April 13, 2007 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >
April 13, 2007 [00:05:11] *** Witukind has quit IRC[00:08:59] *** guma has joined #haiku[00:11:00] *** JamesB192 has joined #haiku[00:11:03] *** JamesB192 has quit IRC[00:11:41] *** JamesB192 has joined #haiku[00:17:44] *** fanskap has joined #haiku[00:18:46] <nonesuch> I see there is (at least) preliminary support for ACPI in Haiku. Couldn't find anything recently posted about it, but is that going to be workable by R1 or is it beyond that?[00:19:32] *** Witukind has joined #haiku[00:19:38] <nonesuch> particularly battery monitor and sleep support...I haven't had a desktop in 5 years, and no interest in ever having one again[00:22:55] *** Fanskapet has quit IRC[00:23:06] <stargater> ACPI cames with the driver api i think[00:28:01] <kokito> nonesuch, I think it was originally planed for post R1 :)[00:35:22] *** kr1stof_ has quit IRC[00:36:33] <cnuke> did anybody boot haiku on a ibm x21 with the lastest revisions successfully?[00:41:07] *** JamesB192 has quit IRC[00:41:49] *** JamesB192 has joined #haiku[00:42:21] *** JamesB192 has quit IRC[00:42:45] *** JamesB192 has joined #haiku[00:47:47] *** DaaT has quit IRC[00:48:18] *** kb7sqi has joined #haiku[00:50:27] *** Witukind has quit IRC[00:52:54] *** jonaskirilla has joined #haiku[00:54:06] *** jonaskirilla has quit IRC[00:57:08] <Yez> anyone around that knows what is up with svn checkout?[00:57:53] <Yez> I would like to checkout some source so I can build the ATI Radeon driver for a buddy using BeOS/Haiku[00:59:13] <stargater> Yez, http://haiku-os.org/documents/dev/getting_the_source_code[00:59:21] <Yez> thanks[00:59:45] <stargater> cnuke, ibm x21 nice , i like the ibm thinkpads[00:59:53] <stargater> 12" 13"[00:59:57] <cnuke> 12"[01:00:01] <cnuke> of course it is nice[01:00:07] <cnuke> but haiku doesnt boot[01:00:25] <Yez> man, I can't use my cool Tortoise GUI interface to get the Haiku source?[01:00:33] <emitrax> hmmm[01:00:36] *** Sarcas has quit IRC[01:00:45] *** guildencrantz has quit IRC[01:01:50] <cnuke> stargater: afaik the lenovo thinkpads are worse than the good old ibm ones[01:02:23] *** guma has quit IRC[01:04:41] *** MauriceK has quit IRC[01:05:50] *** meianoite has joined #haiku[01:06:00] *** whoopsbob has quit IRC[01:06:11] *** bobwhoops has joined #haiku[01:06:17] <meianoite> any assembly whizzos around?[01:07:17] <cnuke> meianoite: x86 or ppc?[01:07:20] <emitrax> can somebody help me understand how is the actual data sent to a USB device? I'm following the USB stack code but I get stuck at the function that fills the queue. I'm not clear on how the data is sent out to the device.[01:07:36] *** rcjsuen has quit IRC[01:07:49] <emitrax> I must say that I haven't read the USB spec yet, so if there is the answer please forgive me :)[01:08:17] <meianoite> cnuke makes absolutely no difference[01:08:56] <stargater> cnuke, true[01:11:56] <stargater> hi emitrax you are from italy , i see in the ML[01:12:05] <emitrax> stargater: yep[01:12:53] <stargater> emitrax, = http://www.itbug.org/ and http://www.iltuosistema.it/ // for relexing[01:13:01] *** guma has joined #haiku[01:13:10] <emitrax> I know both of them already ;-)[01:13:22] <stargater> ah ok :-)[01:14:23] *** MrSunshine has quit IRC[01:14:49] <stargater> so i go sleep , n8[01:14:59] *** stargater has quit IRC[01:15:02] <emitrax> so do I[01:15:03] <emitrax> night[01:15:06] *** emitrax has left #haiku[01:15:10] *** jevin has joined #haiku[01:21:10] *** Mazon is now known as mazon[01:23:51] *** the_faulkenator has left #haiku[01:31:45] *** Anxiety has quit IRC[01:35:21] *** eNGIMa___ has joined #haiku[01:35:42] *** eNGIMa has quit IRC[01:40:40] *** JamesB192 has quit IRC[01:42:32] *** pikapika has quit IRC[01:47:37] <_hugo> meep[01:48:07] <DeadYak> beep[01:48:19] *** avkig has joined #haiku[01:49:52] *** the_webers_inc has joined #haiku[01:51:13] *** the_webers_inc_ has joined #haiku[01:51:15] *** the_webers_inc has quit IRC[01:52:46] *** the_webers_inc has joined #haiku[01:52:47] *** the_webers_inc_ has quit IRC[01:53:14] *** Fanskapet has joined #haiku[01:55:18] *** the_webers_inc_ has joined #haiku[01:57:21] *** fanskap has quit IRC[02:17:13] *** meianoite has quit IRC[02:17:38] *** the_webers_inc has quit IRC[02:19:03] *** the_webers_inc_ has quit IRC[02:19:22] *** MikeW has joined #haiku[02:19:47] <infamy> boing[02:20:31] *** _Lucretia is now known as _Lucretia_[02:26:20] *** Mike2K has joined #haiku[02:33:45] *** Mike2K has quit IRC[02:34:10] *** evdubs has joined #haiku[02:40:19] *** infamy has quit IRC[02:42:52] *** MikeW has quit IRC[02:47:59] *** SiCuTDeUx has joined #haiku[02:50:06] *** cnuke has quit IRC[02:55:57] <CIA-17> hugosantos * r20672 /haiku/trunk/ (7 files in 4 dirs):[02:55:57] <CIA-17> introduced a new helper class DatagramSocket which provides a consistent base interface and functionality for the implementation of datagram-based sockets.[02:55:57] <CIA-17> - made the ipv4 raw, udp and link protocols use DatagramSocket.[03:22:18] *** rcjsuen has joined #haiku[03:22:22] *** kokito has quit IRC[03:25:21] *** Fanskapet has quit IRC[03:39:10] *** TTRanger has quit IRC[04:05:56] *** Bubba_I has joined #haiku[04:08:41] *** guma has quit IRC[04:13:55] *** nonesuch has quit IRC[04:14:10] *** Sloar has quit IRC[04:14:24] *** Bubba_I has left #haiku[04:18:47] *** Anxiety has joined #haiku[04:20:15] *** Kernel86 has quit IRC[04:36:23] *** phrizek has joined #haiku[04:47:14] <Anxiety> Looking at the source code for Haiku, I'm reminded why I loved BeOS so much[04:50:05] <geist> yeah?[04:50:48] <Anxiety> How clean the API is compared to a lot of others at the time... especially compared to MFC[04:51:28] <Anxiety> Sure it had its weak points but it seemed so much cleaner and more intelligible[04:52:07] <Anxiety> The whole OS gave me that feeling[04:52:53] *** nonesuch has joined #Haiku[04:53:00] <Anxiety> It was like a almost perfect blend of simplicity and power[04:53:15] <Anxiety> an*[04:54:09] * _hugo points at 900 pages of Be Bible[04:54:32] * _hugo hugs[04:54:52] <Anxiety> I had the Be Bible in paperback... I'm still hitting myself for losing it[04:56:10] <Anxiety> I lost it in a move some years ago... I must have left a box behind[05:03:45] <Anxiety> I have a heap implementation that is based off the heap from NewOS that I've developed on and off, originally for my own little hobby OS[05:04:15] <Anxiety> I'm going to see about changing it from C++ back into C and put it into Haiku's coding style and see if its worth anything[05:06:41] *** RedLobster has joined #haiku[05:06:43] <RedLobster> http://spreadsheets.google.com/pub?key=pXPf6D8HwIWmRyV6l1MTaAw[05:06:58] <CIA-17> hugosantos * r20673 /haiku/trunk/ (8 files in 7 dirs): added buffer, datalink and socket module references to the stack module so we don't have to load them in each other module.[05:07:26] <RedLobster> Results of Haiku icon set election. 29.25% of voters strategically voted approval style. Humorously, 24.01% bullet voted, plurality style. A _bad_ strategy.[05:07:58] <RedLobster> So only 4.24% were "smart" strategists. :)[05:08:32] <Anxiety> heh[05:09:51] <Anxiety> actually, nevermind, the heap has been significantly improved since I last looked at it :)[05:10:02] <Anxiety> </senseOfImportance>[05:27:32] <evdubs> hey geist[05:28:11] <evdubs> have you guys decided what to do with the VM reentrancy issues?[05:30:01] <evdubs> Anxiety - userland heap?[05:30:11] <Anxiety> no, kernel heap[05:30:43] <evdubs> ooo... so what exactly are you going to change?[05:31:47] <Anxiety> nothing at the moment, I'm going to study the current one more, see if there is even anything I could add to it[05:32:05] <Anxiety> I last checked it quite a long time ago, I guess it was dumb to think it was still the same :)[05:33:40] <evdubs> the slab allocator still has not been implemented[05:34:12] *** meianoite has joined #haiku[05:34:16] <evdubs> axel is going to start at some point porting the FreeBSD stuff[05:34:47] <evdubs> i *ported* the opensolaris implementation but it was decided recently that the licenses would be incompatible[05:35:01] <meianoite> yo[05:35:04] <Anxiety> hmm[05:35:14] <Anxiety> I'll look into that then[05:35:19] <Anxiety> as well, I mean[05:35:27] <RedLobster> So to be clear, as new stuff is added to the Haiku API, say in Haiku 2 or whatever, will it break with the convention of the preceeding b_ ? like would a new function be, "haiku_receive_message"?[05:35:35] <evdubs> so, there are other things that can be worked on[05:35:47] <_hugo> RedLobster: that is irrelevant right now, please don't proceed with that discussion[05:35:53] <evdubs> you should probably talk to geist... right now the heap pages are wired[05:36:20] <evdubs> so it doesn't really do paging right[05:36:45] <evdubs> (with the heap, that is)[05:36:56] <Anxiety> yeah, I got what you meant[05:39:20] <meianoite> RedLobster I can't see why that would be necessary, the Cocoa API still uses NS* prefix[05:40:03] <meianoite> who's Anxiety? a GSoC colleague?[05:40:10] <Anxiety> no[05:40:17] <meianoite> sorry :)[05:40:18] <DeadYak> meep[05:40:26] <Anxiety> just a hobby person :)[05:40:37] * meianoite coyotes DeadYak[05:40:46] <Anxiety> decided to devote some time to Haiku after the whole Zeta thing broke[05:40:53] <DeadYak> :)[05:40:56] <meianoite> I just turned "coyote" into a verb. how cool is that :)[05:41:08] <evdubs> but yea, I would definitely recommend talking to both axel and geist about what exactly needs to be done with the VM/heap[05:41:12] <meianoite> Anxiety you know? I've noticed that behaviour lately[05:41:23] <meianoite> quite some buzz around Haiku since Zeta dropped dead[05:41:41] <DeadYak> well, it doesn't help that it happened right around the GSoC announcements :)[05:41:44] <Anxiety> I wouldn't be surprised... It's pretty much BeOS' last best hope now :)[05:42:01] <meianoite> you people were supposed to be helping Haiku for a long time now :)[05:42:29] <Anxiety> Better late then never[05:42:32] <meianoite> instead of throwing cash at the Zeta crew[05:42:41] <RedLobster> meianoite wow. that's crazy.[05:42:50] <Anxiety> Actually, I've never bought Zeta[05:43:07] <meianoite> (no personal offences toward mmu_man, BGA, and everyone else)[05:43:07] <Anxiety> It wont work with my computer, as far as I can tell anyways[05:43:15] <DeadYak> meianoite: none taken :)[05:43:19] <RedLobster> i'm excited to see what GSoC brings.[05:43:34] <evdubs> hopefully a bunch of completed projects ;)[05:43:42] <meianoite> DeadYak were you with yT as well?[05:44:26] <DeadYak> meianoite: yes[05:44:58] <meianoite> I just hope there won't be any legal consequences towards YOU people[05:45:12] <meianoite> since you had access to the code and whatnot[05:45:23] <DeadYak> doubtful[05:45:31] <DeadYak> until that's clarified though, I'm just a spectator here[05:45:36] <CIA-17> hugosantos * r20674 /haiku/trunk/src/add-ons/kernel/network/protocols/tcp/TCPEndpoint.cpp: don't do MSS segmentation in TCP's SendData(), _SendQueue() already does the appropriate work.[05:45:51] <DeadYak> at least in any way other than general API knowledge[05:46:11] <Anxiety> Well, I hope that clears up in a positive way soon[05:46:16] <DeadYak> should do[05:46:26] <meianoite> I'm glad I NEVER peeked at the Be code[05:46:36] <meianoite> there's some bad juju on it[05:46:44] <DeadYak> meh, even knowing it's more or less useless in the context of haiku[05:46:50] <meianoite> everyone who touches it get screwed ;)[05:46:53] <DeadYak> architecture's too different[05:47:28] <meianoite> on an unrelated note[05:47:35] <meianoite> http://code.google.com/soc/mofo/index.html[05:47:42] <DeadYak> mofo? o.0[05:47:52] <DeadYak> 404[05:47:55] <meianoite> ooooh, finally[05:48:04] <meianoite> I meant http://code.google.com/soc/mofo/about.html[05:48:09] <meianoite> it's goddamn Mozilla[05:48:21] <meianoite> and it was blank up until a few hours ago![05:48:29] <DeadYak> that shortened name is awesome[05:48:32] <meianoite> which was somewhat unbeliavable[05:49:06] <meianoite> isn't it amazing that they got 10 slots, and WE got 8? :D[05:49:07] * JonathanThompson lobs a flaming hairball into the chat room[05:49:20] <meianoite> granted, they have gobs of resources already[05:49:39] <meianoite> but so do the KDE people, who got 40 slots =P[05:49:52] <JonathanThompson> Let's face it: there's probably more people that use Mozilla than is likely that there will ever be Haiku users in the next few years, meianoite.[05:50:10] <meianoite> JonathanThompson I meant that in a VERY positive tone[05:50:13] <DeadYak> JT: I think he's more amazed that we got as many as we did[05:50:18] <JonathanThompson> So the fact that Haiku got 8 and Mozilla only got 10 says something interesting.[05:50:21] <DeadYak> JT: considering that a project many times larger than us got 10[05:50:24] <meianoite> it's like we got 80% the attention spam that Mozilla has[05:50:44] * _hugo wishes he had a Be tshirt[05:50:45] <JonathanThompson> Attention spam? That's a term I've not heard/read before :P[05:50:51] * DeadYak hides his Be tshirt[05:50:57] * _hugo grabs DeadYak's[05:51:00] <meianoite> _hugo *I* have a Be t-shirt ;)[05:51:02] * JonathanThompson unburies DeadYak's tshirt[05:51:03] <DeadYak> that's why I hid it[05:51:21] <meianoite> unfortunately it's pretty worn out already =P[05:51:26] <DeadYak> mine too :/[05:51:28] <meianoite> horrid silking work.[05:51:33] <_hugo> i found my Be Bible, which someone, i don't remember who exactly anymore, bought me back in the day[05:51:36] <_hugo> some 900 pages of love[05:51:42] <JonathanThompson> Hey, if anyone has sufficient experience and are interested in working for Yahoo! in the Seattle area, they're in really short supply of workers :)[05:52:28] <_hugo> isnt Seattle rainy?[05:52:41] <JonathanThompson> Only part of the time.[05:52:49] <JonathanThompson> And it's mostly misty more than anything, not heavy rain.[05:53:02] <_hugo> hm :-)[05:53:06] <Anxiety> I live not far from Seattle, but I dont want to move or commute[05:53:06] <JonathanThompson> It's a tame climate, really: rarely hot or cold and not very humid.[05:53:30] <JonathanThompson> I live on Mercer Island atm, and the Yahoo! office is along I-90 in Bellevue atm.[05:53:41] <_hugo> JonathanThompson: that makes no sense.[05:54:07] <meianoite> JonathanThompson didn't you say something about Detroit last night?[05:54:08] <JonathanThompson> It makes perfect sense, since I-90 goes over Mercer Island.[05:54:24] <JonathanThompson> That's the area I grew up in, meianoite, and I have family around there still.[05:54:25] <meianoite> I assumed you lived there[05:54:30] <meianoite> ah, that explains it.[05:54:31] <_hugo> JonathanThompson: eheh, i was just messing with you[05:54:38] <JonathanThompson> I lived in Rochester, MI until I was 20.[05:54:45] <meianoite> I know nothing about Detroit[05:54:48] <meianoite> except for Robocop ;)[05:54:55] <JonathanThompson> I thought perhaps you were messing with yourself, _hugo :)[05:55:04] <DeadYak> hahahaha[05:55:11] <_hugo> 0:-)[06:00:47] *** ormandj has joined #haiku[06:01:12] <ormandj> hiya :)[06:01:21] <meianoite> hey[06:01:31] <JonathanThompson> Howdy, ormandj.[06:01:39] <ormandj> how goes it?[06:01:51] <JonathanThompson> Almost completed 2 weeks at Yahoo, and signs are sufficiently promising.[06:01:59] * DeadYak prods ormadj[06:02:00] <DeadYak> er[06:02:01] <ormandj> excellent news![06:02:02] <DeadYak> ormandj*[06:02:06] <ormandj> omg, hi bud[06:02:09] <ormandj> ltns :p[06:02:12] <DeadYak> likewise[06:02:18] <ormandj> how're you doing?[06:02:23] <DeadYak> busy busy, yourself?[06:02:28] <ormandj> same man, same[06:02:43] <JonathanThompson> I don't know how long I'll have to wait before they decide one way or the other; a contractor that was hired before me was terminated after about 4 weeks, but I learned he was rather incompetent, and simply was unteachable.[06:04:19] <JonathanThompson> Oh, ormandj, if you're interested in a change of scenery, they're looking for people there :P[06:04:26] <ormandj> JonathanThompson: well, hopefully you'll get sorted ;)[06:04:27] <ormandj> oh[06:04:29] <ormandj> about that[06:04:31] <ormandj> :p[06:04:32] * ormandj pms[06:04:36] <JonathanThompson> They're in a position where there's far more need of people than they can find that are qualified.[06:05:19] <evdubs> this is for a generic software engineer?[06:09:44] *** eNGIMa___ has quit IRC[06:10:38] *** rcjsuen has quit IRC[06:20:49] <CIA-17> hugosantos * r20675 /haiku/trunk/ (2 files in 2 dirs): only deliver frames to IPv4 raw sockets of the specified protocol.[06:26:32] *** Sloar has joined #haiku[06:34:23] <Anxiety> ok... this is going to sound like a dumb question... should I be trying to build Haiku from a x86_64 build of Ubuntu?[06:35:19] <evdubs> you shouldn't notice any problems... i don't think.[06:35:51] <evdubs> gcc shouldn't be building 64-bit haiku binaries[06:36:01] <meianoite> geist you don't happen to be around, do you? :) your assembly expertise is very appreciated[06:36:04] <Anxiety> well, I just got some issues trying to build the cross-compiler... and I want to find out ahead of time before I try and fix the problem[06:36:28] <evdubs> hmm.. i think we still have some old documentation lying around[06:36:59] <evdubs> because i've run into problems building the cross-compiler and i was told i didn't need to do it[06:39:12] <evdubs> you did grab the haiku-compatible jam, right? and you're using http://haiku-os.org/documents/dev/getting_the_source_code for your instructions, right?[06:40:12] <Anxiety> yeah[06:40:19] <Anxiety> jam compiled fine and seems to work[06:41:05] <Anxiety> let me try something real quick[06:50:50] *** nonesuch has quit IRC[06:56:55] <Anxiety> nope, no luck[06:56:56] <Anxiety> hmm[06:59:18] <geist> Anxiety: I wasn't able to get it to work from my x86-64[06:59:25] <geist> some of the build tools fail to build[06:59:32] <geist> and i think it's well understood that it's screwed up[06:59:50] <Anxiety> well then... I'm off to download the i386 iso :)[07:00:12] <Anxiety> bbiab[07:00:12] <geist> i spent some time on it the other day and it didn't look like it was gonna be easy to fix[07:00:38] <Anxiety> Im not going to bother then... Its not like I plan on using Ubuntu that much[07:01:10] *** Anxiety has quit IRC[07:04:12] *** Anxiety has joined #haiku[07:05:05] <Anxiety> Do you think it would be too absurd to run Linux in a VM just for the purpose of compiling haiku?[07:05:32] <ormandj> no[07:05:53] <ormandj> i run solaris in a VM just for the purpose of development[07:05:56] <ormandj> on my macbook[07:06:05] <ormandj> so i don't see why running linux would be any diff[07:06:24] <Anxiety> is there any particular distro that is easiest to work with Haiku?[07:06:33] <geist> i have linux in a VM on my OSX box largely for that purpose[07:07:54] *** kokito has joined #haiku[07:08:42] *** nonesuch has joined #Haiku[07:09:12] <kokito> howdy[07:12:19] *** kr1stof has joined #haiku[07:17:34] <JonathanThompson> Howdy, kokito.[07:18:56] *** wildur has joined #Haiku[07:20:58] <meianoite> hey geist[07:21:06] <meianoite> mind if I ask you some asm stuff? :)[07:23:13] <geist> okay[07:23:53] <meianoite> remember our discussion regarding expanding the bitmap to handle >32 values?[07:24:19] <meianoite> I'm trying to find an efficient way to do it.[07:24:26] <meianoite> so far, here's what I've got[07:24:51] <meianoite> > ] cat bitmap.cpp[07:24:51] <meianoite> extern int prio;[07:24:51] <meianoite> extern int bitmap[4];[07:24:51] <meianoite> /* Bit width of the native register , minus 1*/[07:24:51] <meianoite> #define MASK 31[07:24:52] <meianoite> /* Log2(MASK+1) */[07:24:54] <meianoite> #define SHIFT 5[07:24:57] <meianoite> void set_bit() {[07:24:58] <meianoite> bitmap[prio>>SHIFT] |= 1 << (prio & MASK);[07:25:00] <meianoite> }[07:25:04] <meianoite> void clear_bit_a() {[07:25:06] <meianoite> bitmap[prio>>SHIFT] ^= 1 << (prio & MASK);[07:25:09] <meianoite> }[07:25:10] <meianoite> void clear_bit_b() {[07:25:12] <meianoite> bitmap[prio>>SHIFT] &= (MASK ^ (1 << prio));[07:25:14] <meianoite> }[07:25:16] <meianoite> void clear_bit_c() {[07:25:18] <meianoite> bitmap[prio>>SHIFT] &= ~(1 << (prio & MASK));[07:25:20] <meianoite> }[07:25:39] <meianoite> they're wrapped into functions just to have something compilable, so nevermind that[07:26:22] *** dsg has quit IRC[07:26:31] <geist> mkay[07:26:51] <meianoite> both setting and clearing (all 3 versions) compile into 7 x86 asm instructions[07:26:55] <meianoite> as in gcc -march=pentium-m -O3 -fomit-frame-pointer -c -S bitmap.cpp[07:27:07] <meianoite> is there a more efficient way? =P[07:27:13] <geist> no[07:27:23] <meianoite> feh. =)[07:27:26] <geist> and that is absolutely totally a microoptimization[07:27:32] <geist> use clear_bit_c[07:27:39] <_hugo> yes, that completely doesnt matter[07:27:40] <geist> that's the generally defined way to clear a mask[07:28:03] <meianoite> ok.[07:28:22] <geist> and like I said before, sicne there's a load/store there (at least on non x86s), that'll probably dominate the entire thing anyway[07:28:25] *** Fanskapet has joined #haiku[07:28:28] <meianoite> I actually read how FD_SET and FD_CLR works AFTER writing those[07:28:41] <_hugo> this reminds me of an optimization in Linux, where some guys did several new versions of a cubit root implementation for something that is seldom called[07:28:48] <meianoite> and indeed they resemble clear_bit_c[07:28:53] <_hugo> cubic root even[07:29:13] <_hugo> meianoite: worry about algorithm complexity, not optimizations[07:29:22] <geist> and all of that aside, you want to make mask be configurable[07:29:29] <meianoite> I was actually not worrying about neither[07:29:30] <geist> and calculate shift off of that[07:29:33] <geist> so for a 64-bit machine[07:29:38] <meianoite> just trying to understand HOW to do it...[07:29:39] <geist> it'll use a 64bit thing[07:29:51] <meianoite> geist hence the defines ;)[07:29:59] <meianoite> but that's not production code whatsoever, just a quick test[07:30:05] <geist> to be picky, you can derive one from the other[07:30:06] <meianoite> to see what the compiler would do[07:30:14] <geist> ((1 << SHIFT) - 1) == MASK[07:31:05] <meianoite> well, I DID define one based on the other[07:31:08] <meianoite> but manually ;)[07:31:16] <meianoite> *only manually :)[07:31:21] <geist> anyway[07:31:33] <meianoite> it's not like I was TRYING to optimize anything[07:31:47] <meianoite> actually just figuring something that would WORK...[07:31:54] <meianoite> then which one sucked less ;)[07:32:03] <geist> well, except when you come at it spouting how many instructions it compiled to, that seems to be thinking about optimization :)[07:32:16] <geist> and anyway a good compiler will substitute any of those for another[07:32:19] <meianoite> ah, sorry.[07:32:25] <meianoite> not my intention.[07:32:35] <_hugo> dont worry about having those in functions as well, the compiler inlines it all[07:32:38] <meianoite> it was all about finding the one that sucked less, not bit combing[07:32:41] <geist> http://newos.org/WebSVN/filedetails.php?repname=extern&path=%2Flk%2Ftrunk%2Finclude%2Fbits.h&rev=0&sc=0[07:32:52] <geist> i use these macros for a bunch of stuff occasionally[07:33:19] <meianoite> _hugo: "[02:25] <meianoite> they're wrapped into functions just to have something compilable, so nevermind that"[07:33:23] <geist> mostly used in the arm emulator, where it's masking out chunks of instructions[07:33:24] <meianoite> GMT-3 ;)[07:33:59] <_hugo> meianoite: from that sentence i understood that eventually you wouldnt want them "wrapped into functions"[07:34:47] <meianoite> well, as in... "I would write them as macros, actually; that was just a quick hack so GCC would compile it, since CPP would do nothing at all =P"[07:35:19] <geist> right, and by compiling them into discrete functions you also remove all chance of the compiler being able to interleave it with other code[07:35:25] <_hugo> meianoite: you can write them as functions as well. they are so small the compiler will inline them. it will have the same result as macros[07:35:36] <geist> that eing since, since it involves a read/write from an array, it probably wont schedule it much[07:35:49] <geist> as soon as you do pointer dereferencing, compilers tend to have to serialize everything[07:35:54] <geist> it's the curse of C[07:36:05] *** Anxiety has quit IRC[07:36:10] <meianoite> _hugo I'd wager GCC 2.95.3 is not THAT smart... unless told so. but it may break some stuff along the way; GCC 2.9x was a bit fragile in this regard[07:36:29] <geist> at work I've done some fun scheduling with gcc on the cell to make it unroll stuff[07:36:32] <_hugo> geist: it really depends on the scope. if gcc knows all paths it is usually able to inline it just fine[07:36:37] <geist> but you have to be pretty goo dabout __restrict on pointers[07:37:22] <_hugo> meianoite: well, gcc 2.95 is kinda weak in several regards, yes. my points were more regarding >= gcc 4[07:37:35] <meianoite> I should probably read up on GCC intrinsics..[07:37:53] *** Anxiety has joined #haiku[07:37:59] <geist> there arent a lot i'd recommend in generic code[07:38:03] <meianoite> but as geist said, that's all quite unnecessary microoptimizations[07:38:19] <meianoite> it's just the way I happened to write them in the first place[07:38:19] <geist> for vector stuff and whatnot, it's a big deal, but you're generally platform specific by the time you get to that code[07:38:36] <_hugo> one of the next gcc versions (4.3 or 4.4 i dont remember) will even be able to do whole program optimization, inline stuff from different object files[07:38:37] *** nonesuch has quit IRC[07:38:41] <_hugo> *inlining stuff[07:38:42] <meianoite> so I wasn't wasting time trying to comb bits... it came naturally as such[07:39:18] <_hugo> or maybe after 4.4 :-) dont really remember the version[07:39:30] <geist> _hugo: 4.1 already has some of that[07:39:37] <geist> gcc -combine -whole-program[07:39:47] *** fanskap has joined #haiku[07:39:55] <meianoite> _hugo that was the sad part about splitting the scheduler into several files... GCC up to 4.xx just blindingly ignores inlining opportunities[07:39:56] <_hugo> ah, cool. so it is probably to extend that[07:40:07] <geist> it effectively makes everything static (except symbols you override and some defaults like main())[07:40:11] <_hugo> meianoite: add the stuff to an header then[07:40:26] <geist> it's more handy for embedded stuff, but my experience has been that it's pretty good[07:40:36] <meianoite> I'm almost about to suggest some concatenation magic among scheduler.cpp, thread.c and the ilk =P[07:40:37] <geist> downside is you have to compile all the code at once[07:40:51] <geist> meianoite: the scheduler used to be inside thread.c[07:40:58] <_hugo> geist: good to hear. i haven't tried it yet since i thought it would come in a next version, didn't consider it might already be in, at least part[07:40:59] <meianoite> yeah, you told me so[07:41:15] <meianoite> and geist[07:41:17] <geist> as part of the 'fix' that was put in that screwed it all up, it was moved to it's own file, made to be .cpp, and generally screwed up[07:41:39] <meianoite> axeld is the one to blame for turning the O(1) code from NewOS into O(n) code ;)[07:41:42] <geist> one of the things that generally screwed it up was the fact that it couldn't inline some stuff[07:41:58] <meianoite> I really need to ask him if there was some good reason for doing that[07:42:09] <geist> I do remember that early on some folks were dismayed at some of my big .C files and wanted to break them up because they felt too large[07:42:42] <geist> which i really could care less about. I grouped the stuff based on it's locality, and use of static functions, so I could keep the intermodule dependencies down[07:42:52] <geist> one thing you'll find about newos is there are only a handful of global variables[07:43:12] <geist> meianoite: no he isn't. lillo did it[07:43:13] <meianoite> yeah, that .cpp thing puzzled me as well. I couldn't recognize any C++ idiom in there[07:43:18] <geist> axel actually fixed it a bit[07:43:23] <meianoite> geist lilo tried to unify the lists[07:43:28] <geist> since it initially was actually broken[07:43:33] <meianoite> axel killed the priority queue[07:43:51] <meianoite> stating the tail never got maintained[07:44:02] <geist> well, okay[07:44:03] <meianoite> it's all there in the svn logs =/[07:44:11] <geist> anyway, you know we've talked about it longer than it would take to fix it[07:44:23] <geist> just go in there and fix it[07:44:27] <geist> would be a good first thing[07:44:41] <geist> even if you later redo it[07:44:54] <meianoite> actually, I'm waiting for BGA to return from his trip[07:44:55] <geist> dive in and get your hands dirty, fix the current queue to not be dumb, but stick to the current algorithm[07:44:58] <meianoite> so we can sit down and sicuss[07:45:10] <meianoite> *discuss[07:45:19] *** dsg has joined #haiku[07:45:24] <geist> i actually have a branch around here somewhere where I already rewrote all of that stuff to not e dumb[07:45:27] <meianoite> he's very wary of me screwing something ;)[07:45:48] <meianoite> huh, since the night before last????[07:45:57] <geist> yeah, about 6 months ago or so[07:46:01] <_hugo> meianoite: do the changes locally and post patches to the kernel mailing list for comments[07:46:24] <meianoite> _hugo actually I have to sort some stuff before doing that[07:46:31] <meianoite> like setting a VM build environment[07:46:40] <meianoite> neither zeta nor beos will take my SATA disks[07:46:45] <geist> i had started to completely redo the scheduler about 6 months ago, and the first thing i did was fix thoe stupid queues[07:46:54] <_hugo> meianoite: don't you have a Linux box?[07:46:57] <geist> but since i was going after the whole thing it isn't in a state to check in[07:47:04] *** Fanskapet has quit IRC[07:47:09] <Anxiety> same here, meianoite[07:47:14] <meianoite> _hugo used to. HD failure. never got around it.[07:47:19] <geist> and it was pretty clear it wasn't going to be a one night task, so i never finished it[07:47:56] <_hugo> i would say having a Linux box to build images is the best solution right now considering the turnaround times[07:48:10] <_hugo> you can get everything compiled pretty fast[07:48:13] <geist> absolutely, get a test machine and a linux box[07:48:23] <geist> that works pretty fast. I can turn it around in about 45 seconds[07:48:30] <meianoite> geist considering how we're going the same direction, me with the design I did for OS class, you with LK, I wonder if you'd agree with doing it as array of circular priority queues[07:48:35] <geist> mostly spent starting and shutting down linux on the target machine[07:48:54] <geist> meianoite: I've already pretty much pointed out that I really dont care what the low level data structure is. that's just details[07:48:56] *** kr1stof has left #haiku[07:49:13] <geist> array of circular queues is fine for me. there are smarter things than that, but simple is always nice[07:49:22] <geist> especially at first[07:49:37] <meianoite> well, that'll take care of the O(1) requirement[07:49:37] <geist> there's a whole lot to be said for just getting stuff working[07:49:47] <meianoite> posterior refinements can be made in due time[07:49:58] <geist> well, to be honest I'm not sure what other structure you could do that wouldn't be O(1)[07:50:11] <geist> like, you'd have to work at it to do something dumber than that[07:50:14] <meianoite> the CURRENT one? ;)[07:50:35] <geist> again, it could be fixed in like 5 minutes. it's totally not worth focusing on that much[07:50:45] <meianoite> and remember... even Linux only become O(1) with the 2.5 tree[07:51:03] <geist> it's not that simple. they had a much different deign[07:51:13] <geist> one that had little to do with the data structure and more with the overall design[07:51:29] <geist> in this case the haiku deign is more or less simple, but hasthis time data structure that wrecks it[07:51:31] <meianoite> geist forgive my baby steps, remember I'm not a kernel engineer... but a college student with not a lot of experience in OS design.[07:51:43] <meianoite> what you find trivial took me quite some time to figure out by myself[07:51:46] <geist> ugh, I cna't type today[07:52:10] <geist> that's why I'm trying to push you in the direction of worrying about the overall thing[07:52:13] <geist> not the details at first[07:52:18] <geist> it's too easy to get lost in the details[07:52:36] <geist> I think you're gonna find that once you dive into the SMP parts, it'll be far more difficult than you think[07:52:45] <geist> and there will be some conceptual hurdles to go over[07:52:59] <geist> i want to see you thinking about that stuff as soon as you can, because there will be some work that needs to be done[07:53:05] <meianoite> I hope you'll be there to lend me a hand, time permitting :)[07:53:22] <geist> absolutely, but we can't sit for hours every night talking about linked lists[07:53:33] <Anxiety> geist has always been a huge help[07:53:44] <Anxiety> I've bugged him quite a few times in #osdev[07:53:49] <meianoite> once more... forgive my baby steps[07:53:54] <geist> oh? who are you on osdev?[07:53:55] <meianoite> I'm just trying to get used to it[07:54:08] <Anxiety> I use to go by Kukyona[07:54:11] <geist> ah okay[07:54:14] <Anxiety> I'd pop in from time to time[07:54:25] <meianoite> unlike _hugo, who managed to just dive in and code :) I admire him for that[07:54:39] <meianoite> but I'm still a little too shy[07:54:49] <meianoite> (well, that's an understatement. WAY too shy.)[07:54:50] <geist> yeah hugo has been diving in, though to be honest networking code is pretty standard[07:55:01] <geist> it's generally more or less the same thing everywhere[07:55:21] <geist> in this case the low level scheduler stuff is highly specific to the OS[07:55:34] <_hugo> :-)[07:55:35] <geist> and also in this case, undocumented and poorly written[07:55:54] <geist> though as a plus, it's also very small[07:56:06] <geist> the sum total of the entire scheduler should (hopefully) only be a few hundred lines of code[07:56:10] <geist> net stacks are not :)[07:56:42] *** dsg has quit IRC[07:56:54] <meianoite> I still have to adjust myself to the right mindset[07:57:00] <meianoite> that COULD take some time, you know.[07:57:10] <geist> meianoite: i guess the overall point I'd like you to keep in mind about the scheduler stuff is the choice of the data structure should serve the algorithm[07:57:13] <geist> not the other way around[07:57:26] <geist> find the algorithm, work that out, and the data structure should fall out of it[07:57:42] <meianoite> these are probably wise words, but I don't quite know what you're talking about :)[07:58:05] <meianoite> oh... that's pretty much what I did already in the OS assignment case[07:58:14] <geist> well, basically I've seen you talk about bunch about linked lists and setting bits and whatnot, but I still haven't heard what the *algorithm* is[07:58:25] <meianoite> oh, true[07:58:32] <meianoite> I talked to Ingo about that[07:58:35] <meianoite> you were away :)[07:58:46] <geist> write it down, I'd like to see it[07:58:59] <meianoite> c or pseudocode?[07:59:06] <meianoite> or both :)[07:59:17] <meianoite> I could just paste what I previously had[07:59:24] <meianoite> (and be embarassed about it)[07:59:25] <geist> how long was it?[07:59:35] <meianoite> just a sec, I'll have to check it.[08:01:36] <meianoite> http://homepages.dcc.ufmg.br/~afb/escalonador/escalonador_so/mods/modificacoes.c[08:01:39] <meianoite> the code is here[08:01:49] <meianoite> and the comments are in glorious Portuguese language[08:01:50] <meianoite> =P[08:01:55] <meianoite> btw,[08:02:11] <meianoite> for( tentativas=0; tentativas<LIM_TENTATIVAS; tentativas++ ){[08:02:12] <meianoite> contador = 0; /* (prio>1) para evitar idle list */[08:02:12] <meianoite> while( vazia(lista) && (prio>1) && (contador<LIM_CONTADOR) ) {[08:02:12] <meianoite> lista = &prioridades[--prio];[08:02:12] <meianoite> contador++;[08:02:12] <meianoite> }[08:02:20] <meianoite> THIS LOOP IS DROP DEAD RETARDED[08:02:32] <meianoite> and that's the bottleneck I mentioned having fixed.[08:02:35] <geist> it's kind of like I thought[08:02:42] <geist> you haven't really considered all the cases[08:03:16] <meianoite> no, I have considered NO cases. that was only prototype code. it only handles quantum expiry, not blocking or anything else[08:03:25] <geist> maybe it's just me, but I consider the 'scheduler' to be all the cases where threads are moved into an out of the priority queues, all the different situations, all the cases where priorities are mucked with[08:03:30] <meianoite> not to mention SMP[08:03:42] <geist> ah okay, good. you at least know what you need to look at. that's good[08:03:58] <geist> all of that realy needs to be thought of up front[08:04:14] <meianoite> hence me waiting for BGA to return[08:04:17] <meianoite> so we'll sit down[08:04:23] <geist> i have generally followed the NT model for a lot of that, since it's dirt simple[08:04:25] <meianoite> and he'll just lecture me on how retarded I am ;)[08:04:34] <geist> I used it in dangeros (an os I wrote for a previous company)[08:04:46] <geist> very simple feedback mechanism[08:04:56] *** Master199 has joined #haiku[08:05:08] <meianoite> I'll be really grateful if you could point me to some documentation regarding that[08:05:16] <geist> i read it in a book[08:05:27] <meianoite> namely, what I'm supposed to pay attention to, and what are some classic solutions[08:05:33] <geist> one of the 'NT internals' books[08:05:48] <meianoite> I have a couple of books from Tanenbaum[08:05:58] <meianoite> but that's hardly the pinnacle of modern OS design[08:06:01] <meianoite> =/[08:06:04] <geist> none of the generic os books have very helpful scheduler talks[08:06:20] <geist> they generally give you super generic stuff about scheduling stuff if given infinite knowledge, etc[08:06:23] <meianoite> any papers you're aware of?[08:06:26] <geist> it's handy to know, but not real stuff[08:06:30] <geist> no, I never read papers[08:06:40] <meianoite> books?[08:06:47] <geist> every book I could find[08:07:07] <geist> but frankly very few of them cover modern schedulers that well[08:07:18] <geist> and there are lots of non-modern schedulers out there in the world[08:07:21] *** wildur has quit IRC[08:07:24] *** dsg has joined #haiku[08:07:30] <geist> hence why it took so long for linux to finally get something decent[08:07:37] <geist> and *BSD is still using the classic ones[08:07:44] <meianoite> considering I might give myself a couple weeks to study ONLY that, which (pair of) book(s) would you indicate?[08:08:20] <geist> dont know. it's generally only a page or two out of the entire book that actually talks about the scheduler[08:08:33] <meianoite> (oh, vocabulary, why doth do fail upon me?)[08:08:36] <geist> that's the case with the NT book I have, and the solaris book, and the BSD book, and the linux book[08:08:55] <meianoite> *thou[08:09:04] <Anxiety> http://research.microsoft.com/ur/asia/curriculum/download/BeijingPresentation.ppt[08:09:17] <geist> too bad it's powerpoint[08:09:30] <Anxiety> http://64.233.167.104/search?q=cache:p-K7zPcxwXwJ:research.microsoft.com/ur/asia/curriculum/download/BeijingPresentation.ppt+nt+internals+scheduler&hl=en&ct=clnk&cd=4&gl=us&client=opera[08:09:31] <meianoite> nah, I don't mind. windows box here anyway =P[08:09:59] <meianoite> not that I'm willing to copycat NT anyway[08:09:59] <Anxiety> missing all the layout, so its not as useful in html like that[08:10:21] <geist> yeah, the scheduler part doesn't go into very much detail anyway[08:10:34] <geist> the trick is what the scheduler does for quantum expiration, and how it specifically does it[08:10:41] <geist> and how it deals with priority boosting[08:12:39] <meianoite> what I did was just to roundrobin that given priority's list. easy enough with a circular list. I also could decide to do some boosting or lowering, but I didn't go that far[08:13:52] <geist> i never think in terms of roundrobing[08:14:09] <geist> i always thing in terms of 'does it go to the head or the tail of it's priority's queue'[08:14:33] <geist> so there are a couple of decisions to make after some event (quantum expiration, preemption, blocking, waking up)[08:14:42] <geist> a) what is the new priority (does it get boosted or un-boosted)[08:14:43] <meianoite> that's what I meant by "to roundrobin that given priority's list"[08:14:49] <geist> b) does it go on the head or the tail[08:14:50] <meianoite> if I rotate it, it goes to the tail[08:14:54] <meianoite> if I don't, it's kept on the head[08:15:09] <geist> and then if the scheduler always just picks the first thread on the highest priority queue, it's job is simple[08:15:18] <meianoite> that's exactly what it does[08:15:26] <meianoite> it only skips whole priority levels[08:15:29] <meianoite> not individual threads[08:15:36] <geist> so generally my schedulers are something like this:[08:15:58] <meianoite> and the skipping was controlled by a probability distribuction function[08:16:01] <meianoite> namely:[08:16:04] <geist> 1) if you use up your quantum, you get unboosted one or more priorities, and go to the tail of that queue[08:16:07] <geist> shush listen[08:16:33] <meianoite> sorry :)[08:16:34] <geist> 2) if you block, you are removed from the queue entirely, so it doesn't matter[08:16:45] *** dsg has quit IRC[08:16:46] <geist> 3) when you wake up from block you get boosted one or more levels, and go to the head of the queue[08:17:09] <geist> if you get prempted and you still have quantum left, you go to the head of your queue, and dont get boosted or otherwise[08:17:16] <meianoite> I'd do (3) as tail, not head. care to explain your reasoning?[08:17:35] <meianoite> I'd do (1), (2) and (4) just as you did[08:17:57] <meianoite> ok, 3 boosts responsiveness.[08:18:04] <geist> 4) if you yield you give up the rest of your quantum, so you go to the end of your queue[08:18:20] <meianoite> (4 redux): ok.[08:18:43] <meianoite> cool, we're still on the same page ;)[08:18:46] <geist> right[08:18:57] <geist> as far as 3, it's absolutely critical that the new thread goes to the head of the queue[08:19:07] <geist> it just got woken uyp, you probably want to run it immediately[08:19:12] <meianoite> IYHO ;)[08:19:15] <geist> since there's a good chance it'll just block again anyway[08:19:32] <geist> but of course if it's still lower priority than the current thread it doesn't matter[08:19:38] <geist> the current thread will get selected again[08:19:58] <meianoite> well, you should consider there's a damn QUEUE in there... it doesn't work that way in the real life, why should it behave as that on a thread scheduler? that's what I've thought initially, but I'm sold on the responsiveness argument.[08:20:07] <geist> and since the current thread gets 'preempted' in this case, it wont go to the end of it's queue if it still has quantum left[08:20:22] <geist> so if they're in the same queue, the new thread will get to run and in theory give it right back to the one that woke it up[08:21:12] <geist> oh also, 5) blocking uses up a fraction of your quantum[08:21:24] <meianoite> oh... that's why you keep unqueueing the current thread[08:21:25] <geist> so that frequently blocking threads dont get an infinite quantum[08:21:49] <meianoite> I usually keep it on the queue unless I actually HAVE to pull it out[08:22:10] <geist> right, I generally pull the current thread out of the queue.it's in 'running' state[08:22:23] <geist> the run queue is specifically threads that are ready to run. not ones that are running[08:22:37] <meianoite> I can see how that's supposed to be done on SMP systems[08:22:49] <geist> also, if you left the thread in the queue it would start getting nasty if you have SMP systems using the same set of queues[08:22:50] <meianoite> keep in mind I didn't consider SMP at all with that previous design[08:22:58] <geist> you'd have all sorts of bit turds sitting in the middle of the queues[08:23:03] <geist> big[08:23:31] <meianoite> I'm still considering not sharing the same set of queues[08:23:44] <geist> not sure what you mean[08:23:47] <meianoite> but that argument of yours, regarding unbalacing, is pretty strong against that decision[08:23:55] <meianoite> per-CPU queues, I mean.[08:23:59] <meianoite> not a global one[08:24:00] <geist> oh that's almost required[08:24:16] <geist> the overhead of all the cache thrashing makes it almost a requirement[08:24:21] *** vision has joined #haiku[08:24:27] <geist> the trickyness comes from keeping it still reasonably responsive[08:24:30] <meianoite> figured as much[08:24:40] <geist> of course, it can never be perfectly responsive with that model[08:24:41] <kokito> geist & meianoite: http://www.aequus.net/download/BeijingPresentation.pdf[08:25:06] <meianoite> kokito I already got the PPT ;) not too rich on details, unfortunately[08:25:09] * vision gets a warm feeling : )[08:25:12] <geist> so it's another one of those 'here is the perfect alrogithm that's impossible to do, try to emulate as much of it as possible'[08:25:14] *** vision is now known as mmadia2[08:25:47] <geist> on the other hand, having more cpu than one is still nice, so you're basically already ahead[08:26:28] <meianoite> I also considered having both global and local queues[08:26:45] <Anxiety> strangly... ubuntu runs faster in the emulator... I think its my nic card... the open sourced drive the manufacture released seems to run poorly[08:26:46] <meianoite> the local ones would keep threads with affinity masks set[08:26:59] <geist> the trick is probably deciding pretty efficiently when it's cheaper to push stuff onto other threads[08:27:02] <meianoite> global ones, the threads that wouldn't care about tht[08:27:02] <geist> er other cpus[08:27:29] *** dsg has joined #haiku[08:27:32] <meianoite> linux has a function for that[08:27:39] <geist> trouble is you frequently have the problem of thread a running that releases a sem that wakes up thread b[08:27:40] <meianoite> forgot how's it called[08:27:57] <geist> where does thread b get queued up? you dont know how much more time thread a and b both have[08:28:20] <meianoite> you could guesstimate based on the quantum they already ate[08:28:29] <geist> if thread b is just gonna run for a fraction of a quantum and go back to sleep (a common occurrence) you're better just running it right then and there[08:28:49] <geist> if it's gonna go do a bunch of work, you're probably better off punting it onto another cpu (if another one is more idle than you)[08:29:20] <geist> on the other hand you might do that and A goes to sleep immediately, iun which case you just spent a bunch of time waking up another cpu to run a thread that you may as well have run[08:29:21] <meianoite> we could keep a pair of counters, short_quantum and large_quantum, and then guesstimate based on their ratio[08:29:31] <geist> or you can just run it like a single cpu and let the periodic load balancer deal with it[08:29:47] <geist> or yeah you can try to track their past behavior to see what they tend to do[08:30:16] <geist> all this is because pushing a thread onto another cpu's queue is at best pretty expensive[08:30:26] <geist> you can push it directly into their queue and force a reschedule on em[08:30:50] <meianoite> why force a reschedule?[08:30:50] <geist> or if they're idle you can send an inter cpu message with the thread as a payload, and have the cpu schedule it[08:31:06] <geist> to let the other cpu decide if it wants to run it or not[08:31:18] <meianoite> you could potentially preempt a newly-promoted-to-current thread =P[08:31:21] <geist> obviously if the other cpu is idle you'd want it to immediately reschedule and pick up the new thread[08:31:47] <meianoite> well, IF it's running the idle thread, I'd definitely call reschedule()[08:31:50] <geist> and even if it isn't and you decided to push it anyway, the new thread may be higher priority than the one it's already running[08:31:57] <geist> why would it do that?[08:32:09] <geist> but you can't just call reschedule on another cpu[08:32:11] <meianoite> so I don't preempt that newly-promoted thread for nothing[08:32:35] <geist> to do that you have to fire an interrupt to the other cpu, it has to accept the interrupt, look in a message queue, process the message, return the empty message to the queue[08:32:36] <meianoite> as that could potentially trash the cache[08:32:40] <meianoite> I know, interrupt and such[08:32:47] <geist> hundreds or thousands of cycles[08:32:51] <geist> so it's not something you do lightly[08:33:05] <meianoite> if I have access to the current thread pointer of every cpu[08:33:13] <meianoite> and can determine wheter it's the idle thread or not[08:33:17] <meianoite> I can do just like I described[08:33:25] <meianoite> not fire an interrupt[08:33:35] <meianoite> let the thread just finish it's job[08:33:49] <meianoite> then the scheduling would occur naturally[08:33:57] *** mmadia2 has quit IRC[08:34:03] <geist> what if the new thread was higher priority than the one it is currently running?[08:34:27] <geist> also, if you start poking into the other cpu's data structures, you begin to negate the purpose of the entire exercise[08:34:29] <meianoite> it COULD hamper responsiveness if you just blindingly preempt the running thread[08:34:34] *** SunWuKung has joined #haiku[08:34:35] <geist> because that'll keep flushing the data cache[08:34:41] <meianoite> considering we're skipping high prio threads all the time =/[08:34:56] <geist> I would think the opposite[08:35:01] <meianoite> [03:34] <@geist> because that'll keep flushing the data cache <-- just thought about that. bummer. =P[08:35:12] <geist> reschedule() does not preempt the current thread blindingly[08:35:22] <geist> it only preempts it if there is something higher priority to run[08:35:43] <meianoite> and even then if it passes the random promotion threshold[08:35:51] <geist> it's like a yield but without giving up the current quantum[08:36:30] <meianoite> I missed a "only" between "then" and "if" on my last sentence[08:36:46] <meianoite> geist, http://homepages.dcc.ufmg.br/~afb/escalonador/Relatorio-EscalonadorO1Probabilistico.zip[08:36:51] <meianoite> there's a PDF file inside[08:36:57] <geist> that doesn't sound too good[08:37:08] <geist> considering nearly every word in that url is in portuguese[08:37:13] <meianoite> you won't understand a word of it, but look at the pretty figures[08:37:18] *** SunWuKung has quit IRC[08:37:27] <geist> bleh, I'm not into theoretical stuff[08:37:41] <meianoite> no, I was pointing you to the RESULTS table[08:37:42] <geist> i'm not one to read papers and do theory,[08:38:25] <meianoite> page 7, if you will[08:38:26] <geist> anyway, my point is this is al quite complicated[08:38:34] <geist> and yeah, it may not be worth doing the multi cpu thing[08:38:58] <geist> unforunately having single or multi queues does effect the overall design significantly[08:39:02] <geist> so it's sort of all or nothing[08:39:20] <geist> youcan't just halfway do it and promote to multiple queues without really rethinking the whole thing[08:39:21] <meianoite> fortunately I'll have to opportunity to do multiple designs[08:39:33] <meianoite> and let them compete :)[08:39:45] <geist> well, thsi is another thing that's seriously worth thinking about[08:39:48] <geist> how you're going to test it[08:39:53] <meianoite> the single queue one is mostly complete, only lacking the locking[08:40:10] <geist> that was something i was thinking really hard about, and never realy came up with a great solution[08:40:23] <geist> there are obviously multiple things to test for, responsiveness, efficiency[08:40:29] <meianoite> right...[08:40:31] <geist> and some of those are mututally exclusive[08:40:53] *** dsg has quit IRC[08:41:36] <meianoite> right.[08:42:37] <geist> thats the primary reason i never really finished the scheduler stuff. there wasn't a great way to really tell if progress was being made (also things were a lot more unstable 6 months ago)[08:42:42] <geist> and there were bigger fish to fry[08:43:01] <geist> as lame as the current escheduler is, it's almost identical to beos's, which is mostly the point of haiku[08:44:00] <meianoite> it's still not clear to me whether the BeOS one was O(n) or O(1)[08:44:12] <meianoite> as the current one is pretty much O(n) =P[08:44:24] <geist> dont focus on the data structure that much[08:44:35] <geist> the fact that the list is sorted that way is some dumbness on haiku's part[08:44:45] <geist> assume the list isn't done that way (it wasn't in beos)[08:44:59] <meianoite> k.[08:45:22] <geist> and even then, there's at most a handful of stuff in the queue. yeah it may be O(n) for insertion, but removal is still fast (which is really the one that counts) and tn is pretty small most of the time[08:45:48] <meianoite> I'm having a pretty hard time considering the algorithm separate from the data structure, when the latter (in the case of the current scueduler) almost guarantees it's asymptotic complexity is O(n)[08:45:55] <geist> when folks talk about O(n) schedulers they typically are talking about much more complicated legacy designs where all threads in the system are iterated over every so often[08:46:11] <meianoite> like unsorted queues?[08:46:14] *** infamy has joined #haiku[08:46:34] <geist> and all of that aside, for whatever reason people get really overly excited about 'O()' and 'schedulers'[08:46:35] *** mmadia has quit IRC[08:47:04] <geist> meianoite: like, lets stop and iterate over every thread in the system, decide what slice of the cpu they get in the next second, sort the queue based on that, run[08:47:14] <geist> that's more or less the classic unix scheduler[08:47:52] <geist> hence why when linux and friends finally got off their ass and wrote a real one the last few years, it was a big deal[08:48:00] <geist> mostly because all the previous ones were seriously dumb[08:48:03] <meianoite> yeah, I'm wondering how many people are going to keep both eyes on what I do. some of them are complexity junquies per se, some of them are scheduler junkies, and both camps can be pretty annoying ;)[08:48:38] <geist> so yeah, the current haiku thing is O(n) for *insertion* but the constant in front of n is pretty small, and n is pretty small[08:48:53] <geist> versus having some massive C*N that comes along and blows everything up all the time[08:49:13] <geist> and anyway, like I said before we could have fixed the linked list in the amount of time we talked about this[08:49:19] <geist> replace it with something smarter, done.[08:49:47] <meianoite> even though the whole discussion definitely points the other way, I'm very aware that my real work is to do the "cpu affinity" part, not the "O(1)" part... =P[08:50:11] <geist> what bugs me far more than the insertion thing is the fact that the scelection process on beos and haikuy is effectively O(C*n) where C is pretty small, but still > 1[08:50:29] <geist> because of the random number roll[08:50:32] <meianoite> that random function roll?[08:50:35] <meianoite> figured as much[08:50:52] <geist> and changing that is probably necessary, but would be a pretty big departure from how beos has generally done things, which goes into new territory[08:51:19] <geist> though my guess is that strategy ius a 'fix' for the fact that the beos scheduler is completely static. no priority boosting or depression based on runtime[08:51:37] <meianoite> but how would YOU tackle that? as I told you, I used a linear probability distribution function to choose the *priority* level directly, not how many threads would I skip[08:51:44] *** SunWuKung has joined #haiku[08:51:58] <geist> i generally prefer the selection process to be totally constant[08:52:03] <ormandj> you folks are seriously far too smart for your own good, haha. carry on! ;)[08:52:08] <geist> poick the top of the highest queue[08:52:13] <geist> and have all the real decision making offline[08:52:22] <meianoite> ormandj I'll take that as a compliment ;)[08:52:31] <ormandj> meianoite: it was intended as such :p[08:52:33] <geist> but all in all it's because that's just how I've always done it[08:52:36] <meianoite> by offline you mean...?[08:52:39] <geist> I'm no scheduler junkie[08:52:46] <geist> by offline I mean not at reschedule() time[08:52:55] <geist> when the thread gets woken up, etc[08:52:59] <meianoite> geist could have fooled me!! :D (the not being a junkie part)[08:53:16] *** infamy has quit IRC[08:53:19] *** emitrax has joined #haiku[08:53:25] <geist> schedulers can be massive blackholes of time if you're not careful[08:53:43] <meianoite> I'm starting to see that. lol ;)[08:53:46] <geist> it's super easy to tweak them all the time, but it's really pointless unless you have a measurable thing you want to improve[08:54:04] <geist> and I've generally found myself in the situation of not having something to measure[08:54:22] <geist> or even if I did, not having a clear idea if it's the right thing to tweak for this or that[08:54:31] <ormandj> why don't you just implement the 5.x branch of freebsd's ULE scheduler? *dripping sarcasm*[08:54:39] <geist> so i generally just prefer to be simple, so that at least it's well understood how it works[08:54:50] <geist> i think that was part of the beos's schedulers appeal[08:54:59] <geist> since it was so simple, and static, you can tweak around it[08:55:08] <meianoite> I'm happy enough to have designed something resembling the latest designs of yours, with NO, NO, NO absolutely ZERO previous experience or even mere contact with ANY existing scheduling algorithm. I didn't even read the Minix one =P[08:55:14] <geist> i know a bunch of system code in the app_server, for example, tweaked their priorities constantly[08:55:34] <geist> to work with the scheduler, since it was known pretty well how it was going to behave[08:55:39] <ormandj> geist: out of curiosity, where do you work if you don't mind me asking?[08:55:55] <geist> i recently started working at a game company called Secret Level[08:55:58] <geist> I do PS3 stuff[08:56:15] <ormandj> oh boy, hope you have contingency plans for employment[08:56:22] <geist> and you wanna talk schedulers. man, keeping those SPUs active is fun[08:56:31] <geist> ormandj: it's kind of an experiement. wanted to try something new[08:56:43] <geist> it may or may not work out, but yeah I have plenty of contingency plans[08:56:47] <ormandj> good luck on that one. i hear ps3s are a beast to work on[08:56:55] <geist> by a beast you mean fun[08:56:56] *** dsg has joined #haiku[08:57:03] <geist> if it was easy, there's no challenge[08:57:08] <meianoite> ormandj try not to rain on Sony's parade... I for one hope they'll sell god-damn well. I intend to buy myself one, even ;)[08:57:09] <ormandj> if you're a coding masochist (as you quite obviously are), then yes :p[08:57:26] <geist> nah, actually they're quite easy to use[08:57:31] <ormandj> meianoite: i hate sony post-rootkit and the various other shady things they've done, so i hope it tanks[08:57:32] <geist> the SPUs are incredibly fast[08:57:33] <meianoite> the games already look absolutely gorgeous.[08:57:54] <ormandj> but, unfortunately for sony, the ps3s are selling like dogs[08:58:04] <geist> you can run generic code on em generally faster then the main cpu[08:58:08] <ormandj> geist: you're the first person i've heard praise development on ps3s[08:58:12] <geist> the downside is the 256k of ram[08:58:23] <geist> but you can dma into an out of their ram at will, and it's quite fast[08:58:26] *** myob has joined #haiku[08:58:29] <geist> so you basically have a software controlled cache[08:58:34] <ormandj> gotcha[08:58:51] <geist> and if you actually hand tune vector code for them, they're insanely fast[08:58:56] <ormandj> guess you just have to spend a bit more time designing then[08:58:58] <geist> like holy shit that's fast speed[08:59:18] <geist> yes and no. a lot of game folks are just taking large chunks of their game engine and farming them off as jobs[08:59:22] *** DigitallyStoned has quit IRC[08:59:31] <geist> the trick is you need your game engine to kind of be already working in this multithreaded way[08:59:37] <meianoite> I wonder if someone actually found some math bug on those vector units... like the original Pentium did[08:59:42] <geist> but you kind of have to be doing that anyway, even for the xbox360[08:59:45] <ormandj> yeah, if it's anything like the vector units in the power line, i'm sure it's a heck of a chip[08:59:48] <meianoite> getting fucked up results, but pretty fast ;)[09:00:01] <geist> ormandj: it's like vmx++++[09:00:02] *** RedLobster is now known as weltschmerz[09:00:05] <ormandj> lol[09:00:06] <ormandj> nice[09:00:14] <geist> like, you *cant* not use a vector[09:00:22] <ormandj> meianoite: it's common for all to have such errors[09:00:27] <geist> all of the registers are 128 bits wide (all 128 of em) and all math ops implicitly operate on all 128 bits[09:00:34] <geist> even if you ignore the top 96[09:00:40] <meianoite> ormandj sure, but I havent't read any reports[09:00:41] <ormandj> hah, wow[09:00:49] <ormandj> meianoite: it's generally not publicized[09:00:52] <geist> and of course the spus run at 3.2 ghz[09:00:54] <ormandj> the dev kits probably have workarounds[09:01:00] <geist> and the really cool part is their memory is zero wait state[09:01:04] <geist> no cache, it's just instantly there[09:01:08] <geist> which is why the code is so fast[09:01:12] * kokito thinks Haiku should change all the B_ stuff in the API to H_[09:01:13] <meianoite> ormandj you CAN'T have workarounds when you're that close to the metal[09:01:14] *** mmu_man has joined #haiku[09:01:14] *** ChanServ sets mode: +o mmu_man[09:01:15] <geist> no memory stalls[09:01:28] <meianoite> you MUST get the erratas[09:01:29] <meianoite> period[09:01:45] <myob> kokito: good morning :P[09:01:48] <ormandj> geist: sounds like they made a heck of a machine, too bad they floundered in execution of the rollout[09:01:56] <geist> of course, they're sony[09:02:01] <geist> they always mess it up[09:02:02] <ormandj> :p[09:02:09] <geist> they tacked on this bluray thing which fucked it all up[09:02:30] <meianoite> that's called trojan horseing stuff :)[09:02:33] <myob> and then insisted on the now-outdated regional rollout[09:02:34] <geist> and then on top of that it's relaly slow[09:02:38] *** mazon is now known as Mazon[09:02:42] <meianoite> or, on newspeak, "halo effect"[09:03:05] <geist> so another thing everyone is doing is running zlib on one of the SPUs to decompress everything coming off the disc[09:03:06] <kokito> hi myob :P[09:03:13] <geist> (after another SPU decrypts it, of course)[09:03:28] <meianoite> 5 SPUs left still :)[09:03:36] <ormandj> geist: did they ever fix the mem read speed? i haven't touched a devkit in a while myself[09:03:45] <geist> oh that uses like 10% of the spu to zlib[09:03:59] <geist> surprisingly, stock zlib runs perfectly fine, even in full scalar mode[09:04:11] <ormandj> used to be much slower to read from memory than write[09:04:11] <geist> ormandj: i dunno, what was it before?[09:04:21] <ormandj> was like 16Mbps from memory[09:04:32] <geist> from VRAM, yeah, it's very slow[09:04:36] <geist> you just dont do it[09:04:39] <meianoite> ormandj isn't that GPU stuff?!![09:04:47] <geist> but that's because the VRAM is on the other side of the GPU[09:04:52] <ormandj> write was 4GB/s[09:05:13] <geist> it's fast now[09:05:15] <ormandj> http://www.theinquirer.net/images/articles/PS3_memory_bandwidths.jpg[09:05:17] <geist> 25GB/sec or so[09:05:18] <ormandj> found an old image[09:05:48] <geist> yup, that' spretty accurate[09:05:56] <ormandj> funky[09:05:59] <geist> you basically dont want to read back from video ram[09:06:01] <meianoite> feh... I really wanted to get a 1st gen PS3, only for the builtin PS2 hardware[09:06:06] <ormandj> i can imagine heh[09:06:18] <geist> there are reasons why that is the case. basically you're going through the GPU, and the GPU has it's own memory management[09:06:19] <ormandj> meianoite: why, so it can break? :p[09:06:22] <meianoite> but if they DO fix some memory bandwidth stuff... that pretty much leaves me on the cold. %#!*.[09:06:28] <myob> can you use the PS2 hardware for 'assistance'?[09:06:38] <meianoite> no, read again, the PS2 hardware.[09:06:42] <geist> it actually swizzles some data around in a opaque way, and it has to unswizzle it for you when you read back[09:06:54] <meianoite> no myob, for compatibility. it does run PS2 games, you know.[09:06:55] <ormandj> gotcha geist[09:07:16] *** kokito has quit IRC[09:07:21] <ormandj> interesting topics here in haiku at night, cool[09:07:40] <geist> yeah, can't go into too much more detail of that because it gets into RSX stuff[09:07:48] <geist> which is pretty NDA heavy, since it's basically an nvidia gpu[09:07:49] <ormandj> completely understand[09:07:57] <geist> and it's pretty much wide open it's design[09:08:07] <geist> you're more or less punching commands directly into it's command buffer[09:08:08] <myob> ormandj: its night?[09:08:16] <myob> 08:12 here[09:08:16] <ormandj> myob: i live in hawaii, 2100 here[09:09:16] <myob> ah right, well I'm in UTCland[09:09:19] <meianoite> so on that slide when they say local memory they mean VIDEO memory? well, that makes sense.[09:09:54] <geist> yeah, it's a weird nomenclature[09:10:06] <geist> but video memory is 'local' and regular memory is 'main'[09:10:43] <meianoite> no... video memory is "video memory" and regular memory is "main memory". THAT makes sense. ;)[09:11:27] <meianoite> unless that slide is from a NVIDIA presentation. THEN it makes sense. it's a matter of point-of-view ;)[09:12:06] <geist> that's what they call in in the docs[09:12:32] *** tombhadAC has quit IRC[09:12:43] <ormandj> btw meianoite, i was a bit confused when earlier you said it was not possible for software workarounds to chip bugs[09:12:59] <ormandj> i got distracted and forgot to ask what you meant by that[09:13:23] <meianoite> it of course is[09:13:24] *** slaine_ has joined #haiku[09:13:37] * geist sleeps[09:13:46] <ormandj> goodnight geist, thanks for the nice chat :)[09:14:07] <meianoite> but if you're extracting the last ounce, you wouldn't like to be wrapped up in APIs, you'd program straight in asm[09:14:15] <meianoite> g'nite, geist[09:14:18] *** SunWuKung has quit IRC[09:14:32] <ormandj> meianoite: ahh, you are saying it's not "optimal"[09:14:33] <ormandj> with that i can concur ;)[09:14:34] <meianoite> and if the asm is b0rked... well, you just neet the erratas.[09:14:56] <ormandj> it probably comes along with the rest of the NDA stuff[09:14:59] <ormandj> i doubt it's made public[09:15:19] <ormandj> i'm sure there is a bug or two, though[09:15:39] <meianoite> aaaaah, how good is it to have a blazing fast SPU when you concede to be "not optimal"[09:16:30] <Anxiety> night everyone[09:16:35] <ormandj> i never said that you must concede, merely that it is possible. as to how sony handles the bugs, you are likely correct, i just think the information comes with a nasty NDA[09:16:37] <ormandj> night[09:17:19] <geist> IBM got it right[09:17:27] <geist> it appears to be bug free[09:18:22] <meianoite> am I the only one who finds that an oxymoron? ;)[09:18:24] <ormandj> well, supposedly at least the cell procs being pushed for other uses are relatively bug free, at least insofar as the scientific work that's being performed by them[09:18:41] <geist> IBM knows how to make processors[09:18:51] <ormandj> they also know how to rape customers[09:19:05] <geist> as far as bug free, what sony does to mitigate production costs is eat one of the SPUs[09:19:05] <meianoite> yeah, but the very PEOPLE themselves don't know how to make anything "bug-free"[09:19:12] <ormandj> you should see what they charge for maintenance contracts for the zOS mainframes i run lol[09:19:18] <geist> so the cell processors on ps3s only have 7 active SPUs[09:19:28] <geist> one can be disabled in the factory due to defects[09:19:33] <meianoite> that's for yield purposes, not logic flaws[09:19:45] *** phrizek has quit IRC[09:19:46] <geist> right[09:20:13] <ormandj> time to watch a movie[09:20:16] <ormandj> goodnight all :)[09:20:17] <meianoite> and I meant logic flaws, the one erratas usually address[09:20:29] <ormandj> great talking with you both, hope to see you around[09:20:34] <ormandj> nice meeting you meianoite, btw, i'm david.[09:20:42] <meianoite> hi, David[09:20:50] <meianoite> I know you from osnews :)[09:20:56] <ormandj> ops[09:20:59] <ormandj> hopefully in a good light[09:21:07] <meianoite> no, you damn trol[09:21:10] <meianoite> jk ;)[09:21:10] <ormandj> some do not appreciate my commentary :p[09:21:13] <ormandj> ehe[09:21:27] <ormandj> i'm not a coding god like you two seem to be[09:21:35] <ormandj> just own a data center[09:21:45] <meianoite> 2.35 is a pretty good score[09:21:48] <ormandj> so i might ask you explain things from time to time :p[09:22:01] <meianoite> ME, coding god? jesus holiest.[09:22:12] <ormandj> yeah, i normally do ok in the moderation department. i'm not quite sure how, but i do ok[09:22:21] <ormandj> it must be my lovely nazi comparisons[09:22:24] <ormandj> :p[09:22:25] <meianoite> well, for one, you praise sun hardware[09:22:29] <meianoite> AND software[09:22:38] <ormandj> most of it is pretty good[09:22:39] <meianoite> that brings good karma[09:23:04] <ormandj> i have far less problems with it than i do with the ibm/dell/hp crap[09:23:09] <meianoite> I managed a 2.95 karma so far[09:23:15] <ormandj> how many comments?[09:23:23] <meianoite> very few ;)[09:23:26] <ormandj> that's why[09:23:33] <ormandj> i think i have almost 500 now[09:23:42] <ormandj> if you keep 2.95 over 500, i will give you 10$ :p[09:23:50] <meianoite> and that's only because some people seem to ignore some of my more enlighted comments.[09:24:04] <meianoite> like the PowerPC G5 one[09:24:07] <ormandj> you should do as me, and make references to nazis[09:24:12] <ormandj> it works every time[09:24:12] <ormandj> :p[09:24:12] <meianoite> which got NO plus points[09:24:22] <ormandj> hm well, i have a billion mod points[09:24:26] <ormandj> i will hunt for your comments ;)[09:25:00] <meianoite> nah. I wasn't there, wasn't born there, can't really do comparisons based on stuff I only got to know because of television and History books the "winning" parties wrote.[09:25:18] <ormandj> it's pretty well documented globally my friend[09:25:23] <meianoite> I'd rather bash Microsoft[09:25:28] <ormandj> ahhh[09:25:32] <meianoite> that's something on my timeframe[09:25:34] <meianoite> ;)[09:25:38] <ormandj> fair enough![09:25:46] <ormandj> it makes a great target, i do admit[09:25:56] *** rhu has quit IRC[09:26:09] <ormandj> have a lovely day/evening all of you[09:26:18] <ormandj> ciao![09:26:18] <meianoite> please, I'm not a pro-nazi, holocaust-denying kind of guy[09:26:54] <meianoite> it's just a personal policy on not commenting on stuff I really don't know jack about first hand[09:27:07] <ormandj> as i said, fair enough! :)[09:27:08] <ormandj> ciao :)[09:27:12] <meianoite> bye[09:28:49] <mmu_man> please everyone, don't use those words here[09:33:52] <meianoite> ciao and bye? ;)[09:35:51] *** DigitallyStoned has joined #haiku[09:37:28] <myob> godwins law being invoked here I see?[09:39:03] <ormandj> lol i noticed i didn't hit "yes" for am i sure i want to quit and see this discussion again :p lolol[09:39:18] <ormandj> tschuss![09:39:22] *** ormandj has quit IRC[09:39:25] * mmu_man == op == god[09:39:29] <mmu_man> and yes, god wins :P[09:39:45] <meianoite> bleh. people waste so much time talking about Godwin's law, whereas Hofstadter's law is so much more fun. and edifying.[09:41:52] <myob> meianoite: too much experience of that here to even invoke it ;)[09:42:01] <myob> our 'new' app was meant to be out in August 06[09:42:14] <myob> its not out yet and I can't see it going in to testing until about August 07...[09:45:52] <JonathanThompson> Greets. fellow geeks.[09:46:38] *** Ingenu has joined #Haiku[09:48:53] *** mmu has joined #haiku[09:49:02] *** mmu_man sets mode: +o mmu[09:49:05] *** mmu_man has quit IRC[09:50:32] <myob> early mornings are so intensely boring here,..[09:50:44] <JonathanThompson> It's very early morning here.[09:51:38] <myob> waiting for 9 to take my phone off "nooooo! don't call me!" mode[09:52:27] *** eNGIMa___ has joined #haiku[09:56:38] *** PulkoMandy has joined #haiku[10:02:19] *** meianoite has left #haiku[10:04:58] <myob> 9am comes, customers from hell begin to call...[10:05:29] * JonathanThompson notes that customers from Hell must have horrendous phone bills from talking to myob in Ireland[10:05:40] <JonathanThompson> That's assuming they aren't using VoIP.[10:06:14] <myob> well, considering Hell is in Finland, I'm sure they can...[10:06:25] <JonathanThompson> There's Hell, Michigan, too.[10:06:30] <JonathanThompson> That's the one I was thinking of.[10:06:35] <myob> erm Norway[10:31:21] *** Sloar has quit IRC[10:38:48] *** Jixt has joined #haiku[10:39:01] <Jixt> hi[10:41:01] *** gotaku has joined #haiku[10:41:49] <gotaku> Would it be worth porting Links-2 to Haiku?[10:43:34] <myob> the graphical one? possibly. porting the underlying svgalib would be more interesting, mind[10:43:48] <myob> would allow links2 -g to run as well as other stuff[10:44:22] <myob> however I dunno how tied svgalib is to the philosophy of painting to a fullscreen svga output...[10:46:08] <gotaku> Haiku lacks a browser at the moment and firefox doesn't run well at all.[10:47:28] <myob> not sure it ever intends to 'have a browser' as in have one of its own, tho[10:47:53] <myob> and, well, you can't compare links2 to even net+ really[10:47:56] <mmu> http://nirvana.berlios.de/[10:49:09] <gotaku> That seems a little overkill for what I want.[10:50:21] <myob> the first ever Haiku-ported app was a browser btw ;)[10:50:26] *** JBurton has joined #haiku[10:50:27] *** ChanServ sets mode: +o JBurton[10:50:29] <JBurton> hi all[10:51:00] <gotaku> Wow, 123 commits last week.[10:51:46] <gotaku> Most of them were probably from hugo.[10:53:06] <myob> I remember weeks of a thousand commits :P[10:53:15] <myob> (like when ingo redid the entire build system)[10:54:10] *** GreyGhost has joined #haiku[10:54:37] <gotaku> Maybe we can get that high again with the Google SoC people.[10:55:08] *** GreyGhost has joined #haiku[10:55:14] <JBurton> hopefully :)[10:55:49] <raph_ael> hello[10:56:04] <Ingenu> oh gods, please pity me...[10:56:20] <Ingenu> those guys at work left functions inside the header because they didn't know how to put them inside the cpp[10:56:38] <Ingenu> functions of a struct local to a class[10:56:53] <Ingenu> classname::structname::functionname(...)[10:56:56] <Ingenu> ...[10:57:19] <gotaku> Looks like the BeOS MAX guy is going to work on a Haiku distro.[10:57:26] <Ingenu> someone get Haiku done, quick, so I can have something neat in my life to compensate[10:57:33] <Ingenu> WalterOS[10:57:46] <Ingenu> Haiku based distro, same licence as Haiku[10:57:53] <gotaku> The good news is that he won't release it until after Haiku R1 is released.[10:58:06] <Ingenu> really ?[10:58:46] <gotaku> http://www.beosmax.org/wiki/index.php/WalterOS[11:03:33] <myob> distros = evil[11:04:38] <gotaku> You can't stop them in open source operating systems.[11:05:49] <JBurton> you can, actually[11:06:20] <JBurton> at least, you can stop them using your name, artwork, logo, etc[11:06:24] <JBurton> if you want, of course[11:08:34] <gotaku> That doesn't really stop anything.[11:08:58] <JBurton> it does[11:09:05] <JBurton> for example[11:09:16] <JBurton> let's say Vasper has to come up with his own icons, bitmaps, etc[11:09:18] <JBurton> it's much more work[11:09:25] <JBurton> maybe too much for a single person[11:09:42] <JBurton> yes, he can grab some free stuff around[11:09:45] <JBurton> but it's not the same thing[11:12:22] <myob> it also removes the mindshare, as it doesn't ahve the same name[11:12:27] <JBurton> yeah exactly[11:12:52] <JBurton> there are various example of different restrictions for the use of the name, artwork, etc[11:13:01] <JBurton> see freebsd, openbsd, netbsd, etc[11:13:07] <JBurton> they all have different restrictions[11:13:42] <gotaku> I started learning Erlang recently... it's a very interesting language.[11:21:45] <gotaku> I'm still trying to find some part of Haiku I can help with, but so far I got nothing.[11:22:10] <JBurton> like... you don't like anything, or you can't find anything you CAN do ?[11:22:59] <gotaku> The latter.[11:23:49] *** mmu has quit IRC[11:25:51] <JBurton> oic[11:26:20] <JBurton> do you have any experience with the be api ?[11:26:47] <gotaku> No, but that wouldn't be a problem.[11:27:28] <gotaku> No real experience I mean. I have made a few toy apps.[11:28:03] <JBurton> ah ok[11:28:07] <JBurton> then you DO have :P[11:28:39] <gotaku> Limited.[11:29:08] <gotaku> What applications need work or need to be made?[11:29:53] <myob> erm... serial connect :P[11:30:00] <Ingenu> Visual Studio for Haiku[11:30:17] <Ingenu> OpenGL drivers for nVidia & ATi/AMD hardware[11:30:18] <myob> as goes the user-visible apps in haiku, most of the beos ones have been 'cloned' or improved upon[11:30:40] <Ingenu> (of course that's what I'd need to use Haiku on a daily basis)[11:30:48] <Ingenu> and VMWare Fusion for Haiku[11:31:10] <myob> Ingenu: you won't be using it for some years then...[11:31:10] <gotaku> Oh, so OpenGL doesn't work in Haiku yet?[11:31:22] <myob> gotaku: it does. In software, and hardware for OLD nVidia cards[11:31:47] <myob> the fact that it does hardware is amazing, however.[11:32:19] <gotaku> Maybe I can work on that... I have an older 9600 Radeon card and the xorg drivers seem to work fine.[11:32:54] <myob> isn't that R300 era?[11:33:01] *** Pulko_Mandy has joined #haiku[11:33:10] <myob> my old R100 (!!) had decent OSS 3D drivers but nobody to port them...[11:34:05] <gotaku> RV360 I think.[11:34:55] *** PulkoMandy has quit IRC[11:35:23] <myob> yeah, the R300 drivers are exceptionally slow compared to the Windows ones...[11:37:10] <gotaku> I'm not sure what the V means though.[11:37:16] <myob> faster than 2D though[11:37:19] *** Pulko_Mandy has quit IRC[11:37:25] <myob> Radeon Video xxx[11:37:40] <myob> nothing different to Rxxx...[11:37:47] *** PulkoMandy has joined #haiku[11:38:32] <gotaku> Exceptionally slow?[11:38:59] <Ingenu> well I may consider helping on the OpenGL side, provided there's some sample code[11:39:17] <Ingenu> but that would only be for the new version of OpenGL released in fall, not interested in supporting current API[11:39:39] <gotaku> Is there a guide to installing Haiku to a partition from Linux?[11:41:46] *** GreyGhost has quit IRC[11:42:31] <myob> day is going so sloooowly here[11:44:27] <JBurton> gotaku there are some recent mails about that[11:44:30] <JBurton> on the main list[11:46:50] <gotaku> What was the rational for getting rid of the wiki?[11:50:15] <JonathanThompson> Hi guys.[11:50:26] <JonathanThompson> I hate it when I can't sleep and I need to work later in the morning.[11:52:20] <myob> JT just caffeine your way through it. Oh wait....[11:52:52] <JonathanThompson> For all I know, I may be best off just keeping awake and not trying to sleep this morning.[11:53:03] <myob> with the aid of caffeine, that works[11:53:05] <JonathanThompson> Since once sleep comes, it's hard to wake up again.[11:53:13] <myob> believe me (and about 95% of other geeks...)[11:53:24] <JonathanThompson> Tihng is, that messes up my sugar balance horribly.[11:53:33] <JBurton> hi JonathanThompson[11:53:42] <myob> you know from experience?[11:53:44] <JBurton> gotaku I can't remember[11:53:51] <JonathanThompson> And my body is annoyingly sensitive to sugar levels, or doesn't regulate things well.[11:53:53] <JBurton> gotaku ah yes[11:54:00] <JBurton> gotaku it was because we only wanted quality content[11:54:06] <myob> sugar free irn bru[11:54:33] <JonathanThompson> I do know, myob, that sleep deprivation affects the ability to process/use sugar effectively, and to regulate it.[11:55:25] <myob> yes but do you know the effects of caffeine on you staying awake?[11:55:33] <myob> caffeine comes in many sugar-free formats...[11:55:54] *** eNGIMa___ has quit IRC[11:56:29] <myob> and of them, diet irn bru has very little in the way of carcinogenic sweetners (as its not a sweet drinkk))[11:56:40] <gotaku> JBurton: It's my understanding that the wiki had a lot more information then is now available.[11:56:53] *** JonathanThompson has quit IRC[11:57:15] *** JonathanThompson has joined #Haiku[11:58:29] *** larrythecow has joined #haiku[11:58:33] <JonathanThompson> I seemed to have hit a snag.[11:58:59] *** larrythecow has left #haiku[11:59:00] <JonathanThompson> And even if that may be so for irn bru, I'm not aware of having ever seen that in the US.[11:59:16] <JonathanThompson> Not to mention I don't react well to the artificial stuff, either.[11:59:25] *** tombhadAC has joined #haiku[11:59:30] <JonathanThompson> And caffeine doesn't help me regulate sugar, either, myob.[11:59:41] <myob> irn bru is the only soft drink to outsell coke in any place in the world[12:00:00] <JonathanThompson> I've never heard of it, or seen a commercial or ad for it in my life.[12:00:05] <myob> JonathanThompson: I said it'd keep you awake, not affect your sugar regulation[12:00:10] <JonathanThompson> A few minutes ago is the first time I've ever heard of it here.[12:00:29] <myob> JonathanThompson: well, its insanely massive in Scotland and widely available in Ireland and the UK...[12:00:38] <myob> and outsells Coke in Scotland[12:00:42] <JonathanThompson> If my sugar regulation is too whacked, no amount of caffeine helps me function, myob, because the brain runs on sugar.[12:00:57] <JonathanThompson> But... that has no meaning in the US :)[12:01:32] <myob> its bright orange, not sweet, and has a taste thats... not easily describably[12:01:37] <myob> describable*[12:01:39] <Ingenu> neither in france[12:02:05] <JonathanThompson> Seems Canada has it, according to Wikipedia.[12:02:33] <myob> www.irn-bru-usa.com appears to be their US distributor[12:02:38] <myob> its CLEAR in Canada for some reason[12:02:43] <JonathanThompson> The US isn't listed as having it, and then there's the question of how widely available in Canada it is, since Canada is around the size of the US for area, but less than the population of California.[12:03:16] <myob> where are you in the US?[12:03:22] <JonathanThompson> Seattle area.[12:03:31] <JonathanThompson> A few miles from Seattle, a little farther from Redmond.[12:03:40] <myob> not sold in Washington[12:03:56] <JonathanThompson> I bet it's not sold in Michigan or Indiana, either.[12:04:01] <myob> Florida, Texas, New York, California, North Carolina and, erm, "MA" - what state is that?[12:04:14] <JonathanThompson> MA=Massachusetts.[12:04:19] <myob> thats where its sold[12:04:22] <JonathanThompson> (I think I spelled that wrong)[12:04:40] <gotaku> Where are there no clear instructions for installing Haiku to a hard drive from Linux?[12:04:52] <JonathanThompson> I've spent less than a week total in two of the states that sell it :)[12:04:56] <myob> anyway, want me to ship some over ;)[12:05:36] <JonathanThompson> Maybe a little over a week total if you count California, but I was only there briefly too involved in a business trip in February of 93.[12:05:52] *** bobwhoops has quit IRC[12:06:03] *** bobwhoops has joined #haiku[12:06:09] <JonathanThompson> Or was that 94... either way, I only got to see a little of the California sun.[12:06:45] <JonathanThompson> And that was while I was inside, looking outside from a meeting :([12:09:30] <myob> anyway, do you want some shipped over? :P[12:09:47] <JonathanThompson> It'd take too long anyway :P[12:10:07] <myob> it has a long shelf life[12:10:31] <JonathanThompson> Not to mention, shipping liquid foodstuffs overseas must be about the most wasteful manner of working with things.[12:10:48] <JonathanThompson> So do tweenkies :P[12:11:24] *** larrythecow has joined #haiku[12:11:27] *** larrythecow has left #haiku[12:12:28] <JonathanThompson> Oh, another thing that's made worse with sleep deprivation: asthma and symptoms.[12:12:29] <myob> well, it works for beers... imported heineken is way better than Irish brewed, and so on[12:13:39] <JonathanThompson> I suspect those beers are shipped in very large quantities, and probably largely bottled/canned locally, even if the beer itself is shipped.[12:13:41] * myob notes that the system he's on has a hostname named after an imported beer[12:13:46] <myob> (shell on a friends server)[12:14:14] <myob> JonathanThompson: imported beer in europe = bottled in the country its coming from[12:14:18] <JonathanThompson> Yahoo seems to have a list of words that they randomly string together for internal machine names.[12:14:40] <JonathanThompson> For example, my most recent box is "ourtimerealize"[12:14:55] <myob> werid[12:15:11] <JonathanThompson> Another is "Tomorrowrulez"[12:15:19] <myob> my hostname here is "ed", as the previous user of this desk was called ed and we don't reimage the machiens often...[12:15:31] <JonathanThompson> And the first Linux box (it'll be leaving me soon) is named "eveningchair"[12:16:09] <JonathanThompson> There's a box used by the other guy in the office I'm in named "uncleproud" which seems odd.[12:16:26] <JonathanThompson> And then a server is named "seemedwritten"[12:16:31] <JonathanThompson> (or close to that)[12:16:41] <JonathanThompson> Or was it seemswritten...[12:17:20] <JonathanThompson> I guess they had to figure out something that could scale reasonably well, with probably almost all employees having computers on their intranet.[12:17:28] <myob> beats "PC15656" which was my system in BT[12:18:01] <JonathanThompson> Back when I worked at DADC we named all our machines after looneytoon characters.[12:18:20] <JonathanThompson> The Animaniacs helped extend our naming options at a crucial time :)[12:18:53] <myob> all my PCs at home are named after either new order tracks or dance tracks... laptop is Airwave[12:18:54] *** Debolaz has quit IRC[12:19:18] <JonathanThompson> This machine is currently the one I've named DrMindbender[12:25:26] *** gotaku has quit IRC[12:31:08] *** rhu has joined #haiku[12:33:06] *** axeld has joined #Haiku[12:33:07] *** ChanServ sets mode: +o axeld[12:35:30] <myob> I hate Windows network mounts...[12:41:52] *** ekdahl has joined #haiku[12:45:21] *** rcjsuen has joined #haiku[13:08:36] *** Sloar has joined #haiku[13:16:51] *** karijes has quit IRC[13:21:24] *** JBurton has quit IRC[13:50:09] *** weltschmerz has quit IRC[13:51:08] *** kr1stof has joined #haiku[13:54:37] *** tombhadAC has quit IRC[13:59:18] *** kr1stof has quit IRC[14:07:30] *** ekdahl has left #haiku[14:09:31] *** Sil2100 has joined #haiku[14:31:07] <myob> this is the quietest I've seen this channel in some years[14:36:51] *** wildur has joined #Haiku[14:41:13] *** TTRanger has joined #Haiku[14:42:40] <myob> I'm so bored I'm tempted to post to HNN again :o[14:43:54] <TTRanger> myob A talented guy like you, bored? Hey somebody, put him to work! :-)[14:44:43] <TTRanger> myob What do you know about BeOS Max/Haiku Max?[14:44:54] <myob> TTRanger: I'm getting paid rather a lot to do bugger all thank you very much ;)[14:45:04] <myob> TTRanger: nothing more than the public know[14:45:29] <TTRanger> myob The public probably knows more than I do.[14:45:40] <TTRanger> The word on their site is, though that ACCESS has given them an ok.[14:45:52] <myob> kinda...[14:46:13] <myob> "Lefty" appears to have said he knows of no objects to redistributing the binaries unmodified[14:46:25] <myob> remember now it ships with a modified kernel and zbeos[14:46:53] <TTRanger> So if Lefty gave his permission it might have been without knowlege of those?[14:47:15] <myob> possibly[14:47:18] <TTRanger> mmm[14:47:22] <myob> they're extremely minor mods[14:47:33] <TTRanger> I'm trying to find a "lifeboat" for TuneTracker[14:47:42] <TTRanger> at least until the ZETA thing is straightened out, if iever[14:48:00] <myob> yeah, as you can't really sell it w/o OS[14:48:05] <TTRanger> right[14:48:06] <myob> and R5 is, well.... decrepit[14:48:14] <TTRanger> I know[14:48:43] <TTRanger> If we could find a set of hardware that would run it fine, we could at least keep going.[14:49:00] *** guma has joined #haiku[14:49:01] <myob> but there really isn't for R5, unfortunately...[14:49:06] <TTRanger> mmm[14:49:17] *** GreyGhost has joined #haiku[14:49:28] <TTRanger> So BeOS Max doesn't sound like it's of much use, presently?[14:49:36] <TTRanger> Hi GreyGhost[14:49:39] <myob> have you considering contacting Lefty yourself? You could knock together a PE-based, "new drivers+firefox only" R5 very easily, if it was allowable[14:50:05] <GreyGhost> TTRanger ,hello[14:50:25] *** StylusEater_Work has joined #haiku[14:50:29] <TTRanger> myob That's a disctinct possibility.[14:50:49] <TTRanger> myob Isn't that what "Max" is though?[14:51:10] <TTRanger> Don't they include newer drivers?[14:51:13] <TTRanger> or Firefox?[14:51:14] <myob> TTRanger: no, Max is trying to be Linux On BerOS[14:51:23] <myob> its >400MB of, well, shite in my opinion[14:51:30] <myob> your users don't need 100 SDL games[14:51:52] <TTRanger> But they don't try to keep up to date drivers in it?[14:52:04] <StylusEater_Work> myob: +1[14:52:41] <TTRanger> myob The drivers are the issue for me...to your knowlege they don't keep up to date drivers in it?[14:53:29] *** mmu_man has joined #haiku[14:53:30] *** ChanServ sets mode: +o mmu_man[14:53:35] <TTRanger> mmu_man plop[14:53:41] <myob> they do keep them up to date, as up to date as possibly[14:53:44] <myob> possible*[14:53:50] <TTRanger> ok...[14:53:51] <GreyGhost> IMO .. an OS should come bearbone .. no extra shit than required (a basic package..) but must have easy access to the packages as well .. as in a package manager[14:54:02] <mmu_man> re[14:54:09] <myob> of existant R5 "distros" its the most likely to install/run[14:54:20] <TTRanger> k[14:54:24] <myob> but you have to bear in mind it has all of R5 serious limitations[14:54:43] <TTRanger> including the old network stack, I assume[14:54:58] <myob> >=1GB addressable RAM, 137GB HDD cut-off, 60GB max partitions (unless you use Dano binaries with it), net_server, etc[14:54:59] *** kb7sqi has quit IRC[14:55:21] <TTRanger> oh brother, yeah, forgot about the hard drive limitations[14:55:24] <TTRanger> oof[14:55:47] <myob> R5 is an OS of its era - late 1999 for first release to tests[14:55:50] <Sil2100> Sorry to interrupt your discussion, but could someone check whether Haiku's Trac works right now?[14:56:19] <Sil2100> I'd like to know if it's a problem here or in overall[14:56:34] <myob> Sil2100: up for me[14:56:41] <myob> and this is IRC, interruption is second nature ;)[14:56:48] *** GreyGhost has quit IRC[14:56:50] <Sil2100> Strange then...[14:57:28] <Sil2100> Since I'm getting an internal error everytime, after I tried to log into Trac this morning[14:57:44] <myob> TTRanger: out of vaguely manical interest, have you tried running the Tune* suite on Haiku?[14:58:03] *** GreyGhost has joined #haiku[14:58:05] <myob> I know its not stable, nor supportive of enough hardware to deploy, but... would indicate if its a closer event[14:58:06] <TTRanger> We couldn't yet...SoundPlay needs to be "fixed" first so we can run 4.9.3.[14:58:13] <mmu_man> http://dev.haiku-os.org/ works fine here[14:58:21] <myob> oh yeah, forgot that[14:58:26] <myob> is Marco willing to fix?[14:58:37] <myob> as that private API is unlikely to ever be implemented[14:58:38] <Sil2100> Oh well, anyway, this means that's something wrong here - thanks[14:58:44] <TTRanger> I think so...I think he just needs to hear from some folks who want the latest version.[14:59:02] <myob> hrm, I might buy a licence to poke him a bit...[14:59:12] <mmu_man> I still need to buy one[14:59:19] <myob> as a side note, I presume you use a plugin or scripting to communicate with soundplay[14:59:20] <mmu_man> even though I don't use it much since long ago[14:59:31] <myob> mmu_man: you buy one and I'll buy one, OK ;)[14:59:49] <mmu_man> myob the bin wants a specific symbol from libmedia[14:59:57] <myob> mmu_man: I know...[14:59:58] <mmu_man> somethign to enumerate codecs[15:00:05] <mmu_man> because he wanted to skip FFDecoder[15:00:10] <mmu_man> because he thought it was buggy[15:00:16] <TTRanger> myob Yes, scripting... SoundPlay has a great "back door" for controlling it.[15:00:24] <mmu_man> but he never filed me any bug repport =)[15:00:35] <Yez> heh[15:00:44] <StylusEater_Work> http://dev.haiku-os.org/ <-- works ok for me[15:00:53] <myob> TTRanger: public back door? :P[15:00:54] <Sil2100> After removing the cookies and relogging, Trac works fine again[15:01:22] <myob> TTRanger: as I'm having some terrible issues implimenting a decent playback engine for DJB[15:01:37] <myob> and, what, a tenner a licence for soundplay wouldn't destroy the pricing...[15:02:25] *** MikeW has joined #haiku[15:02:58] <myob> mmu_man: the wonders of knowing|having the source, eh (for marco that is)[15:03:27] <mmu_man> he could also have asked me for the source of FFDecoder and fixed it himself, I wouldn't have mind[15:04:03] <myob> think I'll put some effort in to HNN again now actually...[15:04:14] <myob> but have to make Chris fix some bustications first[15:06:48] *** MikeW has quit IRC[15:09:25] <myob> TTRanger: if you do ask access, you'll find a few people willing to show you how to do a BeOS 5+ style assembly on BeShare/here[15:10:35] <TTRanger> myob Sorry I was on another screen for a moment...reading your posts[15:11:13] <TTRanger> myob Poor terminology...let's say it does "hey" stuff well. :-)[15:11:33] *** tombhadAC has joined #haiku[15:12:07] <myob> TTRanger: "hey soundplay change pitch of track 1 to 98%" style level? :P[15:12:47] <TTRanger> Not sure about pitch, but start, stop, adjust volumes, eject tracks, relay streams[15:13:26] <myob> pitch would be... important[15:13:38] <myob> as would being able to control two seperate instacnes of SP[15:13:48] <myob> might go back to writing me own :P[15:13:51] <TTRanger> myob Regarding creating my own BeOS 5 distro, I'd still be up against the hard drive size thing...that'd kill it for me I'm afraid.[15:14:04] *** guma has quit IRC[15:14:25] *** mats| has joined #haiku[15:14:30] *** mats has quit IRC[15:14:31] <TTRanger> myob Are you sure you need two different SoundPlays? SoundPlay can playback and control multiple tracks at once.[15:14:55] <myob> TTRanger: two outputs[15:14:58] <myob> essential for a DJ[15:15:18] <TTRanger> Then you'd need support for two sound cards too.[15:15:46] <myob> I also need to be able to get tracks to start at specific times, forgot that[15:15:49] <TTRanger> That's actually something we have wrestled with. We need to have an "audition" channel but haven't been able to make it work so far.[15:15:58] <myob> TTRanger: yes, I'm experimenting with channel munging...[15:16:05] <myob> 'audition' channel?[15:16:21] <mmu_man> for the DJ to listen to stuff before it's aired[15:16:24] <TTRanger> "listen to a track in a separate circuit without it going 'over the air'"[15:17:07] <myob> TTRanger: well, thats (close to) exactly what a club DJ needs[15:17:12] <TTRanger> righty[15:17:20] <myob> monitor channel[15:17:23] <TTRanger> yep[15:17:42] <myob> I've a feeling that if you have a second sound card and a second instance of an app you could do decidated routing to that card[15:18:07] <myob> I've two cards in my desktop for that purpose[15:18:22] <myob> just need to get it working now - and selectable from the app frontend[15:18:25] <mmu_man> yes but either the app knows abotu it[15:18:35] <mmu_man> or you'll haev to disconnect it from the system mixer as it's done by default[15:18:40] <mmu_man> and reconnect to the 2nd mixer[15:18:41] <TTRanger> SoudnPlay can be run in multiple instances, though SoundPlay would have to be adapted to allow it to address specific audio hardware.[15:19:01] <myob> mmu_man: well, I'm intending that my app -will- know about it ;)[15:19:02] <TTRanger> The present version of SP only uses the default.[15:20:02] <myob> which is why I may need to write my own playback stuff[15:20:30] <mmu_man> hmm maybe with a different SP output plugin[15:20:41] <mmu_man> rewrite teh Media Kit output :)[15:20:53] <Yez> myob doesn't TuneTracker already do all of this and more?[15:21:55] <TTRanger> Yez We don't use multiple audio cards...that's the problem right now. We digitally mix everything and output through a single sound card.[15:22:20] <Yez> ah[15:23:01] <myob> Yez: TuneTracker is (no offence to TTRanger) entirely unsuited for club DJing[15:23:07] <myob> its like trying to eat soup with a fork[15:23:19] <myob> perfect for another similar purpose (eating food, like "playing sound")[15:23:26] <myob> but completely useless for that situation...[15:23:32] <Yez> ok[15:23:47] <mmu_man> hmm[15:23:54] <mmu_man> SP doesn't have an output plugin API[15:24:02] <mmu_man> thougth so but only CL-Amp does[15:24:03] <TTRanger> no unfortunately[15:24:09] <myob> its massively featured in comparison to what a club DJ needs - but doesn't have some features they do...[15:24:23] <TTRanger> ah, so in CL-Amp you *can* select audio output?[15:24:41] <mmu_man> myob still I suppose reusing code from TT could help a lot[15:24:57] <myob> mmu_man: yeah but I'd need to either buy out or be part of TTsystems for that :P[15:25:01] <mmu_man> TTRanger well there are several output plugins (Media Kit, file output...)[15:25:12] <mmu_man> but it lacks many features from SP[15:25:16] <mmu_man> even multiple tracks[15:25:27] <myob> and its unmaintained. and claes doesn't answer emails about it...[15:25:51] <TTRanger> It's too bad Claes didn't open source it[15:32:47] <myob> especially as it uses a lot of OSS decoder code, etc[15:37:54] <myob> I need to get VNC access to my PC at home, these dead days are getting so incredibly boring...[15:39:32] *** Jixt has quit IRC[15:40:13] *** kb7sqi has joined #haiku[15:44:43] <myob> someones using their callsign....[15:51:11] *** GreyGhost has quit IRC[15:56:22] *** guildencrantz has joined #haiku[15:59:20] *** stargater has joined #haiku[15:59:36] <stargater> hi[15:59:43] <myob> guildencrantz long time no see[16:05:04] *** eNGIMa has joined #haiku[16:06:10] <stargater> cu[16:06:15] *** stargater has quit IRC[16:13:52] *** petterhj has joined #haiku[16:14:43] <myob> petterhj: someone was looking for BeM[16:14:43] *** the_webers_inc has joined #haiku[16:14:48] <myob> BePortable* on here last night[16:15:51] <petterhj> I did abandoned it in favour of WirelessManager, but what about it?[16:16:25] <myob> `well, I can't remember anything else...[16:16:40] <petterhj> I'll check the logs :)[16:16:56] <myob> considering your blog is ~year old I can't even remember wirelessmanager ;)[16:17:24] *** MrSunshine has joined #haiku[16:19:08] <mmu_man> I was just reminding everyone that there were already something to handle wireless connections and other stuff for laptops[16:19:09] <petterhj> myob: to refresh your memory: http://beos.petterhj.net/projects/wireless_manager_beta_2_build_22052005.png .. i did release that, but I kind of abandoned it all togheter.. I'm guessing Haiku will have wifi configurability soon enough[16:19:39] <myob> can't do links from ehre, using irssi :P[16:20:01] <mmu_man> ah I knew I didn't dream it :)[16:20:20] <myob> petterhj: it'll have it sooner if you write it ;)[16:20:20] *** Mazon is now known as mazon[16:20:33] <mmu_man> indeed[16:21:04] <petterhj> hehe.. WirelessManager does only configure the drivers own text files. I'm guessing haiku will have some other kind of integrated system?[16:21:17] <mmu_man> probably yes[16:21:25] <petterhj> at least a standard configurationfile[16:21:35] <myob> yeah but it'll still need a frontend - and a config API...[16:21:46] <myob> intergrated in8 to the new network pref would be great...[16:23:03] <petterhj> yea, thats what I hope.. I wish I could code it, then I would, but I'm far from good enough.. I hope I'll pick up learning C++ again soon though.[16:24:39] <petterhj> guess we have something by the end of the summer though, through SoC :)[16:26:43] *** dr_evil has joined #haiku[16:26:51] <myob> nothing for a preflet in the SoC is there?[16:26:54] *** ChanServ sets mode: +o dr_evil[16:27:01] <dr_evil> hi axeld[16:28:22] *** mmadia has joined #haiku[16:28:34] <petterhj> myob: i thought this was it? http://code.google.com/soc/haiku/appinfo.html?csaid=F651A269990CEF91[16:29:14] <myob> petterhj: remember what I said about links ;)[16:29:33] <myob> but yeah there is actualy one...[16:29:37] <petterhj> :)[16:29:54] <myob> irssi doesn't send links through screen through ssh to firefox[16:33:34] <TTRanger> re[16:35:13][16:35:35] <myob> putty != good at that ;)[16:36:28] <dr_evil> hi mmu_man[16:38:16] <CIA-17> hugosantos * r20676 /haiku/trunk/ (8 files in 7 dirs): use module_dependencies to load the required modules by udp, tcp, ipv4, icmp and arp.[16:45:28] *** MikeW has joined #haiku[16:46:44] <TTRanger> can someone *please* help me with a downloading issue regaring ncftpget (or) wget ?[16:46:48] <guildencrantz> myob howdy. I drop in now and again.[16:47:23] <TTRanger> guildencrantz howdy stranger![16:47:37] <myob> guildencrantz: still making bread (or was that tic?)?[16:47:54] <guildencrantz> myob: both of us do ;-)[16:48:16] <mmu_man> r[16:48:17] <mmu_man> re[16:48:20] <mmu_man> dr_evil :)[16:48:56] *** wildur has quit IRC[16:52:58] <DeadYak> meep[16:56:13] * _hugo pets DeadYak[17:02:20] * DeadYak pets _hugo[17:05:54] *** MrRagga has joined #haiku[17:08:42] *** mazon is now known as Mazon[17:11:28] *** Lelldorin1 has joined #haiku[17:14:39] *** arikato has quit IRC[17:15:54] *** arikato has joined #haiku[17:15:59] *** kr1stof has joined #haiku[17:21:37] *** Kernel86 has joined #Haiku[17:29:08] *** Kernel86 has quit IRC[17:34:14] <myob> DeadYak: I never asked this - when did you stop being evil and, err, die?[17:35:15] *** Sloar has quit IRC[17:36:03] *** Anxiety|Work has joined #haiku[17:36:11] *** Lelldorin1 has quit IRC[17:36:51] *** korli has joined #haiku[17:36:51] *** ChanServ sets mode: +o korli[17:37:05] *** doppeljot has joined #haiku[17:37:43] <korli> hello[17:38:24] *** doppeljot has left #haiku[17:39:21] *** tqh has joined #haiku[17:40:04] <DeadYak> myob: when I started being on IRC entirely via a shell client[17:40:54] <myob> how does that make you either dead or less evil?[17:41:13] <DeadYak> myob: because for a time I had both connected at once and I picked this name for the shell[17:41:40] <myob> oh right[17:41:49] *** GreyGhost has joined #haiku[17:41:57] <myob> well lowercase myob is my shell and uppercase MYOB is home. They're never gonna be online at once...[17:42:04] <myob> far easier way :P[17:43:38] <tic> <3 irssi[17:43:45] <TTRanger> tic![17:43:52] <tic> great scott![17:43:58] <TTRanger> hehe[17:44:18] <tic> how's things flying?[17:44:26] *** MrRagga has quit IRC[17:44:32] <tic> myob, http://svnbox.mine.nu/dump/shots/irssi-20060516-2.png for example[17:44:33] <TTRanger> tic kinda low to the ground right now... http://www.tunetrackersystems.com[17:44:49] <tic> TTRanger, mhm, the Zeta deal. :/[17:44:54] <TTRanger> yeah[17:45:00] <tic> a shame[17:45:01] <TTRanger> I have faith things will work out though[17:45:08] <tic> let's hope so.[17:45:11] <TTRanger> yes[17:45:23] <TTRanger> I was just getting ready to release a whole new product line when this happened![17:45:30] <tic> woo.[17:45:37] <tic> Murphy in action.[17:45:55] <TTRanger> the two came within an hour of each other[17:46:14] <tic> splendid[17:46:15] <TTRanger> Actually it's fortunate the bad news came right before I released rather than just after![17:46:20] <myob> tic I'm in irssi also, so I can't click that link...[17:46:40] <tic> myob, aww.[17:47:00] <myob> highly customised irssi?[17:47:20] <tic> not really. just a config and a theme[17:47:22] <tic> it's a screen[17:47:34] <DeadYak> tic: plain old ircii for me[17:47:38] <tic> two instances running. each irssi has the highlights window on the top[17:47:41] <tic> it's really nifty.[17:48:18] <myob> I just typed "ir[tab][enter] and used whatever opened[17:48:24] <mmadia> any idea why wget ftp://user:pass@IP_ADDRESS:PORT screws up?[17:48:47] <mmadia> *the syntax. wget somehow chokes on port[17:49:04] <myob> I just called someone at the desk beside me thinking they were out of the office :embarassed:[17:49:42] <DeadYak> mmadia: mmm[17:50:01] <myob> must look over cuboid walls more often...[17:50:25] <tic> work work work.[17:50:31] <DeadYak> I'm not sure[17:50:35] <tic> I need to /do/ something.[17:52:41] <tqh> TTRanger Do you do any live broadcasts these days?[17:52:58] <CIA-17> marcusoverhagen * r20677 /haiku/trunk/headers/private/drivers/dvb.h: Interface definition of DVB drivers. Currently private API.[17:53:05] <TTRanger> tqh Not anymore, really...I'd love to still do that but so busy all the time![17:53:42] <myob> tic feel like porting something maybe? :P[17:53:47] <tic> myob, porting what now?[17:53:59] <myob> tic erm. erm. eeeeerm.... bugger[17:54:08] <myob> can't think of much I need/want :P[17:54:13] <tqh> TTRanger ah those were always fun to listen to[17:54:22] <myob> libusb ;)[17:54:26] <TTRanger> :-) thanks...they were fun to do too[17:54:50] <myob> i'm going for a radio DJ job, new station, wonder can I convince them to use TT ;)[17:55:54] <TTRanger> Good luck, on both counts myob! :-)[17:56:03] <GreyGhost> how hard is it to port linux software to Haiku/ BeOS ?[17:56:11] <myob> GreyGhost: depends[17:56:20] <myob> command line no network app = usually very simple[17:56:23] <TTRanger> depending on the software, sometimes it's just a recompile[17:56:26] <GreyGhost> hmm ..[17:56:33] <myob> MonoDevelop/OpenOffice/etc = good fecking luck....[17:56:39] <GreyGhost> myob , ok .. :) ..[17:56:51] <GreyGhost> TTRanger ,yeah .. i would just love that ;)[17:56:52] <myob> what app?[17:57:00] <PulkoMandy> SDL games should'nt give any problem[17:57:49] <GreyGhost> ahh cool :)[17:57:55] <GreyGhost> myob ,have no idea yet ;)[17:58:47] <korli> bbl[17:58:54] *** korli has quit IRC[17:59:26] <PulkoMandy> http://smw.72dpiarmy.com/ :)[18:00:53] *** myob has quit IRC[18:04:16] *** PascalR has joined #haiku[18:07:51] *** PascalR has quit IRC[18:08:51] <guildencrantz> tic! Holy heck, it's like reunion day in here.[18:09:42] <tic> :)[18:11:04] *** emitrax has quit IRC[18:11:21] *** umccullough has joined #haiku[18:12:53] <umccullough> What In Blazes Is Going On?[18:13:27] <mats|> =)[18:16:40] *** mmadia is now known as mmadia_[18:21:20] *** tqh has quit IRC[18:21:20] <CIA-17> marcusoverhagen * r20678 /haiku/trunk/src/add-ons/kernel/drivers/dvb/ (20 files in 2 dirs): Driver for DVB-T cards with cx23882 chipset, like WinTV NOVA-T PCI[18:23:55] *** the_webers_inc has quit IRC[18:27:44] *** MauriceK has joined #haiku[18:27:54] *** Ingenu has quit IRC[18:28:03] *** Kernel86 has joined #Haiku[18:32:21] *** tqh has joined #haiku[18:36:00] <CIA-17> jackburton * r20679 /haiku/trunk/src/apps/terminal/ (10 files):[18:36:00] <CIA-17> Renamed *Dlg to *Window, since the term "Dialog" is alien to the beos[18:36:00] <CIA-17> api. Thanks to Vasilis Kaoutsis for the suggestion. I'll rename the[18:36:00] <CIA-17> classes with the next commit.[18:41:40] *** slaine_ has quit IRC[18:41:46] <CIA-17> jackburton * r20680 /haiku/trunk/src/apps/terminal/ (6 files): Renamed *Dlg classes to *Window, completing my previous commit[18:42:36] *** jiuda_D`arkness has joined #haiku[18:44:26] *** TLF has joined #haiku[18:54:25] *** weltschmerz has joined #haiku[18:54:37] *** dr_evil has quit IRC[18:59:18] *** dr_evil has joined #haiku[18:59:18] *** ChanServ sets mode: +o dr_evil[18:59:59] *** tqh has quit IRC[19:01:25] *** Euver has joined #haiku[19:05:43] *** fanskap has quit IRC[19:06:04] *** Fanskapet has joined #haiku[19:06:20] *** evdubs has quit IRC[19:07:45] <dr_evil> ping axeld[19:07:52] <axeld> Hi dr_evil[19:08:31] <mmu_man> plop[19:08:39] <mmu_man> gahh, stupid cdrecord stuck at 4x[19:08:49] *** lorglas has joined #haiku[19:08:52] <mmu_man> faster! I'm late in 3 minutes![19:10:45] <mats|> burn damnit burn[19:11:14] *** Ingenu has joined #Haiku[19:14:46] *** mmu_man has quit IRC[19:14:49] *** GreyGhost has quit IRC[19:17:12] *** lorglas has quit IRC[19:21:27] *** schmedly3d has joined #haiku[19:22:53] *** GreyGhost-Ubuntu has joined #haiku[19:22:54] *** gotaku has joined #Haiku[19:23:05] *** gotaku has left #Haiku[19:23:11] *** gotaku has joined #haiku[19:24:47] *** rhu has quit IRC[19:28:06] *** rhu has joined #haiku[19:28:06] <GreyGhost-Ubuntu> ok .. i've already built Haiku once .. now i used svn update..[19:28:16] <GreyGhost-Ubuntu> do i use just ./configure[19:28:49] <GreyGhost-Ubuntu> or ./configure --build-cross-tools ../buildtools/ ?[19:28:50] *** rhu has quit IRC[19:29:00] <GreyGhost-Ubuntu> i belive that the second will rebuild gcc ?[19:29:14] <GreyGhost-Ubuntu> but i need to just recompile haiku ..[19:30:54] <GreyGhost-Ubuntu> no matter.. i think i usnderstood ..[19:31:06] <GreyGhost-Ubuntu> just ./confgure gave me an error ;)[19:31:42] *** dr_evil has quit IRC[19:31:46] *** rhu has joined #haiku[19:31:47] *** dr_evil has joined #haiku[19:34:02] * GreyGhost-Ubuntu kicks himself[19:34:11] * GreyGhost-Ubuntu now waits for gcc to rebuild :([19:35:00] *** kokito has joined #haiku[19:35:24] <DeadYak> kokito![19:35:47] *** jevin has quit IRC[19:38:16] <GreyGhost-Ubuntu> how do i specify ./configure to use a precompiled gcc for the next time ?[19:38:21] *** infamy has joined #haiku[19:39:39] <kokito> hey DeadYak[19:39:51] <kokito> lots of people here today :)[19:39:58] <dr_evil> you can't use configure to switch I think, don't know how to solve that[19:40:07] *** schmedly3d has quit IRC[19:40:17] *** dsg has quit IRC[19:40:54] <GreyGhost-Ubuntu> dr_evil , no i dont want to switch .. i mean .. do i have to compile gcc everytime i use ./configure ?[19:41:18] *** JamesB192 has joined #haiku[19:41:32] <dr_evil> can't answer that[19:41:48] *** jiuda_D`arkness has quit IRC[19:41:49] <GreyGhost-Ubuntu> ok .. i'll look around then :)[19:41:50] *** ChanServ sets mode: +o dr_evil[19:43:46] <GreyGhost-Ubuntu> dr_evil ,i think its this option called .. --cross-tools-prefix .. i'll try next time :)[19:47:43] <GreyGhost-Ubuntu> lol .. everyone read the first post on osnews ..[19:47:55] <DeadYak> for?[19:48:03] <DeadYak> ah[19:48:21] <GreyGhost-Ubuntu> for rofl ing !!![19:49:14] <Anxiety|Work> I read that already... the author doesnt make a compelling case, its just a reason to attempt to rehash some of the negative publicity vista has gotten[19:50:38] <{V}> why do they need a reason for that? :D[19:50:55] <GreyGhost-Ubuntu> i havent read it ;) /.. but from what i've heard people who like eye candy like Vista :)[19:51:08] <GreyGhost-Ubuntu> but make it open source ??[19:51:25] * GreyGhost-Ubuntu goes looking at what is WalterOS ..[19:51:32] *** stargater has joined #haiku[19:53:17] <GreyGhost-Ubuntu> who is this Goldman thing ?[19:53:34] <TTRanger> DeadYak Do you know the script trick that tells Terminal to launch a line from the script in a separate instance of Terminal?[19:55:00] <GreyGhost-Ubuntu> O_o .. he means to say that he will live with the same security loop holes with an older version of Windows but not with Vista ..[19:55:25] *** arikato has quit IRC[19:55:49] <GreyGhost-Ubuntu> and crackers create attacks ?? shouldnt that be hackers..[19:56:00] <GreyGhost-Ubuntu> why am i ranting bout this in #haiku though ?[19:56:02] <GreyGhost-Ubuntu> :D[19:58:01] <Sil2100> According to the jargon dictionary, crackers are also the ones who 'hack into' computer systems[19:58:26] <DeadYak> TTRanger: not offhand....[19:58:33] <Sil2100> So I think this word is a bit more appropriate[19:58:40] <GreyGhost-Ubuntu> Sil2100 ,ok .. my bad tech grammar then ;)[19:59:25] <GreyGhost-Ubuntu> oh well .. bed time .. night[19:59:27] <GreyGhost-Ubuntu> *all[19:59:31] <Sil2100> Good night[19:59:38] *** GreyGhost-Ubuntu has left #haiku[20:00:15] <stargater> hi[20:00:21] *** MikeW has quit IRC[20:00:42] <stargater> dr_evil, nice :-)[20:02:55] *** JamesB192 has quit IRC[20:07:20] *** oco has joined #haiku[20:11:08] *** Lelldorin1 has joined #haiku[20:13:50] *** arikato has joined #haiku[20:15:29] *** infamy has quit IRC[20:20:57] *** weltschmerz has quit IRC[20:29:00] *** kb7sqi has quit IRC[20:30:06] *** Lelldorin1 has quit IRC[20:31:33] <_hugo> meep[20:32:50] *** rgb has joined #haiku[20:32:57] <kokito> beep[20:33:37] *** Meanwhile has joined #haiku[20:36:10] *** kb7sqi has joined #haiku[20:44:21] <DeadYak> sqeep[20:44:44] *** evdubs has joined #haiku[20:45:17] *** gareth has joined #haiku[20:46:10] *** MikeW has joined #haiku[20:49:49] <guildencrantz> TTRanger - just incase you're still looking....Terminal -t "foo"[20:50:16] <TTRanger> Ah, thanks guildencrantz ... trying it[20:50:21] <guildencrantz> foo being the window name, append the commands to run in the new terminal[20:51:34] <TTRanger> hah! very cool...thanks![20:51:55] <guildencrantz> That one took some remembering.[20:52:06] <TTRanger> hehe[20:52:24] <TTRanger> I on the other hand would forget it in 10 minutes, so I'll have to add it to my list of "syntaxes to remember" file[21:00:07] *** Lelldorin1 has joined #haiku[21:03:50] <Sil2100> kokito: are there any chances for a Waltercon in Japan in the near future?[21:05:20] <gotaku> Heh, probably not.[21:05:53] <stargater> mähp[21:06:16] <Sil2100> Oh well...[21:08:08] <kokito> Sil2100, I have been entertaining the idea of a WC in Japan for a while.[21:08:31] <kokito> Sil2100, but have no specifics. are you in Japan?[21:08:57] <gotaku> kokito, I thought you quit the Haiku project anyway?[21:09:11] <kokito> yes gotaku, I did.[21:09:40] <gotaku> So why are you still here?[21:09:50] <_hugo> gotaku: why are you here?[21:09:54] <kokito> why not?[21:10:06] <gotaku> I'm just asking...[21:10:09] <Sil2100> kokito: sadly, no - but I would certainly attend such though[21:10:19] <gotaku> So you didn't quit after all.[21:10:19] <_hugo> gotaku: please stop this persecution. anyone who wants to help may do so[21:10:33] *** mmadia_ is now known as mmadia[21:10:34] <kokito> gotaku, I am not an admin anymore.[21:11:00] <gotaku> Oh ok.[21:11:08] <kokito> but if you want, I will stop visiting my friend on this channel. > gotaku[21:11:36] <gotaku> ?[21:13:36] <kokito> gotaku: you seemed annoyed by my presence, so was asking you if you wanted me to leave the channel.[21:13:57] *** gareth has quit IRC[21:14:01] <gotaku> Why would I be annoyed by your presence? I don't even know you.[21:14:38] <kokito> ok, cool then. :)[21:15:17] <gotaku> Although I could see how "So why are you still here?" could be taken the wrong way.[21:16:17] <kokito> Sil2100, the BeOS community is quite small, and have mostly been supporting ZETA. now that zeta is caput, they are starting to shift their attention to Haiku. So 2008 may be a good time to do a WCJ.[21:16:20] <dr_evil> please stay kokito[21:16:29] <kokito> dr_evil!![21:16:52] <dr_evil> hello kokito![21:17:44] *** oco has quit IRC[21:17:56] <kokito> hello dr_evil! nice news about DVB coming to Haiku. ;)[21:18:03] <Sil2100> kokito: that's a splendid idea[21:18:20] <Sil2100> Anyway, I'm looking forward to 2008 then[21:18:39] <[Beta]> hang on, 2007 is going to be good too :P[21:18:43] <dr_evil> kokito yes, i already started today, but can't continue during the weekend[21:19:38] <kokito> Sil2100, but it will really depend on the JP community and Haiku (the organization) willing to make an investment (to bring over at least a couple of core devs.[21:19:39] <Sil2100> [Beta]: where is this year's Waltercon planned?[21:19:42] <kokito> nice dr_evil[21:19:45] <gotaku> Can Haiku reach R1 by the end of 2007?[21:19:51] <DeadYak> dr_evil: does the R5 BT8x8 addon not work with Haiku's Media Kit?[21:19:52] <Sil2100> kokito: I see[21:21:03] <[Beta]> No one has said, but i'm sure it'll exist.[21:21:37] <Sil2100> kokito: Axel seemed interested in visiting Japan sometime, so that might be a good chance[21:22:08] <kokito> yes, we spoke about it when he was in the US[21:22:19] *** DebolazX has joined #haiku[21:22:39] <gotaku> Sil2100: Why are you so interested if you're not in Japan?[21:22:44] <dr_evil> I think i had some trouble when i tried that years ago, it needed some undocumented stuff from libmedia[21:22:55] <DeadYak> aha.[21:23:00] * kokito thinks BGA should be giving his Haiku lecture right now...[21:23:02] <DeadYak> that explains that[21:24:09] <Sil2100> gotaku: maybe because I wanted to visit Japan in the nearest few years anyway, and with Waltercon JP I could both visit the country and attend the conference at one time[21:24:32] <Sil2100> But other than that, no particular reason[21:28:46] <kokito> Sil2100, Japan is fun. lots to do and visit.[21:29:26] *** DebolazY has quit IRC[21:30:39] *** meianoite has joined #haiku[21:31:04] * DeadYak pets meianoite[21:31:12] <meianoite> hi axeld, dr_evil, geist[21:31:13] * meianoite purrs[21:31:26] <meianoite> I'm straight, btw ;)[21:31:48] *** Meanwhile has quit IRC[21:31:53] <DeadYak> wasn't implying otherwise :)[21:32:17] <axeld> Hi meianoite[21:32:19] <meianoite> right... (I hope so ;))[21:32:38] <meianoite> axeld mind if I ask you a couple things?[21:32:42] <kokito> hey meianoite[21:32:43] <meianoite> ok, maybe not just a couple[21:32:48] <meianoite> heya kokito[21:33:44] <axeld> meianoite: No, but I'm afraid I'm away now for today[21:33:54] *** jevin has joined #haiku[21:34:01] <meianoite> it won't take long[21:34:25] <axeld> Hi kokito![21:34:36] <meianoite> and we definitely must agree on some time to meet here :)[21:36:29] <axeld> meianoite: it's mostly bad in the evening, except on monday[21:36:45] <DeadYak> time zones don't help[21:37:09] *** kr1stof_ has joined #haiku[21:37:36] <meianoite> I can see that[21:38:00] <meianoite> the thing is that I just can't be awake at that time[21:38:07] <kokito> hey axeld![21:38:07] <meianoite> it was 4am last night when I logged off[21:38:19] <meianoite> GMT-3 vs GMT+1, you know[21:38:27] <mmadia> there is always email ; )[21:38:28] <meianoite> right now it's 16h37 here[21:38:39] *** emitrax has joined #haiku[21:38:42] <meianoite> mmadia some people tend to hate them, geist for example[21:38:47] <axeld> meianoite: it's DST here, so it's 5 hours[21:39:01] <meianoite> eh, that didn't help ;)[21:39:05] <geist> I HATES TEH EMALE[21:39:06] <axeld> meianoite: so it should work from 9 to 12 for you :-)[21:39:14] <DeadYak> hi geist[21:39:23] <axeld> geist: the females or emails? ;-)[21:39:35] <meianoite> geek humour alert[21:39:36] <meianoite> geek humour alert[21:39:37] <meianoite> geek humour alert[21:39:38] <meianoite> ;)[21:39:47] <MikeW> that's because he uses mail.app? *ducks*[21:40:02] * meianoite crouches and slaps MikeW[21:40:05] * DeadYak has flashbacks to Strong Bad Email[21:40:07] <mmadia> femails = autonomous pornographic emails from "female" senders : P[21:40:57] *** gotaku has quit IRC[21:41:56] <meianoite> holy M.o.G. mmadia.[21:42:01] <meianoite> manners, please ;)[21:42:20] <axeld> night everyone :-)[21:42:20] * meianoite hides his stash of porn[21:42:42] <kokito> good night axeld[21:42:43] <_hugo> see you axeld, have fun[21:42:44] <mmadia> no, no... i'm talking about spam bots.[21:42:56] *** axeld has quit IRC[21:43:01] * mmadia should've clarified better[21:43:02] <meianoite> mmadia ooooooooookay... ;)[21:43:09] <meianoite> let's pretend I believe you[21:43:24] <meianoite> and that you're not part of those Yahoo! groups[21:43:28] <meianoite> heh ;)[21:43:29] *** PulkoMandy has quit IRC[21:43:55] <meianoite> jk, jk, jk please[21:43:56] <mmadia> i'm not.[21:43:56] *** PulkoMandy has joined #haiku[21:43:56] <dr_evil> hmmm, I don't know if it's a good idea to go to 80ties party tonight, considering I slept 3 hours last night, and need to work tomorrow[21:44:09] * mmadia shakes a trout at meianoite : )[21:44:33] * meianoite grabs that trout, grills it, and bakes himself a nice dinner[21:44:39] <meianoite> *cooks[21:44:45] <meianoite> eew, trout cake. no, no =P[21:45:21] <meianoite> dr_evil in case you have trouble sleeping, there's always that NVIDIA SATA thing to tackle ;)[21:45:45] *** kr1stof has quit IRC[21:47:02] <meianoite> anyway. which one should be better to run & debug Haiku (especially the kernel), QEMU, VMWare, something else?[21:47:26] <meianoite> "real hardware" won't do, sorry. (unless the SATA stuff is sorted out)[21:47:31] <dr_evil> vmware for running, qemu for debuggig[21:47:59] <dr_evil> sata will need a few more weeks, i'm sorry[21:48:00] <meianoite> plain QEMU, no kernel accel, right?[21:48:17] <meianoite> seems that kqemu borks Haiku[21:48:20] <dr_evil> no idea, i only use vmware+real hardware[21:48:22] <_hugo> i use vmware for both running and debugging[21:48:29] <meianoite> hey _hugo[21:48:35] <_hugo> hello[21:48:49] <meianoite> you emulate serial console and all?[21:49:05] <meianoite> I'm wondering how am I supposed to read the serial console on VMWare[21:49:12] <_hugo> meianoite: vmware has an option to dump the serial to a file[21:49:12] <meianoite> QEMU is straightforward[21:49:21] <meianoite> hm... that's good to know[21:49:27] <dr_evil> see http://svn.berlios.de/viewcvs/haiku/haiku/trunk/3rdparty/vmware/haiku.vmx?rev=19989&view=markup[21:49:36] <meianoite> even the free VMWare Server one?[21:49:37] <dr_evil> I need to leave now[21:49:43] <_hugo> meianoite: i use vmware player[21:49:45] <_hugo> see you dr_evil[21:49:53] <meianoite> w00t, even the player then! :)[21:49:56] <MikeW> speaking of serial, do any of you use a physical serial cable to get information?[21:50:01] <meianoite> I guess those .vmx files are a good idea afterall[21:50:14] <_hugo> MikeW: that would be from a real machine, yes[21:50:33] * meianoite used to despise metaconfig files[21:50:39] <MikeW> _hugo: Is it serial to serial, serial to usb?[21:50:52] <MikeW> hrm, I mean to db9 or usb[21:51:00] <_hugo> MikeW: i used a null modem cable[21:51:03] <MikeW> oh[21:51:04] <_hugo> so serial to serial[21:51:24] <MikeW> wait, does a null model cable actually plug into a modem?[21:51:24] <meianoite> any plans to get Firewire debugging working?[21:51:34] <_hugo> MikeW: no, it plugs two serial ports[21:51:43] <meianoite> (I guess we'd need the assistance from the guy doing the Firewire GSoC project)[21:51:51] <_hugo> so i used it to connect the serial port from the computer being debugged to another computer[21:52:02] <dr_evil> vmware can dump serial data to a file, you can print that with tail -f[21:52:25] <dr_evil> but I also used a real serial cable like hugo, for real hardware[21:52:37] <MikeW> _hugo: Do you know if there is a null modem USB adaptor? serial -> usb?[21:52:45] <meianoite> aye, I guess I'll have no way around clearing a partition to install Linux... Windows generally blocks those dump files. out of sheer stupidity.[21:52:51] <dr_evil> you can create a vmware image using "jam haiku-vmware-image"[21:52:54] <dr_evil> well, need to go now, bye[21:52:57] * MikeW goes to look for a usb to serial adaptor+emulator[21:52:57] * meianoite looks at the failed hard drive in anger[21:53:00] <PulkoMandy> MikeW: use a usb>serial then a null modem cable plugged on it[21:53:01] <_hugo> MikeW: if you dont have a serial port, you can use a usb serial cable in the other computer, just plug the null modem cable to it[21:53:15] <_hugo> the machine running haiku does need a serial port though[21:53:36] *** dr_evil has quit IRC[21:53:36] * meianoite pets Firewire and nods at DeadYak[21:55:22] * JonathanThompson lights a Firewire and pets DeadYak[21:55:32] * meianoite w00ts[21:55:51][21:56:47] *** petterhj has quit IRC[21:56:57] *** petterhj has joined #haiku[21:57:35] * meianoite tries to find his R5 Pro retail box[21:58:00] * JonathanThompson thinks he still has his from when 5.0 came out[21:59:23] <meianoite> I guess the box is in the basement[21:59:33] <meianoite> I'll have a pretty hard time trying to find it =P[21:59:41] <JonathanThompson> If mine were in the basement, I'd have a real problem :)[21:59:54] * JonathanThompson hasn't lived in a place with a basement since 1991[21:59:55] <meianoite> ?[21:59:58] *** cmlenz has joined #haiku[22:00:09] <stargater> hi cmlenz[22:00:20] <cmlenz> howdy[22:00:41] * JonathanThompson wanders back to work[22:00:44] <JonathanThompson> Seeya[22:01:21] <evdubs> what sort of problems was the vmware video driver causing?[22:01:26] <evdubs> it ran perfectly for me[22:01:39] <meianoite> guess I'll have no way around of burning a PE CD... the basement scares me.[22:02:01] * meianoite is probably letting out unnecessary info about his fobias[22:02:46] <MikeW> Well I found a USB to serial adaptor at http://www.dabs.com/ProductView.aspx?Quicklinx=3Z2W[22:02:51] <MikeW> I just can't tell if its male or female[22:03:30] <meianoite> female to male converter is probably next on the shopping cart then :)[22:04:05] <Grackle> I'm sure the rs-232 port on there is female.[22:04:23] * meianoite is off to the basement. *shivers*[22:04:37] <MikeW> so I'd need a female to male null modem cable[22:05:05] <Grackle> er, I mean male.[22:05:29] <Grackle> Usually stuff on the computer is female, but rs-232 is male.[22:05:42] <geist> yeah it's backwards[22:05:45] <meianoite> there used to be bats living in the freakin' basement[22:05:51] *** wildur has joined #Haiku[22:06:00] <geist> which means you can break the little penisis off on the back of your computer[22:06:03] <geist> instead of the cable[22:06:30] <MikeW> what the hell?[22:06:43] <meianoite> amusing choice of words, geist[22:06:53] <Grackle> haha[22:06:56] <meianoite> well guys, wish me luck. me against the freakin' bats =P[22:06:57] <MikeW> So yeah, http://www.dabs.com/ProductView.aspx?Quicklinx=3Z2W and http://www.dabs.com/ProductView.aspx?Quicklinx=3ZML and http://www.dabs.com/ProductView.aspx?Quicklinx=3ZMK are on my shopping list then[22:11:09] <Grackle> MikeW, is the point of this to get debugging information for that specific hardware? If not, then (I'm pretty sure) qemu can emulate a serial console, which would be cheaper if it served your purpose/[22:11:38] <MikeW> Grackle: Yes, I want to capture debugging information from the hardware[22:11:48] <Grackle> okay[22:12:14] <MikeW> So I'll get both cables just to be on the safe side[22:12:31] <Grackle> sounds good[22:13:00] <MikeW> Is it true that USBa to USBa can fry your computer?[22:13:07] <geist> probably[22:13:10] <Grackle> What sort of information do you get from the debug console? Is it just output, or can you do anything interesting from there?[22:13:18] <DeadYak> USBa = ?[22:13:28] <DeadYak> Grackle: output[22:13:30] <geist> the square plug[22:13:36] <DeadYak> ah.[22:13:38] <DeadYak> what's different about it?[22:13:40] <geist> A is the upstream plug, B is the downstream[22:13:43] <DeadYak> that would make it be able to..[22:13:45] <DeadYak> ah[22:13:47] <geist> and then there are mini A and mini B plug[22:14:06] *** StylusEater_Work has quit IRC[22:14:24] <DeadYak> I always get confused between which mini is which[22:15:30] <Grackle> Heh and now there is a micro- spec.[22:15:44] <DeadYak> I believe the one on my SK is Mini-B but I might be wrong[22:16:16] <Grackle> http://upload.wikimedia.org/wikipedia/commons/a/ad/Mini_usb_AB.jpg[22:16:24] <Grackle> Mini-A is on the left, Mini-B is on the right.[22:16:36] <DeadYak> ah.[22:16:38] <DeadYak> Mini-B[22:16:44] <DeadYak> what's the extra tiny one like on cameras?[22:16:56] *** Lelldorin1 has joined #haiku[22:17:06] <Grackle> I've only ever seen mini-B on cameras.[22:17:18] <DeadYak> the USB connector on my camera is barely a slit[22:17:22] <DeadYak> it's smaller than those two[22:17:25] <Grackle> But I noticed that there is a micro-usb specification, with plugs like this: http://en.wikipedia.org/wiki/Image:MicroUSB_0115.jpg[22:17:31] <Grackle> Ah, must be micro.[22:17:40] <Grackle> That's so tiny.[22:17:40] <DeadYak> yeah, that's it.[22:17:41] *** MikeW has quit IRC[22:18:15] *** jpd has joined #haiku[22:19:25] <meianoite> well, no results on the basement. except for the self-realization that I'm a sissy.[22:19:45] <DeadYak> hey, those vampire bats can be scary[22:20:10] <meianoite> at this time of the day there's NO light inside the basement[22:20:15] <meianoite> no electricity[22:20:22] <meianoite> LOTS of humidity[22:20:25] <meianoite> mold everywhere[22:20:33] <meianoite> and the eventual bat, spider, whatever[22:20:58] <meianoite> I could probably grow some magic mushrooms there with ease[22:21:03] <Grackle> haha[22:23:11] <Sil2100> ^^;[22:23:22] <Sil2100> Interesting basement indeed[22:23:28] <Sil2100> Good night everyone[22:23:41] *** Sil2100 has quit IRC[22:28:31] *** MikeW has joined #haiku[22:31:44] <meianoite> unlike axeld, I go the extra mile NOT to be locked in the basement[22:31:45] <meianoite> ;)[22:31:47] *** rgb has quit IRC[22:32:10] <MikeW> Bought the cables! :)[22:32:36] <meianoite> (chains and iron balls nonwithstanding)[22:33:18] *** mmu_man has joined #haiku[22:33:19] *** ChanServ sets mode: +o mmu_man[22:34:01] * DeadYak pets mmu_man[22:34:12] <mmu_man> re[22:34:26] <DeadYak> meianoite: you say that now, but wait until BGA comes over...[22:34:41] <meianoite> I'd rather be locked inside his place[22:34:46] <meianoite> he's got a Wii, you know ;)[22:34:58] <meianoite> and a large fridge[22:35:12] <meianoite> and a cool stereo heh[22:35:21] <DeadYak> I know :P[22:35:54] <meianoite> mmu_man any plans to port kqemu to BeOS? now that the source became free[22:36:45] <mmu_man> EBUSY :)[22:36:48] *** pikapika has joined #haiku[22:37:02] <meianoite> B_GOD_DAMN_IT =P[22:37:40] <pikapika> hello[22:37:43] <meianoite> zippy QEMU would rock.[22:38:02] <DeadYak> kqemu?[22:38:09] <meianoite> hi, extremely electrically charged pokebeing[22:38:15] <meianoite> DeadYak yep[22:38:20] <DeadYak> the difference being?[22:38:28] *** Euver has quit IRC[22:38:45] <petterhj> "QEMU Accelerator (KQEMU) is a driver allowing the QEMU PC emulator to run much faster when emulating a PC on an x86 host."[22:38:48] <petterhj> had to check myself :p[22:38:53] <DeadYak> ah.[22:39:07] <mmadia> didn't mmlr port qemu?[22:40:02] <MikeW> wow, anyone know why fabrice released it for free?[22:40:39] <_hugo> probably because kvm is in the mainline kernel now[22:41:01] <meianoite> MikeW he stated he eventually would, as long as he had the financial backing of some company interested on its development[22:41:31] <meianoite> so I guess he's standing on a small pile of cash right now ;)[22:41:55] <MikeW> yes, 100 cent coins![22:42:08] <meianoite> still cash ;)[22:43:41] <mmu_man> btw I'm not even sure the BeOS kernel has the required features[22:43:50] <mmu_man> probably needs something to handle LDT and other crap[22:43:55] <meianoite> mmadia I stand corrected, but I think mmu_man was involved as well[22:44:22] <mmu_man> I did some fixes[22:44:55] <meianoite> at least it should be easier now to see why in hades Haiku KDLs right away when kqemu is enabled[22:46:33] *** jpd has quit IRC[22:54:04] *** mmu_man has quit IRC[22:57:24] *** evdubs has quit IRC[22:58:27] *** jevin has quit IRC[22:58:56] *** MikeW has quit IRC[22:59:02] *** jevin has joined #haiku[23:02:31] *** Sloar has joined #haiku[23:05:29] <kokito> does Haiku have a ramfs?[23:05:51] <DeadYak> didn't the userland filesystem stuff include one?[23:05:58] <DeadYak> er wait, that was a network fs[23:05:59] <kokito> hmmm... not sure[23:06:10] <kokito> actaully, you may be right DeadYak[23:06:22] <DeadYak> ah[23:06:24] <DeadYak> Ingo, ever the busy coder, introduced the UserlandFS module as previously widely reported. Including a ReiserFS 3.6 and parts of a RAM FS and BeOS networked FS it promises to be an excellent tool for the development and usage of filesystem drivers alike.[23:06:37] <DeadYak> "parts of" though[23:07:38] <kokito> right. thanks![23:08:06] <DeadYak> I'm not sure if that implies it's functional but incomplete, or what[23:08:13] <DeadYak> but you could try :)[23:08:48] <umccullough> ramfs was the part that was "work-in-progress" - but maybe it's progressed enough to work now ;)[23:10:55] <PulkoMandy> +++[23:12:07] *** eNGIMa has quit IRC[23:12:12] *** PulkoMandy has quit IRC[23:12:30] <meianoite> kokito I seem to remember a RAMfs[23:12:36] <meianoite> by mmu_man[23:12:49] <emitrax> for those of you who is interested in my gsoc project, I opened a blog usbstack.blogspot.com[23:13:11] <meianoite> emitrax didn't you get that email from waldemar?[23:13:25] <meianoite> he'd open a blog account for you @haiku-os.org[23:13:26] <kokito> emitrax, are you not going to blog on the Haiku website then?[23:13:41] <emitrax> hmmm didn't read that I guess[23:13:50] <meianoite> ah, the overexcitement ;)[23:13:55] <kokito> LOL[23:13:56] <emitrax> :)[23:14:01] <meianoite> it's really thrilling to be a GSoC student[23:14:29] <kokito> emitrax, logon to the website, and click on the SUBMIT icon, and see if you see "Blog" in the type of content you can submit.[23:15:27] <emitrax> blog entry[23:16:11] <kokito> that's right[23:16:24] <kokito> so, you can already blog, emitrax :)[23:16:45] <umccullough> yes, the effort to get developer blogs onto the official haiku site was a good move ;)[23:16:57] <umccullough> prevents people from hunting down blogs all over the place :D[23:17:03] <kokito> it will appear here > http://haiku-os.org/blog[23:17:18] <emitrax> I ll give it a try[23:17:36] <umccullough> then your blog entries can be promoted to front page if they're newsworthy ;)[23:18:02] <emitrax> well I was actually using that blog just to post thoughts and doubts about my project[23:18:05] <kokito> emitrax, your own blog would be here: http://haiku-os.org/blog/emitrax[23:18:27] <kokito> and you can also sindicate it; this would be your RSS feed: http://haiku-os.org/blog/1139/feed[23:19:21] <emitrax> meianoite was right... I'm overexcited !! :D[23:19:52] <meianoite> o/[23:20:33] *** Meanwhile has joined #haiku[23:21:27] <meianoite> ooh. logging on as root on IRC.[23:21:32] <meianoite> Meanwhile :)[23:21:44] <meianoite> careful :)[23:22:33] <Meanwhile> heheh I dont know what I'm doing, true[23:22:56] <Meanwhile> I'm on a live CD OpenSolaris([23:25:48] <Meanwhile> getting nervous......bye[23:25:53] *** Meanwhile has left #haiku[23:26:32] <meianoite> lol[23:26:42] <meianoite> I just scared the hell out of him ;)[23:27:54] <emitrax> hmm I'm trying to understand how the actual data is sent to the USB device[23:28:07] <emitrax> I'm looking that ehci driver code[23:28:08] <umccullough> usb_raw?[23:28:14] <umccullough> oh[23:28:19] <emitrax> EHCI::SubmitAsyncTransfer[23:28:31] <umccullough> mmlr is of course the one you probably wanna talk to[23:28:38] <emitrax> because the PeriodicTransfer is not implemnted[23:28:51] <emitrax> he s not here right ?[23:28:53] *** Meanwhile has joined #haiku[23:29:00] <umccullough> don't think so...[23:29:05] <Meanwhile> lol meianoite[23:29:11] <emitrax> I'll post it on my *blog* anyway :)[23:29:20] <meianoite> hehe Meanwhile ;)[23:31:44] <DeadYak> mmlr does not, to my knowledge, join IRC[23:31:49] <DeadYak> you'd be best off emailing him[23:32:29] <meianoite> 18:31] -> *nickserv* info mmlr[23:32:30] <meianoite> [18:31] -NickServ- Nickname: mmlr[23:32:30] <meianoite> [18:31] -NickServ- Registered: 46 weeks 3 days (22h 53m 18s) ago[23:32:30] <meianoite> [18:31] -NickServ- Last Seen: 6 weeks 3 days (20h 33m 50s) ago[23:32:30] <meianoite> [18:31] -NickServ- Last Seen Address: i=mmlr at 145 dot 74.76.83.cust.bluewin.ch[23:32:30] <meianoite> [18:31] -NickServ- Last Seen Quit Msg: "Vision[0.9.7-SF-010705]: i've been blurred!"[23:32:32] <meianoite> [18:31] -NickServ- Nickname Options: Secure, AllowMemos, MemoNotify, MemoSignon[23:32:41] <DeadYak> ah, so he does[23:32:47] <meianoite> well... he logs every couple of months, it seems[23:32:50] <DeadYak> albeit um...very not often[23:33:02] <meianoite> beats my frequency[23:33:20] <meianoite> GSoC prompted me to be around[23:33:30] <meianoite> but the last time I logged onto freenet was 2+ years ago[23:33:33] <CIA-17> hugosantos * r20681 /haiku/trunk/ (4 files in 4 dirs):[23:33:33] <CIA-17> respect SO_BROADCAST for received and sent datagrams.[23:33:33] <CIA-17> - check if the destination address is specified in IPv4's SendData()[23:33:33] <CIA-17> - minor cleanups to IPv4's TRACE()s.[23:35:02] <meianoite> now you had more than enough from me ;)[23:35:39] *** Ingenu has quit IRC[23:38:38] <DeadYak> :P[23:41:42] *** meianoite has left #haiku[23:41:43] <emitrax> stupid question[23:42:38] <emitrax> shouldn't it be there a kind of write function to the usb stack code? ... to send the actual data[23:43:44] <DeadYak> emitrax: I believe they're done as DPCs or something along those lines[23:44:23] *** oco has joined #haiku[23:45:42] <Yez> which network cards/chip sets work in Haiku?[23:46:17] <Yez> I have a 3Com NIC in my Haiku box now but Haiku can't seem to find it[23:49:00] <DeadYak> emitrax: try looking at EHCI::FinishTransfers[23:50:13] <DeadYak> some of what you're looking for might also be in the USB bus manager[23:50:22] <emitrax> DeadYak: I just posted about it on the blog[23:50:32] <emitrax> DeadYak: ok I'll look at it right away[23:50:55] <emitrax> the usb manages leads me to the driver code[23:51:09] <emitrax> that's where I started from[23:53:19] <[Beta]> which blog?[23:55:00] <DeadYak> emitrax: ehci_rh.cpp, EHCIRootHub::ProcessTransfer isn't it, is it?[23:55:28] <emitrax> [Beta]: http://haiku-os.org/blog/emitrax[23:55:39] <DeadYak> hm, maybe not[23:57:32] <emitrax> DeadYak: those memcpy I think does actually do some transfer data[23:57:39] <emitrax> do*[23:57:44] <DeadYak> memory mapped I/O perhaps?[23:58:04] <emitrax> that's what I thought[23:59:29] <emitrax> anyway I'll have a better look tomorrow[23:59:38] <emitrax> thanks DeadYak