24:31] <gdamore> i'm not a file system guy, so bear with me.  i thought zfs was really just for big arrays of drives.
[22:24:42] <astinus> 8 * 250GB and 4 * 200GB .. trying to manage that with Linux means two RAID arrays, both were RAID-5 so I lost 450GB straight away (seems to be less than 450GB with ZFS compression enabled) and even LVM (which worked for >12 months semi-OK) sucks compared to the concept of pools
[22:26:07] <oxygene> gdamore: very flexible filesystem layout - create a big pool, then create dynamically sized filesystems from it.. snapshots help with some experiments, throw-away filesystems (zfs destroy/create to instantly clean a part of the directory hierarchy)
[22:26:45] <gdamore> neat.  what i really need is to be able to support a bunch of different root filesystems.  some S8, S9, S10, SNV, NetBSD, etc.
[22:26:50] *** jcsmith has quit IRC
[22:27:13] <astinus> gdamore, zones?
[22:27:17] *** coffman has left #opensolaris
[22:27:18] <oxygene> gdamore: live upgrades will be dead simple once zfsroot has landed, too: clone the filesystem, update the clone, switch
[22:27:22] <slowhog> wget -r -l1 --no-parent -A.bz2 http://dlc.sun.com/osol/companion/downloads/current/pkgs/sparc does not work
[22:27:34] <slowhog> get 302 response
[22:28:04] <gdamore> nice.
[22:28:15] <gdamore> zones won't work, because they have to run the same OS
[22:28:22] <gdamore> (same kernel)
[22:28:43] <astinus> gdamore, BrandZ ?
[22:28:54] <gdamore> BrandZ == Solaris kernel, Linux userland, IIUC
[22:29:00] <astinus> hm
[22:29:03] <astinus> :(
[22:29:04] <oxygene> yes
[22:29:23] <oxygene> maybe they'll integrate xen or qemu into the zones UI ;)
[22:29:30] <astinus> i hear Xen is coming
[22:29:43] <Error_404> i should try running maple in a branded zone
[22:29:48] <gdamore> qmeu is not ready for prime time.  Xen would be nice.  I have to update my primary desktop to SNV to use it though.
[22:29:51] <jengelh> brandz seems to be something different
[22:29:54] <jengelh> http://en.wikipedia.org/wiki/BrandZ
[22:30:07] <gdamore> what i'd really like to see is  VMware on a Solaris host.
[22:30:32] <jengelh> there's no market for that :p
[22:30:45] <oxygene> gdamore: I don't consider xen to be "ready" in any shape or form either
[22:30:48] *** coffman has joined #opensolaris
[22:31:15] <Error_404> maple is about the only closed-source linux program i'd have any use for brandz with
[22:31:16] <gdamore> xen is much closer than qemu though.
[22:31:28] <slowhog> running wget from Solaris 10 6/06 s10s_u2wos_07a SPARC
[22:31:44] <sickness> gdamore: yeah, I'd really want to see that too
[22:32:09] <oxygene> gdamore: let's see.. it either needs a modified client, or qemu-style peripheral hardware emulation - I fail to see where that should be any better
[22:33:15] <mage2> can you configure vlans on bcme interfaces
[22:34:29] <gdamore> the difference is xen is being actively developed by a larger audience, including big commercial interests like Sun.  so getting a modified client is not impossible.
[22:34:41] <jerome__> it's not a sunpci2 that they gave me for free
[22:34:46] <jerome__> no clue what it is
[22:35:00] <gdamore> and with the support in some newer processors, Xen can run unmodified OS'
[22:35:22] <oxygene> gdamore: as I see it, the trend seems to go towards using cpu virtualization techniques.. and for the next few years it still needs peripheral emulation until the mainboard chipsets are ready for virtualization, too
[22:35:27] <delewis> whereas QEmu was started by the same fellow that started FFmpeg (which should tell you something about the overall quality)
[22:36:13] <jerome__> It runs a mobile athlon 2100+
[22:36:16] <jerome__> what sunpci is that?
[22:36:24] *** Disorganized has joined #OpenSolaris
[22:36:26] <gdamore> probably sunpci3
[22:36:31] <jerome__> no shit
[22:36:37] <jerome__> they gave me a sunpci3 for free??
[22:36:41] <jerome__> that's pretty cool
[22:36:45] <gdamore> who? how did you get one?
[22:37:16] <jerome__> some dude gave it to me for free
[22:37:23] <jerome__> because I bought a blade 150, I was looking for one
[22:37:28] <jerome__> he just gave it to me
[22:37:35] <jerome__> I thought at first I'd not use it
[22:37:40] <jerome__> because I thought it was a sunpci I
[22:37:46] <jerome__> but I'm using it now
[22:37:54] <gdamore> qemu is an interesting project.  i've tried to use it myself.  but it has some major shortcomings.  development on it seems to have been somewhat slow over the past year.
[22:37:54] <jerome__> and I see when it boots: sun PCI III
[22:37:57] <jerome__> with 700MB+
[22:38:09] <gdamore> crap.  has he got any more extras?
[22:38:19] *** stevel has joined #opensolaris
[22:38:20] *** ChanServ sets mode: +o stevel
[22:38:34] <gdamore> (Sun PCi3 is why Bullfrog can support a full size PCI card.)
[22:38:37] <jerome__> no sorry he had only one and no sparc
[22:38:40] <slowhog> seems like a bug for the wget on sparc
[22:38:53] <jerome__> this is indeed very cool I'd say
[22:38:59] <jerome__> can I put a scsi drive in a blade 150?
[22:39:04] <delewis> jerome__: you *could*
[22:39:05] <jerome__> would an adaptec 2940 work?
[22:39:08] <gdamore> if you also put in a controller.
[22:39:13] <delewis> if you find an SCSI controller with the proper Fcode
[22:39:16] <delewis> (if you want it to be bootable)
[22:39:17] <jerome__> adaptec 2940uw ok?
[22:39:22] <jerome__> oh ok
[22:39:25] <jerome__> ebay time then
[22:39:31] <jerome__> what's a good card to use/buy?
[22:39:34] <delewis> jerome__: google for a Sun part number
[22:39:58] <delewis> jerome__: head over to SunSolve and look at the Blade 100/150 system pages, as well as a few others to see if you can find a generic SCSI controller
[22:40:00] <gdamore> it might or might not work without fcode (even without booting).  because some of the sun drivers historically rely on fcode initialization, even when not using the device during boot
[22:40:05] <jerome__> okay
[22:40:11] <delewis> those are guaranteed to have the Fcodes
[22:40:19] <jerome__> uinderstood
[22:40:29] <gdamore> a qlogic controller would be a good idea.  driver support for qlogic chips has been good over the years.
[22:40:32] <delewis> even so, the Blade 150 has one PCI bus
[22:40:35] <delewis> *one*
[22:40:37] <jerome__> damn man a sunpci III for free!!!!
[22:40:38] <jerome__> this is cool
[22:40:52] <astinus> :p
[22:41:03] <delewis> the Ultra 5/10/Blade 100/150 are bottlenecked more ways than I want to know
[22:41:04] * gdamore wonders if a sunpci3 will work in the opteron based U20.
[22:41:07] <jerome__> even is the blade 150 is kinda slow, I'm not complaining
[22:41:16] <jerome__> it's usable
[22:41:22] <delewis> jerome__: try doing some IO :-)
[22:41:26] <jerome__> yeah right
[22:41:27] <delewis> processor speed is *not* the problem.
[22:41:29] <gdamore> heh.  the sunpci3 is probably a bit faster than the host sb150.
[22:41:30] <jerome__> wait till I have scsi :)
[22:41:39] <delewis> jerome__: you still only have one PCI bus
[22:41:45] <delewis> that's shared by *everything*
[22:41:46] <jerome__> yeah well
[22:41:49] <jengelh> why would you need a sunpci-x86 when the host processor is already x86-capable ;-)
[22:41:51] <jerome__> it's a workstation, it's fine
[22:41:54] <delewis> gdamore: absolutely :-)
[22:42:11] <jerome__> man a sunpci 3 :)
[22:42:12] <gdamore> jengelh: because i want to run windows (well need to run windows is a better statement).
[22:42:26] <jengelh> until vmware comes, that's an argument.
[22:42:38] <gdamore> i'd rather have vmware, true.
[22:42:42] <delewis> gdamore: you can run IE in Wine
[22:42:50] <delewis> if you absolutely have to
[22:42:55] <slowhog> uh, the problem is a domain entry in /etc/resolv.conf
[22:43:02] <jerome__> http://www.memoryxsun.com/sunbl150sc.html
[22:43:03] <slowhog> remove the domain entry, the redirection works
[22:43:04] <jerome__> which one is best?
[22:43:12] <gdamore> the apps i need have a bunch of active x controls that probably won't work
[22:43:24] <delewis> gdamore: worth a try :-)
[22:43:30] <delewis> though, Wine takes *forever*
[22:43:34] <delewis> building*
[22:44:09] <gdamore> wine looks "challenging" to configure.  frankly, i've never wanted to try hard enough to look very far into it
[22:44:21] <delewis> gdamore: no, the build process actually works.
[22:44:26] *** gnu2it2 has quit IRC
[22:44:43] <delewis> and this is coming from someone that maintains the MPlayer Solaris port :-)
[22:45:28] <gdamore> well, i got a disk drive loaded with XP Pro for my x86 laptop yesterday, so i don't need to solve the problem right now.
[22:46:11] <icon> what is write perf like on zfs?
[22:46:33] <delewis> icon: not any worse than UFS in most cases
[22:46:43] <icon> hrmm
[22:46:56] <icon> considering converting a few volumes to give it a shot
[22:46:59] <delewis> with compression, write performance tends to be even better.
[22:47:00] <Stric> in my opinion: much better than ufs
[22:47:07] <delewis> given you aren't CPU bound
[22:47:25] <icon> any issues with servering nfs content on zfs volumes?
[22:47:30] <sommerfeld> icon: like most such things, it really depends on workload
[22:47:32] <icon> (including loopback)
[22:47:53] <Error_404> my NFS exports are all zfs
[22:47:54] <Error_404> *shrug*
[22:47:55] <icon> just looking for a mean, assuming workstation and light server load
[22:48:07] <Error_404> apparantly benr was having problems with iscsi target performance
[22:48:09] <astinus> hm, I allocated 35G of my 73G drive to Solaris to start with, but would like to use all 73G now -- any easy way to expand?
[22:48:48] <astinus> its a UFS root filesystem...
[22:49:12] <Stric> icon: write performance on zfs-nfs kinda sucks up until late builds.. because of nfsd issuing fsync() which means sync on the entire pool on zfs
[22:49:17] <delewis> astinus: there's some undocumented option in newfs, IIRC.
[22:49:30] <delewis> I say undocumented, as expanding a UFS filesystem isn't "recommended"
[22:49:39] <icon> hurm
[22:49:49] <Stric> or you can use the command 'growfs'
[22:50:04] <astinus> i think I only partitioned half the drive to be 'SOLARIS' in the setup
[22:50:13] <delewis> ah, I don't think growfs has always been around
[22:50:32] <delewis> doesn't growfs depend on SVM?
[22:50:56] <delewis> hmm, maybe not.
[22:50:57] <icon> hrmm, pooling is a very different concept to me
[22:51:08] <icon> starting to think of how i want to chop up disks
[22:51:22] <icon> i need one ufs part for the global zone
[22:51:47] <icon> throw the rest on zfs... but im scratching my head as to where to place the volumes, and if zfs can even support partial disks
[22:51:50] <icon> rtfm time :)
[22:52:26] <astinus> If I don't want to add a whole disk to a ZFS pool, is that an option?
[22:52:29] * delewis wishes expanding filesystems in Solaris were was easy as expanding them in AIX
[22:52:32] * astinus snickers
[22:52:38] <Stric> astinus: yes, you can add a slice.
[22:52:47] <astinus> Stric, how do I 'reslice' a disk?
[22:52:53] <delewis> chfs -a size=+<number of 512-byte sized blocks> <filesystem>
[22:52:54] <delewis> *poof*
[22:53:01] <Stric> astinus: you use format as always
[22:53:04] *** mage2 has quit IRC
[22:53:08] <astinus> cool :)
[22:53:25] * quasi swears at smf and its whole dependency on filesystem/local - it is bloody hard to get back up
[22:53:27] <Stric> delewis: yeah, the lvm is one of the nicer things in aix
[22:53:34] <quasi> when hba is missing
[22:53:40] <astinus> Stric, i feel wet behind the ears again, its a weird feeling
[22:53:50] <Stric> astinus: grab a towel?
[22:54:22] <astinus> Stric, your towel was good ;) I'm used to fdisk or cfdisk, now I've found format :D
[22:54:48] <Stric> :)
[22:55:29] <astinus> freenode++  I'd be screwed without all you guys (and girls) :)
[22:55:48] <quasi> girls? where, where? ;)
[22:56:16] <astinus> quasi, i'm on a CS course .. we get ~3% girls every year
[22:56:42] <astinus> quasi, of course, most leave <=2 weeks into the course because they thought 'Computer Science' was all spreadsheets .. rather than hacking with Java and C :P
[22:57:12] <quasi> astinus: you write code on a CS course? odd place
[22:57:17] <delewis> memorizing Java APIs is computer science? *ducks*
[22:57:18] <delewis> :-)
[22:57:30] * astinus aimlessly thwaps at people
[22:59:04] <astinus> delewis:  its ironic, the dept claims we're learning "useful stuff" and also claims they're "on the cutting edge" yet we've got 30+ old CRT sunrays driven off a single V240 (dual 1ghz, 3GB RAM, very slow!) and for group projects using version control == out of the question, its paper change control forms or else!
[22:59:48] *** eboutilier has joined #opensolaris
[23:00:10] <quasi> "on the cutting edge of insanity" ;)
[23:00:31] <delewis> ugh.
[23:00:32] <astinus> yeah :S
[23:00:39] <delewis> well I suppose no SCM is better than 50 different SCMs :-)
[23:01:07] <gdamore> heh.  i resemble that remark.
[23:01:36] <astinus> i've got an enlightened member of staff as my "customer" for this years project, so we're using Subversion .. but thats only available because I've got colocated hardware off-site, and I have to hand out pen drives with TortoiseSVN on :/
[23:01:51] <delewis> gdamore: me too -- though, I just have to mess with svn & cvs, and I'm really forcing myself not to mess with Hg, though, it is tempting.
[23:02:00] <gdamore> the CRT sun rays are probably not that bad, nor is the server. its just you need more ram.  a lot more ram.  :-)
[23:02:13] <astinus> gdamore, top doesn't show much being used ever :/
[23:02:17] <Stric> astinus: we have 16 rays on 2xquad v440 w/ 16G ram each :P
[23:02:21] <astinus> gdamore, load averages are frequently 10+
[23:02:25] <delewis> gdamore: top?
[23:02:26] <delewis> *evil*!
[23:02:29] <gdamore> i will be switching to Hg, once i can switch completely from svn.
[23:02:30] <jengelh> CSWtop!
[23:02:36] * astinus grins
[23:02:52] <astinus> i dunno what the problem is, I'm new to Solaris, but it *shouldnt* be that slow, really!!
[23:02:54] <gdamore> delewis: what are your users doing?
[23:02:56] <delewis> gdamore: do you not like svn?
[23:03:09] <gdamore> i don't like having yet another SCM.
[23:03:25] <delewis> gdamore: well, there's my personal repository, and the MPlayer SVN repository, which is what I'm referring to -- plus, the genunix.org SVN repository.
[23:03:35] <gdamore> do you have 16 users each running seti@home or something?
[23:03:38] <delewis> my personal repository is actullay cvs :-)
[23:03:58] <jengelh>  /platform/i86pc/boot_archive is 27 megs. Now don't tell me all that stuff is loaded at once into memory during boot -.-
[23:03:59] <gdamore> i only use svn for ON
[23:04:06] <delewis> which has much *lighter* dependencies than SVN
[23:04:07] <astinus> gdamore, erm its all Java Desktop Environment .. and they run Eclipse, open lots of PDFs and use Firefox with loads of tabs
[23:04:12] <delewis> yeah, that's what I'm using genunix.org for
[23:04:16] <gdamore> i have CVS for almost everything else.
[23:04:20] <delewis> but genunix tends to be out of date by about 2 weeks
[23:04:32] <gdamore> firefox is evil, evil, evil.
[23:04:32] <delewis> they still haven't merged 10061016
[23:04:54] * delewis still has the 10061009 trunk
[23:05:05] <gdamore> i wish we could roll back the calendar and shoot whatever idjiot created flash.
[23:05:10] <jengelh> is there some howto on how to install a second kernel?
[23:05:29] <astinus> gdamore: so basically, they're running lots of RAM-heavy applications and stuff like Flash :P
[23:05:33] <gdamore> yes, though i don't recall exactly where it is
[23:05:38] <astinus> gdamore: and it grinds the little V240 to a bloody standstill
[23:05:39] <delewis> astinus: evil.
[23:05:44] <delewis> pkill -9 firefox
[23:05:47] <delewis> all problems solved :-)
[23:05:51] <gdamore> Flash is a pig.  it is why my firefox is frequently > 500MB
[23:06:02] <astinus> Lecturer: "Could everyone please open the assignment PDF and browse to the course directory to view the animation of how I want your data structures to be created?"
[23:06:06] <delewis> yes, I'm strongly considering killing Flash on my SB1000
[23:06:09] <astinus> ... five seconds later .... *SPLAT*
[23:06:10] <delewis> I'm using it over SSGD remotel
[23:06:19] <delewis> and the animated graphics don't make that very pleasant :-)
[23:06:22] <quasi> gdamore: not flash alone - ff leaks^Wcaches insanely
[23:06:31] <delewis> I always find myself pressing "page down" repeatedly whenever I see Flash :-)
[23:06:40] <gdamore> i have a pref to limit my cache to ~8mb.
[23:06:53] <delewis> actually
[23:06:57] <delewis> I should migrate back to Opera
[23:06:58] <gdamore> but still firefox is huge.  and it seems to grow really big as soon as flash/shockwave get invoked
[23:07:01] <astinus> so what do you guys reckon?  upgrade to 16G of RAM and see how the V240 fares? :P
[23:07:06] <jengelh> I have no clue what I did "wrong" but actually I just happened to successfully compile uts
[23:07:07] <delewis> which *never* has a footprint above 60MB
[23:07:11] <quasi> gdamore: do they allow limiting the memory it uses?
[23:07:19] <gdamore> i don't think so.
[23:07:35] <astinus> you can set how much disk cache you want
[23:07:40] <gdamore> opera works fine, until it crashes.  which happens about once a day for me.  i keep trying the new releases hoping they'll fix it, but not yet
[23:07:41] <astinus> but you can't memlimit
[23:07:49] *** sparc-kly has joined #opensolaris
[23:07:49] *** ChanServ sets mode: +o sparc-kly
[23:08:07] <gdamore> i would definitely throw more memory at your sun ray server.
[23:08:19] <delewis> gdamore: really? interesting, Opera crashes on me *very* rarely.
[23:08:29] <delewis> Opera 8.x was rock-solid.
[23:08:31] <quasi> astinus: the more memory the better - no doubt about that
[23:08:36] <gdamore> on Solaris?
[23:08:38] <delewis> though, I think 9.x has some stability issues, which I haven't used much.
[23:08:39] <delewis> gdamore: yes
[23:09:01] <Stric> astinus: you can set with 'ulimit -d' and 'ulimit -v', but that just means ff will crash when it hits the limit :)
[23:09:01] <gdamore> i don't recall if i tried 8.  I've tried ~every 9.x release so far, though.
[23:09:11] <delewis> I never had any crashes with 8
[23:09:20] <delewis> which is pretty impressive, given how long I keep my browser sessions open
[23:09:23] <gdamore> i just wish i could kill off flash, without nuking all of firefox.
[23:09:27] <delewis> (usually 2 weeks at a time)
[23:09:43] *** hile_ has joined #opensolaris
[23:09:56] <Stric> gdamore: install Click To Play or some similar.. so you have to click on flash thingies to activate them..
[23:10:01] <astinus> Stric: not really conducive to a good learning environment
[23:10:01] <gdamore> i tend to do the same.  unless i try opera, where a daily crash seems to keep me from running it that long
[23:10:10] <astinus> i guess I just need to request they add more RAM
[23:10:17] <astinus> or just remote desktop to my room, which has plenty
[23:10:39] <gdamore> click to flash won't prevent the leaks, it will just annoy me.  when i close the firefox window, i want that instance of the flash application to go _away_
[23:11:24] <gdamore> IIUC flash in firefox is just a thread, and the application uses firefox to alloc memory.  in retrospect a bad design decision
[23:11:26] <jengelh> erk manpage trouble again. Assume I have the file /export/jengelh/WS/usr/src/tools/scripts/Install.1, how do I display that?
[23:12:00] <hile_> i'd use nroff directly :)
[23:12:35] <gdamore> if iit isn't in a directory named "man/man1", you have to use nroff
[23:12:36] <jengelh> nroff Install.1  does not look too good
[23:12:43] <gdamore> nroff -man Install.1
[23:12:46] <gdamore> unless its sgml.
[23:12:54] <jengelh> ah wonderful
[23:13:10] <gdamore> then its harder: /usr/lib/sgml/sgml2roff <args>  | nroff -man -
[23:13:24] <jengelh> When I get this finished, I'll... well what could I?
[23:13:29] <jengelh> throw a party? go on an icecream?
[23:13:41] <gdamore> have a beer!
[23:13:52] <jengelh> which brand? (note I am in Germany)
[23:14:08] <jengelh> nm, I think I know a good one
[23:14:09] <jengelh> :)
[23:14:10] <gdamore> trappist.
[23:14:16] <jengelh> does not look like we have that
[23:14:24] <gdamore> shame.
[23:14:29] <astinus> argh, looks like I need to reinstall :/
[23:14:36] <jengelh> (Would have expected like: Becks Warsteiner Herforder, etc.)
[23:14:37] <gdamore> most of the rest of my favorites are english varieties.
[23:14:38] * quasi grabs a rochefort no. 8
[23:14:39] <astinus> I have two active partitions on my 73GB scsi
[23:14:47] <astinus> one is active, type Solaris2
[23:14:57] <astinus> other isn't active, Solaris2 .. and I want to use it for ZFS :/
[23:17:05] <quasi> gdamore: anything in particular?
[23:17:26] *** gnu2it2 has joined #opensolaris
[23:18:28] <gdamore> bass is good.  but i've had so many others while visiting, its hard to pick just one.
[23:18:59] * quasi is partial to London Porter and Oatmeal Stout
[23:19:07] <gdamore> trappist (sold in the us as chimay) double is still my favorite
[23:19:25] <gdamore> i like darker stuff.  stouts and porters are good.
[23:19:50] *** gisburn has joined #opensolaris
[23:19:55] <quasi> gdamore: trappist is the brewing "method" - bascilly means brewed by monks
[23:20:00] <astinus> c0t6d0p1 is the second partition on that disk, right?
[23:20:01] <gdamore> we have a local variety that i like to: "Stone Arrogant Bastard"
[23:20:05] <gisburn> alanc: ping!
[23:20:15] <quasi> gdamore: one of those monastaries is chimay
[23:20:18] <alanc> pong
[23:20:23] <g4lt-mordant> arrogant bastard is national....
[23:20:26] <gisburn> alanc: Is my "add_install_client" IPv6 bug queued ?
[23:20:31] <g4lt-mordant> double bastard is much better
[23:20:40] <jengelh> c1t3d3p7
[23:20:56] <gisburn> jengelh: c3p0!
[23:20:56] *** bengtf has quit IRC
[23:21:01] <gdamore> ah.   when i was in holland i ordered it as "Trappist Dubbel".  There was a triple, but it was more alcohol and less flavorful  i thought
[23:21:08] <alanc> I don't see any new bugs in the triage queue - how long ago did you file it?
[23:21:18] <gisburn> alanc: 15-20mins
[23:21:37] <gdamore> btw, the stone brewery is in san diego.  i didn't realize you could get it nationally
[23:21:44] <quasi> gdamore: that's another brand
[23:22:08] <gisburn> alanc: Is the bug reporting system eating the reports again ?
[23:22:11] <gdamore> ah.  well, shows you what i know.  :-)
[23:22:14] <jengelh> gisburn: Heh
[23:22:23] <quasi> gdamore: and you have to be careful - the tripel is horrendously bad
[23:22:38] <gdamore> i tried the tripel, but i much preferred the dubbel
[23:22:41] <gisburn> jengelh: C3P0/R2D2!!
[23:22:46] <alanc> I don't know - I don't know how long the delay is supposed to be
[23:23:03] <jengelh> well in that case I prefer c0t0d0s0.org
[23:23:06] * gisburn growls
[23:23:15] *** jcea has joined #opensolaris
[23:23:44] <gisburn> stevel: is there a way to get the diffs for the HG changes in the onnv-notify list, too ?
[23:23:56] <stevel> gisburn: i don't want to do that
[23:24:00] <gisburn> stevel: like attaching the diffs to the mails ?
[23:24:06] <gisburn> stevel: why ?
[23:24:08] <stevel> it's unnecessary
[23:24:15] *** sparc-kly_ has quit IRC
[23:24:17] <gisburn> (except "mails may become big")
[23:24:18] <stevel> and just adds noise to the email
[23:24:19] <alanc> ah - it's appeared now - 6484624: "add_install_client"'s "-i" option does not support IPv6 addresses
[23:24:20] <jengelh> "SunOS Release 5.11 Version WS 32-bit"
[23:24:51] <gisburn> "SunOS Release 6.66 Version WS 128-bit"!!, released in 2660
[23:25:06] <jengelh> 2660? Sounds late.
[23:25:09] <gisburn> yeah
[23:25:12] <gisburn> as usual
[23:25:13] <jengelh> 128-bit is already here
[23:25:31] <gisburn> jengelh: address space ?
[23:25:40] <jengelh> register size
[23:25:46] <jengelh> IIRC some box from HP
[23:25:52] <gisburn> jengelh: which type of CPU ?
[23:26:01] <jengelh> argh i fuckin can't remember :>
[23:26:16] <gisburn> 128bit appears to be the new hip discussion thing in the HPC lists.
[23:26:17] <alanc> and I've moved it from the triage queue to the install_setup category
[23:26:22] <gisburn> 128bit address space!!
[23:26:34] <gisburn> alanc: thanks! :-)
[23:26:50] <jengelh> http://en.wikipedia.org/wiki/128-bit
[23:26:58] <jengelh> IBM S/370
[23:26:59] *** Fish- has joined #opensolaris
[23:27:10] <hali> im sure we'll have 72 or something before that... as we had with the 36 bit "in betweens" the 32 and 64bit
[23:27:15] <jengelh> no
[23:27:26] <g4lt-mordant> they will come out with sparc128 in early 2630, it's just it'll take two revisions to get it right and seven revisions to get it up to the speed of the rest of the market
[23:27:27] <jengelh> if the world needed "in betweens" we would have 48 bits first (in fact - we do)
[23:27:39] <jengelh> LBA 48 for one
[23:27:48] <jengelh> another one is that current Opterons only have 48 bit addressing
[23:28:30] <gisburn> jengelh: The point of an 128bit address space are large clusters with an unified address space. even todays clusters with 16386 nodes can run out of virtual 64bit address space in some cases.
[23:28:35] *** bougie_ has joined #opensolaris
[23:28:45] <jengelh> if your name is CERN, yes
[23:28:52] <sommerfeld> actually, 36 bit architectures largely predated 32-bit.
[23:29:05] <jengelh> 32 is much more convenient because it is a power of 2 :)
[23:29:06] <delewis> 139345 dlewis     46M   30M sleep   47    4   0:00:06 3.1% opera/2
[23:29:08] <delewis> sweet :-)
[23:30:03] <gdamore> i'm now running setup_install_server on a cobalt running NetBSD.  there were a few small annoyances.  i wonder if anyone would like portability fixes for it
[23:30:47] <gisburn> gdamore: if you ask me I think the jumpstart stuff should switch over to use a POSIX shell.
[23:31:02] <gisburn> gdamore: less painfull and much more portable.
[23:31:03] <gdamore> that would go a long way to helping
[23:31:29] <g4lt-mordant> sommerfeld, didn't some DECs have 48-bit DWORDs?
[23:31:35] <gdamore> for example, i see a bunch of /\b\c stuff from the dialer on screen
[23:31:41] <delewis> last time I checked, the NetBSD /bin/sh wasn't POSIX compliant (there were a few incompatibilities)
[23:31:43] <jengelh> gdamore: Any idea why /dev/pts/0 never seems to be used?
[23:31:53] <delewis> though, this occured last year, when the NetBSD was being re-written, IIRC.
[23:32:05] <gdamore> it still isn't.
[23:32:09] <delewis> gdamore: ugh.
[23:32:32] <gdamore> there is pdksh which is a bit better, though
[23:32:40] <gisburn> delewis: if NetBSD people would be interested I could provide a POSIX shell for them... :-)
[23:32:52] <delewis> gisburn: ksh93 is already in ports :-)
[23:32:56] <gdamore> there would probably be interest if you asked.
[23:33:22] <gdamore> but ksh93 is _probably_ too big.  some of thse environments are _tiny_.  think sub 8Mb.
[23:33:24] <delewis> and there's 'dash' supposedly
[23:33:30] <delewis> which is *strictly* POSIX compliant
[23:33:37] <delewis> and very tiny
[23:33:46] <gdamore> never heard of dash before
[23:33:57] <delewis> it's something Debian uses for a POSIX compliant shell
[23:34:06] <gdamore> so probably GPLd.
[23:34:10] <delewis> given bash's POSIX-compatibility mode leaves a lot to be desired, and given how heavy bash is ;-)
[23:34:15] <jengelh> just remove all of the backward compatibilty cruft and have perfect posix compliance
[23:34:16] <delewis> gdamore: most definitely.
[23:34:29] <gdamore> GPL is not BSD friendly.
[23:34:43] *** Fish has quit IRC
[23:34:43] <delewis> gdamore: NetBSD still depends on gcc, does it not? :-P
[23:34:48] <gisburn> gdamore: no, ksh93 is not that big if you look closely what it does. If the OS can mmap() executables then everything is fine.
[23:34:53] <gdamore> yes, but not in base.
[23:35:01] <gdamore> i.e. you can ship a netbsd system without it.
[23:35:05] *** axisys has quit IRC
[23:35:05] <quasi> dash ~ Debian-version of NetBSD's lightweight bourne shell
[23:35:06] <delewis> true
[23:35:20] *** bengtf__ has joined #opensolaris
[23:35:24] *** bengtf__ is now known as bengtf
[23:35:24] <gisburn> gdamore: ksh93 just carries lots of giant strings around, including fully-blown man pages.
[23:35:25] *** bougie_ has quit IRC
[23:36:01] <gdamore> bash is a pig.  i wouldn't want to use that.
[23:36:14] <gisburn> gdamore: and libast can be build-in statically - which would automagically remove all unneccesary functions
[23:36:17] <delewis> instead of man pages, you have info pages :-)
[23:36:19] <delewis> which are infinitely worse
[23:36:30] <gisburn> gdamore: right now ksh93 only uses 1/5 of the functions in libast
[23:36:34] <gdamore> although sh on MIPS netbsd is 219732.  which is till pretty big
[23:36:34] * delewis loves how all GNU man pages suck, and refer to info pages, instead.
[23:36:50] <delewis> even on-line documentation sucks and refers you to info pages
[23:36:51] * gdamore hates how all GNU man pages suck, and refer to the info pages, instead
[23:36:53] * jengelh thanks linux-manpages project for providing "single-page" docs
[23:37:11] <gisburn> yeah
[23:37:14] <gisburn> loosers
[23:37:14] <delewis> well, my problem is not really the man pages sucking
[23:37:17] <gisburn> :-)
[23:37:19] <delewis> it's *everything* revolving around info pages
[23:37:34] <delewis> take Sun's 'cc' man page for example -- I don't expect it to go into C99 compatibility
[23:37:35] <jengelh> i just can't watch info pages with more or less, that's the problem
[23:37:35] <rydis> Well, info might not be man-pages, but info is really nice, IMO.
[23:37:36] <g4lt-mordant> deather_, it's that the info pages suck as well?
[23:37:40] <delewis> I expect *on-line* documentation to do that
[23:37:45] <g4lt-mordant> delewis, ^^^^^^^
[23:37:52] <delewis> so, head over to developers.sun.com and *poof* -- documentation
[23:37:57] * gisburn imagines a new planet system - the GNU-approved view of the universe where the sun and earth are revolving around *info*
[23:38:02] <delewis> head to say 'groff' documentation
[23:38:07] <delewis> and you get referred to the groff info page
[23:38:08] <delewis> !
[23:38:18] <jengelh> "groff info" page, what a pun
[23:38:27] <g4lt-mordant> talk about irony
[23:39:18] <g4lt-mordant> so basically gnu rewrites nroff semi-incompatibly, then decides to abandon it for info pages, and this is a good thing?
[23:39:35] <delewis> actually, groff is superior to nroff in a number of ways, I'll admit.
[23:39:38] <gdamore> holy crap.  setup_install_server calls adb.
[23:39:41] <delewis> virtually anyone since 1990 uses groff over nroff
[23:39:46] <jengelh> talk about understanding GNU
[23:39:47] <gdamore> not me.
[23:39:51] <rydis> The info format is older than GNU, so it's not that strange. :)
[23:39:52] <delewis> (all of Stevens' texts are written in groff)
[23:40:11] <gdamore> not NetBSD.
[23:40:21] <gdamore> (actually, i have to take that back, they are in groff mandoc.
[23:40:26] <delewis> does the NetBSD nroff incorporate groff extensions? :-)
[23:40:30] * gdamore eats words.
[23:40:33] <delewis> hehe
[23:40:37] <g4lt-mordant> rydis, uhm info is based on emacs, which is what was the genesis of gnu...
[23:40:54] <rydis> g4lt: Nope. Sorry. It's an old ITS format.
[23:41:09] <delewis> that explains why I hate it so much.
[23:41:16] <g4lt-mordant> rydis, its, you mena teco-based?
[23:41:24] *** Mazon is now known as mazon
[23:41:35] <g4lt-mordant> and emacs is what, a collection of teco macros?
[23:41:51] <rydis> Well, TECO had an info reader, of course, but there was a stand-alone info reader, too.
[23:42:25] <gdamore> crap.  the only reason setup_install_server calls into adb is to calculate some netmasks.  talking about using a jackhammer to drive a tack.
[23:42:43] *** bougie has quit IRC
[23:45:01] <gisburn> alanc: in which Category/Subcategory does the graphical installer fall ?
[23:45:28] <delewis> "things that should die a horrible death" :-)
[23:45:48] <richlowe> gisburn: consolidation/install?
[23:47:59] *** hile_ has quit IRC
[23:49:38] <alanc> I don't know the bugs.os.o category for it - if you go with consolidation/install that should be close enough
[23:50:09] <alanc> no matter what category you choose it all ends up in the same triage queue for us to re-assign to the real bugster category
[23:50:12] <richlowe> I'm not sure b.o.o does expose the category at all.
[23:50:23] <richlowe> Yes, and that's silly too.
[23:52:59] <jengelh> The Solaris Kernel seems to be two-step, is not it?
[23:53:44] <jengelh> otherwise it's hard to think how integral parts like devfs and specfs could be a module
[23:54:33] <richlowe> the loader loads krtld, krtld loads unix and genunix, and hands off to unix, unix loads the rest.
[23:54:36] <richlowe> iirc.
[23:54:41] <alanc> two-step?  isn't that a dance style done with country music?
[23:54:55] <richlowe> actually, scratch that, I like alanc's answer better.
[23:54:59] <richlowe> :)
[23:55:14] <alanc> (you can tell I don't do kernels)
[23:57:52] <jengelh> hah
[23:58:25] <jengelh> so solaris is hypermodular





top