[00:00:20] <cmihai> flyingparchment, SXCE and SXDE are virtually the same thing (especially when the build correspond). But if you pick the SXDE mode, now you get a different installer, plus it installs /opt stuff (Sun Studio, NetBeans). If you pick SXCE mode, it uses the old installer and you can just install the development tools from the DVD later.
[00:00:25] <CIA-25> damico: 6602489 <unistd.h> typo: __EXTENSIONS_ should be __EXTENSIONS__
[00:00:54] <cmihai> clay, nah, just get the DVD, easier that way. Do a full install + OEM too. Hell, you should have it all setup withing 2 hours or so anyway.
[00:01:23] <cmihai> Oh, and you don't "install" the source, you clone the hg repo.
[00:01:24] <richlowe> cmihai: SX:DE is specifically the first grub menu option on the builds released as such.
[00:01:47] <richlowe> cmihai: so when avoiding the installer and the wrong studio version is the goal, yeah, they differ
[00:02:01] <cmihai> Like "hg clone ssh://anon at hg dot opensolaris.org/hg/onnv/onnv-gate"
[00:02:31] <clay> cmihai: will that bring the deleted_files directory structure over, too?
[00:02:37] <richlowe> sadly, yes.
[00:02:50] <cmihai> Oh, and you don't need to remove studio, just pick the first option (Solaris Express) and it won't install it in the first place.
[00:02:51] <clay> cmihai: the bug I'm trying to correct is in multiboot, which is now relegated to deleted_files
[00:02:53] <richlowe> in the long run, deleted_files will go away, since Hg deals with deletes properly.
[00:03:34] <clay> will  be able to build multiboot within the deleted_files branch?
[00:03:48] <clay> will I ...
[00:04:19] <cmihai> Good news is over a good connection (3-4 Mbit)  it should clone in about 20-30 minutes or so.
[00:04:27] <sommerfeld> i think you'll be better off checking out a pre-multiboot-removal version of the tree
[00:04:28] <alanc> *sigh*
[00:04:33] <sommerfeld> (which you can do with mercurial)
[00:04:36] <alanc> Inception: Memcached 1.2.2 to be included in OpenSolaris (2007/385)
[00:04:40] <alanc> Exposure: closed
[00:04:42] <richlowe> cmihai: or specify the ssh command explicitly, and pass -C
[00:04:48] <richlowe> alanc: Bzzzt, wrong.
[00:04:48] <alanc> syntax error: does not computer
[00:04:52] <alanc> err, compute
[00:05:10] <alanc> yeah, I've just pinged the team
[00:05:12] <richlowe> alanc: (as a suggested response, not a suggestion you are wrong)
[00:05:13] *** nostoi has quit IRC
[00:05:20] <richlowe> you should have ping'd 'em with my text :)
[00:05:25] <sommerfeld> "derail."
[00:05:28] <flyingparchment> include memcached in solaris?  ... why?
[00:05:41] <alanc> assume it was a case filed by one-pager, which I think still defaults to closed
[00:05:45] <cmihai> clay, well, once you have the repo you can revert to older versions or whatever. You mean multiboot as opposed to the new directboot?
[00:06:19] <clay> cmihai: yes. multiboot contains a bug that we are seeing in our environment, and I need to fix it ASAP
[00:06:28] <clay> having to install opensolaris first is quite a chore :)
[00:06:34] <alanc> flyingparchment: because all the cool web2.0 sites use it, so it must be good
[00:07:17] <sommerfeld> I think it's close to "because all the cool web2.0 sites use it, we have to bundle it regardless of whether it's any good"
[00:07:51] <clay> some things belong in an os
[00:07:54] <clay> memcached does not
[00:08:15] <flyingparchment> of all the useful programs i could think of bundling with solaris, memcached is about #nowhere on te list
[00:08:20] <clay> though it should be simple and painless to install it when necessary
[00:08:42] <flyingparchment> although including memcached would mean libevent
[00:08:45] <steleman> flyingparchment: we need a memory hog elephant which does absolutely nothing to improve performance
[00:08:57] <flyingparchment> steleman: it does a lot to improve performance
[00:09:12] <steleman> flyingparchment: not really, compared to the alternatives
[00:09:22] <flyingparchment> which are what?
[00:09:29] <steleman> APC
[00:09:38] <flyingparchment> how do you share APC cache between machines?
[00:09:48] <steleman> flyingparchment: why do you want to ?
[00:10:03] <flyingparchment> because i have 300 web servers and don't want 300 separate copies of everything in the cache
[00:10:03] <steleman> flyingparchment: what are you sharing ?
[00:10:13] <steleman> flyingparchment: what are you sharing ?
[00:10:20] <flyingparchment> cached generated data from a webapp
[00:10:23] <steleman> flyingparchment: SSL seessions ?
[00:10:32] <steleman> flyingparchment: APC. why do you care it's cached in each box ?
[00:10:45] <flyingparchment> because caching it locally means much lower hit rate
[00:10:51] <flyingparchment> (because not nearly as much data can be stored)
[00:10:56] <steleman> flyingparchment: it's cached in each box.
[00:11:14] <flyingparchment> steleman: if i cache locally, i can only cache 100MB of data.  if i use a centralised cache, i can cache 2GB of data.
[00:11:17] <steleman> oh so APC cannot cache what memcached caches ?
[00:11:29] <steleman> flyingparchment: i fail to see the reasoning behind that statement.
[00:11:54] <clay> APC and memcached are not mutually exclusive
[00:11:55] <steleman> flyingparchment: define "centralised cache".
[00:12:02] <clay> we use APC and Coherence at our site
[00:12:03] <steleman> flyingparchment: what is a "centralised cache" ?
[00:12:08] <flyingparchment> steleman: memcached is network accessible, an object is only stored in one place, and all machines access it there
[00:12:10] <clay> they do different things, and work well in tandem
[00:12:20] <flyingparchment> steleman: so, the object is only cached once.  with APC, every server has to store its own copy of the object.
[00:12:49] <steleman> flyingparchment: i see. so, in your theory, memcached which does network i/o every time it stores or retrieves an object, is much better performing than shared memory caches which don't even swap out, and which allow direct memory access. interesting.
[00:12:53] <flyingparchment> we do actually use APC, but not for this :) (rather for the bytecode cache)
[00:13:15] <clay> sommerfeld: any idea how I can find which version of the source to check out? and how to do the checkout? I am not familiar with hg
[00:13:22] <flyingparchment> steleman: yes, because a network access is much cheaper than regenerating the cached data
[00:13:31] <steleman> flyingparchment: doodoo.
[00:13:31] <flyingparchment> steleman: you understand how a non-shared cache has a lower hit rate, right?
[00:13:45] <flyingparchment> excuse me, how many years have you spent studying our application in production?
[00:13:54] *** nachox has quit IRC
[00:13:55] <steleman> flyingparchment: i dont understand how network i/o is better performing than shared memory, but that's just my inherent limitation.
[00:14:03] <steleman> flyingparchment: more than you apparently.
[00:14:07] <flyingparchment> steleman: it's not.  the advantage is the much larger cache capacity.
[00:14:18] *** nachox has joined #opensolaris
[00:14:21] <steleman> flyingparchment: doodoo ^ 2.
[00:14:45] <alanc> R.I.P. dtmail (LSARC/2007/516 - Removal of Dtmail - closed approved fasttrack 9/14/2007)
[00:15:07] <sommerfeld> clay: on esec
[00:15:23] <sommerfeld> one sec, even
[00:15:52] <clay> and the rest of dt with it?
[00:15:59] <cmihai> Noooo
[00:16:12] <cmihai> They've killed dtmail.. you bastards!
[00:16:44] <cmihai> What did it ever do to you? Besides, unlike the stuff in the other UNIX systems, Sun's dtmail actually did IMAP and stuff :-)
[00:16:45] <alanc> clay: soon - they're removing first the pieces no one else depends on, while waiting for dependencies like Install using dtterm and the default login using dtlogin to be resolved
[00:16:52] <flyingparchment> steleman: how do you figure that spending 500ms regenerating data is better than spending 50ms to fetch it from a cache?
[00:17:31] <cmihai> I don't know, there's still plenty of reasons to keep CDE around, like compatibility and the fact JDS is just so heavy and sometimes unstable ;-(
[00:17:36] <sommerfeld> i'd hope you could fetch in-memory data from memory in a a nearby machine faster than you could fetch it from a disk
[00:17:56] <steleman> flyingparchment: who makes you spend 500ms regenerating data ?
[00:18:06] <clay> cmihai: I don't know why Solaris needs a desktop at all
[00:18:12] <clay> cmihai: imho it's a server os
[00:18:36] <cmihai> clay, eh, there's still some people who use it for PLM
[00:18:46] <cmihai> As in Catia and such
[00:18:48] <flyingparchment> steleman: our data is text documents that need to be converted to HTML.  a large one can take that long
[00:18:57] <sommerfeld> clay: direct boot integrated into build 57
[00:18:58] <steleman> flyingparchment: have you heard of TTL ?
[00:19:05] <steleman> flyingparchment: how often do these documents change ?
[00:19:08] <steleman> flyingparchment: ever ?
[00:19:15] <flyingparchment> very frequently
[00:19:15] <sommerfeld> clay: so I'd check out the "onnv_56" tag
[00:19:17] <cmihai> clay, well, UNIX was quite a big deal a few years back in the PLM industry...
[00:19:23] <flyingparchment> well
[00:19:24] <clay> sommerfeld: excellent, thanks
[00:19:27] <cmihai> That whole big bad UNIX workstation deal...
[00:19:29] <flyingparchment> it varies, some very frequentry, some less
[00:19:39] <clay> cmihai: what's PLM?
[00:19:42] <steleman> flyingparchment: ok. therein is your answer.
[00:19:55] <flyingparchment> steleman: not really.  they're accessed more frequently than they change
[00:19:57] <sommerfeld> I *think* the command you want is "hg update -r onnv_56"
[00:20:02] <steleman> flyingparchment: then cache them in shmem
[00:20:13] <flyingparchment> steleman: why?
[00:20:15] <clay> sommerfeld: where do I tell hg what server to use, etc?
[00:20:18] <cmihai> Besides, CDE had some pretty cool stuff too, dtterm was nice, and dtksh was awesome ;-\
[00:20:18] <richlowe> sommerfeld: no -r
[00:20:21] <steleman> flyingparchment: i have already explained why.
[00:20:37] <cmihai> PLM - Product lifecycle management
[00:20:38] <flyingparchment> no you haven't, all you've said is that shared memory is fast.  yes, it is, but its capacity is much lower.
[00:20:49] <sommerfeld> richlowe: serves me right for not waiting for my test "clone" to complete so I can try it.
[00:20:54] <steleman> i said shared memory access is faster than network i/o.
[00:21:03] <flyingparchment> i agree, it is
[00:21:10] <steleman> you said capacity is lower, which is false.
[00:21:14] <flyingparchment> no it isn't.
[00:21:18] <steleman> ok.
[00:21:26] <steleman> doodoo ^ 3.
[00:21:38] <cmihai> You know, people doing CAD/CAM/CAE/MPM/PDM and using stuff like Catia and such. I've seen Solaris used quite often as a workstation platform for PLM software and circuit design software and such.
[00:21:39] <flyingparchment> say i have 100MB of cache on each machine.  i want to store a 1 MB object.  with APC, i have to store it once on each machine, so i can only store it 100 times.
[00:21:51] <flyingparchment> rather, only store 100 1MB objects
[00:21:59] <steleman> flyingparchment: i dont follow.
[00:22:10] <flyingparchment> steleman: each machine has to store its own copy of each object
[00:22:14] <flyingparchment> (because the cache is not shared)
[00:22:17] <steleman> flyingparchment: yes. so ?
[00:22:36] <richlowe> so you cache the same object N times locally, rather than once slightly-less-locally
[00:22:37] <clay> w/ 300 web servers w/ independent caches, to populate all 300 caches with one object requires 300 DB queries
[00:22:38] <steleman> flyingparchment: you can't afford RAM ?
[00:22:46] <clay> to populate a shared cache takes one DB query
[00:22:49] <flyingparchment> with memcached, i have a 30,000 MB shared cache (assuming 300 machines).  so, i can store 30,000 1MB objects
[00:22:56] <clay> there is no way distributed caches are faster
[00:23:10] <steleman> flyingparchment: where do you have a 300000MB shared cache ? where does it live ?
[00:23:11] <flyingparchment> you don't see how storing 30,000 objects gives a better hit rate than only storing 300?
[00:23:16] <steleman> flyingparchment: in the ether ?
[00:23:21] <flyingparchment> steleman: memcached runs on each machine, using the 100MB that APC would use if we used that
[00:23:30] <steleman> where does your shared cache live ?
[00:23:35] <flyingparchment> on each machine
[00:23:40] <steleman> AHA
[00:23:43] <flyingparchment> it uses hashing to know which memcached instance to ask for the object
[00:23:44] <cmihai> clay, sure there is, Infiniband
[00:23:46] <steleman> and how is this different from APC then ?
[00:23:52] *** victori__ is now known as victori
[00:24:10] <steleman> other than the fact that each time you talk to your "centralized cache" monster you do network i/o ?
[00:24:15] <flyingparchment> steleman: because with memcached, if object with the key "foo" is hashed to machine 1, every other machine asks machine 1 for the object
[00:24:20] <steleman> and in your theory this is faster ?
[00:24:25] <flyingparchment> steleman: machine 1 is the only system storing that object
[00:24:33] <steleman> great!
[00:24:38] <flyingparchment> steleman: don't you see how having 30,000 objects cached instead of 300 improves cache hit rate?
[00:24:42] <steleman> you have a N-> contention
[00:25:26] <steleman> i dont follow.
[00:25:27] *** eboutilier has quit IRC
[00:25:27] <flyingparchment> the load is quite acceptable, memcached is pretty efficient
[00:25:42] <flyingparchment> higher hit rate -> less time spent regenerating data -> lower cpu use
[00:25:51] <steleman> flyingparchment: i dont need slogans. i need facts.
[00:26:44] <flyingparchment> okay, i checked one server, memcached is using 1% cpu
[00:26:48] <clay> does solaris express come w/ hg?
[00:26:54] <flyingparchment> clay: yes
[00:27:06] <elektronkind> it comes wth hugs
[00:27:09] <steleman> flyingparchment: on my box here memcached uses 0% cpu because it's not running.
[00:27:19] <clay> hugs?
[00:27:21] <clay> and kisses?
[00:27:26] <flyingparchment> steleman: yes, so instead you're spending all your CPU regenerating the data you can't cache
[00:27:29] <elektronkind> and lots of love, damnit
[00:28:09] <steleman> flyingparchment: no, because i can set TTL on the cache. depending on how often my objects change. and if i can actually write PHP code, i can actually cache PHP objects differently depending on my expectations on how they will be accessed.
[00:30:40] <sommerfeld> clay: try "hg clone -U ssh://anon at hg dot opensolaris.org/hg/onnv/onnv-gate myws; cd myws; hg update onnv_56"
[00:31:21] <flyingparchment> steleman: TTL doesn't increase your cache capacity.  if your working set is 1TB, and you have 300MB of cache, you're not going to be able to store all of it.
[00:31:48] <steleman> flyingparchment: and you are storing 1TB in memcached ?
[00:31:53] <steleman> doodoo ^4
[00:32:02] <flyingparchment> of course not, but we're storing more than we could with APC (a lot more, as i showed above)
[00:32:06] <flyingparchment> so, the hit rate is much higher
[00:32:19] <sommerfeld> stelman: it depends on your working set.  sometimes a bigger, farther away cache is better than a tiny close cache.
[00:32:22] <steleman> flyingparchment: you have shown absolutely nada. you have repeated slogans.
[00:32:36] <sommerfeld> (excuse me for misspelling steleman)
[00:32:37] <steleman> "shared memory has limitations"
[00:32:38] <flyingparchment> steleman: um, i gave you exact figures from our environment.  300 machines, 100 MB per machine.
[00:32:49] <flyingparchment> do the maths, it's not hard.
[00:32:50] <clay> sommerfeld: thanks
[00:32:53] <steleman> "cache hit ratio is better when there are less objects in cache"
[00:33:08] <elektronkind> less?
[00:33:13] <flyingparchment> who said hit ratio is better when fewer objects are cached?  that makes no sense
[00:33:20] <steleman> you said it.
[00:33:33] <flyingparchment> no i didn't.  i said it's better when each object is only cached once.
[00:33:45] <flyingparchment> caching each distinct object fewer times means more total objects can be cached.
[00:33:56] <steleman> <flyingparchment> steleman: don't you see how having 30,000 objects cached instead of 300 improves cache hit rate?
[00:34:05] <flyingparchment> 30,000 is more than 300, not less.
[00:34:21] <steleman> but how is it cached only once ?
[00:34:27] <steleman> explain that to me.
[00:34:28] <steleman> how ?
[00:34:29] <flyingparchment> why would it need to be cached more than once?
[00:34:38] <steleman> where does it live ?
[00:34:39] <flyingparchment> any machine can contact any other machine and ask memcached on that machine for the object
[00:34:52] <flyingparchment> it lives on a machine calculated by a hash of the object's key
[00:34:53] <steleman> how is it cached only once ?
[00:35:25] <flyingparchment> the object with key "foo" lives on machine 1.  the object with key "bar" lives in machine 2.  the object with key "quux" lives in machine 3.
[00:35:37] <flyingparchment> if machine 3 needs the object with key "foo", it contacts memcached on machine 1
[00:35:42] <steleman> ok.
[00:35:48] <steleman> so you have an incoherent cache.
[00:35:50] <flyingparchment> whereas with APC, machine 3 would need to cache "foo" itself, so now both 1 & 2 cache it
[00:35:54] <steleman> you have no idea where an instance actually resides.
[00:36:05] <flyingparchment> i just told you how i know where it lives
[00:36:06] <steleman> so, every time you have to look for an instance, you have to find out where it is, and then go fetch it.
[00:36:10] <flyingparchment> it's calculated from a hash of the key
[00:36:21] <flyingparchment> there's no finding out, it takes like 0.05ms
[00:36:24] <sommerfeld> no, because hash("foo") = 1, hash("quux")=3, etc.,
[00:36:28] <flyingparchment> (random figure i just made up)
[00:36:31] <steleman> how often do you rehash ?
[00:36:36] <flyingparchment> never
[00:36:37] <steleman> do you know ?
[00:36:38] <sommerfeld> i'd assume never
[00:36:41] <steleman> oh really ?
[00:36:42] <steleman> why is that ?
[00:36:58] <clay> get a room :)
[00:37:22] *** Fish has quit IRC
[00:37:23] <sommerfeld> there may already be a #memcached
[00:37:37] <clay> hah
[00:37:57] <steleman> sommerfeld: i am quite certain you know that rehashing has nothing to do with that.
[00:38:28] <clay> sommerfeld: I downloaded the onnv tarball earlier today
[00:38:51] <clay> sommerfeld: can I extract that and use it as a base for the update? or would it be better to start from scratch?
[00:39:03] <stevel> the onnv tarball doesn't have any source history with it
[00:39:13] <stevel> you wouldn't be able to do any hg operations against it
[00:39:18] <clay> ahh
[00:39:44] <clay> well that was a waste of bandwidth, heh
[00:48:05] <FireflyST> So, I'm going to have a celebration of SCO's bankruptcy this weekend
[00:51:58] *** bunker has quit IRC
[00:54:45] *** pfa3rh has quit IRC
[00:56:12] *** perlmonk has quit IRC
[00:57:21] *** Mdx4 has quit IRC
[00:58:23] *** sahafeez has quit IRC
[01:00:56] *** preek has quit IRC
[01:01:28] *** nachox has quit IRC
[01:01:55] *** sommerfeld has quit IRC
[01:03:07] <CIA-25> gtb: 6573019 mit 1.4 sub-glue layer resync
[01:03:52] *** yarihm has quit IRC
[01:11:10] *** andy_ has quit IRC
[01:11:25] *** andy_ has joined #opensolaris
[01:12:48] <flyingparchment> has anyone managed to access the v40z megaraid bios over the sp console?
[01:14:01] <boyd> Looks like Alan Cox has the end-to-end fever :)
[01:15:26] <dlg> v40z has megaraid?
[01:15:33] <flyingparchment> its an option
[01:15:35] <dlg> boyd: but he thought of it all by himself
[01:15:48] <gdamore> btw, still looking for reviewers for the C code at http:/cr.opensolaris.org/~gdamore/dmfe-x86/   :-)
[01:16:15] <flyingparchment> alternatively, is there any way i can destroy a megaraid array from software?  (i want to use svm instead)
[01:18:42] *** cmang_ has joined #opensolaris
[01:19:11] <gdamore> flyingparchment: why not just use ZFS with raidz?
[01:19:30] <flyingparchment> zfs paniced on me.  now i'm scared of it.  :)
[01:20:01] <boyd> Sounds like it's scared of *you*
[01:20:08] <jbk> haha
[01:20:19] <flyingparchment> (i need svm for root anyway)
[01:24:20] *** axisys has quit IRC
[01:25:08] *** cmang has quit IRC
[01:25:13] *** Kush- has quit IRC
[01:26:01] *** cmihai has quit IRC
[01:26:02] <alanc> wow, actual stable links to docs.sun.com pages...may wonders never cease
[01:26:35] <alanc> (and to think that was one of the first web usabilty tenets established by Jakob Nielsen when he was working for Sun in the early web days)
[01:26:50] *** spackest has joined #opensolaris
[01:27:07] <flyingparchment> "MegaRAID Storage Manager-Integrated RAID (MSM-IR) configuration setup application enables you to configure, monitor, and maintain storage configurations on SAS106x Integrated RAID controllers." .. i wonder if that works on other controllers
[01:27:34] <flyingparchment> oh wait, that only works on windows
[01:30:26] <jbk> hmm
[01:30:33] <jbk> there seem to be issues with swing + compiz
[01:32:14] *** sommerfeld has joined #opensolaris
[01:32:14] *** ChanServ sets mode: +o sommerfeld
[01:34:21] *** Fullmoon has joined #opensolaris
[01:39:26] *** Vorbis_w48 has quit IRC
[01:39:32] *** yarihm has joined #OpenSolaris
[01:47:30] *** Pietro_S has quit IRC
[01:51:52] *** r3wt__ has joined #opensolaris
[01:51:55] <r3wt__> elo
[01:52:27] *** spackest has left #opensolaris
[01:52:30] <r3wt__> i m trying to us emy usb external drive but don't know where to start ... iread the dmesg message but can't find the exact device name ...
[01:53:18] <jbk> what build (uname -a) are you using?
[01:53:31] <jbk> if it's recent, should just be able to plug it in and look in /media
[01:55:04] <r3wt__> SunOS unknown 5.10 Generic_120012-14 i86pc i386 i86pc
[01:55:35] <r3wt__> nothing :(
[01:55:38] <r3wt__> but in dmesg i have
[01:55:40] <r3wt__> hid2 is /pci@0,0/pci1043,80a6@1d,3/input@1
[01:55:52] <r3wt__> i just want to know what is the device name in /dev/dsk :(
[01:57:09] <r3wt__>  usba: [ID 912658 kern.info] USB 1.10 device (usb3f0,40c)
[01:57:29] *** jamesd has quit IRC
[01:57:36] *** Gekkko has joined #opensolaris
[01:58:11] <nrubsig> next madness is ready: http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/piano.ksh (prototype code)
[01:59:16] <r3wt__> i realy don't understand :(
[01:59:30] <r3wt__> i tried to mount /dev/usb/hid2
[01:59:35] <r3wt__> but no result
[02:00:18] <CIA-25> wh94709: 6585396 sun4v huron panics in odd ways
[02:00:48] <nrubsig> r3wt__: how did you try to mount it ?
[02:01:32] <r3wt__> ok i tryied
[02:01:44] <r3wt__> mount -F hsfs /dev/usb/hid2 /media
[02:01:51] <r3wt__> and the same with pcfs
[02:02:05] <r3wt__> but i m not sure it is the correct device nam
[02:02:33] *** Fullmoon has quit IRC
[02:02:58] <flyingparchment> can i put svm metadata on a vxvm device?
[02:03:54] <r3wt__> very strange
[02:03:57] <r3wt__> it wasn't working
[02:04:04] <r3wt__> and after the iostat -En command
[02:04:26] <r3wt__> it is completely recognized in dmesg and available on the desktop
[02:08:32] *** alanc is now known as alanc-away
[02:09:34] *** Jondice has joined #opensolaris
[02:12:55] *** r3wt__ has left #opensolaris
[02:14:22] *** piwi has quit IRC
[02:15:28] *** stevel has quit IRC
[02:30:41] *** sbahra has joined #opensolaris
[02:31:46] *** jwk404 has joined #opensolaris
[02:34:33] *** tg has quit IRC
[02:41:56] *** aglio2 has joined #opensolaris
[02:42:13] <aglio2> anyone know if there's a way to get scriptreplay to work on opensolaris?
[02:42:31] <aglio2> the only place I seem to be able to find it is linux-utils
[02:44:35] *** johnlev has quit IRC
[02:49:39] *** tg has joined #opensolaris
[02:59:36] *** vmlemon has quit IRC
[03:05:32] *** yarihm has quit IRC
[03:05:49] *** m0le has quit IRC
[03:06:25] *** unixware has joined #opensolaris
[03:07:25] *** jwk404 has quit IRC
[03:07:33] <unixware> good friday night every one
[03:08:02] <jbk> hello
[03:08:25] <unixware> :)
[03:09:03] <unixware> there is nothing better than chat on  friday night :P
[03:09:12] <jbk> well i'm on call for work
[03:09:17] <jbk> so can't really do too much
[03:09:22] <clay> same here
[03:09:23] <unixware> a ok
[03:09:27] <unixware> xD
[03:10:15] <jbk> had to do it this week so i can go to the os developer's conference next month :P
[03:10:27] <unixware> :O
[03:12:56] *** sommerfeld has quit IRC
[03:15:26] *** Shiv__ has joined #opensolaris
[03:21:44] *** vmlemon has joined #opensolaris
[03:34:50] *** quenell1 has joined #opensolaris
[03:36:16] *** BatonT has quit IRC
[03:39:11] *** comay has quit IRC
[03:42:26] *** boyd_ has joined #opensolaris
[03:44:05] *** troff has joined #opensolaris
[03:45:21] *** Shiv__ has quit IRC
[03:46:28] *** aglio2 has quit IRC
[03:46:39] *** Shiv__ has joined #opensolaris
[03:50:02] *** laca has quit IRC
[03:51:05] *** _Hunger- has quit IRC
[03:51:29] *** _Hunger- has joined #opensolaris
[03:59:43] *** boyd has quit IRC
[04:00:14] <CIA-25> dr146992: 6588495 IP can use the wrong interface for filtering/qos, 6599516 locking in fr_natderef causes lock contention and performance drop
[04:02:07] *** deedaw has quit IRC
[04:04:01] *** binarycrusader has joined #opensolaris
[04:06:38] *** vmlemon has quit IRC
[04:08:37] *** sycofly has joined #opensolaris
[04:10:03] *** alobbs has quit IRC
[04:10:26] *** alobbs has joined #opensolaris
[04:13:30] *** Andy_Pease has joined #opensolaris
[04:13:39] *** slowhog has joined #opensolaris
[04:13:59] *** andy_ has quit IRC
[04:21:35] <clay> is opensolaris supposed to come with hg?
[04:22:07] <jbk> sxce does
[04:22:51] <g4lt-sb100> what, didn't like the answer last time?  [1625] clay does solaris express come w/ hg?
[04:22:52] <g4lt-sb100> [1625] flyingparchment clay: yes
[04:23:16] <clay> no ... I spent 5 hours downloading and installing and can't find it, so thought I'd ask again :)
[04:24:05] <clay> where is it installed?
[04:24:12] <g4lt-sb100> /bin/hg
[04:24:15] <g4lt-sb100>  /exec -o which hg
[04:25:27] *** slowhog has quit IRC
[04:25:30] <clay> I wonder which CD it's on
[04:25:42] <g4lt-sb100> wait, you didn't do a full install?
[04:25:55] <clay> I did the developer install
[04:26:05] *** family has joined #opensolaris
[04:27:09] * g4lt-sb100 notes that his question wasn't answerd
[04:28:05] <clay> the installer offered a few choices; one was full, one was developer; I selected developer
[04:28:32] <family> Anyone used a Rosewill or Promise PCI IDE controller card ?
[04:28:45] <g4lt-sb100> ohhhh.  you're boned then
[04:29:01] <g4lt-sb100> ALWAYS do a full_OEM installation
[04:29:16] <clay> g4lt-sb100: how silly of me to think the developer tools would be in the developer installation :)
[04:29:29] <g4lt-sb100> yes, in this case, you were
[04:30:34] <clay> any idea which cd it's on?
[04:31:01] <g4lt-sb100> nope
[04:31:15] <jbk> look for SUNWmercurial
[04:32:06] <clay> of course now I need to figure out how to mount the cd after installing vmware tools
[04:32:28] <binarycrusader> Oi.
[04:32:33] *** troff has quit IRC
[04:33:20] <g4lt-sb100> again, ALWAYS do a full+OEM installation
[04:33:54] <clay> ok
[04:33:55] <jbk> hopefully that won't have to be the case in the future though..
[04:34:00] <binarycrusader> If you install "Solaris Express Developer Edition" this will be taken care of too.
[04:34:27] <binarycrusader> The old installer's installation profiles are -- somewhat lacking.
[04:35:08] <clay> binarycrusader: someone on here was saying sxde and sxce are essentially the same?
[04:35:25] <binarycrusader> They are, except choosing SXDE will install a lot more by default.
[04:35:38] <binarycrusader> And if you choose SXDE in b70 or newer, it will use the spiffy new installer.
[04:35:50] <binarycrusader> (maybe earlier than b70? but you get the point...)
[04:35:57] <e^ipi> s/spiffy/fucking retarded
[04:36:05] <binarycrusader> You're welcome to that opinion.
[04:36:21] <clay> well one thing's for sure; the old one sucks
[04:36:24] <binarycrusader> However, the majority consensus has been quite the opposite.
[04:36:27] <clay> or whichever one I used tonight
[04:36:34] <binarycrusader> That's the old one. It really isn't that bad.
[04:36:51] <jbk> i thought it was a nice improvement over the very old installer
[04:36:51] <clay> it might work ok, but it looks and feels awful
[04:37:01] <binarycrusader> Yep.
[04:37:16] <g4lt-sb100> binarycrusader, until it's actually seen the light of day, leet's not get our hopes up on hte new installer
[04:37:17] <jbk> clay: well it's probably 10-15 years old
[04:37:24] <binarycrusader> g4lt-sb100: what do you mean light of day?
[04:37:29] <e^ipi> i'll take something that works over something that looks pretty any day
[04:37:32] <binarycrusader> g4lt-sb100: it's already here in Solaris Express builds
[04:37:34] <jbk> the thing is, for most of sun's traditional customers, it wasn't really used
[04:37:41] <jbk> most tend to do automated installs (jumpstart)
[04:37:43] <binarycrusader> e^ipi: it works fine for many
[04:37:48] <jbk> so there wasn't much incentive to do a lot with it
[04:37:59] <binarycrusader> e^ipi: I don't know why you're better, but unless you have something less vague to complain about, your input isn't useful.
[04:38:03] <g4lt-sb100> binarycrusader, obviously it's not, since clay just got a SXCE set today and he's using the old installer
[04:38:04] <binarycrusader> s/better/bitter/
[04:38:11] <clay> jbk: *nod* I'm primarily a jumpstarter
[04:38:23] <binarycrusader> g4lt-sb100: No, it is here.
[04:38:33] <binarycrusader> g4lt-sb100: It depends on which boot option you choose whether or not it uses it.
[04:38:51] <binarycrusader> g4lt-sb100: I know it's here because I used the b72 ISO to install and the new installer was used.
[04:38:52] <clay> jbk: in fact, the reason I'm downloading opensolaris is so I can build multiboot to correct a problem in it that's preventing us from jumpstarting a couple hundred X2200s
[04:39:15] <jbk> ahh
[04:39:25] <jbk> neat
[04:39:36] <jbk> i'm stuck using hp x86 boxes at work :(
[04:40:33] <clay> aww
[04:40:49] <clay> I never understood HP
[04:41:03] <binarycrusader> g4lt-sb100: See the project page: http://www.opensolaris.org/os/project/caiman/Dwarf/
[04:41:59] <binarycrusader> screnshots: http://www.gnome.org/~gman/gui-install/
[04:42:06] <binarycrusader> s/screnshots/screenshots/
[04:42:46] <g4lt-sb100> binarycrusader, don't tell ME about it, I couldn't give a flying fark about the installer, clay used the old installer on today's ISO
[04:44:02] <binarycrusader> g4lt-sb100: As I said, both installers are there, it depends on which boot option you choose that determines which installer is used
[04:44:07] <g4lt-sb100> and my advice stands with the old installer, and quite probably the new one, since if the dwarf/caiman group shows the same progress on secondary installation that they did on the installer, I'll di of old age before it happns
[04:44:08] <binarycrusader> g4lt-sb100: I used today's ISO too
[04:44:28] <binarycrusader> g4lt-sb100: There is no secondary installation for the new installer if you're talking about the post-boot installation
[04:44:49] <clay> there is no SUNWmercurial in the Product directory on any of the 5 CDs
[04:44:50] <g4lt-sb100> so why did you give the boneheaded advice to not use full+OEM?
[04:44:50] <e^ipi> g4lt-sb100: and the new installer is like the old one but with a prettier GUI and less features
[04:44:58] <binarycrusader> g4lt-sb100: Because that doesn't exist in the new installer
[04:45:08] *** ballchalk has joined #opensolaris
[04:45:14] <binarycrusader> g4lt-sb100: The new installer only does full, period, for the moment
[04:45:37] <steleman> e^ipi: do you happen to have the link to the developer conference proposal submittal thingy ?
[04:45:45] <e^ipi> umm...
[04:45:46] <e^ipi> no
[04:45:49] <g4lt-sb100> binarycrusader, and this differs from my advice how?
[04:45:56] <steleman> e^ipi: bummer
[04:45:58] <binarycrusader> steleman: you mean this? http://www.genunix.org/wiki/index.php/OpenSolaris_Summit
[04:45:59] <e^ipi> i have the link to the wiki
[04:46:07] <binarycrusader> steleman: The topic suggestion area is at the bottom
[04:46:09] <steleman> binarycrusader: lemme look hold on
[04:46:10] <clay> so do I have to reinstall to get mercurial to show up?
[04:46:15] <binarycrusader> clay: no
[04:46:27] <clay> I just want to fix this bug and move on
[04:46:38] <clay> downloading and installing open solaris has been a huge detour
[04:46:39] <g4lt-sb100> clay not really, but I'd advise it as easier than poring through the entire CD set
[04:46:40] <steleman> binarycrusader: yah that looks like it
[04:46:43] <steleman> binarycrusader: ty
[04:46:46] <clay> at this point I just want it to work
[04:46:48] <binarycrusader> steleman: are you going to be there?
[04:47:02] <steleman> binarycrusader: 90% yes
[04:47:11] <binarycrusader> steleman: I will be, booked the hotel, etc. today
[04:47:11] *** unixware has quit IRC
[04:47:20] <steleman> binarycrusader: 10% i dont feel like flying for 6 hours :-)
[04:47:35] <binarycrusader> steleman: Yes, same here. The total flight time from KC to San Jose ain't so pretty
[04:47:36] <clay> where is it?
[04:47:48] <steleman> is it in San Jose ?
[04:47:49] <clay> KC, MO?
[04:47:54] <binarycrusader> steleman: No, Santa Cruz
[04:47:59] <binarycrusader> steleman: But closest airport is San Jose
[04:48:07] <clay> that's not so bad
[04:48:08] <jbk> binarycrusader: you're in kc?
[04:48:09] <binarycrusader> clay: KC, KS
[04:48:10] <steleman> oh i can drive around 101 then
[04:48:13] <binarycrusader> jbk: Aye
[04:48:14] <steleman> get stuck in traffic
[04:48:16] <clay> I fly ATL to SJC weekly
[04:48:17] <steleman> give the finger
[04:48:18] <jbk> no shit
[04:48:19] <steleman> smoke pot
[04:48:21] <binarycrusader> jbk: Yep.
[04:48:22] <jbk> i just moved from olathe
[04:48:25] <steleman> all these commute activities
[04:48:30] <steleman> :-P
[04:48:58] <binarycrusader> Nothing beats flying to Melbourne, AU though...oh the pain :P
[04:49:07] <clay> so, is the package possibly called something else? or not in the Product directory?
[04:49:13] <binarycrusader> clay: let me look I have a full install
[04:49:14] *** ballchalk has quit IRC
[04:49:17] <binarycrusader> clay: I'll see if I can find the package
[04:49:23] <clay> excellent, thanks
[04:49:44] <binarycrusader> SUNWmercurial
[04:49:47] <binarycrusader> is definitely the name
[04:50:08] <jbk> my stress levels dropped dramatically since i moved :)
[04:50:14] <binarycrusader> jbk: where to?
[04:50:19] <jbk> houston
[04:50:23] <jbk> new job
[04:50:27] <binarycrusader> must be the job
[04:50:32] <jbk> yes
[04:50:33] <binarycrusader> because I never hear anything but complaints about houston's traffic
[04:50:36] <binarycrusader> :)
[04:50:42] <jbk> *shrug* depends where you live
[04:51:03] <jbk> i can take the bus from in front of my apartment to work in 20 mins typically
[04:51:16] <jbk> the problem is houston grew out, not up, so it's _very_ spread out
[04:51:37] <clay> same with atlanta
[04:51:39] <jbk> the city itself is as big as nyc, dc, boston, miami, and i think seattle combined
[04:51:51] <jbk> like 600 sq mi
[04:52:10] <steleman> how do i add my blabber on this page do i just edit the wiki ?
[04:52:17] <jbk> but yeah... you can probably guess who my former employer was :)
[04:52:21] <jbk> steleman: i believe so
[04:52:23] <binarycrusader> clay, i'm looking
[04:52:34] <steleman> oh cool i'll put myself up every day for 3 hours :-)
[04:52:39] <binarycrusader> jbk: no, not really
[04:52:57] <binarycrusader> jbk: unless you're talking about seattle
[04:53:16] <jbk> a large company based in the kc area known for overworking and overstressing it's employees? :)
[04:53:25] <binarycrusader> Sprint?
[04:53:42] <jbk> did i make it too easy? :)
[04:53:51] <Shiv__> A person has reported that memory clean in the solaris C libs doesnt happen
[04:53:54] <binarycrusader> clay:
[04:53:54] <binarycrusader> No, I have a friend that works for them as a contractor
[04:54:06] <Shiv__>  (malloc & free repeatedly shows only a increase in mem, no increase/decrease alternating)
[04:54:09] <binarycrusader> clay: cd /media/SOL_11_X86/Solaris_11/Product
[04:54:16] <binarycrusader> clay: assuming you  have the CD inserted
[04:54:39] <Shiv__> I checked just now and seems to be true, anybody has a better insight ?
[04:54:45] <clay> err
[04:54:48] <binarycrusader> clay: then as root, pkgadd -d . SUNWmercurial
[04:55:22] <clay> binarycrusader: I have 5 CDs :)  and they are Solaris 10?
[04:55:33] <binarycrusader> clay: Oh you're using Solaris 10 Update 4?
[04:55:48] <jbk> Shiv__: that could be, though how are you testing -- is it still increasing even in the face of balanced malloc/free pairs?
[04:55:48] <clay> omg
[04:56:05] <clay> I *wanted* SXCE ... I *got* S10U4
[04:56:08] <jbk> i can see why it might grow to match the peak memory usage and stay there
[04:56:11] <binarycrusader> clay: That would explain a few things :)
[04:56:18] <jbk> possibly :)
[04:56:21] <clay> !@%$#^!%@$%^
[04:56:25] <binarycrusader> g4lt-sb100: there's your explanation partially :)
[04:56:26] <clay> the web site lied
[04:56:36] <binarycrusader> It did? :{
[04:57:00] <binarycrusader> http://www.opensolaris.org/os/downloads/sol_ex_cd
[04:57:03] <Shiv__> Yes. I simple C prog that does 20 mallocs and then 20 frees. Parallely prstat is dumping the mem build up.
[04:57:06] <binarycrusader> That's the download link for CD images to latest SXCE
[04:57:31] <Shiv__> After 20 mallocs, mem is built up, but doesnt reduce on freeing. The RSS stays.
[04:57:44] <steleman> i am *so* there :-)
[04:57:47] <clay> Shiv__: free() does not return memory to the kernel
[04:57:50] <binarycrusader> steleman: eh?
[04:57:58] <jbk> Shiv__: if you do it again, does it stay the same, or increase?
[04:58:31] <sbahra> clay, you cannot say this.
[04:58:49] <sbahra> clay, you cannot guarantee it actually unmaps though.
[04:59:02] <binarycrusader> clay: look at the man page for free
[04:59:05] <binarycrusader> er...
[04:59:06] <binarycrusader> Shiv__:
[04:59:12] <binarycrusader> Shiv__: look at the man page for free:
[04:59:13] <jbk> well there's no definite relationship between when memory is released to the kernel and calls to free()
[04:59:16] <Shiv__> Am checking.
[04:59:17] <clay> sbahra: in his case, it probably wouldn't return the memory
[04:59:20] <sbahra> jbk, exactly.
[04:59:25] <binarycrusader> shiv__: After free()
[04:59:26] <binarycrusader>      is executed, this space is made available for further  allo-
[04:59:26] <binarycrusader>      cation  by  the application, though not returned to the sys-
[04:59:26] <binarycrusader>      tem. Memory is returned to the system only upon  termination
[04:59:26] <binarycrusader>      of  the  application.
[04:59:47] <sbahra> That's not a portable assumption.
[04:59:53] <binarycrusader> Right, but it is true for Solaris.
[05:00:02] <clay> Shiv__: truss the program and watch the brk() system call ... notice there is no unbrk()  :)
[05:00:26] <jbk> well in general, i don't think you can make an assumption either way -- it's largely an implementation detail
[05:00:29] <Shiv__> Yeah, I saw that the mem doesn't increase for subsequent allocation.
[05:00:34] <binarycrusader> jbk: bingo
[05:01:14] <clay> ooh, SXCE is only 3 GB
[05:01:23] <binarycrusader> clay: there are DVD isos too
[05:01:28] <sbahra> binarycrusader, I think the bingo was belated by a dozen lines.
[05:01:29] <sbahra> heh
[05:01:32] <binarycrusader> clay: and you should be able to mount the iso image directly into vmware
[05:01:33] <clay> that's what I'm lookin at :)
[05:02:00] <clay> yeah, I've been doing that ... although it's a little quirky about disk change notification
[05:02:06] *** ballchalk has joined #opensolaris
[05:02:07] <binarycrusader> sbahra: :}
[05:02:23] <sbahra> binarycrusader, ;-p
[05:02:23] <binarycrusader> clay: yep, which is why I like the single DVD iso
[05:02:32] <Shiv__> What if my app builds up memory for some processing and then it has to make way for another app...
[05:02:41] <Shiv__> How do I give up the mem I hoarded
[05:02:57] <clay> Shiv__: exit(0)  :-D
[05:03:10] <Shiv__> Hey, exit is no solution.
[05:03:15] <clay> Shiv__: why do you want to give up the memory?
[05:03:30] <clay> Shiv__: if other applications needs it, the kernel will reclaim those pages
[05:03:33] <sbahra> Shiv__, use another malloc implementation which will unmap the memory :-)
[05:03:39] <clay> that's what virtual memory is all about
[05:03:53] <sbahra> clay, not if the appropriate pages are not unmapped.
[05:04:07] <clay> sbahra: even if they are unmapped .. it will just page them to swap
[05:04:24] <clay> err, still mapped, I mean
[05:04:38] <sbahra> Not really, unless it is an idle low priority process.
[05:05:03] <clay> sbahra: has nothing to do with the idleness or priority of the *process* ... only with the pages in question
[05:05:26] <sbahra> clay, hrm.
[05:05:27] <clay> sbahra: and if the app has free()ed all of its objects, those pages will be "idle"
[05:05:31] <jbk> it should be based on use -- if it's memory that's been free'd but just being kept around by libc, the pages should be pretty inactive and a candidate for pageout
[05:05:32] <Shiv__> There is a fixed swap/VM/RAM, if there are multiple apps vying for the memory, the apps/underlying-libs need to handle deallocation well.
[05:05:43] <binarycrusader> Shiv__: look at libumem
[05:05:46] <sbahra> clay, does the page scanner actually do this accounting page-level?
[05:05:54] <sbahra> jbk, ic
[05:05:55] <sbahra> Ok.
[05:06:01] <jbk> there's a scanner that runs
[05:06:08] <clay> sbahra: that's what the page scanner does .. looks for "idle" pages that can be paged out
[05:06:09] <jbk> it actually sets some bits on each page
[05:06:18] <nrubsig> Shiv__: what about allocating with mmap(MMAP_NORESERVE) ?
[05:06:20] <jbk> then there's another process that follows it sometime later
[05:06:21] <sbahra> Yeah, I forgot. I stopped my BPF stuff and I read Solaris Internals a month ago, haven't picked it up since with no coding for Solaris kernel ;]
[05:06:40] <jbk> if the pages are changed inbetween, the bits for the page entry will be modified and it knows they're active
[05:06:41] <sbahra> clay, and more. Though I thought it did take into account the owners of those pages.
[05:07:02] <clay> sbahra: I might have simplified. it might take process priority into account
[05:07:05] <Shiv__> mmap != ansi lib
[05:07:28] <jbk> well i think it is aware to a small degree what are in the pages (kernel, kernel cache, user, libraries/code)
[05:07:50] <sbahra> jbk, simply with the question of whether the page is wired or not.
[05:07:53] <jbk> and tries to be a bit selective about which pages it chooses if it needs one
[05:08:00] <jbk> wired?
[05:08:05] <clay> DAMNIT
[05:08:08] <clay> safari is braindead
[05:08:12] <sbahra> jbk, yes. Pinned.
[05:08:17] <sbahra> clay, so is the Mac OS 10 VM :-P
[05:08:27] <sbahra> WTF
[05:08:32] <sbahra> Cray XD1 is going haywire now.
[05:08:38] <sbahra> So loud.
[05:08:45] <clay> Cray?
[05:08:46] <jbk> by default, most pages aren't pinned in memory
[05:08:51] <clay> like
[05:08:55] <clay> Y-MP
[05:08:58] <sbahra> jbk, yes.
[05:09:01] *** jamesd has joined #opensolaris
[05:09:10] <jbk> except kernel pages... i think the app has to explicitly do it, and even then, i'm not sure if it's priviledged or not
[05:09:32] <sbahra> clay, loose cluster, 2 AMD64's per FPGA (XC2VP50)
[05:09:41] <sbahra> clay, have 6 of them on this one
[05:10:07] <binarycrusader> clay: make sure you choose the Developer Express install option, the new installer should be better for you, and you will get all development tools and then some (Sun Studio 12, etc.)
[05:10:20] <clay> err
[05:10:25] <clay> I'm downloading community?
[05:10:29] <binarycrusader> clay: yep
[05:10:32] <binarycrusader> clay: I know it's confusing
[05:10:34] <Shiv__> binarycrusader: How do I say use libumem, the app is linked to only libc & libm.
[05:10:39] <clay> and installing developer
[05:10:40] <clay> gotcha
[05:10:43] <binarycrusader> clay: When you boot the SXCE edition, there will be a "Developer Express" boot option
[05:10:46] <jbk> LD_PRELOAD=libumem.so ./run_app
[05:10:57] <binarycrusader> clay: when grub comes up
[05:10:58] <clay> I'm rapidly learning not to expect the obvious w/ OpenSolaris
[05:11:07] <sbahra> jbk, I plan on playing more with the Solaris kernel. I lost internet at home which is where my Solaris box is.
[05:11:10] <binarycrusader> clay: Project Indiana aims to make this far less confusing
[05:11:14] *** SunTzuTech has joined #opensolaris
[05:11:17] <sbahra> jbk, should be up this weekend :-D
[05:11:20] <binarycrusader> clay: and the download far smaller
[05:11:24] <clay> binarycrusader: I'm subscribed to that list .. but haven't had a chance to contribute
[05:11:55] <jbk> clay: some of the naming sucks, but the OS itself isn't too bad.. just there are differences from linux or fbsd, not necessarly right or wrong or better or worse, but just different
[05:11:58] <binarycrusader> clay: that's ok, by the way, once you use the new installer, post your feedback on caiman-discuss if you would or installer-discuss
[05:12:11] <sbahra> clay, what operating system do you typically use?
[05:12:36] *** johnlev has joined #opensolaris
[05:12:39] <clay> sbahra: we're primarily a solaris server / mac workstation shop
[05:12:45] <sbahra> ic
[05:12:49] <binarycrusader> sounds like a winning combo to me
[05:12:59] <sbahra> Mac OS 10 still needs work :-)
[05:13:04] <binarycrusader> what with DTrace and ZFS, etc. coming to Mac ;}
[05:13:05] <sbahra> Next release will be very nice though.
[05:13:10] <jbk> sounds like a nice (technology-wise) place
[05:13:11] <sbahra> They added MAC, ZFS and dtrace.
[05:13:13] <sbahra> binarycrusader, yeah.
[05:13:22] <sbahra> They need to dump the Mach VM though.
[05:13:24] <clay> sbahra: on my mac I have Solaris 10 U3, Windows XP, CentOS 5, and soon to be OpenSolaris 72
[05:13:39] <sbahra> On demand zeroing ;-p heh
[05:13:43] <sbahra> clay, ic
[05:13:52] <sbahra> clay, never took a look at FreeBSD?
[05:14:09] <jbk> i'm hoping xen support will be putback soon
[05:14:15] <clay> sbahra: I love some things about FreeBSD, but haven't had the need to install it
[05:14:59] <sbahra> jbk, you know...they might integrate a lot of L4 things pretty soon.
[05:15:02] <sbahra> jbk, that should be interesting
[05:15:08] <Shiv__> Behaviour of malloc/free with libumem is same as before
[05:15:25] <Shiv__> Just  checked.
[05:15:49] <clay> Shiv__: rest assured that your memory is safe with Solaris
[05:15:50] <Shiv__> LD_PRELOAD & built, ldd shows that app is linking to libumem
[05:15:57] <e^ipi> leopard's got mandatory access control?
[05:16:02] <jbk> you can also try bsdmalloc, mapmalloc, and mtmalloc
[05:16:05] <e^ipi> rockin' for them then I guess
[05:16:12] * e^ipi *hugs* RBAC
[05:16:21] <sbahra> e^ipi, yeah.
[05:16:23] <clay> Shiv__: is there a specific reason you need your app to release memory?
[05:16:36] <sbahra> e^ipi, they have this new thing called "Seatbelt"
[05:16:53] <sbahra> Which sounds sort of annoying :-P But it is a major feature they will tout completely based on MAC.
[05:17:11] <Shiv__> This is not for a reason I have at hand. Someone else had reported this on a mailing list and I was poking around to get to know it better.
[05:17:18] <sbahra> e^ipi, rwatson is also currently backporting a lot of the API changes made in Jaguar to ease the progress of integrating any new code worth importing come time.
[05:17:35] <sbahra> Leopard, whatever
[05:17:36] <Shiv__> But I have come across *many situations* where it  makes sense to release memory.
[05:17:43] <clay> Shiv__: ahh. it's a case of knowing just enough to cause trouble :)
[05:17:56] <binarycrusader> Shiv__: You might try posting about it on opensolaris-code, I'm sure one of the osol engineers will have better input
[05:17:59] <e^ipi> backporting to what? tiger?
[05:18:06] <binarycrusader> Shiv__: there might also be something on docs.sun.com buried somewhere
[05:18:08] <sbahra> e^ipi, to FreeBSD MAC.
[05:18:15] <e^ipi> oh, cool
[05:18:24] <sbahra> Yeah.
[05:18:38] <clay> Shiv__: Solaris Internals would be a good reference
[05:18:46] <binarycrusader> Oh, good idea. I have that...
[05:18:49] *** ballchalk has quit IRC
[05:18:54] <e^ipi> as I understand it, freebsd's MAC subsystem kinna sucks at the moment
[05:19:02] <sbahra> jbk, hey
[05:19:03] <e^ipi> more a tech demo than a useful thing
[05:19:06] <sbahra> jbk, what is "bsdmalloc"?
[05:19:07] <jbk> yes?
[05:19:09] *** ballchalk has joined #opensolaris
[05:19:11] <Shiv__> binarycrusader: Yeah. Am looking at some of the docs.
[05:19:12] *** quenell1 has quit IRC
[05:19:29] *** sfire||mouse has joined #opensolaris
[05:19:31] <sbahra> jbk, sounds similar to phkmalloc from man page
[05:21:03] <jbk> just a different implementation, performs a bit better at the cost of extra memory overhead
[05:21:17] *** ballchalk has quit IRC
[05:21:23] <sbahra> Doesn't say much ;-p
[05:21:41] <binarycrusader> Shiv__: according to p462 of Solaris internals, the memory usage for the process will not decrease until there is a shortage
[05:21:59] <jbk> i think it might be based off the old sunos alg, but not sure
[05:22:08] <binarycrusader> No, this is the new book
[05:22:15] <binarycrusader> for Sol10+
[05:22:24] <sbahra> jbk, yeah, looking at the code now
[05:22:46] <sbahra> jbk, phkmalloc is similar to this, but manages tables much better.
[05:23:42] <binarycrusader> Shiv__: the short answer is that if you want memory to be reclaimable much faster, to use an mmap based method for memory
[05:23:47] <binarycrusader> Shiv__: you can use libumem to do that
[05:24:09] <binarycrusader> Shiv__: if you look at "man umem_free" you'll notice that one of the options is backend=mmap
[05:24:12] <Shiv__> Hmm
[05:24:36] <sbahra> jbk, http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libbsdmalloc/common/malloc.bsd43.c if you're interested
[05:24:45] <clay> bryan cantrill spends some time talking about why mmap(/dev/zero)ing for anonymous memory isn't such a good idea
[05:24:54] <clay> ... in his google tech talk on dtrace
[05:25:01] <binarycrusader> right
[05:25:05] <binarycrusader> but it's ok if you let libumem do it
[05:25:12] <binarycrusader> and it's also ok as long as you fulfill certain provisio
[05:25:16] <jbk> but if they absolutely need to do that, you can always write your own malloc/free implementation and load it via LD_PRELOAD
[05:25:18] <clay> I don't remember the why, though, hah
[05:25:21] <binarycrusader> there was one specific case that Bryan didn't like it
[05:25:32] <clay> binarycrusader: was it when the user picks the address?
[05:25:35] <binarycrusader> yes!
[05:25:40] <binarycrusader> that was it
[05:25:42] <jbk> sbahra: ahh so really old code :)
[05:25:47] <sbahra> ya ;]
[05:25:50] <binarycrusader> something to do with page translation in the tlb if I remember right
[05:26:02] <clay> funnily enough, that's similar to the bug in multiboot i'm setting off to fix
[05:26:07] <binarycrusader> hah
[05:26:26] <tsp> Is there a lsof for solaris?
[05:26:29] <binarycrusader> pfiles
[05:26:34] <tsp> thanks
[05:26:37] <binarycrusader> close enough anyway
[05:26:40] <clay> which finally prints the name of the file!
[05:26:40] <jbk> tsp: you can compile it
[05:26:53] <jbk> but yeah, pfiles does a decent amount of the functionality
[05:26:57] <jbk> though it's not 100%
[05:27:10] <tsp> I just need to know waht rsync is currently thinking about :)
[05:27:23] <binarycrusader> tsp: fuser too
[05:27:37] <e^ipi> jbk: wow, that /is/ old code... tagged from 1982
[05:27:51] <Shiv__> tsp: lsof can be built from source. I had built it on Sol9 it had worked on Sol10 as well.
[05:27:59] <binarycrusader> or installed from blastwave
[05:28:04] <Shiv__> blastwave has it.
[05:28:06] <jbk> probably older than some of the people in here :)
[05:28:13] <clay> I try to do things the solaris way
[05:28:22] <clay> it takes some getting used to if you've developed on linux
[05:28:27] <clay> but the solaris toolset is pretty nice
[05:28:33] <binarycrusader> clay: it's not too bad, i only started using Solaris a few years ago
[05:28:34] <clay> and now w/ dtrace it kills everything else
[05:28:51] <jbk> clay: unfortunately, i don't think enough is done to advertise that
[05:29:11] <binarycrusader> clay: you might find my blog useful for dev tips: http://binarycrusader.blogspot.com/ since your new to Solaris
[05:29:20] <jbk> p* commands, mdb, dtrace, libumem (can also be used for diagnosing memory leaks)
[05:29:34] <tsp> I wouldn't be surprised if old solaris binaries still worked on S11 :)
[05:29:37] <e^ipi> way to plug your own blog shawn...
[05:29:38] <clay> binarycrusader: I'm actually not new to solaris ... been administering it since 2.5.1
[05:29:46] <clay> binarycrusader: just new to opensolaris :)
[05:29:55] <binarycrusader> e^ipi: I wrote it specifically for people new to solaris like I was when I started
[05:29:56] <tsp> since it seems to be backwards compatible with everything else
[05:30:02] <binarycrusader> e^ipi: I have no compunction whatsoever in "plugging it"
[05:30:06] <jbk> tsp: I believe someone took a sunos 4.x emacs binary and ran it on a beta copy of solaris 10
[05:30:10] <jbk> and it ran fine
[05:30:21] <tsp> what year is 4.1 again?
[05:30:37] <clay> jdk: yeah, it kills me when commercial software claims to support only "Solaris 9" or whatever
[05:30:42] <jbk> pre 1992
[05:30:48] <jbk> clay: yeah, that's mostly a paper tiger
[05:31:07] <binarycrusader> clay: just remember if you're planning to build opensolaris from source that you'll have to install sun studio 11 and not use the sun studio 12 that comes with SXCE
[05:31:10] <jbk> sun actually catagorizes their interfaces
[05:31:39] <jbk> basically unless you're using something undocumented or classified as 'unstable' it's gonna work on future versions
[05:31:52] <binarycrusader> jbk: or private, etc.
[05:31:57] <sbahra> e^ipi, I've been curious
[05:32:16] <e^ipi> re: ?
[05:32:22] <sbahra> e^ipi, what got you into Solaris? I mean, they did offer the source-code for a short while recently ('98 I had a copy of Solaris 8 source code).
[05:32:37] <sbahra> e^ipi, why didn't you go into Linux (ok, a bit obvious maybe)? or FreeBSD? or some other open-source operating system
[05:33:15] <e^ipi> I used linux for ages, and freebsd off & on
[05:33:35] *** Sup3rkiddo has quit IRC
[05:33:46] <sbahra> ic
[05:34:13] <sbahra> Anyways, I think I'll go back to reading the QNX sourcecode. Maybe tomorrow I'll be motivated to continue work
[05:34:28] <sbahra> Hate how TLS structure is private in freebsd ;[
[05:34:32] * sbahra poofs
[05:34:32] <clay> QNX?
[05:34:35] <sbahra> clay, ya
[05:34:42] <clay> old OS?
[05:34:45] <sbahra> no
[05:34:54] <sbahra> clay, qnx.com
[05:34:55] <family> anyone use a Rosewill RC-208 PCI IDE Controller ?
[05:34:56] <sbahra> clay, RTOS
[05:35:00] <clay> ahh
[05:35:01] <clay> neat
[05:35:04] <sbahra> clay, good one too. Can't say the same for the code ;-p
[05:35:07] <clay> building a lunar rover?
[05:35:16] <e^ipi> part of it was OS wanderlust (getting bored with old stuff & wanting something new ) and part of it was because Solaris doesn't make compromises on stability & reliability to appease the desktop users
[05:35:32] <jbk> heh
[05:35:33] <e^ipi> which linux does far too often
[05:35:34] <sbahra> e^ipi, neither did FreeBSD, or NetBSD
[05:35:44] <jbk> what you think it should ignore null pointer dereferences? :)
[05:36:00] <clay> they are so annoying
[05:36:04] <sbahra> BTW: QNX has a nice pthreads implementation
[05:36:05] <clay> I mean, who wants a blue screen?
[05:36:06] <family> e^ipi, not true, they are working gnome into solaris as we speak :P
[05:36:09] <sbahra> Very very simple, it is kind of cool
[05:36:20] <nrubsig> ok... stupid question: is twitter like some kind of syslog for human activities or what ?
[05:36:27] <jbk> nrubsig: no idea
[05:36:29] <e^ipi> family: yeah, and indiana seems to want to murder what makes solaris good too
[05:36:31] * sbahra poofs
[05:36:35] <jbk> i guess i'm getting old as i don't 'get it' myself
[05:36:44] <family> e^ipi, if they went for stability and resource consumption xfce4 would have been the choice :P
[05:36:55] <binarycrusader> nrubsig: best I can tell twitter is the "here's what I'm doing every second of the day via text msg"
[05:36:56] <e^ipi> because you can't attract drooling idiots with stability... you need spinny desktop cubes
[05:37:00] <clay> it's a page view generator
[05:37:16] <Gekkko> e^ipi: I was attracted by stability
[05:37:17] <Gekkko> >_>
[05:37:23] * nrubsig adds shtwitter to his ToDo demo list
[05:37:24] <Gekkko> oh right
[05:37:28] <binarycrusader> e^ipi: I think you'll be surprised to learn that Indiana is not about murdering Solaris
[05:37:29] *** jamesd_ has joined #opensolaris
[05:37:29] <Gekkko> I wasn't a drooling idiot.
[05:37:40] <e^ipi> why you would want to attract drooling idiots is left as an exercise to the reader
[05:37:49] <binarycrusader> binarycrusader: not everyone is a drooling idiot as you seem to imply
[05:37:56] <Gekkko> LOL
[05:38:00] <family> e^ipi, I'm a linux idiot, and the spinny cubes, and bouncing windows make me sick to my stomach
[05:38:03] *** sfire||mouse has quit IRC
[05:38:04] <Gekkko> I'm sorry but thats a quoute
[05:38:07] <Gekkko> quote*
[05:38:10] *** nrubsig is now known as droolingidiot
[05:38:11] <Gekkko> [13:38:08] <binarycrusader> binarycrusader: not everyone is a drooling idiot as you seem to imply
[05:38:16] <droolingidiot> binarycrusader: blabla
[05:38:23] <droolingidiot> donkdonk
[05:38:26] <family> e^ipi, I had the beryl effects on, and had to take pills for motion sickness after 5min!
[05:38:29] *** droolingidiot is now known as xjajs
[05:38:55] <binarycrusader> Gekkko: I was just pointing out that I felt as though I was talking to myself with e^ipi ;)
[05:39:03] <Gekkko> yeah bullshit.
[05:39:04] <Gekkko> lol
[05:39:13] *** xjajs is now known as nrubsig
[05:39:14] <binarycrusader> Gekkko: Nonetheless, the quote is true regardless.
[05:39:21] <e^ipi> family: yeah, I played with it once, thought it was stupid & jettisoned it
[05:39:39] <clay> nonetheless and regardless are redundant :-P
[05:39:54] <binarycrusader> Welcome to the department of redundancy department :}
[05:40:07] <clay> hah
[05:40:14] <clay> 50% downloaded
[05:40:23] <binarycrusader> It would be a grievous error at best to imply that users wanting a aesthetically pleasing operating system or better package management are "drooling idiots."
[05:41:05] *** sycofly has quit IRC
[05:41:36] <Gekkko> I have nothing against Compiz, beryl or whatever it calls itself right now.  It can be a waste or resources is my only issue, but if you have the resources to waste, why nto.
[05:41:52] <Gekkko> it can also be toned down, there's some nice features of it like the pop up of all active windows
[05:42:01] <binarycrusader> Right, an idle processor is an idle processor.
[05:42:04] <e^ipi> no, but not understanding the negative ramifications of a poorly-maintained moving target of a package management system such as every one employed to date & the value of standards-compliant userspace... does
[05:42:27] <e^ipi> if you want a pretty desktop, use windows or OSX
[05:42:32] <clay> binarycrusader: favorite books: design & implementation of FreeBSD
[05:42:33] <clay> really?
[05:42:37] <binarycrusader> clay: Yep, have it
[05:42:42] <clay> no, no
[05:42:43] <clay> not "have it"
[05:42:46] <clay> not "read it"
[05:42:47] <binarycrusader> e^ipi: right, but nowhere has it been indicated that Project Indiana lacks that undersatnding
[05:42:48] <clay> but favorite?
[05:42:52] <binarycrusader> clay: I've read some of it
[05:43:10] <clay> if you've only read some of it, it shouldn't count as a favorite :)
[05:43:15] <binarycrusader> clay: it would be hard to call it my favourite over the APU book though
[05:43:24] <clay> well, you did
[05:43:26] <e^ipi> binarycrusader: with the notable exception of every discussion involving it in any way shape or form
[05:43:31] <clay> I'm reading your blogspot profile :)
[05:43:49] <binarycrusader> clay: I don't think I've updated that since I got APU
[05:43:54] <Doc> and lets not forget the March 1978 edition of Playboy - i challenge anyone not to consider that their favourite book!
[05:43:54] <binarycrusader> clay: I have the 2nd edition of APU
[05:43:58] <clay> APUE?
[05:44:01] <binarycrusader> Aye.
[05:44:02] <clay> *old*
[05:44:09] <e^ipi> Doc: why march '78?
[05:44:10] <binarycrusader> The one from last year?
[05:44:12] <binarycrusader> clay
[05:44:15] <clay> Doc: I probably don't
[05:44:28] <clay> binarycrusader: there's a new one? Stevens is dead :(
[05:44:35] <binarycrusader> clay: yes, there's a new one, someone updated his content
[05:44:41] <jbk> e^ipi: i must have comprehension issues then
[05:44:54] <binarycrusader> clay: http://www.linuxjournal.com/article/2656
[05:45:19] <clay> binarycrusader: I have that seriues
[05:45:23] <clay> series, even
[05:45:26] <binarycrusader> clay oh wait
[05:45:28] <binarycrusader> clay: not that one
[05:45:31] <clay> and an original APUE
[05:45:38] <jbk> what i saw was: get rid of the pieces that today prevent sxce from being able to be freely distributed, and improve the default user experience (which I see as improved defaults, not throwing out binary compatability)
[05:46:01] <binarycrusader> clay: http://www.amazon.com/Programming-Environment-Addison-Wesley-Professional-Computing/dp/0201433079/ref=pd_bbs_sr_1/102-6579962-9428100?ie=UTF8&s=books&qid=1189827949&sr=8-1
[05:46:04] <binarycrusader> clay: that one
[05:46:15] <jbk> i still maintain if indiana breaks things as badly as some of the naysayers think, it will die and go away quickly
[05:46:22] <Shiv__> binarycrusader: Updated by Steve Rago, one of the authors of SVR4 I suppose.
[05:46:50] <e^ipi> jbk: how do you figure? if it breaks solaris pretty badly, it'll probably still attract a bunch of linux users
[05:47:11] <binarycrusader> e^ipi: if you think that all of the solaris engineers are just going to lie down and let compatibility die you have another thing coming
[05:47:18] <e^ipi> netting more users, at the expense of making solaris in to a pile of shit
[05:47:25] <jbk> you think it'll do it so fast as to compensate for the mass exodus of existing users that'd happen?
[05:47:50] <binarycrusader> e^ipi: i'm sorry you believe that, but isn't true
[05:48:32] <e^ipi> how do you figgure?
[05:48:50] <jbk> it's a pain in the ass to have to go through the download center -- no torrents, no 'pop in cd, point at mirror of choice, install' is a huge problem
[05:48:51] <binarycrusader> e^ipi: because the project hasn't fully defined itself it yet, plans are still being formulated
[05:49:08] <jbk> however it's there due to legal issues with certain bits shipped
[05:49:15] <binarycrusader> e^ipi: it's like judging a hafl-finished painting
[05:49:22] <jbk>  /sbin/sh is extremely user-unfriendly
[05:49:49] <jbk> having to manually setup $PATH to get stuff you expect to be there is unfriendly
[05:50:43] <clay> I second that
[05:50:58] <clay> PATH is generally stupid
[05:51:07] <e^ipi> so if all it needs is a different default $PATH, why have a project at all? why not just commit it to ON and be done with it?
[05:51:11] <clay> so many old concepts don't make sense anymore
[05:51:21] <binarycrusader> because it needs more than that e^ipi
[05:51:59] <Shiv__> e^ipi: Network based SW install/upgrade not available. Blastwave is the closest and
[05:52:08] <jbk> because you want to be sure that all the implications of that have been explored
[05:52:37] <jbk> i mean, if their goal was to really 'break' solaris, they could have accomplished that and shipped something a long time ago
[05:52:45] <Shiv__> Bits are not re-distributable
[05:53:17] <binarycrusader> e^ipi: read this http://www.genunix.org/wiki/index.php/Indiana_FAQ
[05:53:25] <binarycrusader> e^ipi: if you want to know a few things
[05:53:28] <jbk> i think the bigger issue is ian just shot his mouth off and inadverently hit a _really_ sensitive nerve, and people haven't gotten over it since
[05:53:38] <binarycrusader> jbk: how so?
[05:53:47] <e^ipi> I still don't trust the guy
[05:54:06] <binarycrusader> well, trust is earned
[05:54:17] <binarycrusader> so, it will take time
[05:54:20] <jbk> well maybe not a bigger issue, but i think the root cause of all the distrust
[05:55:15] <Shiv__> He was in a situation where linux camp called him a traitor and Solaris camp didnt trust him, very unenvious situation to be in...
[05:55:32] <jbk> 'oh he's gonna turn solaris into linux' ... says a few poorly worded things... fear confirmed (note: i'm not saying i believe that, just what I think caused the skepticism)
[05:57:09] <jbk> i actually look forward to seeing what can be produced
[05:57:12] <e^ipi> I figure he either wants to destroy solaris because he's a long time linux zealot, or destroy the OSol community because he's a friendly face that's loyal to Sun middle management ( why bother with indiana? why not just help moinak with BeleniX ? )
[05:57:51] <e^ipi> either possibility doesn't bode well ...
[05:58:11] <Shiv__> Me too. More than the product characterisitics, further opening up of the development is one thing I am looking forward to.
[05:58:14] <binarycrusader> e^ipi: neither is true
[05:58:34] <e^ipi> binarycrusader: can I see in to your crystal ball? because I'd like to know next week's lottery numbers
[05:58:51] <e^ipi> how the hell would you know what murdock's plans are?
[05:58:52] <binarycrusader> e^ipi: no, it's just a logical conclusion
[05:58:56] <e^ipi> no, it's really not
[05:58:57] <jbk> or perhaps sun thinks they want to drive greater developer adoption, saw he had success w/ debian and wanted him to be able to do the same (drive developer adoptance) with solaris
[06:00:00] <e^ipi> and what's in it for him other than a paycheque? actively fighting something that you've been an integral part of for over a decade isn't something that comes easily to most people
[06:00:27] <Shiv__> Skepticism is not good starting point :). Accept in good faith.
[06:00:32] <binarycrusader> e^ipi: this isn't about a war, open source is about collaboration and healthy competition
[06:00:48] <Shiv__> Solaris camp *assumes* GNU/linux is an enemy
[06:00:58] <binarycrusader> no, some members of the solaris community do
[06:01:04] <Shiv__> Yeah.
[06:01:08] <binarycrusader> e^ipi: you are correct that I have no crystal ball, but something you seem to be missing is that the solaris engineers have been the one controlling solaris for a long time
[06:01:16] <Shiv__> And this is what is unhealthy
[06:01:24] <binarycrusader> e^ipi: murdock may be able to make suggestions, but he cannot dictate the direction of solaris explicitly
[06:01:43] <binarycrusader> e^ipi: he doesn't make the technical decisions, he's the chief OS platform strategist, not the gatekeeper, not the ARC, etc.
[06:01:49] <e^ipi> yes, of course not... ask what desktop environment is shipping with Indiana ...
[06:01:50] <Shiv__> Probably no single person can dictate things unilaterally.
[06:01:56] <e^ipi> ( hint: Ian uses GNOME )
[06:02:08] <binarycrusader> e^ipi: and so does so Sun in Solaris
[06:02:09] <Shiv__> JDS was there before Ian came in
[06:02:21] <binarycrusader> e^ipi: Sun announced years ago that GNOME was their new direction and that CDE was being phased out
[06:02:24] <binarycrusader> e^ipi: where have you been?
[06:02:26] <e^ipi> and now's a chance to find out what the community wants, but they won't
[06:02:33] <binarycrusader> e^ipi: there is no choice
[06:02:39] <binarycrusader> e^ipi: CDE is being phased out
[06:02:43] <e^ipi> they won't even vote on it... it's just come down by executive fiat that it /will/ be gnome
[06:02:56] <binarycrusader> e^ipi: project indiana is a project
[06:03:01] <e^ipi> binarycrusader: so why not XFCE? or KDE? or flux, or whatever
[06:03:22] <e^ipi> flux would be nice, it'd strip down the CD a lot
[06:03:30] <binarycrusader> e^ipi: because no one has stepped forward to deal with the ARC process, etc. for doing that
[06:03:31] <e^ipi> you could save almost 100 megs by using it
[06:03:48] <binarycrusader> e^ipi: and let's face it, none of those are really a practical option for business or accessibility except KDE and it has it's own problems
[06:04:03] <Shiv__> You can expect a KDE support for Indiana (if mode of contribution is made easy enough)
[06:04:09] <binarycrusader> e^ipi: KDE is not yet fully integrated into the Solaris platform, people need to step up and finish porting all the tools
[06:04:18] <binarycrusader> e^ipi: Ian has already stated he'd like to see a Kindiana
[06:04:22] <jbk> http://www.opensolaris.org/os/project/xfce/
[06:04:27] <Shiv__> Integrating KDE into Solaris is a non-trivial effort.
[06:04:30] <binarycrusader> e^ipi: You could ask why the Ubuntu project chose GNOME as their primary desktop
[06:04:37] <jbk> how do you know that when that project's in a decent state, they won't include it?
[06:04:40] <Shiv__> Other desktops are not as mature as gnome & kde
[06:04:50] <e^ipi> binarycrusader: why does it need the qualifier ? why not have Indiana be KDE, and make a Gindiana ?
[06:04:51] <binarycrusader> e^ipi: and why Mark Shuttleworth has "chosen by fiat" to make GNOME their default and primary desktop
[06:05:03] <e^ipi> binarycrusader: exactly.
[06:05:12] <e^ipi> ubuntu is shuttleworth's distro
[06:05:21] <e^ipi> iandiana is murdocks
[06:05:23] <Shiv__> There is a kubuntu, xubuntu
[06:05:28] <e^ipi> neither are community distros
[06:05:29] <binarycrusader> e^ipi: because someone that wants KDE has to do the work of proposing it
[06:05:37] <Shiv__> And ON belongs to SUN
[06:05:41] <e^ipi> binarycrusader: I have, I was shot down
[06:05:47] <binarycrusader> e^ipi: no, you suggested an iea
[06:05:49] <e^ipi> because god forbid we have a vote on the issue
[06:05:51] <binarycrusader> e^ipi: you didn't offer to do the work
[06:06:14] <e^ipi> voting implies community input
[06:06:23] <binarycrusader> again
[06:06:26] <binarycrusader> voting means nothing
[06:06:31] <binarycrusader> without people there to do the work
[06:06:35] <e^ipi> and that'd take power away from His Holiness Who Shall Not Be Questioned Ian Murdock
[06:06:36] <binarycrusader> e^ipi: Sun has already sunk millions of dollars into GNOME, it makes no financial sense to choose KDE
[06:06:52] <binarycrusader> e^ipi: and from a business standpoint, GNOME has better licensing from Sun's perspective
[06:06:53] <Tempt> ooh
[06:06:58] <Tempt> people are testy today.
[06:07:24] <jbk> but still, have you proposed shipping kde w/ indiana?
[06:07:24] <Shiv__> Qt libs has restrictions on commercial use
[06:07:25] <binarycrusader> e^ipi: nonetheless, if a community member is willing to propose a KDE integration project (which I believe has already been done) and be part of the process, it can happen
[06:07:30] <jbk> or did you propose shipping it with sxce?
[06:07:48] <g4lt-sb100> binarycrusader, when in the "sunk millions of dollars into gnom" was community input asked for?
[06:07:55] <e^ipi> jbk: I proposed it in both actually. I was shot down with SXCE, and fair enough
[06:08:04] <binarycrusader> g4lt-sb100: that process was started well before this project so you can't phrase it that way
[06:08:09] <binarycrusader> g4lt-sb100: and be fair
[06:08:14] <e^ipi> but Indiana pretends to be a community driven distro
[06:08:15] <Shiv__> Moinak had mentioned that he is committed to providing a KDE based distro(belenix)
[06:08:39] <Shiv__> This might find its way into Indiana as additional desktops available off a n/w install
[06:08:46] <binarycrusader> e^ipi: since I'm going to the OpenSolaris developer summit, I will get an answer from Ian for that question
[06:09:00] <jbk> or he can ask himself when he's there :)
[06:09:00] <binarycrusader> e^ipi: I will personally ask him or ensure the he gets asked, will that make you happy?
[06:09:09] <e^ipi> Shiv__: and all the help, docu, etc will all pretend it doesn't exist, just like KUbuntu
[06:09:11] <binarycrusader> e^ipi: Assuming you're not there
[06:09:22] <e^ipi> binarycrusader: I'm there, for emancipation
[06:09:31] <binarycrusader> e^ipi: You are john s then?
[06:09:31] <Shiv__> Till now Indiana has been *closed development*
[06:09:34] <jbk> i still maintain if things are deliberately broken, there's gonna be a line of engineers with torches going down i65 :)
[06:09:36] <e^ipi> binarycrusader: yeah
[06:09:39] <g4lt-sb100> binarycrusader, so where did "fair" come in?  certainly not on the community's side,as I already pointed out they were never consulted
[06:09:46] <Shiv__> But am willing to wait for till the first release
[06:09:54] <Shiv__> and the immediate next steps
[06:10:10] <binarycrusader> g4lt-sb100: not even the Linux kernel project is a democracy, it's a community project, with a benevolent dictator
[06:10:19] <e^ipi> Shiv__: s/closed/no ... it's mostly just a jerk-session on how great linux is, and wouldn't it be great if we were just like them
[06:10:25] <binarycrusader> g4lt-sb100: until we are told otherwise, I think it's safe to assume this will be similar
[06:10:38] <g4lt-sb100> binarycrusader, yes, but the bnevolent dictator in that case had somthing to do with the inception
[06:11:00] <g4lt-sb100> pleas tell the class what ian murdock did for solaris 1.1
[06:11:05] <binarycrusader> g4lt-sb100: and Ian had something to do with the inception of Debian, so I think he's familiar with community project
[06:11:16] <e^ipi> binarycrusader: linux is run by a benevolent dictator, and it's a mess.
[06:11:23] <Shiv__> e^ipi: I wouldnt jump to such conclusions so fast.
[06:11:26] <binarycrusader> e^ipi: but a successful one
[06:11:33] <e^ipi> freebsd is run by a committee, and it's doing much better tech. wise
[06:11:38] <Shiv__> It is 2-3months to be given.
[06:11:53] <g4lt-sb100> binarycrusader, waht does debian have to do with opensolaris?
[06:12:08] <e^ipi> success is in how you measure it
[06:12:19] <Shiv__> The experience of community development (as seen by ppl who got him in)
[06:12:19] <binarycrusader> e^ipi: everyone has different measures of success, you know that
[06:12:26] <binarycrusader> g4lt-sb100: what shiv said
[06:12:27] <Shiv__> And there is nothing wrong in this.
[06:13:01] <g4lt-sb100> so basically the critcism of PI of "it will make solaris into another Debian" is true then?
[06:13:20] <Shiv__> NO.
[06:13:25] <e^ipi> I don't see lots of loud users at the expense of using a piece of shit OS to be successful
[06:13:29] <binarycrusader> e^ipi: let's put it this way, at this early state, offering a choice of desktops is pointless until KDE is ready to be used as a fully-functional desktop on Solaris
[06:13:32] <Shiv__> Look at the principles espoused by Debian.
[06:13:47] <binarycrusader> e^ipi: GNOME and KDE are the only two complete desktops and KDE isn't yet fully ready for Solaris
[06:13:47] <Shiv__> All projects based on those principles need not become Debian.
[06:13:58] <binarycrusader> e^ipi: They are the only two desktops with good accessibility, and the only two mature ones
[06:14:03] <binarycrusader> e^ipi: as far open source desktops go
[06:14:12] <e^ipi> kde4 will be ready before indiana version 1 is out
[06:14:18] <binarycrusader> e^ipi: and it will be neither ready nor mature
[06:14:23] <binarycrusader> e^ipi: for solaris more than likely
[06:14:28] <g4lt-sb100> then why is murdock's experience with debian relevant to my point that torvalds is a dictator because he started linux, not because he was brought in later
[06:14:29] <e^ipi> and adrian & stefan are upstreaming some code changes to make them compile fabulously on solaris
[06:14:35] <e^ipi> like, right now
[06:14:40] <binarycrusader> e^ipi: compiling and having full functionality are two different things
[06:14:57] <binarycrusader> e^ipi: even Stefan has noted that there are still many programs that need some changes for them to be fully functional
[06:14:58] <Shiv__> It is not yet compiling.
[06:15:09] <Shiv__> It is in the works.
[06:15:15] <binarycrusader> e^ipi: Even the KDE project is not expecting 4.0 to be great for users
[06:15:23] <Gekkko> how so?
[06:15:45] <e^ipi> Shiv__: all the dependencies up to kdelibs compile fine
[06:15:46] <binarycrusader> sorry, one moment, bfu build is slowing me down
[06:15:52] <e^ipi> the current target is kdebase IIRC
[06:15:53] * binarycrusader digs up the link
[06:15:58] <g4lt-sb100> binarycrusader, so perhaps you can tell the class also why gnome-pilot replaced stdpdasync, given that you LOSE functionality in the transition?
[06:16:09] <Tempt> Why does KDE need to be integrated?
[06:16:13] <Tempt> Why not just have packages to install?
[06:16:16] <binarycrusader> g4lt-sb100: could you please stop trolling with the whole 3rd person angle thing?
[06:16:22] <e^ipi> Tempt: ask the same thing about GNOME
[06:16:31] <Shiv__> e^ipi: A desktop is not ready to be shipped as soon as it gets built !
[06:16:38] <binarycrusader> g4lt-sb100: as for what you asked, I have no idea what you're talking about as I'm not familiar with stpdasync
[06:16:46] <binarycrusader> g4lt-sb100: I've only ever used gnome-pilot
[06:17:23] <binarycrusader> Tempt: there will be KDE packages to install, it just won't be the default desktop until a proposal is made and accepted
[06:17:26] <Tempt> I already think GNOME is a roaring waste of time
[06:17:35] <g4lt-sb100> !seen useful PI
[06:17:50] <Tempt> Then again, I'm not alone in thinking Indiana is a waste of time
[06:18:06] <Shiv__> Some ppl prefer shell only via serial access :)
[06:18:08] <Tempt> Indiana was voted the least interesting Solaris/OpenSolaris happening as MSOSUG
[06:18:19] <binarycrusader> popular votes don't always give us the best results
[06:18:26] <binarycrusader> just look who got elected as president for the USA :)
[06:18:33] <binarycrusader> by popular vote i might add...
[06:18:34] <Shiv__> So remove all the desktops !?
[06:18:43] <g4lt-sb100> ...but ignoring them gains you eveen worse results
[06:18:51] <binarycrusader> there is no proof of that
[06:19:00] <Shiv__> Democracy often results in the least common denominator
[06:19:15] <Shiv__> Dictatorship in an efficient decisions...
[06:19:17] <g4lt-sb100> binarycrusader, that's because nobody is stupid enough to try
[06:19:21] <Shiv__> bad things happen fast...
[06:19:29] <Shiv__> good things also happen fast....
[06:20:00] <g4lt-sb100> dronee, you slacker!
[06:20:37] <binarycrusader> Gekkko: http://arstechnica.com/journals/linux.ars/2007/09/07/kde-4-0-release-schedule-revised-libraries-to-be-released-early-to-speed-up-porting
[06:20:51] <Gekkko> what does this mean to me.
[06:21:02] <binarycrusader> Gekkko: not all application will be ready by the time the initial release is here, basically, 4.0 is coming out earlier to give developers a chance to ramp up their applciations
[06:21:12] <binarycrusader> Gekkko: to avoid the chicken and egg problem with the platform not being there, developers won't adopt, etc.
[06:21:20] <Gekkko> oh rihgt
[06:21:23] <binarycrusader> Gekkko: so many applications will come after the initial release
[06:21:33] <Gekkko> Yeah I understand.
[06:21:34] <Gekkko> >_>
[06:21:47] <binarycrusader> Gekkko: in addition Trolltech is doing a big change to allow flicker-free rendering in a later Qt version than 4.0 will ship with
[06:21:50] <Tempt> It means people will dribble shit about this for at least two more years.
[06:22:07] <binarycrusader> Gekkko: thus KDE 4.1 or 4.2 will likely be the best for users and the most complete
[06:22:12] <binarycrusader> as far as KDE4 is concerned
[06:22:48] <Shiv__> Hmm. So KDE4 final release is not available before Dec.
[06:22:59] <binarycrusader> At this point, though that may change again.
[06:23:20] *** alexw has joined #opensolaris
[06:24:01] *** alexw is now known as alexwaft
[06:24:10] <Shiv__> Anyone here uses Solaris as their desktops at home but not at work ?
[06:24:12] <binarycrusader> e^ipi: All I can tell you is that the backwards compatibility story of Solaris is very important to me personally and to *many* people in the community. I don't see Ian destroying that.
[06:24:17] <binarycrusader> Shiv__: me
[06:24:28] <alexwaft> Shiv__: I've just installed solaris 10
[06:24:31] <binarycrusader> Shiv__: Stuck on Windows XP at work
[06:24:36] <Shiv__> me too.
[06:24:47] <Shiv__> I work with RHEL at work
[06:24:52] <binarycrusader> Although I spend all day in an ssh terminal to RHEL boxes
[06:24:54] <jbk> Shiv__: same here
[06:25:00] <binarycrusader> and some SCO UNIX boxes occassionally :D
[06:25:25] <Shiv__> In my earlier life I have been a hardcore windows programmer
[06:25:34] <binarycrusader> Never been a windows programmer, at least not professionally.
[06:25:35] <Shiv__> (about 4years back)
[06:25:44] <binarycrusader> Done some hobby hacking on the GtkRadiant and DarkPlaces projects though.
[06:25:55] <binarycrusader> (under Windows)
[06:26:21] <Shiv__> Windows: MFC, COM, VC++ and such...
[06:27:54] <Shiv__> Linux offers just too many flexibilities at varying levels for organizations shipping their own H/W and custom distros with tools...
[06:28:35] <Shiv__> Ex: one of the projects has a highly customized 16core box with a modified gcc to suit the needs of the HW vendor.
[06:28:55] <Shiv__> This is a pizza box telecom platform that we are experimenting with
[06:29:37] <Shiv__> For ON to reach this space, ppl should gain experience in creating their own distros, customizing it for different needs, etc...
[06:30:36] <binarycrusader> yep, and that is where the livemedia and caiman projects come in
[06:30:52] <binarycrusader> distribution constructor is looking as if it will be great
[06:31:01] <jbk> Shiv__: and lack of binary compatability kills it there
[06:31:13] <Shiv__> Yes. I am keenly watching the DC project.
[06:31:22] <jbk> i remember when my previous employer tried to do linux on egenera
[06:32:02] <jbk> couldn't use any of the tools used on the other unix platforms because of all the differences w/ the version of linux that ran on egenera
[06:32:06] <Shiv__> Ofcurse it has its pain.
[06:34:07] <Shiv__> But there is considerable number of ppl who have developed considerable experience and in a corporate setting can deal with such issues(lack of binary compatibility) effectively
[06:35:11] <Shiv__> Ex: Our org has a central team that provides a IT service to all the centres across the world. It is a customized RHEL with a huge stack (proprietory+open source)
[06:35:20] *** yokelbombardier has joined #opensolaris
[06:35:29] <Shiv__> The team already has rich experience with that platform.
[06:35:49] <Shiv__> Even if there are *more* pain points, they can deal with it.
[06:35:52] <Tempt> Sad
[06:36:01] <Tempt> Should have made the right decision on day one.
[06:36:26] <Shiv__> Day 1 was much before ON happened :)
[06:36:41] <Shiv__> So what do you think were the choices?
[06:37:35] <jbk> i do think such things blast away any 'tco' argument for linux :)
[06:37:38] <clay> so is the Solaris Express Developer Edition Installer the new or old installer?
[06:37:53] <jbk> does it have pretty icons for disks and such?
[06:38:02] <Tempt> Hmm
[06:38:04] <binarycrusader> clay: new
[06:38:04] <clay> depends what you mean by pretty
[06:38:18] <Tempt> I seem to remember there being a Solaris long before there was a Red Hat anything.
[06:38:20] <clay> compared to the solaris 10 installer I suffered through earlier tonight, these are prettier :)
[06:38:42] <binarycrusader> clay: it isn't finished yet so some things will change
[06:38:53] <clay> software is never finished
[06:38:59] <Shiv__> There is a massive IT body of knowledge that is there, system support that works extremely well, high service levels to all internal business units already happening....
[06:39:00] <clay> just like a model railroad
[06:39:11] <binarycrusader> clay: yes, but this version hasn't yet met all the project goals
[06:39:16] <binarycrusader> clay: it's just an initial delivery
[06:39:25] <binarycrusader> clay: it's only been in the builds since b70a
[06:39:31] <binarycrusader> clay: so not very long
[06:39:35] <jbk> or preview if you want to think of it that way
[06:39:42] <Drone> I've never seen useful pi talk in #opensolaris.
[06:40:10] <binarycrusader> e^ipi is doing great work with emancipation, he's just very passionate about the KDE issue
[06:40:14] <clay> all OS installers have pictures and text while installing
[06:40:22] <clay> they should give you what you really want ... a web browser
[06:40:25] <binarycrusader> clay: not all :}
[06:40:26] <clay> or chat client
[06:40:35] <binarycrusader> clay: but i'm being facetious and you know that
[06:40:36] <jbk> http://www.opensolaris.org/os/project/caiman/ <-- does it look like that?
[06:40:41] <binarycrusader> jbk: yep
[06:40:56] <jbk> clay: you mean like older versions of solaris did? :)
[06:40:57] <clay> indeed
[06:41:05] <binarycrusader> jbk: slackware, gentoo...
[06:41:08] <binarycrusader> jbk: ;}
[06:41:20] <binarycrusader> jbk: SCO Unix, Irix
[06:41:33] <binarycrusader> jbk: vms? :P
[06:41:51] <clay> fear
[06:42:19] <Shiv__> Whatever improvements have happened between Sol7-Sol9 are not very noticeable for the end user I suppose...(exclude sysads/server maintenance personnel)
[06:42:20] <clay> I tried learning Irix when I though SGI was cool
[06:42:37] <Shiv__> Sol10 & SX is a major leap.
[06:42:40] <clay> and not just expenseive
[06:43:02] <jbk> Shiv__: yes, even between sucessive sxce releases
[06:43:12] <Shiv__> I installed SX when my ubuntu upgrade broke and I just fell in love with the clean interfaces.
[06:43:17] <jbk> i'm running b62 on my laptop & b72 on my desktop.. and the difference is impressive
[06:43:56] <binarycrusader> hrm, took user  1:46:17.6 to build b72
[06:44:05] <binarycrusader> bleh
[06:44:20] <jbk> i think the next 6-12 months will end up with something _really_ impressive
[06:44:27] <clay> a make world never took so long ;)
[06:44:36] <binarycrusader> clay: I know, I know .. ;P
[06:44:42] <jbk> clay: that's going to be fixed :)
[06:44:43] <Shiv__> Not bad at all.....on my celeron 256mb system (my initial solaris laptop) it takes over 2days :)
[06:44:51] <jbk> at least if i have anything to do with it
[06:44:52] <binarycrusader> clay: yep, there's a project to improve it
[06:45:03] <binarycrusader> http://www.opensolaris.org/os/project/onnv/onnv_build/faster_builds/
[06:45:06] <clay> what are they going to do? delete half the source code? :)
[06:45:16] <binarycrusader> clay: nope, there's just a lot of inefficiencies in the build process
[06:45:42] <binarycrusader> clay: makefiles, files that get processed multiple times, possible change of shell
[06:46:08] <jbk> cases where it unnecessairly waits on stuff
[06:46:14] <clay> building software is harder than writing it
[06:46:23] <binarycrusader> clay: not enough parallelism during the build, etc.
[06:46:29] <Shiv__> Also, they'll assume all systems have 4cores and increase the parallelism to the limit that my system becomes unusable for builds ;)   (kidding)
[06:46:34] <clay> I avoided this (fixing multiboot) because I didn't want to have to learn the ins and outs of building solaris components
[06:46:54] <binarycrusader> clay: plus right now, it actually builds the entire tree twice using both gcc and sun studio compiler
[06:47:01] <clay> binarycrusader: wtf?
[06:47:08] <Shiv__> gcc can be disabled even now
[06:47:14] <binarycrusader> yes
[06:47:19] <clay> I don't want gcc anything
[06:47:28] <binarycrusader> clay: well, gcc is part of solaris 10
[06:47:29] <Shiv__> The build is to ensure it is build clean with gcc.
[06:47:33] <clay> can I build w/out gcc?
[06:47:36] <binarycrusader> clay: yes
[06:47:37] <clay> rm -rf /usr/sfw
[06:47:42] <Shiv__> But only binary bits from sunpro remain
[06:47:45] <binarycrusader> clay: bada idea :P
[06:47:53] <Shiv__> by default
[06:48:06] <clay> cp /bin/true /usr/sfw/bin/gcc
[06:48:41] <nrubsig> How can I add other people to my twitter "with others"-list ?
[06:49:01] <binarycrusader> clay: http://www.opensolaris.org/os/community/tools/gcc/shadow/;jsessionid=2ADEA54F180F16F29B5857CC54FA8D0C
[06:49:15] <binarycrusader> clay: there's a much easier way
[06:49:18] <binarycrusader> clay: read the devref or that page
[06:50:07] <Shiv__> Are there any efforts to open source sunpro?
[06:50:20] <Shiv__> Havent seen much of discussion about that?
[06:50:31] *** alexwaft has quit IRC
[06:50:33] <binarycrusader> Shiv__: Jonathan said in 2005 that everything sun does will be open sourced eventually
[06:50:47] <binarycrusader> Shiv__: considering the costs and time involved, I think it will happen, it's just taking time
[06:50:53] <binarycrusader> Shiv__: it has been mentioned on the lists before
[06:51:17] <clay> so I want CW_NO_SHADOW?
[06:51:21] <binarycrusader> yep
[06:51:33] <clay> wtf would anyway *want* to shadow?
[06:51:42] <clay> s/anyway/anyone
[06:51:47] <clay> it's late
[06:51:54] <binarycrusader> clay: because it ensures the build will work with both compilers
[06:52:05] <clay> who cares
[06:52:06] <clay> hah
[06:52:13] <binarycrusader> clay: a lot of people wanted to be able to have a freely available and redistributable compiler they could use to build opensolaris
[06:52:32] <binarycrusader> clay: since sunspro isn't freely redistributable (and open source) yet, gcc was the other option
[06:52:41] <binarycrusader> clay: the community wanted it, got involved, and working with Sun, made it happen
[06:53:29] <Shiv__> Other potential benefit of gcc support is building the source on HWs other than x86/SPARC...
[06:53:34] <clay> does sunpro result in a faster set of binaries?
[06:53:41] <binarycrusader> currently, yes
[06:53:46] <binarycrusader> especially on SPARC
[06:53:55] <binarycrusader> the gcc project has refused to accept patches from Sun that would increase performance on the SPARC platform
[06:54:01] <binarycrusader> why I don't know, some disagreement
[06:54:30] <clay> yeah I imagine sunpro would blow gcc away on niagra
[06:55:30] <Shiv__> In the roadmap, I see unit test repository available by 8/2007.
[06:55:42] <binarycrusader> yep, and it is there, a lot of test suites are out already
[06:55:49] <Shiv__> Is this integrated into ON or is it a tool set outside it.
[06:56:01] <binarycrusader> tool set outside of it
[06:56:02] <binarycrusader> http://blogs.sun.com/timf/entry/zfs_test_suite_released
[06:56:03] <Shiv__> (Does build run unit tests as well)
[06:56:08] <binarycrusader> no, build does not
[06:56:32] <binarycrusader> See this:http://www.opensolaris.org/os/community/testing/selftest/
[06:56:33] <binarycrusader> http://www.opensolaris.org/os/community/testing/testfarm/
[06:56:38] <binarycrusader> As well.
[06:59:34] <jamesbrink> whats the best way to stream live video ?
[07:00:38] <sbahra> jamesbrink, not sure if it is the best, but I was using ffmpeg for a while.
[07:00:44] <sbahra> jamesbrink, I found it really flexible for my purposes.
[07:00:53] <clay> this is taking
[07:00:55] <clay> forever
[07:00:56] <clay> to install
[07:01:13] <jbk> yeah.. that's another issue being worked on :)
[07:01:15] <jamesbrink> sbahra do you need special software to stream it on the internet
[07:01:31] <jamesbrink> somone wants me to help them stream on a webpage
[07:01:43] <jbk> basically the packages are bzip2'd, which is slow
[07:01:48] <clay> could be due to vmware and a slow ata laptop disk too
[07:01:57] <clay> hmm
[07:02:04] <clay> stopping iTunes seems to have sped it up ...
[07:02:13] <clay> hah
[07:02:15] <sbahra> jamesbrink, research it yourself.
[07:02:22] <sbahra> jamesbrink, ffmpeg is the server itself.
[07:02:41] <jamesbrink> ahh ok thanks
[07:02:50] <jamesbrink> i was looking at the darwin streaming server
[07:03:34] <sbahra> ffmpeg can stream in various formats.
[07:03:37] *** Shiv__ has left #opensolaris
[07:06:07] <jamesbrink> oh ok its called ffserver
[07:07:19] <steleman> that'd be the ffmpeg server
[07:07:27] <jamesbrink> yeah
[07:07:40] <steleman> it's quite a piece of software
[07:08:10] <steleman> i think icecast has a plugin for it and you can stream through icecast as well
[07:08:47] <sbahra> Yeah, ffmpeg is pretty huge.
[07:10:41] <jamesbrink> wow the documentation on it is very brief
[07:10:46] <steleman> now let's think for a moment: not only you are descrambling encrypted DVD's, you are also streaming them on the internet. :-P
[07:11:02] <steleman> the MPAA will have a coronary
[07:14:48] <binarycrusader> steleman: I think I just my machine a coronary when I tried to move the build directory between zfs filesystems on the same pool ;}
[07:14:56] <binarycrusader> steleman: (nightly build of b72)
[07:15:10] <steleman> ouch
[07:15:15] <steleman> did you lose anything ?
[07:15:24] <binarycrusader> no, it was a temporary coronary
[07:15:29] <binarycrusader> it's just running dog slow
[07:15:40] <binarycrusader> althoug I have had my fair share of kernel dumps the past few weeks
[07:15:47] <binarycrusader> b72 fixed most of them though
[07:15:55] <binarycrusader> well, all of them so far I should say
[07:16:16] <steleman> im still on S10U3
[07:16:18] <steleman> :-P
[07:16:25] <binarycrusader> Hey I was too for a long time
[07:16:38] <steleman> and im not moving until i get a download of U4
[07:16:49] <binarycrusader> steleman: how's the solaris integration going for KDE? I know last time I heard you said there was a lot of programs that still needed to have their functionality ported, etc.
[07:17:07] <binarycrusader> steleman: like control panel programs that didn't have their functionality implemented, etc.
[07:17:39] <steleman> kde4 is still in beta. well, yes, beta.
[07:18:00] <binarycrusader> steleman: oh so you've moved on to 4 then?
[07:18:08] <steleman> 3.5.7 Solaris i will do very soon, because i'm 80% done with all the dependencies and with the bug fixing.
[07:18:10] <steleman> not yet
[07:18:21] <steleman> 4 is not ready yet. not by far.
[07:18:24] *** Vorbis_w48 has joined #opensolaris
[07:18:25] <binarycrusader> steleman: that's what I hear
[07:18:35] <steleman> yah it still has a while to go
[07:18:37] <binarycrusader> steleman: but that seems to be a general assessment, not solaris specific
[07:18:57] <steleman> oh yah
[07:19:02] <jamesbrink> sbahra will ffserver work with a pipe?
[07:19:05] <steleman> not solaris specific, it's just not ready yet
[07:19:20] <steleman> certain apps work very well though.
[07:19:23] <steleman> Marble for example.
[07:19:24] <binarycrusader> steleman: now, as far as 3.5.7 I think you said the other day that some programs hadn't had their functionality ported
[07:19:27] <steleman> that one already works
[07:19:39] <binarycrusader> steleman: like some of the control panels weren't fully implemented on Solaris, etc.?
[07:19:44] <steleman> oh no 3.5.7 should be all there its no really different than 3.4.3
[07:20:05] <steleman> well there was some stuff missing in Solaris simply because the info is just not there
[07:20:12] <binarycrusader> steleman: such as?
[07:20:17] <steleman> but i wrote most of the Solaris specfic parts in 3.4.1
[07:20:29] <steleman> for example info about your SPARC board :-)
[07:20:34] <steleman> no clue how to get it :-)
[07:20:37] <binarycrusader> steleman: :)
[07:20:46] <steleman> or your PCI bridge
[07:20:49] <binarycrusader> steleman: so hardware browser, and stuff like that
[07:20:56] <steleman> well really specific low level hardware
[07:20:58] <binarycrusader> steleman: may not be "fully flshed out"
[07:21:12] <steleman> but the useful info is all in KControl
[07:21:15] <steleman> for Solaris
[07:21:22] <binarycrusader> steleman: what about sound output? work good with OSS arts, etc.?
[07:21:28] <binarycrusader> steleman: or is aRts dead yet?
[07:21:31] <steleman> OSS == JUST SAY NO
[07:21:37] <steleman> aRts is dead for KDE4
[07:21:39] <steleman> yay!
[07:21:44] <binarycrusader> steleman: oss is my only option so don't say no
[07:21:46] <steleman> KDE4 uses NMM
[07:22:09] <steleman> oss jus panic'ed my S10U3 the other day and completely FUCKED my audio810 which was running flawlessly since 11/06
[07:22:10] <binarycrusader> steleman: besides, given that Sun is supposed to be shipping OSS someday soon, I would hope that NO would be a YES
[07:22:15] <steleman> which i had to reinstall from scratch
[07:22:16] <steleman> so
[07:22:22] <steleman> that's my opinion of OSS at this point :-)
[07:22:48] <binarycrusader> steleman: well, I don't really have a choice
[07:23:00] <binarycrusader> steleman: my sound hardware is only supported on Solaris by OSS
[07:23:10] <steleman> what kind of card do you have
[07:23:15] <binarycrusader> Special revision of Audigy 2
[07:23:24] <steleman> blah.
[07:23:35] <steleman> get a normal Audigy or Live! for USD $25 :-)
[07:23:44] <steleman> and use Juergen's drivers *which work*
[07:23:51] <binarycrusader> Yes, I know, I used to use them
[07:23:52] <steleman> and *do not* panic the kernel
[07:24:01] <binarycrusader> Well, so far, OSSv4 hasn't panic'd my kernel yet either
[07:24:13] <steleman> you're lucky
[07:24:23] <binarycrusader> Besides, I bought a license, so I intend to squeeze all the support I can out of them if I have problems :}
[07:24:28] <steleman> i have not had a single panic since 11/06
[07:24:37] <steleman> until two days ago that is
[07:25:24] <binarycrusader> oh goody, it's finally finished moving the build
[07:25:29] <binarycrusader> I can alt-tab again, hallelujah!
[07:26:23] <binarycrusader> steleman: so is the kde work being done on blastwave or for Sun?
[07:26:35] *** yokelbombardier has quit IRC
[07:26:36] <binarycrusader> steleman: I ask because I was wondering if the code repository was public
[07:26:37] <steleman> it's being done at OpenSolaris KDE
[07:26:53] <steleman> well we have a repository at CVSDude
[07:27:07] <steleman> which is fully public anonymous access is enabled
[07:27:33] <steleman> and once the full port of KDE4 + all the deps is done there, it will move into KDE SVN
[07:27:34] <binarycrusader> steleman: https://svn2.cvsdude.com/kdesolaris/trunk/?
[07:27:39] <steleman> yup
[07:28:02] <steleman> which reminds me i have to check in about 12 things tonight :-)
[07:28:13] <binarycrusader> steleman: so do you have a build document or anything anywhere?
[07:28:23] <steleman> including a fully working version of BOOST 1.34.1
[07:28:25] <binarycrusader> steleman: a list of things left to do, etc?
[07:28:32] <steleman> i think it's under Docs
[07:28:39] <binarycrusader> steleman: I assume it works with sun studio 11 then?
[07:28:42] <steleman> if you do a svn list <bla>
[07:28:45] <steleman> we moved to 12
[07:28:48] <binarycrusader> bah
[07:28:52] <binarycrusader> this compiler madness is driving me insane
[07:28:53] <steleman> because of BOOST and KDE4
[07:29:03] <steleman> my compiler is bigger and badder than your compiler :-)
[07:29:10] <binarycrusader> Yeah, well tell that ONNV
[07:29:18] <steleman> i sent a bitchgram
[07:29:31] <binarycrusader> It annoys me to no end that SXDE comes with Studio 12
[07:29:36] <steleman> why ?
[07:29:37] <binarycrusader> but you have to install 11 to build it
[07:29:42] <steleman> ohh that
[07:30:04] <binarycrusader> it makes setting up my development workspace for on rather tedious
[07:30:14] <steleman> for KDE4 and BOOST and stdc++ there were things which you simply cannot do with Studio 11
[07:30:18] <binarycrusader> I've ended up just installing both 11 and 12 and switching my SUNWspro directory symlink as needed
[07:30:27] <binarycrusader> steleman: it would be good to see you cover those in your talk
[07:30:34] <binarycrusader> steleman: if it gets approved as a topic
[07:30:37] <steleman> i also have 11 and 12 installed
[07:30:47] *** nostoi has joined #opensolaris
[07:31:06] <binarycrusader> steleman: so is there a todo list for the KDE stuff somewhere?
[07:31:34] <steleman> binarycrusader: not really
[07:31:39] <steleman> it's in our heads so to speak
[07:32:13] <binarycrusader> I hope for KDE4 you will consider using a bug tracker?
[07:32:19] <steleman> if you ask on kde-discuss@opensolaris and you want to do some work, i'm sure we'll find something for you :-)
[07:32:26] <steleman> oh we have a bugzilla thingy at CVSDude
[07:32:38] <binarycrusader> So this is the only buil doc at the moment? https://svn2.cvsdude.com/kdesolaris/trunk/Build/README
[07:32:39] <steleman> right now we're still porting.
[07:33:01] *** theRealballchalk has joined #opensolaris
[07:33:33] <binarycrusader> steleman: is all of your work going to be accepted upstream hopefully?
[07:33:54] <steleman> yes
[07:34:05] <steleman> we made the decision to switch Standard C++ Libraries
[07:34:06] <binarycrusader> and that link I pasted above is the only build doc at the moment?
[07:34:09] <steleman> so we are using apache/roguewave
[07:34:37] <steleman> and because of this the number of studio specific patches is much smaller
[07:34:42] <steleman> probably is
[07:34:56] <steleman> dude if you wanna get involved, there's #kde-solaris on IRC and there's the mailing lists
[07:35:08] <binarycrusader> steleman: just testing the waters, I'll look around, and thanks for the info :}
[07:35:09] <steleman> drop by and you'll soon have work on your hands :-)
[07:35:31] <steleman> right now it's not very "corporate" because it's just 3 of us basically
[07:35:47] <binarycrusader> you are somehow connected with the Sun Studio team?
[07:36:06] <steleman> well i am on very friendly terms with them
[07:36:21] <binarycrusader> Are they doing anything about the deficiencies that made you adopt a different C++ library?
[07:36:33] <steleman> (by the way they are really really nice people)
[07:36:38] <binarycrusader> I have no reason to doubt that.
[07:36:41] <steleman> the problem is: they can't.
[07:36:54] <binarycrusader> What's the short version?
[07:37:03] <steleman> the current libCstd.so.1 *must* stay binary compatible with Workshop 5 from 1997.
[07:37:13] <binarycrusader> ...until?
[07:37:16] <steleman> so, it must stay crippled.
[07:37:34] <steleman> until there's a newer version, which will be incompatible, but fully standards compliant.
[07:37:45] <binarycrusader>  I take it they are years away from a major new version?
[07:37:50] <binarycrusader> so, essentially, they can't for now
[07:38:05] <steleman> actually Studio 12 compiled RogueWave's STDCXX without a glitch
[07:38:11] <binarycrusader> That's nice
[07:38:14] <steleman> so, with stdcxx, Studio 12 is compliant.
[07:38:18] <steleman> fully compliant.
[07:38:23] <binarycrusader> gcc too?
[07:38:30] <steleman> uhm.
[07:38:36] <binarycrusader> or no, because of the whole gcc ABI thang
[07:38:37] <binarycrusader> ?
[07:38:48] <steleman> there is a secret flag which makes Studio 12 understand GCC's ABI.
[07:39:10] <binarycrusader> No, I meant, could gcc in combo with RogueWave be considered fully compliant
[07:39:19] <steleman> gcc + roguewave ?
[07:39:23] <binarycrusader> Aye
[07:39:24] <binarycrusader> Can you even do that
[07:39:29] <steleman> gcc 4.x yes
[07:39:34] <steleman> yes you can build RW with GCC
[07:39:35] <steleman> definitely
[07:39:49] <binarycrusader> It sounds like your proposed topic would definitely be interesting
[07:40:23] <binarycrusader> I've always wanted to see C++ take off on *NIX platforms, but it has so many issues on the most popular *NIX-like platform (GNU/Linux) that I've always hated it.
[07:40:28] <steleman> it's about studio 12 and roguewave and boost and how everything fit together quite nicely
[07:41:17] <binarycrusader> Studio 12 runs on GNU/Linux stuff, so sounds like a great dev environment whether you're on Solaris or GNU/Linux...
[07:41:22] <steleman> C++ is a very nice language. it's just not very fitted for every single task. there are things you would never do in C++.
[07:41:29] <steleman> Studio 12 on Linux rocks.
[07:41:39] <binarycrusader> Oh, I know. I've maintained a Linux port of a Windows game runtime engine for years that's written in C++.
[07:41:40] <steleman> i demo'ed QT and QT/Jambi on Linux at the KDE conference
[07:41:49] <steleman> built with Studio 12.
[07:41:55] <steleman> it was about 2 times faster than GCC :-)
[07:42:00] <e^ipi> Qt is v. nice to work with
[07:42:06] <binarycrusader> Let's just say my personal experience with C++ and GNU/Linux has stunk bad.
[07:42:14] <steleman> Qt is a dream.
[07:42:15] <e^ipi> complaints about moc aside
[07:42:27] <binarycrusader> The ABI issues are especially irritating on the GNU/Linux platform.
[07:42:28] *** FurryCat has joined #opensolaris
[07:42:38] <steleman> well yeah. moc is a leftover from the 90's when meta compiling was so popular
[07:43:10] <steleman> binarycrusader: the gcc guys have a fundamentally different approach to compiler writing than the studio team.
[07:43:15] <FurryCat> hi everyone.  Is it safe to upgrade from 11/06 to 08/07 without doing a backup?
[07:43:17] <binarycrusader> steleman: it shows :P
[07:43:22] <steleman> the gcc guys want to be compliant first, and ABI compatible maybe.
[07:43:37] <steleman> the studio guys are the exact opposite. ABI compatibility first, compliance in time.
[07:44:06] <binarycrusader> which I'd rather have, because then, at least I could build one C++ binary that worked on multiple GNU/Linux distros regardless of version going forwards
[07:44:27] <steleman> i would dare anyone link something compiled with GCC 4.x against a C++ library compiled with gcc 3
[07:44:37] <binarycrusader> indeed
[07:45:11] <binarycrusader> I gave up on my dream of one platform, one binary last year and just told everyone I was only going to support Ubuntu
[07:45:14] <steleman> on the other hand, you can link libraries built with Workshop 5 in Studio 12.
[07:45:32] <e^ipi> binarycrusader: Solaris has a C++ abi defined... GNU just chooses to ignore it
[07:45:37] <binarycrusader> I know
[07:45:43] <e^ipi> or so stoustrup informs me
[07:46:09] <steleman> the Studio guys just proved that they do. linking against libraries built 7 compiler releases ago ?
[07:46:14] <steleman> and they are still 100% compatible.
[07:46:26] <binarycrusader> yep, gotta love it
[07:46:49] <binarycrusader> The development tools were the reason I switched to Solaris at home a few years ago
[07:46:57] <binarycrusader> DTrace, Sun Studio, etc. best place for me.
[07:47:00] <theRealballchalk> e^ipi: ooh stoustrup
[07:47:12] <steleman> i just dont' get this "We Must GCC Now."
[07:47:13] <e^ipi> theRealballchalk: if anyone'd know it'd be him
[07:47:53] <steleman> yeah he knows a thing or two about C++
[07:48:43] *** binarycrusader has quit IRC
[07:48:58] *** binarycrusader has joined #opensolaris
[07:52:17] <trochej> yup, definately
[07:52:20] <trochej> a thing or two
[07:52:21] *** libkeise1 has joined #opensolaris
[07:53:21] <trochej> Fok
[07:53:24] <trochej> I can't spell
[07:54:31] *** libkeiser has quit IRC
[07:54:43] *** libkeise1 is now known as libkeiser
[07:57:01] <nrubsig> Who again asked for a twitter client ?
[07:57:03] <nrubsig> http://twitter.com/gisburn
[08:03:33] <steleman> wut is twitter
[08:04:05] <nrubsig> steleman: do yo blog ?
[08:04:12] <binarycrusader> steleman: a way to post a text message every single moment of the day best I can tell to a webpage
[08:04:13] <nrubsig> s/yo/ you/
[08:04:27] <binarycrusader> "A global community of friends and strangers answering one simple question: What are you doing? Answer on your phone, IM, or right here on the web!"
[08:04:31] <binarycrusader> Sounds..."so exciting"
[08:05:06] <steleman> nrubsig: i rarely blog
[08:05:23] <steleman> nrubsig: i have nothing to say, and i'm saying it.
[08:10:48] <jbk> heh that's pretty much my feeling on blogging :)
[08:12:04] <binarycrusader> I keep a blog, but strictly as a place to record technical information that I or others might use
[08:12:09] <binarycrusader> I don't do personal blogging
[08:12:18] <binarycrusader> so I may go months without making an entry
[08:13:06] *** supernerd has quit IRC
[08:24:51] <trochej> I write something small every two days about this and that in OS
[08:24:59] <trochej> Just not to forget where I keep my blog :)
[08:33:22] *** Biovore has joined #opensolaris
[08:33:42] <trochej> http://www.flickr.com/photos/kebu/558014001/
[08:33:43] <trochej> Khehehe
[08:33:52] <trochej> Sign says: "Live children"
[08:33:53] <trochej> :)
[08:33:55] *** Sup3rkiddo has joined #opensolaris
[08:36:28] <flyingparchment> does vim have an ascii art mode, like emacs?
[08:37:07] <trochej> [d]
[08:40:02] *** binarycrusader has left #opensolaris
[08:49:50] *** supernerd has joined #opensolaris
[08:56:40] *** dlynes_laptop has joined #opensolaris
[08:57:00] *** dlynes_laptop has quit IRC
[08:57:04] *** dlynes_ has joined #opensolaris
[08:59:27] *** Fish- has joined #opensolaris
[09:00:12] *** Fish- is now known as Fish
[09:00:38] *** derchris has quit IRC
[09:01:06] *** FurryCat has quit IRC
[09:07:45] *** Shiv__ has joined #opensolaris
[09:09:41] *** FurryCat has joined #opensolaris
[09:10:03] <FurryCat> what's the difference between solaris 8/07 and opensolaris?
[09:14:23] <logic_> http://www.opensolaris.org/os/about/faq/general_faq/#opensolaris-solaris
[09:14:30] <Atomdrache> OpenSolaris is largely derived from Solaris 10, so yes.
[09:14:36] <Atomdrache> (correct me if I'm wrong)
[09:14:49] <flyingparchment> yes?  i didn't see a question :)
[09:14:59] <flyingparchment> or not a yes/no one
[09:16:28] <FurryCat> i guess, will solaris 8/07 be missing a lot of cool features that are in opensolaris?
[09:16:47] <flyingparchment> most of the interesting features currently in opensolaris got backported to solaris 10
[09:17:19] *** mikefut has joined #opensolaris
[09:17:36] *** nrubsig has quit IRC
[09:18:10] <Atomdrache> (I perhaps misunderstand what you mean by 8/07, too.)
[09:26:49] <FurryCat> if I do an upgrade from 11/06 to 8/07, do I have to worry that my /home directories are zfs?
[09:27:00] <FurryCat> my opt directory is zfs too
[09:27:42] *** Mdx4 has joined #opensolaris
[09:29:24] *** Doc has quit IRC
[09:31:00] <Bartman007> if you have to ask the question, stick with solaris 8/07
[09:32:16] <Bartman007> (oops, that is regarding the solaris vs opensolaris question)
[09:32:26] *** nostoi has quit IRC
[09:34:41] *** FurryCat has quit IRC
[09:45:08] *** cypromis has quit IRC
[09:48:08] *** jfndi has joined #opensolaris
[09:50:01] *** Mdx4 has quit IRC
[09:51:12] *** supernerd has quit IRC
[09:52:11] *** bengtf has quit IRC
[09:55:55] *** JWheeler has quit IRC
[09:57:08] *** Shiv__ has quit IRC
[09:57:54] *** estibi has joined #opensolaris
[10:00:07] *** dmarker has quit IRC
[10:00:24] *** dmarker has joined #opensolaris
[10:02:02] <richlowe> gah.
[10:02:24] <richlowe> I guess it's the KMF putback that lints so far from cleanly?
[10:06:55] *** JWheeler has joined #opensolaris
[10:11:50] *** Doc has joined #opensolaris
[10:19:58] *** bengtf has joined #opensolaris
[10:31:39] *** Pietro_S has joined #opensolaris
[10:36:05] *** sycofly has joined #opensolaris
[10:37:10] <JWheeler> Is there meant to be an example named.conf tucked away somewhere for dns/server? Or are we all meant to be unix uber-geeks, that just enjoy making them by hand?
[10:40:25] <Doc> http://justfuckinggoogleit.com/
[10:43:47] <JWheeler> yes. I did
[10:44:08] <JWheeler> the bigadmin articles that I've found so far suggest that you must create one, which I hope is simply outdated
[10:46:39] <quasi> man suggests looking at "BIND 9 Administrator  Reference  Manual"
[10:47:00] <JWheeler> Which will be the ISC one presumably
[10:47:46] <quasi> yeah - it even points to www.bind9.net
[10:48:03] <Tempt> named.conf is pretty simple
[10:48:12] *** logic has joined #opensolaris
[10:48:13] <JWheeler> Oh well, I guess it's a good chance to start from scratch. I might try and find one of the nice secure templates
[10:48:21] <flyingparchment> pdns.conf is even simpler
[10:48:22] <flyingparchment> ;-)
[10:48:55] <JWheeler> Tempt, I know it can be, I'm just really surprised solaris doesn't come with a named.conf.default, or similar. Every other nix that I've used, has - even gentoo!
[10:49:14] <Tempt> Yes, but is it a useful example?
[10:49:37] <JWheeler> enough to get you working, with some helpful comments. Very.
[10:50:53] <Tempt> Well, there you go. I always build bind from source, anyway.
[10:52:01] <JWheeler> Fair enough. I'm a bit out of date on the bind front, I must confess. I see solaris ships with 9.3, but 9.4 is out. Are they many good reasons for the common man to need 9.4 over 9.3?
[10:52:16] <JWheeler> *there
[10:52:17] <Tempt> I think there was a bug recently
[10:52:19] <Tempt> or a 'sploit.
[10:53:03] <JWheeler> O_o
[10:55:12] *** mikefut has quit IRC
[11:00:22] *** cypromis has joined #opensolaris
[11:00:56] *** Vorbis_w48 has quit IRC
[11:01:27] *** logic_ has quit IRC
[11:01:28] *** bunker has joined #opensolaris
[11:04:35] *** Vorbis_w48 has joined #opensolaris
[11:07:14] <asyd> and 9.4 have some nice features
[11:08:02] <JWheeler> I'm just tossing up if I should split my ns and dns servers
[11:08:12] <JWheeler> with zones.... it's just so easy to add another host
[11:08:35] <JWheeler> I already have views, but that doesn't entirely solve all of my issues
[11:08:42] <JWheeler> a proper split would be better
[11:08:52] *** andyshack has joined #opensolaris
[11:09:22] <Tempt> Well, zones make it very easy.
[11:09:58] <JWheeler> I seem to remember some nutter pushing the limits a bit with zones, he had 600+ on one box
[11:10:11] <JWheeler> apache servers I think, awesome... I had a good chuckle
[11:10:20] <Tempt> imagine patching that.
[11:10:22] <Tempt> 600 zones
[11:10:51] <Tempt> it'd take weeks per patch
[11:11:11] <JWheeler> Good times :)
[11:12:07] <Tempt> Indeed.
[11:12:46] <andyshack> Tempt : Glad you're here. I'm trying to nut out building a console cable for my v240 to see if its a lemon or not. I interpret sun docs to say "use the rj45 not the db9 serial port" and assume that to be true as I cant get output on the db9 using a null modem cable..
[11:12:46] <estibi> :)
[11:12:50] <andyshack> Is that correct ?
[11:13:07] <Tempt> The DB9 is serial port B
[11:13:11] <Tempt> Not the ALOM port.
[11:13:14] <Tempt> So yes, use the RJ45.
[11:13:41] <Tempt> (you can use the db9 for your dialup modem ;)
[11:13:57] <andyshack> Yeah, I havent used the ALOM as I'm not sure how its configured at the moment. I did try to see if it would talk to the DHCP although it didnt so I left it alone for now.
[11:14:14] <Tempt> well, the LOM is the way into those machines
[11:14:34] <Tempt> And the DB9 is definately the second serial port.
[11:15:03] <Tempt> So use the RJ45 (which is the ALOM) with a standard Cisco console cable and you'll be Just Fine(tm).
[11:16:01] <andyshack> Yeah. I've tried a few cables and I'm not getting any output, so I'm just building one to make sure the server isnt toast.
[11:16:43] <andyshack> Oh, I tried a few cisco DCE's.
[11:17:32] <Tempt> If you're not getting any response from the alom port with a cisco  cable, there are two common problems:
[11:17:43] <Tempt> a) Your laptop or whatever is fscked
[11:17:46] <Tempt> b) Your 240 is fscked.
[11:18:03] <quasi> c) power is off
[11:18:20] <andyshack> Sorry, I've been trying with the rj45 serial mangment port. Not the ALOM. I should be using the ALOM ?
[11:18:34] <Tempt> The RJ45 serial management port IS the ALOM.
[11:18:57] <andyshack> I'll gp back to the docs. cheers.
[11:19:09] <Tempt> Try dropping it an escape
[11:19:12] <Tempt> #.
[11:20:02] <andyshack> I was more thinking just try dropping it.
[11:20:29] <Tempt> hah.
[11:20:50] <Tempt> test your serial setup on something else in case it turns out to be something silly.
[11:21:15] <andyshack> Oh, im 99% sure this is pebak related.
[11:21:25] <andyshack> +c
[11:29:09] *** derchris has joined #opensolaris
[11:33:18] *** andyshack has quit IRC
[11:38:44] *** jfndi has quit IRC
[11:39:39] *** thezerox has joined #opensolaris
[11:40:34] <WickedWicky> YO
[11:43:18] <Tempt> YO.
[11:43:43] *** estibi_ has joined #opensolaris
[11:52:05] *** estibi has quit IRC
[11:52:11] *** estibi_ is now known as estibi
[11:53:02] <WickedWicky> allrighty
[11:53:05] <WickedWicky> lunch time
[11:53:08] <WickedWicky> see ya all in some hours
[11:53:11] <Gekkko> once you learn vim it's quite nice.
[11:55:25] <JWheeler> ha... I may have done something a bit stupid
[11:55:38] <JWheeler> Will a ntp server work in a non-global zone?
[11:56:07] <JWheeler> I'm thinking it won't be allowed to change the time, only the global would
[11:56:14] <JWheeler> at least, that's the case with Xen
[11:57:28] <libkeiser> if a child zone could control clock drift, that would be a serious security problem
[11:57:40] <richlowe> aye.
[11:57:57] <richlowe> or at least give them the ability to make life in the other zones royally suck.
[11:58:00] <JWheeler> thanks, just needed a sounding board for that one
[11:58:09] <richlowe> ever seen the clock change drastically while make was running?
[11:58:18] <JWheeler> well, I guess I'll have to cross that one off the consolidation list
[11:58:20] <richlowe> especially when combined with autoconf, where it can cause that crap to loop indefinitely.
[12:04:07] *** derchris has quit IRC
[12:06:23] *** thezerox has quit IRC
[12:11:17] *** cmihai has joined #OpenSolaris
[12:11:35] <Berny> hmm, can anyone tell me, where (and how) di_node_t is defined?
[12:11:44] *** Sup3rkiddo has quit IRC
[12:14:36] <richlowe> it's an allegedly (but seemingly not really) struct di_node *
[12:14:43] <richlowe> uh, allegedly *opaque*
[12:15:47] <Berny> yep that i found too... but what's inside that struct?
[12:16:08] *** derchris has joined #opensolaris
[12:16:19] <richlowe> if you're intented to treat it as opaque...
[12:16:28] <richlowe> otherwise, it's in sys/libdevinfo_impl.h
[12:16:36] <richlowe> uh, except no, because I can't type what I mean.
[12:16:45] <richlowe> devinfo_impl.h
[12:17:35] <Berny> ah cheers
[12:21:57] <richlowe> also, it appears that src.o.o's definition search is entirely busted.
[12:22:08] <richlowe> it'll find nothing, but it'll suggest others (which do exist), and then also not find them.
[12:22:26] * jmcp frequently reverts to echo "::print -t $typename"|mdb -k
[12:22:43] <richlowe> jmcp: Yeah, I was at the point of cscope'ing for it when I realized where it'd be.
[12:26:01] * Berny is trying to find out where biosdev fails... at the moment it looks like it just stops searching  through the di_node tree before it comes to the targets i have...
[12:26:39] <jmcp> Berny: if you come up with a solution, please let me know
[12:27:29] <Berny> right now i wonder if the bios might report something different from the stuff the kernel has in its structures... but then again why do these data differ?
[12:27:47] <Berny> i sure will
[12:27:50] <jmcp> because bios is the suX0r
[12:27:51] <madhatter> Hi. Is there a way to force detection of sata disks? I just added my new sata hd to the onboard controller, but 'format' does not shows it.
[12:28:04] <jmcp> madhatter: depends on what your onboard sata controller is
[12:28:44] <jmcp> madhatter: what sata controller do you have your disks connected to?
[12:28:47] <madhatter> jmcp: I talked about it here a couple of days ago. It is a silicon image chip 3112a, but onboard, so I don't know if it really is supported
[12:28:54] <jmcp> ah, right
[12:29:08] <jmcp> it should work ok with the ata driver (ie, in legacy parallel ATA mode)
[12:29:31] <madhatter> jmcp: Do I have to load that manually first?
[12:29:36] <jmcp> unlikely
[12:29:41] <jmcp> it's probably loaded already with your cd/dvd
[12:29:46] <Berny> jmcp: may well be... for me bios reports channel 2 and 3... but the search routine stop after 0,1, because i finds the next sibling to be DI_NODE_NIL :-\
[12:29:52] <jmcp> madhatter: please use a handy pastebin site and show me the complete output of     prtpicl -v
[12:29:57] <jmcp> Berny: damn
[12:30:07] <madhatter> jmcp: Ok, I will do that
[12:30:08] <dlg> what is picl?
[12:30:11] <jmcp> Berny: could you show me your test code please?
[12:30:18] <jmcp> Platform Information and Control Library
[12:30:23] <jmcp> came in with s8 and the serengeti range
[12:30:24] <dlg> whats underneath it?
[12:30:34] <jmcp> uh .... da kernel?
[12:30:43] * jmcp never bothered to lift the covers
[12:30:44] <dlg> the lom stuff on the t1 105 is behind it
[12:30:56] <dlg> and noone knows below the kernel layer
[12:31:13] <Berny> not much to see yet... just more debug info in a few places jmcp
[12:31:28] <jmcp> Berny: I'd be interested in having a crack at it, just quickly
[12:31:36] <jmcp> my u20 has this biosdev issue too
[12:34:17] <richlowe> dlg: a whole bunch of stuff is vaguely involved with it.
[12:34:42] <madhatter> jmcp: Too much output for term history. I will look for a way to put up a txt file somewhere instead
[12:35:00] <jmcp> ok
[12:35:08] <jmcp> prtpicl -v > /tmp/prtpicl-v.out
[12:35:09] <jmcp> etc
[12:35:12] <richlowe> dlg: $SRC/cmd/picl is handy to dig through, as well as libpicl/libpicltree/libfru
[12:35:30] <richlowe> dlg: the plat plugins especially.
[12:36:06] <clay> I've just installed b72 under VMware Fusion. Any idea why, after loading all the SMF manifests and generating the SSH keys, I would get nothing but a blank screen?
[12:36:54] <clay> oh
[12:36:58] <clay> I just need to be patient
[12:37:24] <clay> my poor CPU
[12:37:27] <clay> I think it's going to melt
[12:37:42] <madhatter> jmcp: http://www.box.net/shared/y0mgyb5a35
[12:37:54] <dlg> richlowe: sounds hard, and my head hurts
[12:40:07] <richlowe> why the hell do we conditionalize on the DDI revision?
[12:40:12] <richlowe> we know the DDI specifics, it's in the same damn tree.
[12:40:35] <richlowe> oh, "compatibility definitions for backport to solaris 8"
[12:41:30] <richlowe> so do it in the on28 gate, damnit.
[12:41:32] *** derchris has quit IRC
[12:45:48] <jmcp> madhatter: thanks, looking at it now
[12:47:23] <jmcp> madhatter: you could try this as root::                    update_drv -a -i ' "pci1095,3112" ' ata
[12:47:35] <jmcp> note the quotes  -   single double .... double single
[12:48:11] <dlg> jmcp: nice trick ;)
[12:48:40] <jmcp> dlg: I should blog it, too
[12:48:45] <jmcp> as well as get that rfe integrated
[12:48:56] <jmcp> Berny: incidentally, sys/biosdisk.h isn't delivered via a package
[12:49:16] <dlg> is my sii the raid type?
[12:49:33] <jmcp> yeah
[12:49:40] <dlg> handy
[12:51:44] <jmcp> Berny: http://paste.lisp.org/display/47768
[12:51:49] *** ichigo has joined #opensolaris
[12:52:42] <jmcp> Berny: if I don't run it with -d, then I get some errors on failure to open my usb devices (which aren't plugged in anyway), and then     0x83 /pci@0,0/pci10de,370@6/pci1000,1000@9/sd@5,0
[13:04:35] *** Mdx4 has joined #opensolaris
[13:04:47] <jmcp> madhatter: did that command line assist you at all?
[13:05:20] <Berny> jmcp: you get a match on a disk that isnT plugged in?
[13:06:05] <jmcp> no, all the disks which get an 0x8[0...] are plugged in
[13:06:11] <jmcp> it's the usb storage that isn't plugged in
[13:06:45] <Berny> ah okay
[13:06:46] *** program has quit IRC
[13:08:07] *** deather_ has joined #opensolaris
[13:13:32] <Gekkko> hmm, vi doesn't actually exist, it's ex -v :P
[13:15:07] <madhatter> jmcp: Sorry, I had to do some other stuff here. Postman delivery
[13:15:18] <madhatter> jmcp: I will do that now
[13:16:02] <jmcp> righto
[13:16:05] <jmcp> when you're ready
[13:16:13] <madhatter> jmcp: Hm, nothing happened so far ;)
[13:16:24] <jmcp> what response was printed to your terminal?
[13:16:34] <madhatter> jmcp: Nothing
[13:16:42] <jmcp> ok
[13:16:57] <jmcp> tail /var/adm/messages   for me, let me know whether you've got anything there
[13:17:32] <madhatter> jmcp: Some UDMA mode selections
[13:17:36] <jmcp> ok
[13:17:41] <jmcp> run      devfsadm -v
[13:17:42] <madhatter> Maybe that is from calling format again
[13:17:56] <madhatter> No output
[13:18:05] <jmcp> ok, now run format again
[13:18:28] <madhatter> still the old three disks
[13:18:39] <jmcp> hmm
[13:18:45] <jmcp> I expected there to be some output
[13:18:57] <jmcp> did you type the update_drv command exactly as I'd written?
[13:19:13] <madhatter> jmcp: yes, copy & pasted it
[13:19:27] <madhatter> jmcp: as root
[13:19:32] <jmcp> right
[13:19:39] <jmcp> hmmm
[13:20:01] <jmcp> try     devfsadm -v -i ata
[13:20:33] <madhatter> jmcp: Maybe it's because that controller seems to be ment for hardware raid... I tried to deactivate that but there was not option
[13:20:45] <jmcp> oh. that's a pain
[13:21:02] <madhatter> jmcp: Did no output either
[13:21:03] <jmcp> not sure what else to do to try to help you then, sorry
[13:21:20] <madhatter> jmcp: THank you
[13:21:36] *** deather has quit IRC
[13:21:37] *** deather_ is now known as deather
[13:21:48] <madhatter> jmcp: Maybe I should just get that promise (?) sata controller then
[13:22:20] <jmcp> perhaps go for one that's listed on the hcl already? sun.com/bigadmin/hcl
[13:22:25] <clay> how can I fetch a pre-dboot copy of the ON source code?
[13:22:56] *** nostoi has joined #opensolaris
[13:22:59] <madhatter> jmcp: I did not find that much in there, but read about people using that Promise S-ATA-Controller 300 TX4 successfully
[13:23:16] <sickness> madhatter: the hcl is almost non-existant... :/
[13:23:17] <madhatter> jmcp: I have the problem that most controllers are for pci-x
[13:23:31] <jmcp> madhatter: and you want pci, or pci-e?
[13:23:46] <madhatter> jmcp: My old board only has pci slot afaik
[13:27:12] <sickness> madhatter: any no-brand sil3114 pci will work fine, but you'll have to flash the firmware to make it a jbod...
[13:27:41] <madhatter> sickness: jbod?
[13:27:51] <sickness> madhatter: well yeah, see the disks
[13:27:53] <sickness> directly
[13:28:00] <madhatter> oic
[13:28:27] <madhatter> sickness: Is there any chance of a list what cards have those chips?
[13:28:35] <jmcp> madhatter: if you get a sil311[234]-based card with raid firmware, the prtpicl output will show it with "pciclass,010800"
[13:28:59] <jmcp> madhatter: so the same update_drv trick should work ok for you
[13:29:42] <madhatter> In the hcl, does serial ide mean the same as serial ata?
[13:29:46] <clay> has the URI to the ON Mercurial repository changed?
[13:29:49] <sickness> madhatter: heh, that's not easy, you need to be sure of the chip that's on them :/
[13:30:25] <clay> I get "There is no Mercurial repository here" when I try: hg tags ssh://anon at hg dot opensolaris.org/hg/onnv/onnv-gate
[13:31:32] <madhatter> jmcp: I see, I will note that somewhere for later
[13:32:22] <cmihai> clay, do a local clone first
[13:32:30] <jmcp> though if you get one and put it into this machine that you just ran that command on, it might work immediately since you ran the command
[13:32:33] <clay> cmihai: thanks
[13:32:34] <cmihai> Make sure it's on ZFS, it's like a zillion times faster to deal with after that
[13:33:38] <madhatter> Are VIA VT6421 chipsets supported?
[13:33:55] <jmcp> dunno
[13:33:57] <cmihai> Once you've got an hg clone ssh:// etc all other operations takes minutes. Hell, on local ZFS, takes about a minute to clone :-)
[13:34:10] <jmcp> madhatter: is that a nic or (S)ATA ?
[13:34:21] <madhatter> jmcp: Sata
[13:34:28] <jmcp> gimme a sec
[13:34:42] <madhatter> jmcp: But I just found one with Silicon Image  SATA Link Sil3114CT176
[13:35:11] <jmcp> I'd go with that Sil3114
[13:36:09] <madhatter> jmcp: It's mentioned to support Raid 0, 1 and 10. But no hint if raid can be deactivated
[13:36:24] <jmcp> no idea
[13:36:29] <sickness> madhatter: you'll need to find a bios to flash it
[13:36:30] <jmcp> I'm sure it can be with MS-Windows
[13:36:41] <jmcp> if you had the vt8251 then you could use the ahci driver
[13:36:52] <cmihai> madhatter, you got a SIL?
[13:37:04] <cmihai> madhatter, SIL is supported by Solaris...
[13:37:19] <madhatter> cmihai: At the moment I have an onboard sil
[13:37:30] <madhatter> cmihai: But that does not seem to work
[13:37:41] <cmihai> What Solaris
[13:37:46] <cmihai> http://www.sun.com/io_technologies/vendor/silicon_image__inc_.html
[13:37:47] <madhatter> cmihai: That why I am looking for another one
[13:37:55] *** Fullmoon has joined #opensolaris
[13:38:02] <madhatter> cmihai: sxce nv b70
[13:38:30] <cmihai> Oh, thought it was some old S10. What controller / revision?
[13:38:51] <cmihai> Oh, sorry, I missed that.
[13:38:54] <cmihai> 3114
[13:39:07] <cmihai> That should be _fully_ supported and "verified" according to sun
[13:39:14] <madhatter> cmihai: That's the one I am thinking about to buy
[13:39:37] <madhatter> cmihai: What about hte *CT176? ;)
[13:39:37] <cmihai> Oh, right. Well, sounds good.
[13:40:35] <madhatter> cmihai: Is there any difference between the 3114 chipsets
[13:40:58] <cmihai> No idea.
[13:41:14] <cmihai> Try asking on the lists
[13:41:22] <madhatter> cmihai: THat 's the only thing I am afraid of
[13:41:41] <cmihai> That, and the fact they don't support storage targets > 2TB :-)
[13:41:53] <madhatter> cmihai: Okay, I will. Is there a special hardware list for stuff like that?
[13:42:22] <madhatter> cmihai: Ouh, too bad. I am planning a 10TB pr0n server over here :)
[13:42:23] <cmihai> http://www.sun.com/io_technologies/sata/SATA0.html - that's pretty much it. The list is bigger for SXCE and stuff that _may_ work, but that's all for the official stuff.
[13:42:45] <cmihai> Nah, I think that just means individual disks mate
[13:43:02] <cmihai> As in you can't use 3TB disks or something
[13:43:17] *** infidel has joined #opensolaris
[13:43:30] <infidel> is there hot keys on cde to bring up a terminal?
[13:44:25] <madhatter> cmihai: Oh, I see. But I am starting with one single 500g disk here for now and wanting to extend the zpool now and then
[13:45:38] *** Vorbis_w48 has quit IRC
[13:46:16] <cmihai> infidel, no
[13:46:49] <infidel> ok
[13:55:05] *** Vorbis_w48 has joined #opensolaris
[13:55:56] <madhatter> Somebody mentioned jblod here
[13:56:39] <madhatter> Just found a dawicontrol card that seems to have sil3114 chipset, too. And there is 'Raid JBLOD' listed
[14:05:27] *** nivox has joined #opensolaris
[14:08:24] <madhatter> SOmething else: are there enlightenment packages for sxce available somewhere? I found a link to pmpkg on the opensolaris website, but as far as I understood that just lets you build your own packages
[14:10:25] <quasi> you could try blastwave
[14:11:47] <quasi> http://www.blastwave.org/packages.php/enlightenment
[14:11:49] <madhatter> quasi: I have been thinking about that but I read about issues with live upgrading
[14:13:10] <quasi> madhatter: oh? I haven't heard about that, but then I've never felt the urge to use anything off blastwave
[14:14:12] *** Gekkko[PDA] has joined #opensolaris
[14:14:19] *** master_of_master has joined #opensolaris
[14:14:30] <madhatter> quasi: On the other hand, I could just compile it myself, as I did with mutt, vim and other stuff as well...
[14:15:00] <quasi> madhatter: you could - although I think there's quite a few dependencies
[14:15:32] *** vmlemon has joined #opensolaris
[14:16:11] *** Gekkko[PDA] has quit IRC
[14:18:01] *** Gekkko[PDA] has joined #opensolaris
[14:18:47] *** Gekkko has quit IRC
[14:19:00] <madhatter> quasi: Ya, just found SPARC packages so far
[14:21:49] <quasi> I like pmpkg - even if it does have you building things, it does take care of all the tedious bits - but I don't remember if it has enlightenment
[14:22:41] <madhatter> quasi: It is mentioned here: http://www.opensolaris.org/os/community/desktop/communities/enlightenment/
[14:23:00] <madhatter> quasi: I will read the readme and see how that tool works
[14:23:44] <LeftWing> infidel: You can define hotkeys to invoke a terminal in CDE, though.
[14:24:22] <infidel> LeftWing, oh? how?
[14:25:07] <LeftWing> Good question.  In the settings panels someplace you can configure hotkeys, one of the available actions is start an arbitrary program.
[14:25:13] <LeftWing> I can'
[14:25:15] <LeftWing> t remember the specifics.
[14:25:37] *** mikefut has joined #opensolaris
[14:25:45] <infidel> ok i'll check into it
[14:30:45] *** master_o1_master has quit IRC
[14:37:09] *** Mdx4 has quit IRC
[14:39:57] *** cypromis has left #opensolaris
[14:41:40] *** bunker has quit IRC
[14:45:16] *** bunker has joined #opensolaris
[14:49:32] *** sycofly has quit IRC
[15:11:33] *** Mdx4 has joined #opensolaris
[15:14:02] *** bunker has quit IRC
[15:17:47] *** mega has joined #opensolaris
[15:18:24] *** bunker has joined #opensolaris
[15:19:50] *** bunker has quit IRC
[15:21:15] *** bunker has joined #opensolaris
[15:22:41] <clay> grr
[15:24:13] <clay> sh: /export/home/clay/onnv-gate/usr/src/tools/proto/opt/onbld/bin/i386/cw: not found
[15:26:38] *** shpchp has joined #opensolaris
[15:32:27] *** perlmonk has joined #opensolaris
[15:34:51] <cmihai> Eh?
[15:34:56] <cmihai> What are oyu trying to do clay
[15:35:26] <clay> trying to build multiboot
[15:35:59] <clay> I think running "dmake setup" might have helped
[15:36:00] <cmihai> Right. But what gave you that error? You started the build?
[15:36:09] <clay> yeah
[15:36:22] <cmihai> I supposed you reverted to a previous state of the source
[15:36:30] <cmihai> One where multiboot was still around?
[15:36:40] <clay> following, the readme, I copied opensolaris.sh to my home dir, modified it, ran 'bldenv -d ./opensolaris.sh'
[15:36:43] <clay> yeah, onnv_50
[15:37:04] <cmihai> Just make sure it's one that actually builds ;-)
[15:37:05] <flyingparchment> stop: Failed to stop /SYS
[15:37:06] <clay> then I cd'ed into usr/src/psm/stand/boot/i386 and ran 'dmake all'
[15:37:09] <flyingparchment> awesome.
[15:37:14] <clay> stop -f /SYS
[15:37:21] <flyingparchment> same error
[15:37:30] <clay> pull power cord
[15:37:33] <flyingparchment> can't :)
[15:37:34] <clay> err
[15:37:38] <clay> pull /power/cord
[15:37:39] <cmihai> clay, you could just build the whole thing following the instructions on opensolaris.org
[15:37:52] <cmihai> flyingparchment, LOM :-)
[15:38:01] <flyingparchment> cmihai: that _is_ the lom
[15:38:12] <clay> flyingparchment: yeah, you need smart power strips to reboot the flaky loms in these things
[15:38:14] <cmihai> lol
[15:38:23] <cmihai> Or a slave.
[15:38:26] <cmihai> Slaves rock.
[15:38:33] <cmihai> Erm, I meant interns.
[15:39:51] <flyingparchment> "Error establishing IPMI Session : Communication error 33 : Invalid Pong"
[15:39:54] <flyingparchment> ?!
[15:39:56] <clay> cmihai: oh. apparently I have to 'dmake all' in usr/src/tools  :-P
[15:40:05] <clay> lol
[15:42:26] <clay> yay
[15:42:31] <clay> multiboot is building
[15:42:44] <clay> s/building/built/
[15:42:53] *** family has quit IRC
[15:42:56] <cmihai> Neatto. Patched it yourself?
[15:43:05] <clay> that's pre-patch
[15:43:05] <cmihai> What was the issue btw
[15:43:06] <clay> now comes the patch
[15:43:35] <clay> it doesn't properly check to make sure memory isn't already allocated before trying to map it
[15:43:40] <clay> this is not a problem under pxegrub
[15:43:58] <clay> but under etherboot on some platforms, and under Linux's kexec() on all platforms I've tried, it causes a panic
[15:44:45] <cmihai> Oh, that kind of sucks.
[15:45:18] <clay> yeah, but at least it's open source
[15:45:31] <clay> there's no way this would get fixed w/ any other commercial os
[15:45:38] <clay> esp. since dboot came to town
[15:45:49] <clay> (even though nobody's actually running that in production)
[15:46:23] <cmihai> Well, there's people who run SXDE / SXCE in production.
[15:46:52] <clay> *blink*
[15:47:14] <cmihai> If you have the skill and don't mind the responsability, it's not all bad.
[15:47:59] <clay> I'm meeting up with some of the folks from Joyent for dinner this week ... they run bleeding edge SXCE, I'll have to ask how that's going
[15:48:19] <bda> I thought Joyent ran real OSOL builds?
[15:48:31] <bda> benr@ spammed a resume req last week, mentioned b64, I think.
[15:48:42] <clay> what is OSOL?
[15:48:58] <cmihai> I think he means ON
[15:49:07] <bda> OpenSolaris. OSOL. ONNV.
[15:49:09] <bda> URFACE.
[15:49:10] <bda> ;)
[15:49:11] <clay> I don't know what any of that emeans :)
[15:49:15] <cmihai> You get SXCE and build the latest ON on top of that.
[15:49:29] <bda> TMFA.
[15:49:46] <clay> so OSOL would potentially be bloodier bleeding edge than SXCE
[15:49:56] <cmihai> SXCE - Solaris Express Community Edition. SXDE - Solaris Express Developer Edition. ON - OS/NET (kernel && stuff).
[15:50:10] <cmihai> Well, you can go Solaris - stable
[15:50:21] <cmihai> Developer is like the tested bleeding edge...
[15:50:24] <cmihai> SXCE is the bleeding edge distro
[15:50:26] <bda> clay: Or not, depending on how their certification works. b64 would be how old now?
[15:50:27] <cmihai> but you can go further
[15:50:41] <clay> bda: *nod*
[15:50:42] <cmihai> As in install the ON snapshots.. it's like using Debian Sid
[15:50:51] <cmihai> With a later kernel from kernel.org
[15:50:58] <flyingparchment> or you can take onnv straight from hg :)
[15:51:05] <cmihai> But you can go even further and use actual Mercurial builds, straight from hg
[15:51:07] <cmihai> :-)
[15:51:09] <WickedWicky> yea
[15:51:14] <WickedWicky> compile the tip branch
[15:51:15] <bda> I am real interested in how Joyent's build and deployment system works. :)
[15:51:25] <clay> bda: me too :)
[15:51:29] <flyingparchment> hmm, i can't reset the SP either
[15:51:30] <cmihai> And you can get the latest Vermillion too
[15:51:31] <clay> I'm building the build system for my company
[15:51:37] <cmihai> That's the latest Gnome/JDS...
[15:51:39] <clay> hence why I need etherboot/kexec() support
[15:51:51] <cmihai> There's also BFU's... oh god, it's a mess ;-P
[15:51:58] <clay> what's BFU?
[15:51:59] <clay> bif fucking unix?
[15:52:02] <clay> err big
[15:52:09] <flyingparchment> bonwick-faulkner upgrade
[15:52:19] <flyingparchment> it's how you install a new ON build onto an existing system
[15:52:20] <cmihai> BFU Archives, fast updates for ON usually...
[15:52:30] <flyingparchment> (if you're not using packages)
[15:52:35] <cmihai> Blindingly Fast Upgrade
[15:52:56] <cmihai> You basically slap on a new OS snapshot (pre-built binaries) via a cpio based tool
[15:53:00] <bda> clay: Tell benr@ to blog about it! :)
[15:53:27] <flyingparchment> is it safe to just 'reboot' after logging into the ILAM as sunservice?
[15:53:37] <bda> cmihai: What, like `cd / ; tar -xzpf $RELEASE/foo.tar` ;P
[15:53:45] <cmihai> No, I said cpio :-)
[15:53:50] <bda> ;)
[15:54:01] <WickedWicky> cpio -id <blerg.cpio
[15:54:02] <cmihai> But that's pretty much the concept...
[15:54:19] <WickedWicky> oh, samba music outside
[15:54:21] <bda> Yea, that's how OpenBSD do.
[15:54:40] <cmihai> Yeah, then you fuck arourd trying to mix and match /etc
[15:55:07] <bda> OpenBSD upgrade docs are thankfully pretty explicit. They also have diffs for files not typically touched.
[15:55:15] <bda> Otherwise you're stuck with mergemaster.
[15:55:28] <bda> Or gods forfend how Gentoo used to "upgrade" /etc. Or Debian "Hey, do you want to..."
[15:55:55] <cmihai> Still, it's pretty much a PITA, and a great way to brickify your OS
[15:56:06] <cmihai> Especially if you don't have LOM or local access...
[15:56:14] <bda> Never once had an issue with OpenBSD system upgrades, honestly.
[15:56:20] <bda> Not in four or five years.
[15:56:21] <flyingparchment> mergemaster ~= acr
[15:56:35] <bda> Package upgrades used to be a huge pain in the butt, but the last two years they've gotten more and more trivial.
[15:56:46] <cmihai> Yeah, -u actually works.
[15:56:55] <bda> Sweet ain't it. :P
[15:57:21] <cmihai> Still, in terms of features OpenBSD is falling behind. I really want to see more advanced things like filesystem snapshots ffs
[15:57:35] <cmihai> I hate taking the system down just to take a... ehem.. proper dump
[15:57:55] <bda> Does FFS2 support that? I haven't been paying much attention lately.
[15:58:01] <WickedWicky> fortunatly I was able to read the context of your last sentence
[15:58:08] <cmihai> Yeah, FreeBSD has snapshots in UFS2... \
[15:58:16] <cmihai> But I don't think OpenBSD is porting that in what they're doing...
[15:58:25] <cmihai> NetBSD doesn't seem to have them in FFSv2 either.
[15:58:35] <bda> ah well.
[16:00:41] <clay> ok (possibly) stupid question
[16:00:45] <clay> how can this possibly work:
[16:00:46] <clay> #define	dprintf	if (debug & D_MBOOT) printf
[16:00:52] <clay> if D_MBOOT is never defined anywhere?
[16:01:10] <clay> (err, insert spaces where the tabs were)
[16:03:05] *** jfndi has joined #opensolaris
[16:03:36] <clay> oh, well, it's never used anywhere, so nevermind
[16:04:00] *** perlmonk has quit IRC
[16:05:50] <bda> The box where I get all my alerts is so aptly named: "vex".
[16:28:27] *** kaiwai has joined #opensolaris
[16:30:04] <flyingparchment> hmm, vxfs seems to be a fair bit slower than QFS for mysql workload
[16:33:04] *** bengtf has quit IRC
[16:41:30] *** supernerd has joined #opensolaris
[16:42:00] *** nostoi has quit IRC
[16:50:53] *** Doc has quit IRC
[16:51:27] *** Doc has joined #opensolaris
[16:51:50] *** clay_ has joined #opensolaris
[16:53:12] *** bengtf has joined #opensolaris
[17:00:17] <CIA-25> kucharsk: 6604295 BrandZ installer log messages may not be as clear as needed in error legs
[17:06:16] *** jcsmith has quit IRC
[17:07:38] *** clay has quit IRC
[17:12:42] *** Mdx4 has quit IRC
[17:13:03] *** Sup3rkiddo has joined #opensolaris
[17:14:46] *** mega has quit IRC
[17:15:46] *** binarycrusader has joined #opensolaris
[17:28:45] <kaiwai> hmm
[17:39:21] <WickedWicky> hmmmmmmmmmmmmmmmm
[17:41:14] <kaiwai> things that make you go hmmmmmm
[17:41:34] <kaiwai> hopefully that kernel panic issue with the ata driver will be fixed soon
[17:42:30] <kaiwai> on a good side, Solaris not supports frequency scaling on my lpatop
[17:45:12] <kaiwai> apparently there is a move to make GIMP suck less
[17:46:20] <binarycrusader> panic issue with the ata driver?
[17:46:26] <binarycrusader> that appears to be fixed in b72, at least it is for me
[17:46:59] <kaiwai> binarycrusader: na, B72/B73 breaks detecting my cd/dvd drive
[17:47:19] <kaiwai> I have been given a driver by a developer at sun, however, the whole computer dies when it is used - doesn't even boot
[17:47:28] <binarycrusader> doh
[17:47:39] <binarycrusader> well, b72 fixed my panics and created yours
[17:47:43] <binarycrusader> I suppose you win some, lose some
[17:47:52] <kaiwai> yeah, kind of annoying, lord knows why they changed the cd/dvd detection code
[17:47:56] <WickedWicky> the ace of spades \m/
[17:48:11] <kaiwai> WickedWicky: dear god, still listening to that stuff
[17:48:25] <WickedWicky> Motorhead rules
[17:48:36] <emergo> how do I restart apache2 -> svcadmin ... apache2 restart ?
[17:48:46] <WickedWicky> but it's radiohead at the moment, binarycrusader's "win some, lose some" reminded me
[17:48:51] <WickedWicky> svcadm restart apache2
[17:49:00] <emergo> WickedWicky: thank you
[17:49:02] <kaiwai> WickedWicky: are you sure that is motorhead?
[17:49:06] <WickedWicky> yes
[17:49:22] <kaiwai> hmm, I thought it was another band
[17:49:32] <emergo> WickedWicky: how can I see if its restarted ?
[17:49:43] <emergo> svcadm ... somehting
[17:49:50] <WickedWicky> http://www.darklyrics.com/lyrics/motorhead/aceofspades.html
[17:49:53] <emergo> dont remember to much linux in the head
[17:50:00] <binarycrusader> emergo: less /var/svc/log/network-http:apache2.log
[17:50:08] <WickedWicky> If you like to gamble, I tell you I'm your man,
[17:50:10] <WickedWicky> You win some, lose some, all the same to me,
[17:50:25] <kaiwai> ah, got confused with another band
[17:50:56] <kaiwai> his
[17:51:03] <kaiwai> his 'power stache" :P
[17:51:12] <emergo> binarycrusader: thank you but I used to use svadm -a or something to list all services that failed etc
[17:51:21] <binarycrusader> emergo: well, that would be different
[17:51:24] <emergo> anyway it works fine now
[17:51:27] <binarycrusader> emergo: svcs -xv apache2
[17:52:14] <emergo> binarycrusader: ty, thats the one. it says how long its been running there
[18:06:10] *** Sup3rkiddo is now known as Lap_64
[18:06:30] *** Lap_64 is now known as Sup3rkiddo
[18:19:37] <WickedWicky> cara
[18:19:40] <WickedWicky> viu isso?
[18:19:41] <WickedWicky> http://apple.slashdot.org/article.pl?sid=07/09/14/1831236
[18:19:56] <binarycrusader> *shrug*
[18:20:03] <binarycrusader> Tempest in a teacup
[18:20:20] <binarycrusader> Headline should read: "Developers relied on private Apple API, it changed, and now they need to update"
[18:20:53] <WickedWicky> oh ops
[18:20:57] <WickedWicky> that wasnt even for here
[18:21:02] <WickedWicky> sorry for the poruguese
[18:21:31] <WickedWicky> portuguese even
[18:22:15] <kaiwai> hmm, I find that rather funny
[18:22:24] <kaiwai> thank goodness I don't own an ipod
[18:23:23] *** dlynes_ has quit IRC
[18:28:20] *** nostoi has joined #opensolaris
[18:41:56] *** vmlemon has quit IRC
[18:42:56] *** vmlemon has joined #opensolaris
[18:44:18] *** jamesd_ has quit IRC
[18:49:07] *** helpmehome has quit IRC
[18:51:36] *** binarycrusader has left #opensolaris
[18:52:33] *** Triskelios has joined #opensolaris
[18:55:39] <madhatter> quasi: Is there some good manual for pmpkg you know of? I read the usage, but it pmpkg seems to stop after config and tells me about install. But I can't get it going.
[18:56:07] <madhatter> Oh, seems he gone fishin' ;)
[18:59:15] *** jamesd has quit IRC
[19:00:13] <CIA-25> willard: 6603942 fmtopo -v should show NAC names pciex slots instead of other strings
[19:00:28] <madhatter> Anybody else familiar with pmpkg?
[19:01:02] <kaiwai> nope, sorry, never used it before
[19:01:58] *** axisys has joined #opensolaris
[19:07:17] *** jamesd_ has joined #opensolaris
[19:08:00] *** jamesd has joined #opensolaris
[19:11:42] *** ashner has joined #opensolaris
[19:11:51] *** jfndi has quit IRC
[19:13:10] <madhatter> I got the source pulled and it looked like configure ran, but now I am stuck.
[19:14:40] *** Murmuria has joined #opensolaris
[19:15:11] *** Murmuria has left #opensolaris
[19:15:12] *** axisys has joined #opensolaris
[19:19:18] *** boyd has joined #opensolaris
[19:20:25] <quasi> madhatter: ehrm, I'm pretty sure it is just a question of running the command - lemme check
[19:21:18] <madhatter> quasi: I get two messages about undefined variables also. But I have no idea which variables that might be. I now set the gcc_path already
[19:22:06] *** Fish- has joined #opensolaris
[19:22:10] *** shpchp has quit IRC
[19:23:10] *** mlh has quit IRC
[19:23:32] <quasi> ehrm, trying to remember - I haven't booted my desktop box today
[19:23:33] *** binarycrusader has joined #opensolaris
[19:24:49] <Pietro_S> madhatter: it uses shell script to build software/package ..., I suggest you to take look on them and maybe try to invoke it manually ...
[19:24:56] <Pietro_S> or step by step
[19:26:35] <madhatter> Pietro_S: But which? I have files, install and work directories. But found no scripts around
[19:28:11] *** polk__ has joined #opensolaris
[19:28:12] <quasi> madhatter: see the USAGE file
[19:28:19] <madhatter> quasi: I did
[19:28:25] <madhatter> quasi: That does not work
[19:28:30] <quasi> madhatter: share/pmpkg
[19:28:50] <madhatter> quasi: Did that. It stops after the configure process
[19:29:23] <quasi> madhatter: add /path/to/PMp/share to your path, cd into one of the pkg dirs and run pmpkg install
[19:29:28] <quasi> madhatter: works for me
[19:30:10] <quasi> but as Pietro_S suggests - take a look at the pmpkg script
[19:30:14] *** nostoi has quit IRC
[19:31:09] *** SirFunk has quit IRC
[19:32:33] <madhatter> quasi: When doing that it prints out the PATH and LD_LIBRARY both ending with 'Command not found'
[19:32:45] *** boyd_ has quit IRC
[19:33:07] <quasi> madhatter: come to think of it, I think I always called it directly
[19:33:20] *** estibi_ has joined #opensolaris
[19:34:11] <madhatter> quasi: What do you mean?
[19:35:14] <quasi> cd into one of the pkg dirs and ../../share/pmpkg
[19:35:59] <madhatter> quasi: That's what I did before. Then it stops after a while
[19:36:16] *** mlh has joined #opensolaris
[19:37:20] <Pietro_S> madhatter: what shell are you using?
[19:37:26] <madhatter> Pietro_S: tcsh
[19:37:38] <Pietro_S> madhatter: ty to run it under bash or ksh
[19:37:48] <Pietro_S> s/ty/try
[19:38:38] <madhatter> Pietro_S: Would it be enough to start bash and then try again?
[19:38:51] *** Fish has quit IRC
[19:38:53] <Pietro_S> yes, it should be fine
[19:39:13] <madhatter> Pietro_S: Does the same. Prints out the both variables
[19:40:20] <Pietro_S> did you install pmpkg properly?
[19:40:25] *** yarihm has joined #OpenSolaris
[19:40:41] * Pietro_S summons oxygene
[19:41:00] <madhatter> Pietro_S: What is there to install? I moved it to /opt and put /opt/PMp/share in my path
[19:46:20] *** estibi has quit IRC
[19:46:20] <madhatter> It seems to be a problem calling make or something. I tried scummvm now, and it downloads the tarball, configures and then prints again my PATH and LD_LIBRARY_PATH variables
[19:47:38] *** estibi_ is now known as estibi
[19:49:35] *** SirFunk has joined #opensolaris
[19:50:24] <madhatter> Seems that gmake and gcc cant be assigned to the variables in the script
[19:50:37] <quasi> sounds about right
[19:51:54] <madhatter> Put the direct paths in it now and it seems to work
[19:54:51] *** Mdx4 has joined #opensolaris
[19:56:57] *** dnilsson has joined #opensolaris
[20:00:00] *** kaiwai has quit IRC
[20:01:17] <madhatter> So, builds done, now the next error on install :/
[20:04:18] <madhatter> Okay, I am giving up. Can't be harder to install from scratch ;)
[20:04:56] <madhatter> What a successful day
[20:09:09] *** Shiv__ has joined #opensolaris
[20:19:43] *** jamesd__ has joined #opensolaris
[20:21:32] *** Shiv__ has left #opensolaris
[20:21:37] *** jamesd_ has quit IRC
[20:22:11] *** jamesd has quit IRC
[20:30:49] *** Jondice has quit IRC
[20:47:18] *** Shiv__ has joined #opensolaris
[20:49:32] <Pietro_S> madhatter: what sf are ou trying to build/install?
[20:59:10] *** Vorbis_w48 has quit IRC
[21:02:29] *** GoodKarma has joined #OpenSolaris
[21:13:24] *** Vorbis_w48 has joined #opensolaris
[21:15:51] *** BadKarma has quit IRC
[21:15:52] *** GoodKarma is now known as BadKarma
[21:36:20] *** infidel has quit IRC
[21:36:38] *** Mdx4 has quit IRC
[21:41:03] *** Shiv has joined #opensolaris
[21:41:04] *** Shiv__ has quit IRC
[21:42:50] *** Shiv has quit IRC
[21:53:24] *** kadath has quit IRC
[21:53:51] *** kadath has joined #opensolaris
[21:54:55] <madhatter> Pietro_S: Enlightenment
[21:54:57] *** iustin23 has joined #opensolaris
[22:02:23] <iustin23> My idea is to develop a virtual machine ( MS Windows) on a nice good server and this to serve via VNC some other pc`s
[22:02:24] <iustin23> can a solaris container support such a virtual machine ?
[22:04:15] <delewis> if you mean can you run Windows in a zone, the answer is no.
[22:05:16] <binarycrusader> solaris containers currently only support GNU/Linux or Solaris
[22:05:50] <johnlev> iustin23: a solution involving xen is what you want.
[22:05:59] <richlowe> johnlev: you would say that.
[22:06:06] <johnlev> naturally.
[22:06:08] <delewis> iustin23: don't think of zones/containers as virtual machines.
[22:06:20] <richlowe> johnlev: is that your answer to most questions, at this point? :)
[22:06:26] <johnlev> naturally.
[22:06:29] <richlowe> "Lunch?  A solution involving xen..."
[22:06:42] <johnlev> the last few lunches have been pretty much like that yes
[22:09:20] *** thalin has quit IRC
[22:09:22] *** thalin has joined #opensolaris
[22:14:56] <g4lt-sb100> virtually lunch?
[22:15:05] <WickedWicky> I thought the solution for everything was beer or vodka
[22:15:21] <g4lt-sb100> you forgot rum and tequila
[22:15:24] <WickedWicky> oh yeah
[22:15:43] <g4lt-sb100> which reminds me, so did I, last time I bought liquor
[22:15:55] * g4lt-sb100 needs to fox that
[22:17:42] *** henriknj has joined #opensolaris
[22:18:13] <Pietro_S> madhatter: well, the copyfiles.sh (in files dir) looks strange, I guess that you have a bit different platform than the creator of package, sparc?
[22:20:06] *** iustin23 has quit IRC
[22:20:59] <Pietro_S> wait, sry I was wrong - the files contain only some png and txt, which should be independent on B/L endieness, looks like some dependency is missing on your workstation
[22:34:58] *** iustin23 has joined #opensolaris
[22:40:04] *** iustin23 has quit IRC
[22:40:42] <madhatter> Pietro_S: I don't have autoconf installed here
[22:41:11] <madhatter> Pietro_S: automake, sorry
[22:47:18] *** mikefut has quit IRC
[22:51:08] *** nrubsig has joined #opensolaris
[22:51:08] *** ChanServ sets mode: +o nrubsig
[22:52:29] <nrubsig> Anyone familar with twitter ?
[22:54:05] * g4lt-sb100 shudders
[22:54:55] <nrubsig> g4lt-sb100: http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/shtwitter.ksh
[22:55:16] <g4lt-sb100> nrubsig, you are now thrice damned to hell
[22:55:46] <richlowe> dude, you did that but didn't call it 'shitter?'
[22:56:59] <nrubsig> richlowe: is the script that bad ? ;-(
[22:57:30] <binarycrusader> no, the script is fine, i believe the enemity is being reserved for twitter itself
[22:57:57] <richlowe> I just can't imagine resisting a crude pun when the opportunity arises.
[23:00:40] *** Jondice has joined #opensolaris
[23:08:15] <g4lt-sb100> nrubsig, twitter is
[23:08:43] <g4lt-sb100> binarycrusader ++
[23:10:47] <nrubsig> bah
[23:10:50] *** Gman has joined #opensolaris
[23:11:09] <nrubsig> I have the killer application for twitter: Internet version of /var/adm/messages !
[23:11:37] <binarycrusader> I hope you are being sarcastic :)
[23:11:42] <jbk> hey gman
[23:12:14] <Gman> hi jbk
[23:13:52] <richlowe> nrubsig: that's what twitter is, pretty much.
[23:14:01] <richlowe> a constant stream of badly formed unparsable, mostly uninteresting messages.
[23:14:04] <richlowe> the parallels are obvious.
[23:17:03] <cmihai> "Friends of OpenSolaris"...
[23:17:07] <cmihai> you've got to be fucking kidding me
[23:17:23] <binarycrusader> cmihai: positive attitude cmihai
[23:20:12] <binarycrusader> irony, an article about the myths of opensolaris makes an interesting statement: "OpenSolaris users still have to remap the backspace key because it?s not supported in the operating system by default."
[23:20:17] <binarycrusader> (http://www.networkworld.com/news/2007/091207-sun-opensolaris.html)
[23:20:29] <binarycrusader> I just love tech reporting...
[23:22:01] <richlowe> if you remove the comma between "myths" and "bugs", suddenly their examples make sense.
[23:22:10] <nrubsig> http://twitter.com/syslog_sxb72x86
[23:22:36] <binarycrusader> nrubsig: seriously, that's not a very secure thing to do
[23:22:57] <nrubsig> I know.
[23:22:58] <jbk> Gman: you probably know this off the top of your head -- there's an annoying bug w/ the gnome clock applet w/ the timezone button -- it seems to randomly disappear -- where would one log a bug on that?
[23:23:03] <richlowe> though you could make it only publish a specific facilitie and level, then use logger as an interface.
[23:23:29] <jbk> or is it a known issue already?
[23:23:43] <richlowe> jbk: very known.
[23:23:53] <richlowe> knowing JDS, they'll tell you it's fixed.
[23:24:10] <binarycrusader> Hrm...speaking of which, the timezones popup shows <span> tags, neat
[23:24:24] <binarycrusader> Chicago <span gravity="south"> 4:24 PM </span>
[23:24:26] <Gman> jbk, just on the phone, one sec
[23:24:44] <richlowe> actually, it's still 1-Dispatched
[23:24:54] <jbk> ahh so there is already a bug assigned then
[23:24:54] <richlowe> jbk: 6499300
[23:24:57] *** Mdx4 has joined #opensolaris
[23:25:01] <richlowe> jbk: several duplicates of that one.
[23:25:20] <richlowe> though there's a more efficient workaround.
[23:25:25] <richlowe> drag the panel to another side of the screen and back
[23:26:11] <nrubsig> http://twitter.com/syslog_sxb72x86
[23:29:04] *** Sieghard has quit IRC
[23:30:03] *** derchris has joined #opensolaris
[23:32:03] *** bunker has quit IRC
[23:43:56] *** dme has quit IRC
[23:45:57] *** estibi has quit IRC
[23:55:20] *** Marv|LG has left #opensolaris
[23:58:03] *** Mdx4 has quit IRC
[23:58:58] *** linux_user400354 has joined #opensolaris
[23:59:03] *** Chihan has quit IRC





top