[00:10:36] *** Lebuzzer has joined #haiku
[00:13:01] *** liquidboy has joined #haiku
[00:13:39] <@mmu_man> -rw-r--r--   1 web site     1961 Jul 25  2000 index.html
[00:13:45] <@mmu_man> 5 <@mmu_man>	5 years here you beat me :)
[00:13:55] <AnEvilYak> 1961? o.0
[00:14:19] <Konrad> mmu_man so Iam waiting in vain for you to update your window98 theme then?
[00:14:45] <AnEvilYak> windows 98 theme?
[00:15:19] <@mmu_man> lol
[00:16:01] <Konrad> mmu_man give rene the link
[00:17:27] <Konrad> AnEvilYak http://clapcrest.free.fr/revol/theme.htm
[00:18:04] * AnEvilYak looks at mmu_man
[00:18:05] *** liquidboy has quit IRC
[00:18:26] *** mmu_man has quit IRC
[00:18:32] <Konrad> Hehe
[00:18:39] <Konrad> I would also hide
[00:18:40] <AnEvilYak> apparently I'm Medusa...
[00:19:06] *** Fanskapet has quit IRC
[00:19:13] *** mmu_man has joined #haiku
[00:19:13] *** ChanServ sets mode: +o mmu_man
[00:20:10] *** Korli has joined #haiku
[00:20:11] *** ChanServ sets mode: +o Korli
[00:24:41] <Begasus> testing fase? :P
[00:25:04] <Begasus> hoops :P
[00:27:36] <Dr_Evil> just choppping out the video and audio packets of one program from trom mpeg2 ts and saveing them to file creates nothing that is mpeg2 compatible
[00:27:48] <Dr_Evil> :-(
[00:28:25] * tqh don't know the inner workings of the transport stream
[00:30:30] *** mmu_man has quit IRC
[00:31:08] *** mmu_man has joined #haiku
[00:31:09] *** fyysik has joined #haiku
[00:31:09] *** ChanServ sets mode: +o mmu_man
[00:31:15] <fyysik> hi again
[00:32:28] <fyysik> mmadia?
[00:32:37] <mmadia> hi fyysik
[00:32:58] <fyysik> did you complain about locking of last mozilla 1.7a ?
[00:33:03] <fyysik> or was that tigerdog?
[00:33:24] <mmadia> not me.  i haven't been able to reliably reproduce it.
[00:33:40] <fyysik> someone with dial processor machine
[00:33:44] <fyysik> dual
[00:33:50] <mmadia> i've also a dual.
[00:34:47] <mmadia> fyysik i've gotta run out of a bit. leave atrus a message if you need or email me
[00:34:56] *** mmu has joined #haiku
[00:35:14] *** mmadia is now known as mmadia_out
[00:35:19] *** mmu_man has quit IRC
[00:35:48] <Dr_Evil> need too look into the mpeg2 specifications on monday
[00:40:00] <tqh> Dr_Evil There are two interesting documents at the end of this page: http://www.cdaniel.de/elektronik-dvb.phtml
[00:41:51] <tqh> Also replex might fit your needs: http://www.metzlerbros.org/dvb/
[00:43:06] <Dr_Evil> hmm, remuxing PID 0, 1, 16, 17, 18, 20, 21, 336, 337, 338 into one stream gives mit RTL television :)
[00:43:30] <Dr_Evil> 337, 338 is RTL audio/video, but the others seem to be needed, too, especially 336
[00:43:48] <slaad> RTL?
[00:44:05] <Dr_Evil> tqh thanks
[00:44:25] <Dr_Evil> slaad german T
[00:44:27] <Dr_Evil> TV
[00:44:49] <slaad> Ah.
[00:45:10] <fyysik> is at this channel also RTL-Shop availabable, with famous TV-start B.T.K. ?:)
[00:45:20] <fyysik> TV-star
[00:45:42] <fyysik> or that is different special shopping channel?
[00:46:08] <Dr_Evil> different, but same company
[00:46:43] *** oco has quit IRC
[00:47:28] <tqh> Anyone here using a AMD64 computer with BeOS?
[00:48:02] * tqh is thinking about trying to speed up his build-times.
[00:48:20] <@BGA> tqh: I am.
[00:48:25] <@BGA> A>	A laptop, actually.
[00:48:38] <tic> tqh, get a Pentium M system instead, and you can have a really silent box!
[00:48:47] <@geist> run linux
[00:49:10] <slaad> Hehe.
[00:49:26] <tqh> tic The winchester AMD64 don't use more than 35W and only 3W idle so it's much cheaper to buy those.
[00:49:43] <tic> tqh, using Cool'n'Quiet, yes.  But without that, i.e., software-support, it's far worse.
[00:49:59] * tqh was wondering if nforce3 will work
[00:50:16] <tqh> well pentium m ain't an option for me.
[00:50:20] <tic> why?
[00:50:27] <tqh> expensive
[00:50:40] <tic> but not much!  1000 kronor more expensive.
[00:50:42] <tic> (roughly)
[00:51:17] <tqh> well, mainboard is also expensive
[00:51:36] <tic> No, I meant mainboard + CPU is 1000 kr more expensive than an equivalent AMD64 plus mobo
[00:52:36] <@geist> you should get a mac
[00:52:39] * geist ducks
[00:52:48] * fyysik noticed that mozilla starts in beos much faster than under windoze. probably due missing antivirus programs
[00:52:48] <tqh> well I rather use that money on performance.
[00:52:55] <tic> o/~ RISC chip too, now (I want a mac!). Smoot transition o/~
[00:53:01] <tic> quack!
[00:53:09] * tqh would get a mac laptop if he could build mozilla and run BeOS on it.
[00:53:15] <tic> fyysik, oh? I tend to see the opposite
[00:54:14] <fyysik> tic - numbers. ~1.5 sec in BeOS and 12 sec in Windows at 2GHz Celeron.     3.5 sec in BeOS and 25 under windows at PIII-550
[00:54:24] <tic> fyysik, yow.
[00:54:41] *** mmu is now known as mmu_man
[00:54:44] <fyysik> thoug, in windoze you have preload option for Mozilla
[00:54:54] *** ChanServ sets mode: +o mmu_man
[00:54:57] <fyysik> if you use it, start time is short
[00:54:59] <tic> 119W max power consumption for the Athlon64 at 0 dot 09
[00:55:01] <tic> yeah.
[00:55:14] <tic> fyysik, heh, what do you think the start time in BeOS would be if you pre-loaded moz? 1 sec?
[00:55:18] <tqh> Naha, look here: http://www6.tomshardware.com/cpu/20041115/pentium4_570-20.html
[00:55:29] <fyysik> tic - immediate:)
[00:55:35] <@geist> negative time
[00:55:41] <@geist> it would start before you clicked on it
[00:55:52] <fyysik> hehe, prefetching user action feature:)
[00:56:05] <tic> ah, I was looking at a overclocked one: http://www.hothardware.com/viewarticle.cfm?articleid=613
[00:56:10] <fyysik> idea fro Boll Gates
[00:56:17] <fyysik> for Bill Gates
[00:56:18] <@geist> I was thinking it would just corrupt your bfs partition immediately
[00:56:28] <@geist> prefetching the corruption that will eventually happen
[00:56:35] <tic> Heh.
[00:57:28] <tic> tqh, lemme dig up another article..
[00:57:49] <tic> http://www.gamepc.com/labs/view_content.asp?id=dothangaming&page=4
[00:58:25] <tic> thing is, the Dothan runs a lot cooler.
[00:58:35] <tic> which is a good thing, no fans needed if you carefully plan your cooling.
[00:58:51] <tqh> tic, but that article you posted first is total systam wattage. 'we used a Seasonic Power Angel power meter to monitor total system wattage.'
[00:59:04] <tic> tqh, oh crud. :) hehe. sorry!
[00:59:11] <tic> tqh, okay, so look at the gamepc one.
[00:59:33] <fyysik> tqh
[00:59:44] <tqh> fyysik yes
[00:59:55] <fyysik> tqh, did you notice that there are two options in configure:
[01:00:11] <fyysik> ac_add_option --enable-nspr-threads
[01:00:11] <fyysik> and
[01:00:12] <fyysik> ac_add_option --with_bthreads
[01:00:32] <fyysik> wondering what is actual default and if any of that has some effect
[01:00:56] <fyysik> in last build i made today i put those explicitly for sake in mozconfig
[01:01:01] <tqh> fyysik I have no clue. Sound interesting though.
[01:01:39] <fyysik> tqh - can you look into your config.status and config.cache, if there any traces of those options there?
[01:02:16] <tqh> sure
[01:03:39] <fyysik> ik>	i think that i got with that more threads visible in process-controller and more responsive behaviour but it my be illusion
[01:03:56] <tqh> fyysik searchin on thread didn't give anything interesting.
[01:04:39] <tqh> tic didn't find anything about power consumption...
[01:05:07] <fyysik> tqh - actually i think that maybe even only nspr recompilation is required fi you reconfigure it with those options
[01:05:32] <tic> tqh, okay..
[01:05:38] *** emwe has quit IRC
[01:05:59] *** AnEvilYak has quit IRC
[01:06:25] <tqh> and it doesn't look like they had 'winchester' AMD 64's
[01:07:36] <tic> very well.  Still, the difference in power usage is important for me, as I plan to make a completely passively cooled box. :)
[01:08:14] <tic> eek, the A64s are expensive!
[01:08:50] <tqh> well, they are much cheaper than any petium m's I've seen
[01:08:56] <tqh> pentium
[01:09:04] <tic> 1.7 GHz costs 2050 at shg.biz
[01:09:22] <tic> eqiv to approx athlon64-3200?
[01:09:48] *** AnEvilYak has joined #haiku
[01:11:06] <Konrad> good night
[01:11:06] *** Konrad has quit IRC
[01:11:27] <Dr_Evil> night
[01:11:39] <tqh> tic and that one costs 1895.00 at datorbutiken.se
[01:12:32] <tic> tqh, yeah, but you get a) a less power-hungry CPU, and b) a CPU where you can use Intel parts.  I don't really trust VIA and SiS for BeOS compatiibility
[01:13:37] <tqh> no I want nforce3
[01:13:47] <tic> is it beos compatible?
[01:14:20] <tqh> dunno, that's why I asked, but dual channel to memory is goood for building.
[01:14:24] <tic> I'd prefer an aopen w/ the i855 chipset plus dual marvell (gbit/100/10) LAN, 1394, VGA & audio.  But to each his own. :)
[01:14:39] <tic> dunno really.. the Pentium M is as fast as any other system at games and such.
[01:14:43] <tic> and that probably pushes even more memory.
[01:15:17] <tqh> Eew, I'd never use Intels graphics. 'Extreme graphics' is very unextreme.
[01:16:07] <tic> Me neither, I'll actually be using my LeadTek FX5700LE (dual-head, 2x19" TFTs.. ;) in the AGP slot.
[01:16:30] <tic> but it's nice to have it, just in case.  And, having audio, 1394 and LAN in addition to USB-2.0 is very nice.
[01:17:00] <tqh> yes, btw is pentium m 64 bit?
[01:17:13] <tic> Nope.
[01:17:18] <tic> it's a regular 32 bit CPU
[01:17:31] <tic> and I frankly don't understand the argument for using a 64-bit CPU
[01:17:43] <tic> it's not like you're really using up 3-4 GB RAM anyway :)
[01:18:25] <tqh> well, it's not just for adressing memory, it's also used in processing :)
[01:18:41] <tic> I guess you could do some mad memcpy()s with that.
[01:18:57] <tic> but hey, you might just as well use the FPU for pushing data on an Intel CPU. or the SSE2 registers.
[01:19:10] *** CIA-6 has quit IRC
[01:19:34] <AnEvilYak> tic: 64-bit wide registers are nice.
[01:19:42] <AnEvilYak> tic: as are the extra registers on x86-64 in general.
[01:20:08] <tqh> yes, there is much fun stuff in the 64 bit processors.
[01:20:09] <tic> AnEvilYak, but most people don't code on bare metal these days, they let a thing called the "compiler" mess with register allocation. ;)
[01:20:40] <AnEvilYak> tic: but the compiler won't deal with 64-bit register allocation if it's not allowed to operate in 64-bit mode ;p
[01:21:25] <tic> AnEvilYak, of course it won't  :)  But I mean, what's the real-life benefit? Can't be that much faster.
[01:21:43] <tqh> You can run 64 bit Linux :)
[01:22:24] <AnEvilYak> tic: well, consider for instance anytime you do math with an off_t ...
[01:23:47] *** TLF has quit IRC
[01:24:44] <tic> AnEvilYak, I don't! :P  but yeah, I see your point.  but other than that, what -is- the point?
[01:24:59] <AnEvilYak> tic: also simplifies VM management a lot afaik
[01:25:10] <AnEvilYak> tic: no need for 3GB/1GB hacks for the memory space, etc.
[01:26:29] <tic> AnEvilYak, *nod*
[01:34:56] <@mmu_man> well current 64bit OSes use 32bit vm for processes AFAIK
[01:35:11] <@geist> depends
[01:35:22] <@geist> some architectures only run in 64-bit
[01:35:27] <@geist> some allow two user modes
[01:35:41] <@geist> in all cases the kernel is always running in 64bit mode, which is a pretty big win
[01:35:47] <@mmu_man> yup
[01:35:54] <@geist> since the kernel has a lot of space to play with memory caches and whatnot in it's address space
[01:36:11] <@geist> you dont have to fool with memory views and whatnot
[01:36:21] <@mmu_man> well 64bit doesn't mean 2^64 bytes of RAM :)
[01:36:35] <@mmu_man> but it's easier to map things in yeah
[01:37:01] <@geist> but newere machines are hitting the wall pretty hard right now
[01:37:13] <@mmu_man> less things like a physical_ram area in a 2GB kernel map in the vm
[01:37:14] <@geist> even having 1GB of ram is already introducing some complexity
[01:37:20] <@geist> since most kernel spaces aren't that big
[01:37:36] <@geist> right, linux for example has a 800MB memory hole that maps the first 800MB of ram
[01:37:48] <@mmu_man> it still maps the ram ?
[01:37:49] <@geist> more memory than that requires a on demand memory mapper
[01:37:53] <@mmu_man> thought it was "fixed"
[01:37:59] <@geist> which is why linux vm has the concept of "low" and "high" memory
[01:38:05] <@mmu_man> oh god
[01:38:27] <@mmu_man> good old conventional/ext mem like things
[01:38:29] <@geist> on a 64bit machine you can probably just map it all
[01:38:41] <@geist> well it's a really big win to have it always mapped
[01:38:50] <@geist> simplifies a lot of stuff
[01:39:04] <@mmu_man> yes it's much easier to do things like virt to phys calculation
[01:39:16] <@geist> much more efficient, but of course you then bump into a memory limit (a la beos) unless you have an alternatie mechanism to get to higher ram
[01:39:37] <@mmu_man> still I thought they fixed it cleanly wqhen the did that in linux
[01:39:46] <@geist> for example, linux doesn't have a "kernel heap"
[01:39:55] <@geist> they simply run a slab allocator on top of the phys mem region
[01:40:02] <@geist> and the slab allocator allocates physical pages
[01:40:21] <@mmu_man> hmm yeah
[01:40:23] <@geist> at first I was pretty grossed out by that, but it's actually pretty elegant
[01:40:32] <@mmu_man> so it can use the >800Megs for kernel
[01:40:36] <AnEvilYak> what differentiates a slab allocator anyhow?
[01:41:21] <@geist> well, on a x86 it sets the max limit to 800MB, but it would be that way anyway if it had a traditional heap region
[01:41:53] <@geist> on 64-bit arches that have more address space you have no limit to how much ram the kernel can use for internal data structures
[01:42:05] * mmu_man recalls lilo args like linux mem=64M
[01:42:13] *** BGA has quit IRC
[01:42:14] <@geist> and virt to phys and phys to virt calcs are trivial
[01:42:15] <@mmu_man> was handy to work with a broken 128M stick
[01:42:20] * tic listens to French music
[01:42:45] <@geist> the newos/haiku vm just on demand maps *all* pages as they're needed and has a traditional heap
[01:42:57] *** dipp has quit IRC
[01:43:02] <@geist> which means it'll always work to get to whatever ram you have, even more than 4GB on a x86 if you want
[01:43:10] <@geist> at the expense of speed and complexity
[01:43:32] * mmu_man pets PAE
[01:43:42] <@mmu_man> with NT you need to use a different kernel
[01:43:56] <@mmu_man> and another one if you want a 3/1GB split
[01:43:59] <AnEvilYak> that's an optimization I thought.
[01:44:00] <@geist> heh, speaking of. I'm deep in the middle of arm mmu code right now
[01:44:01] <AnEvilYak> mmu_man: mm, no
[01:44:03] <tic> geist, so the NewOS VM is a slow one, then?
[01:44:07] <AnEvilYak> mmu_man: /3gb is a switch for regular kernel
[01:44:14] <@geist> well, it takes a hit there for correctness
[01:44:17] <AnEvilYak> mmu_man: just have to add it to the boot line in boot.ini
[01:44:26] <@mmu_man> hmm I think the switch is used by the bootloader to use a diffferent kernel
[01:44:36] <@mmu_man> well I do'nt use NT anyway
[01:44:40] <AnEvilYak> mmu_man: nope, only ntoskrnl, ntkrnlmp, and ntkrnlpa
[01:44:54] <@mmu_man> oh right
[01:44:57] <@geist> and on a amd64 it always runs with PAE on
[01:44:58] <AnEvilYak> /3gb is a parameter like /sos
[01:45:01] <@mmu_man> I recalled there were 3 of them
[01:45:07] <tic> yum, 8.3
[01:45:08] <@mmu_man> sos lol
[01:45:11] <@geist> since that's the only way to get to the x bit
[01:45:20] <AnEvilYak> /sos == verbose boot
[01:45:29] <tic> so what does the future hold for the NewOS/Haiku VM?
[01:45:32] <@mmu_man> and /lol ?
[01:45:35] <AnEvilYak> haha.
[01:45:42] <AnEvilYak> /sol would be funny ;p
[01:45:43] <@mmu_man> to get meaningless jokes in BSOD ?
[01:45:43] <@geist> tic: dunno, it's probably not coming from me
[01:45:50] <@geist> axeld is taking it pretty far now
[01:45:57] <AnEvilYak> I thought mphipps was doing the new vm
[01:46:00] <AnEvilYak> but I might be wrong.
[01:46:01] <@geist> I approve of his work, he's following my design
[01:46:06] <@geist> hah, yeah that's a disaster
[01:46:24] <tic> mphipps doesn't really have any time to do anything right now, or at least that's my impression.
[01:46:25] <@geist> ist>	i was quite unimpressed with the mphipps vm design effort
[01:46:36] <@geist> it was a class case of "I dont understand this so I'll do my own version"
[01:46:51] <@geist> the reason no one understood it was they didn't know how vms work
[01:47:07] <@mmu_man> now I still can't cope with C++ in the kernel
[01:47:17] <@geist> axeld however is pretty well skooled in the ways of modern os design
[01:47:18] <@mmu_man> I get big red buttons
[01:47:22] <@geist> he understood my vision on it
[01:47:25] <tic> lately axelds submit have been a bunch of user-space stuff.. Why isn't he doing kernel work?! :P
[01:47:41] <@geist> he checked in a bunch of file cache stuff the other day
[01:47:46] <@geist> which is something I hadn't even started on
[01:47:57] <@geist> though I had a general idea for it, and the vm was designed around it
[01:47:59] <tic> ah.
[01:48:00] <AnEvilYak> geist: I thought caching was kind of implicit in the VM design?
[01:48:30] <@geist> if I started over, I might change the vm to be physically scanned instead of virtually scanned
[01:48:31] <@mmu_man> I like C++ in userland, but I really don't like not having control of the flow in kernel land
[01:48:35] <@geist> the virtual scanning is very complex
[01:48:39] <@mmu_man> just feels bad
[01:48:43] <@geist> only NT uses it as far as I can tell
[01:49:05] <AnEvilYak> geist: can you clarify what you mean virtual vs physical scanning?
[01:49:20] *** liquidboy has joined #haiku
[01:49:27] * tic needs a detractable monitor so he can sit in bed and read. :)
[01:49:40] <@geist> for page replacement you need to "scan" through either virtual memory or physical memory to find pages to be paged out
[01:49:46] <AnEvilYak> oh.
[01:49:50] <AnEvilYak> right.
[01:49:56] <@geist> physical memory is much easier, but isn't necessarilly as fair
[01:50:08] <@geist> virtual scanning lets you do more advanced things like working set and whatnot
[01:50:24] <@geist> but is much more complex, since there are usually many more times the number of virtual pages in use than physical pages
[01:50:35] <@geist> the current vm scans virtual
[01:50:43] * AnEvilYak nods
[01:50:56] <@mmu_man> like paging out pages from teh same process you page in, so vm taking 4GB won't page out a 1kb app ?
[01:51:06] <@geist> seems that most vms do a physical scan + a bunch of hacks to make it more fair
[01:51:13] *** liquidboy has quit IRC
[01:51:40] <@geist> yeah. virtual scanning tends to punish bigger apps more, since they have a larger footprint for it to scan through
[01:51:44] <@geist> which in general is what you want
[01:51:50] <@mmu_man> yep
[01:51:54] <@mmu_man> more logical
[01:52:10] <AnEvilYak> presumably virtual scan is somewhat slower?
[01:52:14] <@mmu_man> *slap* you big memory hog
[01:52:25] <@geist> the newos/haiku vm has a bunch of complexity where I try to track the current working set, which then lets the scanner decide if it wants to scan through a process or not
[01:52:48] <@geist> ie, if it's over it's allocated pages, it gets scanned h"harder"
[01:53:05] <AnEvilYak> harder! better! faster! stronger!
[01:53:05] <@geist> and the page faulting behavior modifies the current working set
[01:53:08] <AnEvilYak> sorry, had to
[01:53:10] <@geist> apps that fault more get more working set
[01:53:44] <@mmu_man> now
[01:54:04] <@mmu_man> on related topic, how will you handle the kernel fd semantic ?
[01:54:13] <@geist> pretty standard stuff, but it's tough to implement efficiently
[01:54:22] <@geist> what do you mean?
[01:54:26] <@mmu_man> like "ugh I'm in that driver foo_read(), but how do I tell if it's for a kernel fd or the ap's" ?
[01:54:55] <@geist> ist>	i forget, I think there's a flag somewhere
[01:55:04] <@geist> the driver doesn't see a "fs" by that point
[01:55:08] <@mmu_man> for both things like sending SIGPIPE or not (you don't want an app to get a sig for an action the kernel did on its own fd
[01:55:11] <@geist> it's already been translated to an open file handle
[01:55:19] <@geist> er "fd"
[01:55:20] <@mmu_man> and is_valid_range() like stuff
[01:55:54] <@mmu_man> linux has a field in the current thread desc that tells the vm address limit that is allowed to be used
[01:56:02] <@geist> ist>	i did something to deal with that, but I think that's an area that axeld has rewritten
[01:56:06] <@geist> so it might be different in jaiku
[01:56:08] <@mmu_man> whent eh kernel calls things it puts ffffffff in there
[01:56:16] <@mmu_man> and put the old back in afterwards
[01:56:17] <@geist> sorry for the typoes, my connection is laggy
[01:56:20] <@geist> it's about a sentence behind
[01:56:34] <@mmu_man> I'll have to do something about that in Zeta as well
[01:56:41] <@mmu_man> as currently there is nothing that I know of
[01:56:51] <@mmu_man> now if only driver writers called is_valid_range()
[01:56:52] <AnEvilYak> geist: so, who do I have to nag to get you guys to release the sidekick IRC client? ;p
[01:56:57] <@geist> the beos way ot pinning the memory beforehand is fundamentally broken
[01:57:01] <@geist> which is why I did it that way
[01:57:14] <@mmu_man> even with wrong args (they dont know if it's for the kernel or not)
[01:57:25] <@mmu_man> geist well linux does it also
[01:57:30] <@mmu_man> it does user_memcpy()
[01:57:32] <Dr_Evil> is_valid_range() is a stupid idea
[01:57:40] <@geist> yes, it does user_memcpy
[01:57:42] <@mmu_man> but user_memcpy() does a check_range() anyway
[01:57:43] <Dr_Evil> since the kernel has no global lock
[01:57:44] <@geist> which is what I do in newos
[01:57:58] <@geist> beos used to pin the memory ahead of time
[01:58:03] <@geist> which of course doesn't work for strings
[01:58:21] <@geist> the simplest way to crash the kernel is to pass a point to a string that walks off the end of an area
[01:58:21] <@mmu_man> ?
[01:58:38] <@mmu_man> don't you know strnlen ?
[01:58:46] <@geist> doesn't work
[01:58:55] <@mmu_man> or strllen or whichever is supposed to be better
[01:58:56] <@geist> because to implement strnlen,you have to actually read the memory
[01:59:07] <@geist> which is what crashes the kernel (faulting on invalid memory)
[01:59:34] <@mmu_man> aren't you supposed to have called is_valid_range() before ?
[01:59:44] <@geist> yes, but what is the "range" of a string?
[01:59:49] <@mmu_man> well noone does it anyway in ioctl in drivers
[01:59:52] <@geist> you dont know how long it actually is until you read it
[02:00:00] <@geist> but you can't read it without risking a fault
[02:00:16] <@mmu_man> you should never pass a ptr to a string to a syscall anyway
[02:00:18] <@geist> and thus you can't really do anything about it but make sure the intial char is okay
[02:00:24] <@mmu_man> not a string that has a MAXFOO anyay
[02:00:31] <@mmu_man> paths have MAXPATH
[02:00:32] <@geist> what about open()?
[02:00:34] <ShackaN> windows has a IsBadReadPtr() thing to know that ahead of time..
[02:00:44] <@mmu_man> other stuff have B_OS_NAME_LENGTH
[02:01:04] <@mmu_man> open takes a path
[02:01:13] <@geist> what I do in newos is simply do the read/write to memory and recover from a fault
[02:01:17] <@mmu_man> is_valid_range(p, MAXPATH, URD);
[02:01:31] <@geist> solves the problem elegantly, since you dont have to do any checks up front
[02:01:34] <@geist> you only take a hit on failure
[02:01:38] <@mmu_man> yeah, but anyway you must check the prot anyway
[02:01:48] <AnEvilYak> isn't it possible for an operation on another CPU to cause the range to become invalid after you all it though?
[02:01:53] <AnEvilYak> er
[02:01:55] <AnEvilYak> s/all/call/
[02:02:06] <@geist> yes, that's another problem with the range checking idea
[02:02:09] <@mmu_man> what do you do if someone ioctl() a device with a ptr to a kernel data ?
[02:02:20] <@mmu_man> you won't crash
[02:02:25] <@geist> the driver is responsible for using user_memcpy
[02:02:25] <@mmu_man> but you'll clober your data
[02:02:33] <@mmu_man> now sometimes the kernel does ioctl()
[02:02:37] <@mmu_man> on its own fds
[02:02:46] <@mmu_man> *then* it's allowed to pass ptr to kernel
[02:02:51] <ShackaN> is it easy for user programs to crash the kernel?
[02:02:52] <@geist> also, it's a better design because some architectures have physically different address spaces for user and kernel data
[02:02:57] <@mmu_man> which is why you need a way to differentiate that
[02:03:02] <@geist> ie, you shouldn't really assume that user space is within the same namespace
[02:03:08] * Dr_Evil uses "#define user_memcpy memcpy" in his latest driver
[02:03:10] <@geist> ie, ultrasparc
[02:03:11] <@mmu_man> yeah tha'ts right
[02:03:23] <@mmu_man> anyway
[02:03:38] <@geist> and that alone is why linux has to do it, all other reasons aside
[02:03:49] <@mmu_man> linux does both check_range() for address checks and user_*
[02:03:55] <@mmu_man> yep
[02:04:05] <@geist> in bsd and linux they also dont assume that iospace is memory mapped
[02:04:09] <ShackaN> what does linux have to so ?
[02:04:14] <@geist> they do all io via functions
[02:04:18] <ShackaN> *to do
[02:04:24] <@geist> which may just do memory accesses
[02:04:32] *** Teknomancer has joined #haiku
[02:04:38] <@mmu_man> well we do have an isa bus manager :p
[02:04:59] <@mmu_man> now
[02:05:06] <@geist> correct, but the idea of memory mapping some devices registers and referencing via that is broken
[02:05:20] <@geist> though it's pretty hard to find a system where that's not valid
[02:05:20] <ShackaN> why?
[02:05:24] <tic> mmu_man, you heard the man. B_FIX :)
[02:05:25] <Teknomancer> i have a bunch of static int32 threads, should i replace them by BLoopers instead ?? :-) I know its a silly question, but just to get more opinions. .. :-)
[02:05:27] <@mmu_man> yeah, some drivers do that without asking
[02:05:29] <@geist> there are some systems that dont let you do that
[02:05:55] <AnEvilYak> Teknomancer: not unless you need message loops
[02:06:00] <AnEvilYak> Teknomancer: and all the overhead that entails
[02:06:09] <Teknomancer> AnEvilYak ah
[02:06:19] <@mmu_man> like ich
[02:06:23] * mmu_man pets Dr_Evil
[02:06:33] <@geist> anyway, back to this damn mmu code
[02:06:38] <Teknomancer> AnEvilYak okay but i have about 7 such static threads.. i thot it would be 'neater' to put all those threads in a BLooper...
[02:06:41] <@geist> trying to get the arm mmu working, but somethigns not right
[02:06:42] <@mmu_man> ya call me ?
[02:06:47] <AnEvilYak> Teknomancer: neater how?
[02:06:48] <@mmu_man> :)
[02:06:56] <Teknomancer> AnEvilYak in a seperate class
[02:07:01] <Teknomancer> AnEvilYak rather than functions of Mainwindow
[02:07:06] <AnEvilYak> Teknomancer: oh
[02:07:11] <AnEvilYak> Teknomancer: what're the threads doing?
[02:07:17] <ShackaN> geist, for newos ?
[02:07:29] <@mmu_man> actually I think user_memcpy() in linux is not that safe on x86
[02:07:36] <tic> Teknomancer, you know those threads can access the Mainwindow data directly, right? just send this as data and cast it to a Mainwindow and run another function within.
[02:07:39] <@mmu_man> it's really a check_range(); memcpy()
[02:07:40] <Teknomancer> AnEvilYak adding files, deleting files, extracting files, etc. from an archive - hopefully they should never run simultaneously
[02:07:48] <Dr_Evil> geist why is the idea of memory mapping device registers broken?
[02:07:50] <Teknomancer> tic thats what i'm doing now...
[02:07:51] <@mmu_man> since latest kernels are preemptibles ...
[02:07:58] <@mmu_man> like beos
[02:07:58] <tic> Teknomancer, "hopefully", heh.  You need locking! semaphores.
[02:08:11] <AnEvilYak> don't make that assumption....
[02:08:17] <@mmu_man> Dr_Evil because some arch don't allow that
[02:08:25] <Teknomancer> tic :) i don't exactly lock 'across' those threads.. i just make sure the MainWindow doesn't spawn them simultaneously
[02:08:32] <@mmu_man> they use different address spaces for PCI io
[02:08:37] <Teknomancer> AnEvilYak 2 mainwindows can spawn the threads...
[02:08:46] <Teknomancer> AnEvilYak but not from the same window ..
[02:08:47] <AnEvilYak> Teknomancer: ahh, but how does it ensure that one thread's gone before starting another?
[02:08:49] <tic> Teknomancer, that's nasty. read up on the Semaphore section of the Kernel Kit in the Be Book
[02:08:53] <AnEvilYak> Teknomancer: timing sensitive check there.
[02:09:07] <Dr_Evil> mmu_man yeah, so what? I have this DVB card, with 16 MEGABYTE of memory space, where registers are located, what should I do with it?
[02:09:07] <Teknomancer> AnEvilYak i block using some kind of modal window
[02:09:12] <AnEvilYak> ew
[02:09:13] <tic> good thing geist left, I really need to sleep now :P
[02:09:37] <AnEvilYak> modal = evil
[02:09:52] <Teknomancer> AnEvilYak :) ?
[02:10:08] * mmu_man pets Dr_Evil
[02:10:28] <AnEvilYak> Teknomancer: I tend to hate when apps block like that.
[02:10:38] <AnEvilYak> brb, need to walk dog
[02:10:42] <Teknomancer> AnEvilYak hehe
[02:11:07] * Dr_Evil likes memory mapping.
[02:11:18] * tqh likes to sleep
[02:11:18] * mmu_man likes 68k
[02:11:28] *** tqh has quit IRC
[02:11:29] <Teknomancer> how else do i block threads from a mainwindow ? ;-P
[02:11:30] <@mmu_man> no bloody inb outb
[02:11:31] <ShackaN> everybody likes memory mapping :D
[02:11:34] <Teknomancer> using volatile bool ?
[02:12:21] <@mmu_man> I really need to try that crashtest in linux
[02:13:09] <ShackaN> mmu_man, what test ?
[02:13:15] <Dr_Evil> in beos it's very easy to crash the kernel
[02:13:29] <Dr_Evil> there are plenty of drivers that will crash on ioctl
[02:13:39] <@mmu_man> ShackaN some black magic
[02:13:46] <@mmu_man> Dr_Evil of course, noone uses is_valid_range()
[02:13:51] <Dr_Evil> and if you are root in linux, you can open /dev/mem and write() something
[02:14:07] <ShackaN> mmu_man, calling syscalls with trashed pointers ?
[02:14:10] <@mmu_man> Dr_Evil well you are root in BeOS, so it's not a bug it's a feature
[02:14:20] <Dr_Evil> mmu_man is_valid_range() is not documented in the BeBook
[02:14:33] <@mmu_man> I recall hearing abuot some big arguments about whether root should be allowed to destroy the hardware or not
[02:14:40] <@mmu_man> some said it was bad
[02:14:49] <ShackaN> only SOME ?
[02:14:56] <@mmu_man> others said it should be allowed since root has all rights
[02:15:03] <@mmu_man> ShackaN yeh only
[02:15:10] <Dr_Evil> it should be allowed
[02:15:16] <@mmu_man> root is supposed to have all powers
[02:15:19] <@mmu_man> including that one
[02:15:27] <@mmu_man> and I agree with that :)
[02:15:40] <@mmu_man> Dr_Evil yeah that's a bug of the bebook
[02:15:59] <@mmu_man> actually it's documented elsewhere
[02:16:37] <@mmu_man> volume_III/Issue27.html uses it but doesn't give the proto
[02:16:52] <@mmu_man> I recall something at driven.be-in
[02:17:03] <ShackaN> I think the almighty root should not exist at all
[02:17:26] <Dr_Evil> my new DVB driver uses user_memcpy (defined to memcpy) on all ioctl arguments, maps the data read-only into userspace, and the 16 MB only accessible by kernele
[02:17:30] <AnEvilYak> it's not anywhere in headers
[02:17:54] <Dr_Evil> perhaps I'll add is_valid_range to the user_memcpy define should be 100% save
[02:18:53] <Dr_Evil> if only mpeg2 de-/remultiplexing wasn't that complicated :-(
[02:19:09] <@mmu_man> hmm yeah, well you need
[02:19:10] <@mmu_man> #define B_USER_READ_AREA    0x00000004  /* user can read */
[02:19:11] <@mmu_man> #define B_USER_WRITE_AREA    0x00000008  /* user can write */
[02:19:14] <Dr_Evil> but thats in userspace, the driver is 100% finished and working
[02:19:18] <@mmu_man> /* why isn't that in KernelExport ??? */
[02:19:18] <@mmu_man> extern bool is_valid_range(void *start, size_t len, uint32 prot);
[02:19:34] <@mmu_man> my own guess
[02:19:37] <Dr_Evil> B_READ_AREA is has the same effect
[02:19:52] <@mmu_man> no
[02:20:03] <@mmu_man> B_READ_AREA actually means B_KERNEL_READ_AREA
[02:20:20] <@mmu_man> meaning "is mapped" & "*kernel* can read it"
[02:20:21] <Dr_Evil> when create_area is called from the kernel, using 0 as protection makes it read/writeable by the kernel
[02:20:35] <Dr_Evil> and B_READ_AREA makes it readbale by userspace
[02:20:42] <Dr_Evil> I'm using it that way, and it works
[02:21:05] <Dr_Evil> haiku has some more
[02:21:43] <@mmu_man> #define PROT(kern,rw) ( rw ? (kern ? (B_WRITE_AREA | B_READ_AREA) : \
[02:21:43] <@mmu_man> (B_USER_WRITE_AREA | B_USER_READ_AREA)) : \
[02:21:44] <@mmu_man> (kern ? B_READ_AREA : B_USER_READ_AREA) )
[02:21:55] <@mmu_man> Dr_Evil yes it does
[02:22:10] <@mmu_man> but it will eat things like passing 0x10000000
[02:22:20] <@mmu_man> (base for physical_ram area)
[02:22:23] <@mmu_man> which is in kspace
[02:22:29] <@mmu_man> and which user shouldn't mess up with
[02:22:48] <@mmu_man> well it's more important for write actually
[02:23:03] <@mmu_man> being mapped and writable is != from being mapped and user-writable
[02:23:44] <Dr_Evil> haiku is better :) http://cvs.sourceforge.net/viewcvs.py/open-beos/current/src/kernel/core/vm/vm.c?rev=1.95&view=auto
[02:24:00] <@mmu_man> currently I use kern = true since the kernel doesn't maintain any state wether user_foo or the kernel foo was called
[02:24:09] <@mmu_man> yeah I saw those
[02:24:39] <@mmu_man> the defines are not the same
[02:24:42] <ShackaN> uh, a new vm?
[02:24:57] <@mmu_man> well those weren't public, so I guess noone could guess :)
[02:25:02] <ShackaN> it wasn't there last time I checked the cvs (some months ago)...
[02:26:31] <@mmu_man> anyway #define malloc(n) ({ void *addr; int fd = open("/dev/zero", O_RDWR); if (mmap(fd, MAP_PRIV, &addr) < 0) NULL; addr })
[02:26:36] <@mmu_man> :D
[02:26:47] <@mmu_man> now we might even be able to do that in haiku :)
[02:27:06] <ShackaN> uh?
[02:27:10] <ShackaN> an 08 01:27:10 <ShackaN>	a fake malloc ?
[02:27:16] <ShackaN> neat...
[02:27:40] <Dr_Evil> thats a stupid linux hack
[02:27:47] <Dr_Evil> to get large areas
[02:27:56] <@mmu_man> yep
[02:27:58] <ShackaN> I didn't know that stupid hack..
[02:28:17] <AnEvilYak> seems like it'd be nice that it implicitly zero fills though
[02:28:18] <@mmu_man> cause they don't have a cleaner way like create_area
[02:28:45] * Dr_Evil hates schedule_timeout(foobar/HZ):
[02:28:50] <@mmu_man> ShackaN yeah linux freaks like to mmap() everything
[02:28:58] <@mmu_man> quite painful when poritng stuff
[02:29:07] <@mmu_man> they mmap() shared libs in the elf loader
[02:29:19] <@mmu_man> /dev/zero to get ram
[02:29:30] <@mmu_man> /dev/framebuffer
[02:29:50] <@mmu_man> Dr_Evil yeah that HZ sux
[02:29:56] <ShackaN> but mmap si powerful, and it's a shame beos doesn't have it
[02:30:18] <@mmu_man> well you can try libmoreposix but it doesn't carry all the semantics
[02:30:22] <Dr3w> didn't someone write an mmap driver once?
[02:30:27] <@mmu_man> but at least it really mmaps an fd if it's a file
[02:30:33] <@mmu_man> JBQ
[02:30:40] <@mmu_man> though it couldn't be used from an fd
[02:30:45] <@mmu_man> I fixed it
[02:30:49] <@mmu_man> but it's a big hack
[02:30:53] <ShackaN> wow
[02:31:02] <@mmu_man> it doesn't correctly write back dirty pages
[02:31:12] <@mmu_man> and doesn't work with MAP_SHARED
[02:31:16] <ShackaN> uhm, too bad...
[02:31:34] <@mmu_man> (that could be fixed by cloning the area at_fork())
[02:31:45] <ShackaN> why are be areas so different from unix shared memory ?
[02:31:52] <@mmu_man> flushing pages requires cooperation from teh vm
[02:32:19] <Dr3w> moving to laptop!
[02:32:20] *** Dr3w has quit IRC
[02:32:21] <@mmu_man> probably because they were written for nukernel
[02:32:33] <ShackaN> nu?
[02:32:33] 
[02:32:37] 
[02:32:40] <@mmu_man> micro
[02:32:41] <AnEvilYak> microkernel
[02:32:41] <ShackaN> ok
[02:32:44] *** Dr3w has joined #haiku
[02:32:56] <ShackaN> that's mu, not nu :D
[02:33:04] <@mmu_man> mmap() is between vm and vfs
[02:33:13] *** Teknomancer has left #haiku
[02:33:21] <@mmu_man> the vfs (fsil) isn't really part of that nukernel
[02:33:29] <AnEvilYak> http://www.bedope.com/archive/041398.html
[02:33:36] <ShackaN> so beos is based on a MUkernel ?
[02:33:49] <AnEvilYak> keyword: based on, it's not very micro nowadays
[02:33:54] <AnEvilYak> tons of stuff living in kspace
[02:34:20] <ShackaN> why?
[02:34:28] * mmu_man wants Zeta on N64 and XBox
[02:34:31] <Dr3w> speed I think.
[02:34:43] <AnEvilYak> generally speaking microkernels offer a trade off between reliability and performance
[02:34:44] <ShackaN> it  would be nice to leave it in uspace and avoid many KDL :D
[02:34:50] <ShackaN> i do know
[02:34:54] <AnEvilYak> espeically on a CPU like x86 where context switch == slow as fuck
[02:34:56] <Dr_Evil> doesn't matter
[02:35:08] <Dr3w> mmu_man N64 would be good! I have a Z64, lets you load files from Zip disk into memory, then boot the memory on the N64.
[02:35:10] <AnEvilYak> and besides, for a lot of things, them crashing in userspace really wouldn't be a huge benefit.
[02:35:14] <Dr_Evil> when something important in userspace crashes, you are dead, just like KDL
[02:35:23] <AnEvilYak> indeed.
[02:35:24] <ShackaN> well, yes..
[02:35:26] * Dr_Evil pets AnEvilYak
[02:35:30] *** mmadia_out is now known as mmadia
[02:35:38] * mmu_man pets his restart_app_server script
[02:35:42] <ShackaN> but it would be cooler anyway :P
[02:36:04] <AnEvilYak> i.e. if, say, your driver_server or whatever crashes, having to restart all that stuff and everything else with dependencies on it wouldn't be significantly faster than just doing a reboot anyhow
[02:36:20] <AnEvilYak> unless you're one of those linux dorks that seems to equate uptime with penis size or something
[02:36:31] <Dr3w> you mean it doesn't??!!
[02:36:41] * Dr3w restarts in that case...
[02:36:45] <ShackaN> ohhhhhh
[02:36:50] <ShackaN> hahahahaha
[02:37:52] <@mmu_man> AnEvilYak the same linux freaks who anyway run khttpd ?
[02:37:53] <slaad> "Man, my app_server's been dead for months... but I was afraid to restart on account of penis shrinkage"
[02:38:01] <AnEvilYak> mmu_man: ROFL
[02:38:11] <@mmu_man> I think it's even officially part of teh kernel
[02:38:14] <slaad> khttpd? A web server in the kernel?
[02:38:23] <@mmu_man> hmm I need to fix the beos one btw
[02:38:25] <@mmu_man> slaad yeah
[02:38:34] <slaad> Scary.
[02:38:43] <@mmu_man> I wrote one for beos but it does bad things for now
[02:38:52] <ShackaN> uhm
[02:38:55] <AnEvilYak> mmu_man: you need to write knethangmand
[02:39:03] <ShackaN> about webservers...
[02:39:10] <@mmu_man> well first I want to finish kshell
[02:39:11] <ShackaN> the netstack lie in ks too ?
[02:39:16] <AnEvilYak> ShackaN: BONE's does.
[02:39:25] <AnEvilYak> ShackaN: R5/net_server does not.
[02:39:32] <Dr3w> mmu_man: http://www.softidea.com/Z64/
[02:39:42] <@mmu_man> so I can easily add commands
[02:39:42] <ShackaN> and the haiku one ? :)
[02:39:48] <AnEvilYak> but net_server is complete crap anyhow so let's not talk about that.
[02:39:55] <AnEvilYak> ShackaN: haiku's afaik using a kernel stack too.
[02:40:01] <ShackaN> kaN>	k
[02:40:12] <AnEvilYak> unless people're still having huge arguments about that
[02:40:43] <ShackaN> 'night, gonna read a book
[02:40:44] <Dr_Evil> night
[02:40:51] <AnEvilYak> see ya
[02:40:57] <@mmu_man> aw 2:30
[02:41:07] <Dr_Evil> 2:40
[02:41:16] <Dr3w> 1:42 :)
[02:41:18] <ShackaN> 2:44 here
[02:41:21] <AnEvilYak> 18:41
[02:43:31] 
[02:44:20] <@mmu_man> shutdown -h now
[02:44:22] *** mmu_man has quit IRC
[02:49:04] *** xingshter has joined #haiku
[02:49:09] <xingshter> esta sheen?
[02:51:41] <xingshter> esta sheen?
[02:52:04] *** xingshter has left #haiku
[03:01:35] *** Dr3w has quit IRC
[03:02:41] *** Begasus has quit IRC
[03:05:24] *** MikeW has joined #haiku
[03:06:29] <MikeW> waaaahooooooo
[03:06:38] <MikeW> It has been absolutely calm for the past 5 hours
[03:07:00] <MikeW> at 2 mins past two, whoosh, hugely loud from the wind blowing outside
[03:07:06] <AnEvilYak> nice
[03:07:08] <AnEvilYak> where?
[03:07:11] <MikeW> forecast 90mph
[03:07:15] <MikeW> west of Ireland
[03:08:59] *** Korli has quit IRC
[03:09:39] *** MikeW has quit IRC
[03:14:50] *** Teknomancer has joined #haiku
[03:17:47] <Teknomancer> why is blocking a thread using a modal window bad ?
[03:18:23] <AnEvilYak> not very friendly UI-wise.
[03:18:34] <Teknomancer> but
[03:18:38] <Teknomancer> the user sees a Progress window
[03:18:42] <Teknomancer> and can cancel the operation
[03:19:10] <AnEvilYak> that works then
[03:19:12] <Teknomancer> and once the thread is 'cancelled' it will post message to main window signalling the cancel, ...
[03:19:17] <Teknomancer> ah :)
[03:19:20] <AnEvilYak> often see modal windows very badly abused though
[03:19:29] <Teknomancer> so i'm not tooo baaad a programmer i guess :) YAY !!
[03:19:37] <Teknomancer> thx AnEvilYak :)
[03:19:46] <Teknomancer> i thot u will convince me to rewrite code :)
[03:19:47] <AnEvilYak> np.
[03:19:48] <Teknomancer> hehe
[03:20:04] <Teknomancer> ah, a big relief :)
[03:20:10] <Teknomancer> i thot i had done it wrong once again...
[03:20:35] <Teknomancer> there always seems to be a better way to do things in beos programming then what i do .. :-(
[03:20:48] <Teknomancer> but luckily this time, in this project atleast .. i hope i have done most things right.
[03:21:03] <AnEvilYak> *shrug* take any programming problem and there's a bazillion ways to solve it
[03:21:14] <AnEvilYak> what's the "correct" way is often pretty subjective.
[03:22:10] <Teknomancer> guess that's correct, what seems an ugly way to others may seem easier for me ...
[03:22:29] <AnEvilYak> well, that having been said, there're certainly lots of approaches that might be less efficient...
[03:22:39] <Teknomancer> yes
[03:22:42] <AnEvilYak> so you could say there's better ways..
[03:22:45] <AnEvilYak> but "best"
[03:22:50] <AnEvilYak> is a different story.
[03:23:09] <Teknomancer> i even have heard arguments of whether to use SendMessage or PostMessage
[03:23:10] <Teknomancer> lol
[03:23:23] <AnEvilYak> mm
[03:23:29] <AnEvilYak> that's a different story
[03:23:39] <AnEvilYak> there used to be various edge cases with PostMessage that would cause deadlocks
[03:23:43] <AnEvilYak> those've since been fixed.
[03:24:02] <Teknomancer> in R5 ?
[03:24:05] <AnEvilYak> you'll notice the R5 BeBook had a warning in the PostMessage section
[03:24:09] <Teknomancer> yes
[03:24:11] <Teknomancer> i remember
[03:24:17] <Teknomancer> it said SendMessage is safer
[03:24:23] <AnEvilYak> *nod*
[03:24:29] <Teknomancer> but whats the point?? I need to make  BMesseneger each time and use SendMessage ?
[03:24:35] <AnEvilYak> BMessage/BMessenger/etc. were rewritten from scratch since then.
[03:24:45] <AnEvilYak> well, either that or make one BMessenger and keep it around.
[03:24:53] <Teknomancer> so i need to replace all my PostMessage with SendMessage ?
[03:25:01] <AnEvilYak> depends how you're doing it.
[03:25:04] <Teknomancer> that would take forever :)
[03:25:18] <Teknomancer> i use only 1 or 2 PostMessage that needs a reply.
[03:25:25] <AnEvilYak> it's been long enough since I've done much R5 stuff that I forget exactly what case causes a deadlock.
[03:25:34] <AnEvilYak> anyways, brb dinner.
[03:25:47] <Teknomancer> ah no
[03:25:51] <Teknomancer> i use SendMessage for reply
[03:25:54] <Teknomancer> l8r AnEvilYak ;)
[03:36:07] *** Teknomancer has quit IRC
[03:44:46] *** brennanOS has joined #haiku
[03:48:43] *** Potn has quit IRC
[03:48:52] <AnEvilYak> back
[03:49:00] *** Potn has joined #haiku
[03:52:15] <@geist> whew
[03:52:23] <AnEvilYak> wb geist
[03:52:51] <@geist> ist>	i was always right here
[03:53:01] <slaad> That's what they all say.
[03:53:08] <AnEvilYak> your body was, your mind wasn't ;p
[03:53:28] <slaad> That's not as scary as when the body wasn't but the mind is.
[03:53:37] <AnEvilYak> hahahaha.
[03:53:39] <AnEvilYak> true.
[04:04:02] *** LupusMic1aelis has joined #haiku
[04:20:38] *** ChanServ has quit IRC
[04:22:27] *** LupusMichaelis has quit IRC
[04:27:25] <@AndrewBachmann> well geist I have committed to buying a sidekick
[04:27:39] <@AndrewBachmann> unless I change my mind and am willing to give up my $20 deposit, I'll pick it up on monday
[04:28:06] <@AndrewBachmann> congrats, now you can join the marketing department :-)
[04:30:04] *** illissius- has joined #haiku
[04:34:40] *** ChanServ has joined #haiku
[04:34:40] *** irc.freenode.net sets mode: +o ChanServ
[04:35:28] <@geist> haha
[04:36:00] <@AndrewBachmann> I will ended up pay $50 since I didn't want to order online
[04:36:14] <@AndrewBachmann> anyway, they are fairly sure that it will work in china
[04:36:21] <@AndrewBachmann> although I still have my doubts :-)
[04:37:58] <@AndrewBachmann> wow I got a spam through gmail
[04:38:25] *** illissius[sleep] has quit IRC
[04:39:13] *** fyysik has quit IRC
[04:53:06] *** Kernel86 has joined #haiku
[05:02:25] *** ShackaN has quit IRC
[05:22:48] <@AndrewBachmann> voidref I think I decided to get a gateway 3520gz even though I can't get the graphics working in beos yet
[05:23:43] *** mmadia has quit IRC
[05:23:53] *** mmadia has joined #haiku
[05:36:06] *** NathanW has quit IRC
[05:36:40] *** NathanW has joined #haiku
[05:43:35] *** Fanskapet has joined #haiku
[05:44:07] *** Fanskapet has quit IRC
[05:45:17] *** Potn has quit IRC
[05:45:59] *** Potn has joined #haiku
[05:49:13] *** Fanskapet has joined #haiku
[06:03:58] *** mmadia has quit IRC
[06:03:59] *** mmadia42 has joined #haiku
[06:04:12] *** Potn has quit IRC
[06:04:55] *** Potn has joined #haiku
[06:20:56] *** AndrewBachmann has quit IRC
[06:47:26] *** mmu_man has joined #haiku
[06:47:26] *** ChanServ sets mode: +o mmu_man
[06:51:30] *** mmadia42 has quit IRC
[07:01:05] *** brennanOS has quit IRC
[08:08:14] *** nPHYN1T3 has joined #haiku
[08:33:34] *** AndrewBachmann has joined #haiku
[08:33:34] *** ChanServ sets mode: +o AndrewBachmann
[08:45:35] *** Potn has quit IRC
[08:47:14] *** Potn has joined #haiku
[08:53:27] *** Potn_ has joined #haiku
[08:53:32] *** Potn has quit IRC
[09:07:53] *** Potn_ has quit IRC
[09:08:43] *** Potn_ has joined #haiku
[09:15:02] *** Potn_ has quit IRC
[09:15:51] *** Potn_ has joined #haiku
[09:22:04] *** [Potn] has joined #haiku
[09:22:39] *** Potn_ has quit IRC
[09:26:30] *** Potn has joined #haiku
[09:26:46] *** [Potn] has quit IRC
[09:27:58] *** Potn has quit IRC
[09:28:56] *** Potn has joined #haiku
[09:36:31] *** Potn has quit IRC
[09:36:32] *** Potn_ has joined #haiku
[09:46:51] *** Potn has joined #haiku
[09:48:30] *** Potn has quit IRC
[09:49:53] *** Potn has joined #haiku
[09:53:01] *** Potn_ has quit IRC
[09:53:49] *** Potn has quit IRC
[09:54:18] *** Potn has joined #haiku
[09:55:01] *** lechu_mys has joined #haiku
[09:55:08] <lechu_mys> hi all
[09:55:09] <lechu_mys> ;]
[09:58:13] *** bs0 has joined #haiku
[09:59:47] *** Loppan has joined #haiku
[10:01:13] *** Loppan has quit IRC
[10:01:14] *** Potn has quit IRC
[10:02:05] *** Potn has joined #haiku
[10:05:14] *** Loppan has joined #haiku
[10:13:31] *** Potn has quit IRC
[10:16:39] *** Teknomancer has joined #haiku
[10:18:43] *** tqh has joined #haiku
[10:21:08] *** lechu_mys has quit IRC
[10:25:43] *** dipp has joined #haiku
[10:27:44] *** Karina`` has quit IRC
[10:28:16] *** Karina`` has joined #haiku
[10:32:29] *** Potn has joined #haiku
[10:40:04] *** Potn has quit IRC
[10:40:22] *** Potn has joined #haiku
[10:45:29] *** Karina`` has quit IRC
[10:46:04] *** Karina`` has joined #haiku
[10:46:06] *** Potn has quit IRC
[10:47:23] *** Potn has joined #haiku
[10:48:35] *** Potn has quit IRC
[10:50:27] *** Potn has joined #haiku
[10:52:32] *** BeMonni has quit IRC
[10:55:34] * tqh can't find any info on anyone running or tried to run BeOS with nforce3 or 4
[11:18:25] *** frankps has joined #haiku
[11:19:20] *** Teknomancer has quit IRC
[11:24:57] *** mmu_man has quit IRC
[11:36:42] *** Potn has quit IRC
[11:37:41] *** Potn has joined #haiku
[11:38:27] *** hUMUNGUs has joined #haiku
[11:39:11] *** hUMUNGUs has quit IRC
[11:39:11] *** Potn has quit IRC
[11:40:18] *** Potn has joined #haiku
[11:43:33] *** LebuzzerJan has quit IRC
[12:02:25] *** Begasus has joined #haiku
[12:16:25] *** TuneTracker has joined #haiku
[12:16:57] <TuneTracker> tic
[12:20:37] <Begasus> tac
[12:21:47] *** ShackaN has joined #haiku
[12:25:25] <TuneTracker> toe
[12:26:47] *** Konrad has joined #haiku
[12:27:49] *** Begasus has quit IRC
[12:29:33] *** Begasus has joined #haiku
[12:47:12] *** Greyhawk_Dragon has joined #haiku
[12:48:59] *** Potn has quit IRC
[12:51:03] <tic> mooo
[12:57:36] *** mmu_man has joined #haiku
[12:57:36] *** ChanServ sets mode: +o mmu_man
[13:00:55] *** Potn has joined #haiku
[13:04:48] *** emwe has joined #haiku
[13:06:22] <@mmu_man> 3h here
[13:06:25] <@mmu_man> tired too :)
[13:10:31] *** Greyhawk_Dragon has quit IRC
[13:15:33] *** bs0 has quit IRC
[13:16:50] *** Leonardo_Lopes has joined #haiku
[13:17:10] <Leonardo_Lopes> where can I download BONE?
[13:18:54] <@mmu_man> not here
[13:19:23] *** BGA has joined #haiku
[13:19:23] *** ChanServ sets mode: +o BGA
[13:20:51] <Leonardo_Lopes> mmu_man, I'm not asking if I can download here, I'm asking if anyone of here nows where I can download BONE
[13:22:08] <Dr_Evil> you don't expect an answer, do you?
[13:22:22] <Leonardo_Lopes> Dr_Evil, Why not?
[13:23:38] <Dr_Evil> its not legal
[13:25:06] <tic> Leonardo_Lopes, buy YellowTAB's Zeta if you want BONE.
[13:25:31] *** dipp has quit IRC
[13:26:07] <Leonardo_Lopes> isn't legal? shit. Someone said to me download it =(
[13:26:33] *** Methe has joined #haiku
[13:26:53] <[Beta]> w00t, the english version of http://insidebeos.de/art/cover2-2000big.gif just arrived :o
[13:27:11] <tic> yay
[13:27:18] <[Beta]> lo Methe
[13:27:18] <Leonardo_Lopes> how can I put my adsl to work in a Legal way?
[13:28:09] <tic> might be a PPPoE driver for net_server on BeBits.
[13:28:42] <[Beta]> it's lucky I got a modem+router.. or i'd need to write a PPPoA driver :(
[13:29:04] <Leonardo_Lopes> It doen't works, it connects but some seconds after, the connection drop
[13:29:21] <[Beta]> Lopes - and it works with other o/s ?
[13:30:03] <Leonardo_Lopes> Windows, Linux
[13:30:24] <Leonardo_Lopes> FreeBSD
[13:32:45] <[Beta]> Worms Armageddon for Be ?
[13:33:02] *** Greyhawk_Dragon has joined #haiku
[13:33:31] <[Beta]> Lopes; as tic said, you could buy Zeta
[13:34:28] *** nPHYN1T3 has quit IRC
[13:36:14] *** mahlzeit has joined #haiku
[13:36:14] *** ChanServ sets mode: +o mahlzeit
[13:36:50] * Methe jumps in the sea to make a wave on mahlzeit
[13:36:55] <tic> mahlzeit
[13:36:59] *** Korli has joined #haiku
[13:36:59] *** ChanServ sets mode: +o Korli
[13:37:05] * Methe jumps in the sea to make a wave on Korli
[13:37:43] <[Beta]> taste?
[13:39:16] *** hUMUNGUs has joined #haiku
[13:40:20] *** hUMUNGUs has quit IRC
[13:43:43] *** Loppan has quit IRC
[13:47:01] *** Loppan has joined #haiku
[13:51:30] *** Begasus has quit IRC
[13:58:21] <@mahlzeit> hi there
[13:59:12] <@Korli> hey mahlzeit
[14:03:40] *** Fanskapet has quit IRC
[14:28:25] *** [Beta] has quit IRC
[14:50:32] *** [Beta] has joined #haiku
[14:53:53] *** fyysik has quit IRC
[15:01:37] *** SirMik has joined #haiku
[15:12:17] *** Lebuzzer has joined #haiku
[15:13:21] *** Lebuzzer has quit IRC
[15:14:01] *** Potn has quit IRC
[15:15:19] *** nielx has joined #haiku
[15:24:30] *** Fanskapet has joined #haiku
[15:29:34] *** Loppan has quit IRC
[15:30:28] *** tqh has quit IRC
[16:01:18] *** ConneX has quit IRC
[16:05:42] *** Loppan has joined #haiku
[16:06:49] *** TuneTracker has joined #haiku
[16:24:21] *** thaflo has joined #haiku
[16:25:36] *** tqh has joined #haiku
[16:29:17] *** frankps has joined #haiku
[16:29:31] *** phoudoin has joined #haiku
[16:30:45] <phoudoin> hi guys
[16:30:52] <@mahlzeit> hey phoudoin
[16:31:23] <phoudoin> I'm stuck with a missing makedepend tool to build latest Mesa CVS...
[16:33:36] <SirMik> ah mahlzeit, je hangt hier ook rond :)
[16:33:41] <SirMik> hey phoudoin
[16:33:49] <@mahlzeit> hey SirMik
[16:34:03] <SirMik> mahlzeit jij hebt rc toch geschreven he?
[16:34:16] <@mahlzeit> yep
[16:34:42] <SirMik> krijg hem niet gecompileerd...heb je een binary bij de hand?
[16:35:11] <@mahlzeit> wat is de fout?
[16:35:12] <SirMik> (mijn buildsystem is echt een puinhoop, missing m4sugar enz.)
[16:35:20] <@mahlzeit> oh
[16:35:22] <SirMik> nah hij vindt parser.hpp niet
[16:35:40] <@mahlzeit> die moet hij automatisch aanmaken
[16:35:48] <@mmu_man> plop
[16:35:51] <SirMik> terwijl m4sugar wel op mijn systeem aanwezig is....maar wordt gek van die unix troep
[16:35:54] <@mahlzeit> anyway, http://home.tiscali.nl/mahlzeit/rc-1_1.zip
[16:36:08] <SirMik> thnks mahlzeit
[16:36:29] <@mahlzeit> probeer je haiku te compilen?
[16:36:53] <SirMik> ja
[16:37:00] <SirMik> vooral de input kit eigenlijk
[16:37:00] <@mahlzeit> met welke jam?
[16:37:06] <SirMik> uhm ff kijken
[16:37:22] <SirMik> 2.4
[16:37:34] <@mahlzeit> ik bedoel: waar heb je die jam vandaan? :-)
[16:37:38] <SirMik> inputkit werkt niet
[16:37:43] <SirMik> uh haiku site meen ik
[16:37:58] <SirMik> (staat er al een hele tijd op)
[16:38:00] <@mahlzeit> je kunt het beste de jam uit /current/src/tools/jam zelf compilen
[16:38:27] <SirMik> hehe als dat lukt....zoals ik al zei: automake, autoconf en die hele zeik werken allemaal niet
[16:38:43] <@mahlzeit> ik geloof niet dat je die nodig hebt...
[16:39:35] <SirMik> ah ik zal het proberen dan :)
[16:43:51] <SirMik> werkt niet echt...maar kan het zijn dat het komt doordat ik op een andere partitie probeer te compilen?
[16:44:10] <@mahlzeit> wat werkt niet?
[16:45:02] <tic> Sorry for being off-topic, but...I've installed a fresh Deb system using 'debootstrap', rebooted, and after that it only reaches 'crond', and then nothing. I want a console running login.  How do I go about doing that?
[16:46:00] <SirMik> nee, wacht ff probeer zo op deze partitie....hij miste rc zowieso...maar die heb ik geinstalleerd :S
[16:47:58] *** TuneTracker has quit IRC
[16:48:41] <SirMik> ole....720mb kopieren met 0.35mb/s
[16:56:31] *** dipp has quit IRC
[16:59:33] *** emwe has quit IRC
[17:02:41] *** mmu_man has quit IRC
[17:08:05] *** voidref has quit IRC
[17:09:00] *** Konrad has quit IRC
[17:09:08] *** Potn has joined #haiku
[17:10:11] *** Konrad has joined #haiku
[17:11:43] 
[17:11:49] <@mahlzeit> woot!
[17:12:04] <SirMik> jah weet ook niet waarom het zo sloom was
[17:12:26] <SirMik> ene keer gaat het met 17mb/s nu met 0.35
[17:13:01] <@mahlzeit> veel kleine files, denk ik
[17:13:07] <SirMik> nee krijg hem dus niet gecompileerd
[17:13:12] <SirMik> argghhhh
[17:13:30] <SirMik> weer m4 sugar
[17:13:42] *** fyysik has joined #haiku
[17:13:44] <@mahlzeit> wat probeer je nu te compileren?
[17:13:46] <SirMik> hier: /bin/m4: m4sugar/m4sugar.m4: No such file or directory
[17:13:47] <fyysik> hi ppl
[17:13:47] <SirMik> jam
[17:14:09] <fyysik> question about Haiku print Kit - does it update also Transport and ports?
[17:14:14] <@mahlzeit> voor jam hoef je alleen het volgende te doen:
[17:14:15] <@mahlzeit> cd ~/current/src/tools/jam
[17:14:15] <@mahlzeit> make
[17:14:15] <@mahlzeit> cp bin.beosx86/g/jam ~/config/bin
[17:14:25] <@mahlzeit> niks geen autoconf en die shit
[17:14:55] <@mahlzeit> misschien is je compiler een beetje in de war :-)
[17:15:20] <SirMik> ah heb jam gedaan voor jam
[17:15:27] <fyysik> R5 printing sucks heavily...dunno what is guilty, port driver or trnasport add-on
[17:16:54] <SirMik> mahlzeit gcc: jamgram.c: No such file or directory
[17:17:10] <SirMik> en dan een hele riedel nog
[17:17:11] <@mahlzeit> welke versie van beos heb je?
[17:17:18] <SirMik> r5 + BONE7a
[17:17:24] <@mahlzeit> hmm
[17:17:45] <SirMik> origineel de dev edition geloof ik
[17:17:58] <SirMik> maar heb er een van de laatste gcc's op draaien
[17:18:11] <@mahlzeit> dat is het probleem denk ik
[17:18:18] <@mahlzeit> heb je flex en bison?
[17:19:12] <SirMik> geen idee
[17:19:27] *** Potn has quit IRC
[17:19:30] <@mahlzeit> easy to find out :-)
[17:19:32] <SirMik> flex version 2.5.4
[17:19:51] <SirMik> bison 1.75
[17:20:09] <@mahlzeit> die heeft hij namelijk nog om jam te compileren (en rc ook), zover ik weet
[17:20:16] *** dipp has joined #haiku
[17:20:19] *** Potn has joined #haiku
[17:21:50] <SirMik> ja maar die heb ik dus...:S
[17:22:30] <nielx> probeer te compileren met: LEX=flex jam
[17:22:40] <nielx> bij mij wil hij flex ook nooit vinden
[17:22:46] <nielx> en dan gaan er dingen fout
[17:23:22] <SirMik> ah nielx moet ik dat er aan toevoegen op de command line?
[17:23:29] <nielx> jep
[17:23:35] <nielx> misschien helpt het
[17:23:46] <SirMik> ok bedankt...ff proberen
[17:24:57] <SirMik> nee dus...begint weer over m4sugar te zeiken
[17:25:26] <SirMik> lijkt wel of ie helemaal niet in config/bin config/lib kijkt
[17:26:34] *** Leonardo_Lopes has quit IRC
[17:29:39] *** mmu_man has joined #haiku
[17:29:40] *** ChanServ sets mode: +o mmu_man
[17:31:36] *** emwe has joined #haiku
[17:32:46] <SirMik> naja ik meel korli wel als de site weer online is...
[17:34:38] <SirMik> in ieder geval bedankt voor de hulp...ben er weer vandoor
[17:34:52] *** SirMik has quit IRC
[17:42:53] <tqh> fyysik Firefox from CVS leaks pretty badly. I used refcount tools to and saw that nsImageBeOS seems to be the only BeOS class I could see.
[17:46:28] *** thies has quit IRC
[17:51:28] <fyysik> ?
[17:51:47] <fyysik> do you think that nsImage leaks?
[17:52:16] *** thies has joined #haiku
[17:53:12] <fyysik> did you apply any patches there?
[17:54:42] <fyysik> tqh - and is Simon's leak patch already in CVS?
[17:55:28] <tqh> fyysik I don't know havn't really used the leak tool enoughs, and the maketree script don't work for me..
[17:55:47] *** TuneTracker has joined #haiku
[17:57:04] <fyysik> look in nsDrawingSurface.cpp beos at surface destructor, tqh
[17:58:04] <tqh> fyysik No it's not in HEAD at leas.
[17:58:22] <tqh> Oh, wait it is
[17:58:46] <fyysik> so, no RemoveChild in destructor anymore?
[17:59:25] <tqh> it's gone.
[17:59:50] <tqh> So it's something else. I used this http://www.mozilla.org/performance/refcnt-balancer.html but maketree don't seem to work for me.
[18:01:55] <tqh> the log file got quite big 141MB
[18:02:32] <fyysik> ohhh
[18:03:55] <fyysik> ik>	i will look at nsImage again
[18:04:37] <tqh> Well there are a lot of non BeOS classes listed so not sure it's BeOS
[18:04:41] <fyysik> there were recently changes in generic nsImage, to improve clipping
[18:05:07] <fyysik> to not draw regions which are outside clipping
[18:05:16] <tqh> The list of objs in leak is only 138K if you are interested
[18:05:27] <fyysik> so maybe they introduced leak, or we didn't implement something related
[18:05:56] <tqh> maybe, windows and mac tinderbox don't leak at least.
[18:06:27] <tqh> 5 min of surfing: => mFreeCount:           18359  --  LEAKED 220 !!!
[18:06:32] *** thies_ has joined #haiku
[18:06:47] *** thies has quit IRC
[18:07:51] *** phoudoin has quit IRC
[18:10:35] *** vaste has quit IRC
[18:17:36] <tic> tja vaste
[18:17:39] <tic> d'oh.
[18:18:06] <tqh> fyysik Added som info at Bezilla blog.
[18:18:54] <tqh> oh, blog didn't like < and > :(
[18:19:25] <fyysik> yup
[18:20:48] <tqh> updated
[18:26:50] <fyysik> tqh, do you think that getter_AddRefs() is quite intelligent to release object by self, withourt explicit NS_RELEASE?
[18:27:37] <fyysik> tqh - http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&subdir=mozilla/gfx/src/shared&command=DIFF_FRAMESET&file=nsRenderingContextImpl.cpp&rev1=3.50&rev2=3.51&root=/cvsroot
[18:35:37] *** MikeW has joined #haiku
[18:39:16] *** bs0 has joined #haiku
[18:39:17] <tqh> fyysik don't know.
[18:39:36] *** bs0 has quit IRC
[18:41:31] <@Korli> http://www2.b3ta.com/heyhey16k/heyhey16k.swf
[18:42:06] *** voidref has joined #haiku
[18:42:06] *** ChanServ sets mode: +o voidref
[18:42:21] <@AndrewBachmann> hello voidref
[18:42:25] <@voidref> hi
[18:42:31] <@voidref> hower things going?
[18:43:14] <@AndrewBachmann> I couldn't get the display issues worked out but I think I am going to get the gateway 3520gz anyway
[18:43:30] <@AndrewBachmann> perhaps I can continue to be your guinea pig
[18:43:30] <@voidref> what video is it?
[18:43:39] <@AndrewBachmann> wait, I wrote it down
[18:43:41] <fyysik> question about Haiku print Kit - does it update also Transport and ports?
[18:44:39] <@voidref> brb
[18:44:39] *** tic has quit IRC
[18:44:41] <@AndrewBachmann> Intel(R) 82852/82855 GM/GME - Vendor Intel (0x8086) - Device ID : 0x3582 - Revision 2 - Base 3 : Subtype 80 : Interface 0
[18:47:58] <tqh> fyysik this bug might be related to this. https://bugzilla.mozilla.org/show_bug.cgi?id=213391
[18:48:10] <tqh> Anyway, got to go.
[18:48:13] *** tqh has quit IRC
[18:48:46] *** tic has joined #haiku
[18:50:58] *** TuneTracker has quit IRC
[18:53:56] *** thies_ has quit IRC
[19:13:09] *** [Beta] has quit IRC
[22:01:51] *** [Beta] has joined #haiku
[22:12:52] *** Greyhawk_Dragon has joined #haiku
[22:16:35] *** Loppan has quit IRC
[22:18:07] <@mmu_man> hmm
[22:18:08] <@mmu_man> [revol@patrick ~]$ imlucky openweb
[22:18:09] <@mmu_man> /google/openweb 00001/Openweb.eu.org - Accueil : string : http://openweb.eu.org/
[22:18:09] <@mmu_man> /bin/xargs: catattr: terminated by signal 7
[22:18:29] *** ShackaN has joined #haiku
[22:27:00] *** thies has quit IRC
[22:40:39] *** thies has joined #haiku
[22:43:36] *** nielx has joined #haiku
[22:43:50] <nielx> Is there anyone that actually works in the ict sector?
[22:45:40] <[Beta]> quite a few I imagine
[22:46:17] <nielx> I actually meant to place 'here' in the sentence
[22:46:33] <nielx> but I guess it's floating in the air somewhere
[22:46:39] <nielx> damned wireless ;-)
[22:46:51] <[Beta]> I applied for a position at my local school - they were wanting one :)
[22:47:11] <[Beta]> it probably flew out the window.. did you shut all of them ?
[22:48:06] <tic> mmu_man, pretty nifty still.
[22:49:11] <nielx> I've been thinking about business models that should make it possible to incorporate developers working on haiku development fulltime
[22:49:29] <nielx> but as I'm not in the sector myself, I don't have a complete picture of the market and the required services
[22:49:54] *** TLF has joined #haiku
[22:50:30] <tic> nielx, how?
[22:50:30] <[Beta]> web dev would give you time ?
[22:51:07] *** Fanskapet has quit IRC
[22:51:14] *** Fanskapet has joined #haiku
[22:51:27] <nielx> Well, it's obvious that donations by itself will not provide a constant stream of income for the Haiku organisation itself
[22:51:40] <nielx> to sustain one or more permanent (full-time) developers
[22:52:25] <nielx> So you would economically need a solid stream of revenue to sustain developers
[22:53:00] <tic> yes.
[22:53:01] <nielx> The most logical step is to provide ICT services for companies
[22:53:05] <tic> what is ICT?
[22:53:17] <nielx> is it dutch?
[22:53:30] <nielx> information communication technology
[22:53:37] <nielx> (or something like that)
[22:53:43] <tic> okay, basically IT. :)
[22:53:48] <nielx> yep
[22:53:49] <tic> so, what -is- it? :)
[22:54:16] <nielx> well, I've been thinking about it. We can basically divide the businesses in three types based on size
[22:54:29] <nielx> four actually
[22:54:34] <tic> go ahead
[22:55:29] <nielx> one-man businesses, small-size businesses (I won't number it, because I don't realy envision it, but let's say 2-30 employees), medium size businesses, and large corporations
[22:55:55] <tic> yes.  and in which way does this has to do with Haiku?
[22:55:58] <nielx> the one-man businesses are of no use- these basically can co-ordinate all their it-work themselves
[22:56:28] <nielx> medium size businesses and large businesses have specific requirements in regard to it-needs
[22:56:46] <nielx> The small-size businesses fall into a gap
[22:57:20] <nielx> it might be a big overhead to hire external work to do maintainence on your pc's (security, helpdesk, etc)
[22:57:40] <nielx> on the otherhand, if you have about 15 machines to take care off, it becomes a rather big job
[22:57:57] <tic> yes.
[22:58:32] <nielx> And I believe that Haiku would basically cater such companies (ease of use, low maintainence, good desktop, etc)
[22:59:05] <tic> but then all of a sudden we realize that they use Office for everything.
[22:59:18] *** thies_ has joined #haiku
[22:59:19] <tic> and especially Outlook. While MDR is good, it's nowhere as capable.
[22:59:21] <tic> hey thies
[22:59:58] <nielx> So the trick is to come up with a business model where you supply and maintain the workstations for your customers from a central position, keep them updated and maintained, and provide a single point of it-supplier
[23:00:05] <nielx> (hardware, but also software)
[23:00:19] <nielx> That all for a reasonable price (free software)
[23:00:45] *** brennanOS has left #haiku
[23:01:02] <tic> yes.
[23:01:05] *** thies has quit IRC
[23:01:08] <tic> it's indeed a good idea.
[23:01:17] <nielx> When you actually come to think of combining it-services into tight packages - so individual companies basically can pick any of the standard services they might rquire - you keep the overhead low
[23:01:20] <tic> the key points that needs to be solved are:  Office, Outlook and Windows Networking.
[23:01:27] *** thies_ is now known as thies
[23:01:28] <NathanW> This is true
[23:01:31] <NathanW> Especially Office
[23:01:40] <nielx> okay
[23:01:41] <NathanW> Simply because the software is larger, and thus more complicated
[23:01:44] <nielx> I'll answer those
[23:01:54] <NathanW> And we have no good replacement at the moment
[23:02:08] <tic> note: it needn't be the exact thing, but they need to have the same functionality, and ease of use. For example, Outlook - a calendar, inbox and contacts.
[23:02:22] <nielx> First of all: Windows Networking - the idea is to have a complete managed it-solution for every company
[23:02:43] <tic> nielx, yes. But we need something like this, anyway.  The computers need to communicate. FTP is not good enough.
[23:02:46] <tic> and it needs to be as fast
[23:02:46] <nielx> so you basically replace windows networking with a haiku equivalent
[23:02:53] <tic> and it has to support auto-discovery.
[23:02:53] <tic> sure.
[23:03:05] <tic> it has to open all kinds of attachments in e-mail without hazzle
[23:03:24] <nielx> true
[23:03:26] <nielx> all true
[23:03:50] <tic> Don't get me wrong here; I've been using BeOS 100% for six years no.
[23:03:51] <tic> now.
[23:03:53] <nielx> so the problem is the chicken and the egg one
[23:03:57] <tic> not really.
[23:04:06] <tic> developers needn't have users before they start working on an application, such as Office
[23:04:12] <tic> except they won't get payed. ;)
[23:04:20] <tic> paid
[23:04:25] <nielx> There's the rub
[23:04:34] <NathanW> tic: What do you mean about the attachment thing?
[23:04:46] <NathanW> Just that we should support all relevant formats of word docs/etc.
[23:04:47] <NathanW> ?
[23:05:26] <nielx> I guess that's what he meant
[23:06:02] <tic> nielx, exactly. Sorry, wasn't clear. Incoming e-mail, simply.
[23:06:12] <tic> the ease of use we already have, once you've learned it.
[23:06:21] *** BGA has quit IRC
[23:06:22] <nielx> Anyway, I need to find out how badly small businesses need these kinds of solutions
[23:06:25] <tic> mhm.
[23:06:41] 
[23:06:43] <tic> Documents.
[23:06:43] <nielx> and how to get to the point of actually being able to offer these solutions
[23:07:00] <tic> needn't be as brain-dead w/ regard to My Images, My Music, but users generally tend to wanting to be held in their hands.
[23:07:06] <tic> (+grammar)
[23:08:08] <nielx> That's going into details, I guess that when you have two or three codemonkeys playing with haiku, oo, and firefox you'll have things what your clients require
[23:08:25] <nielx> so I can basically identify three problems
[23:08:45] <nielx> 1. I need to know what markets there are and how viable the idea is
[23:09:23] <nielx> 2. Is the level of technology predictable enough to be able to supply these kinds of solutions profitably (I know it definately won't be with windows)
[23:09:43] <nielx> 3. How to get to a point where you actually can offer the services
[23:10:40] <nielx> Anyway, I'm going for a bath (thinking about it)
[23:10:53] <nielx> bbl
[23:11:55] <tic> an 08 22:11:55 <tic>	a first step is probably to get Haiku working, at all..
[23:12:23] <tic> or if you by haiku mean the components, then somehow license Zeta and specific Haiku components (heh, that aren't already in Zeta, that is.)
[23:14:41] <NathanW> yeah
[23:14:45] <NathanW> This is true
[23:15:14] <NathanW> I think Haiku has a shot (and a good one) at being what desktop Linux wants to be
[23:15:39] <NathanW> If we can get that far, we're in good shape, I think
[23:17:52] <tic> *nods*
[23:18:46] <tic> computer hardware is cheap nowadays, and one can use even cheaper hardware when dealing with BeOS/Haiku
[23:19:00] <NathanW> yeah
[23:19:03] <tic> who knows, in the future it might even be feasible to have some kind of BeOS Terminal Server and extremely thin clients?
[23:19:13] <tic> (even for small businesses)
[23:24:54] *** fyysik has joined #haiku
[23:25:12] <NathanW> hi fyysik
[23:25:20] <fyysik> hi NathanW
[23:27:24] <ShackaN> http://linux.slashdot.org/article.pl?sid=05/01/08/0311257&from=rss
[23:42:59] <nielx> re
[23:43:06] <nielx> I actually don't think Zeta is a good choice
[23:43:21] <nielx> (but again, that's implementation)
[23:43:51] <fyysik> Zeta? for what?
[23:44:49] <nielx> Provide generic subscription based it-services to small companies
[23:45:03] <nielx> And, in the process, be able to hire some haiku developers fulltime
[23:45:09] <tic> nielx, why not Zeta?
[23:45:39] <nielx> I don't think it will provide a clean upgrade path to Haiku
[23:46:02] <nielx> plus the questionable legality
[23:47:03] <@mmu_man> legality of what ?
[23:51:02] *** Tenzin has joined #haiku
[23:53:44] <Tenzin> Karina pyCube updated htmlchat and it works very nicely in any browser with javascript and css
[23:53:44] <Tenzin> http://beshare.escholaracademy.net:8000 ... thought you might want to try it in OS X ... I've been running it in gnustep suse
[23:53:44] *** slaad has joined #haiku
[23:56:32] *** Greyhawk_Dragon has quit IRC
[23:57:10] <fyysik> damn, why any soft which tends to be commercials, turns into unusable slow grady "cools" software?
[23:57:18] <tic> moo slaad.
[23:57:30] <tic> fyysik, that's the way things are.
[23:57:36] <tic> eye candy appeals to customers.
[23:57:42] <tic> that they can't actually use the software is another thing.
[23:57:50] <slaad> Meep, meep.
[23:58:03] <fyysik> ik>	i needed some ICQ client for Windows, recalling that i used Trillan years ago, mamma mia!
[23:58:08] <fyysik> Horrorful!
[23:58:46] <Fanskapet> miranda-icq!!
[23:58:48] <fyysik> lamost reinstalled Windows as result, lost old Trillain contacts etc tec
[23:59:00] <fyysik> yeah, resulted in open-source miranda
[23:59:01] <@Korli> i>	i like Trillian
[23:59:10] <fyysik> Compact, adequate
[23:59:16] <Fanskapet> yuck.. Trillian is grose..
[23:59:22] <Fanskapet> miranda is light and easy
[23:59:34] <fyysik> Korli - not on 300MHz/128MB laptop
[23:59:56] <@Korli> fyysik it's not aimed at PocketPC you know





top