Switch to DuckDuckGo Search
   April 12, 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 | >

Toggle Join/Part | bottom
[00:01:00] *** Mazon is now known as mazon

[00:03:46] *** BuildFactory has joined #haiku

[00:07:43] *** BuildFactory has quit IRC

[00:07:54] *** BuildFactory has joined #haiku

[00:08:20] *** MikeW has joined #haiku

[00:11:42] <[Katisu]> SiCuTDeUx, what is the model of you cardbus card?

[00:11:49] *** petterhj has quit IRC

[00:21:14] <[Katisu]> have to reboot, bbiab

[00:21:17] *** [Katisu] has quit IRC

[00:24:45] <CIA-17> hugosantos * r20660 /haiku/trunk/ (10 files in 6 dirs):

[00:24:45] <CIA-17> moved address selection logic to a new 'get_buffer_route'.

[00:24:45] <CIA-17>  - IPv4 now assumes the addresses it is supplied in send_routed_data are already the appropriate ones.

[00:24:45] <CIA-17>  - made the Data(), operator* and operator-> methods in NetBufferFieldReader const so we can use them in the same expression as the constructor.

[00:24:45] <CIA-17>  - fixed an issue with UDP where the wrong source address could be used in the calculating the checksum.

[00:24:49] <CIA-17>  - changed ipv4_print_address to use the more common 0.0.0.0 format.

[00:24:50] *** kr1stof_ has quit IRC

[00:26:13] *** avkig has quit IRC

[00:28:59] *** guma has joined #haiku

[00:34:27] *** guma has quit IRC

[00:39:30] *** avkig has joined #haiku

[00:41:31] *** [Katisu] has joined #haiku

[00:56:52] *** phoudoin has joined #haiku

[00:58:34] *** oco has quit IRC

[01:02:08] *** guildencrantz has quit IRC

[01:09:29] *** phoudoin has quit IRC

[01:09:42] *** Yez has quit IRC

[01:09:55] *** guma has joined #haiku

[01:10:23] *** Yez has joined #Haiku

[01:13:48] *** wildur_ has quit IRC

[01:22:56] *** meianoite has joined #haiku

[01:22:58] *** ul1984_ has joined #haiku

[01:23:49] <meianoite> geist, are you here?

[01:28:37] *** Hummin has joined #haiku

[01:28:39] <Hummin> hi!

[01:28:46] <Hummin> damn.. who is this hugo dude?

[01:28:51] <Hummin> he's on fire though..

[01:28:56] <Hummin> that's for sure!

[01:33:52] <_hugo> Hummin: hello

[01:34:48] *** ul1984 has quit IRC

[01:35:07] <Hummin> oh.. hi hugo!

[01:35:14] <Hummin> damn..you're a workhorse

[01:35:32] <_hugo> mostly bug fixes, trying to make the life better for haiku users

[01:35:47] <Hummin> hehe.. great work!

[01:36:16] <_hugo> thanks, happy i can help

[01:36:27] <Hummin> hehe.. humble you.

[01:36:40] <Hummin> are you one of the GSOC dudes?

[01:37:05] <meianoite> Hummin hopefully he will be

[01:37:11] <meianoite> we'll know tomorrow ;)

[01:37:17] <meianoite> actually

[01:37:33] <meianoite> I guess that it will be unveiled in... 30 minutes!!

[01:37:39] <_hugo> Hummin: like meianoite said, still waiting on the list

[01:37:48] <_hugo> meianoite: if there are no delays :-)

[01:37:56] <meianoite> _hugo, shut the hell up ;)

[01:38:09] <meianoite> that negativity attracts bad luck :D

[01:38:22] <_hugo> being realistic, they kept delaying it before so.. :-)

[01:38:26] <meianoite> let's just be cheerful! weeeeeeeeeeeeeee

[01:38:27] *** emitrax has joined #haiku

[01:39:20] <Hummin> still though.. it's nice to se someone so active

[01:39:52] <Hummin> I've been lurking.. haven't contributed much yet though.. thought I'd start working on some killer apps this summer.. =)

[01:41:13] <_hugo> Hummin: that sounds good.

[01:41:49] <Hummin> I guess.. I'll see if I can port my minimalistic game engine too.. =)

[01:42:05] <_hugo> :-)

[01:42:31] <Hummin> anyways.. just wanted to drop by and say "yay" and seems like the "renaming desktop icons to long names" no longer freaks out

[01:42:44] <Hummin> although it doesn't wrap, and looks a little odd

[01:43:27] <_hugo> Hummin: it is open for you to improve :-)

[01:43:52] <Hummin> Hehe.. maybe I will!

[01:44:05] <Hummin> but for now.. It's sleepy time!

[01:44:07] <Hummin> peace!

[01:44:11] *** Hummin has quit IRC

[01:44:15] <meianoite> _hugo, you'd be surprised by how many people refrain from contributing out of pure shyness

[01:44:35] <meianoite> they think their skills are not up to the task

[01:44:39] <_hugo> meianoite: i've been developing open source software for some years now, so i have an idea :-)

[01:44:45] <meianoite> their coding habilities is not on par

[01:44:47] <meianoite> and so on

[01:44:53] <meianoite> cool. I'm on the first group

[01:44:56] <meianoite> the shy ones ;)

[01:45:17] <meianoite> hopefully I'll start getting my feet wet with the GSoC

[01:45:26] <meianoite> if the powers that be so permit ;)

[01:46:04] <_hugo> of course they'll permit. just listen, learn and use the knowledge you have. im sure everything will be ok

[01:46:25] <meianoite> I mean being selected.

[01:46:32] <meianoite> I did apply, of course

[01:46:41] <meianoite> if selected, I'm all set.

[01:47:00] <meianoite> the task I chose to work on is something I'm comfortable with

[01:47:58] <meianoite> the chat I had with geist last night was VERY enlightening

[01:48:16] <meianoite> I learned a few things that really excited me

[01:48:34] <_hugo> good :-)

[01:49:10] <meianoite> for one, the fact that the design I did is pretty much what geist envisioned for Haiku

[01:49:53] <meianoite> I mean... he's an accomplished kernel engineer, for chrissake.

[01:49:58] <meianoite> I'm an undergraduate student

[01:50:27] <meianoite> how large a boost is it to know you're right there with extremely experienced people, and you had similar ideas?

[01:58:30] *** mmadia has joined #haiku

[02:12:43] *** emitrax has quit IRC

[02:14:19] *** SiCuTDeU1 has joined #haiku

[02:21:53] *** sarcas has left #haiku

[02:24:10] <meianoite> _hugo

[02:24:18] <meianoite> you were right

[02:24:20] <meianoite> =P

[02:25:35] <_hugo> :-)

[02:26:31] <MikeW> he was accepted?

[02:27:34] <MikeW> meianoite: What did you design?

[02:27:41] *** guma has quit IRC

[02:30:32] <meianoite> MikeW, http://dev.haiku-os.org/ticket/1069

[02:30:42] <meianoite> check that out and you'll have a fairly good idea

[02:31:19] <MikeW> ah wait, you posted to the mailing list

[02:31:50] <meianoite> that, too

[02:31:56] 

[02:32:17] *** SiCuTDeUx has quit IRC

[02:32:31] <_hugo> MikeW: i was right in the sense that the lists are delayed

[02:34:20] *** tombhadAC has quit IRC

[02:35:59] <meianoite> MikeW, not that the similarity is immediate

[02:36:05] <meianoite> you see

[02:36:23] <meianoite> my nick and my name DO share the letters "a", "e"...

[02:36:30] <meianoite> even the "n"!

[02:36:57] <MikeW> My nick and my name share all the letters :)

[02:37:13] <MikeW> First name Mike, Surname starts with W

[02:37:14] <meianoite> chicks like the mystery

[02:37:18] *** keller- has joined #haiku

[02:40:04] <MikeW> Blimey, thats some conversation you had with geist last night

[02:42:12] *** MikeW has quit IRC

[02:44:19] *** tombhadAC has joined #haiku

[02:47:11] *** meianoite has quit IRC

[02:47:34] *** meianoite has joined #haiku

[02:52:55] *** the_webers_inc has joined #haiku

[03:01:19] *** Sloar has quit IRC

[03:10:53] *** mmadia has quit IRC

[03:17:13] *** absabs has joined #haiku

[03:28:30] *** ablyss has joined #haiku

[03:29:31] *** M199 has joined #haiku

[03:37:12] *** keller-zZzz has joined #haiku

[03:37:17] *** keller- has quit IRC

[03:40:38] *** the_webers_inc has quit IRC

[03:41:31] *** M_199 has quit IRC

[03:44:26] *** SiCuTDeU1 has quit IRC

[03:50:27] *** bonefish has joined #haiku

[03:50:27] *** ChanServ sets mode: +o bonefish

[03:50:36] <bonefish> Howdy!

[03:50:47] <meianoite> hi, Ingo

[03:51:12] <_hugo> hey Ingo

[04:17:07] *** keller-zZzz has quit IRC

[04:27:37] *** kokito has joined #haiku

[04:46:26] *** rcjsuen has quit IRC

[04:49:55] *** ablyss has quit IRC

[04:57:26] <meianoite> yay

[04:57:47] <meianoite> http://code.google.com/soc/haiku/about.html

[04:57:54] <meianoite> _hugo, congrats

[04:58:08] * meianoite congratulates himself

[04:59:11] <meianoite> salvatore, JiSheng, Andre, Lukasz, Ivo, Krishna, congrats :)

[04:59:12] *** jevin has quit IRC

[04:59:48] <meianoite> not only that, but we managed to get EIGHT!!!!!!!!!!! slots!!

[05:00:01] <meianoite> Google is really making up for rejecting us last year

[05:17:49] *** guma has joined #haiku

[05:20:21] <kokito> congrats meianoite & _hugo :)

[05:20:30] <meianoite> thanks, kokito!

[05:20:34] <meianoite> muchas gracias!

[05:20:46] <meianoite> domo arigato

[05:20:59] <meianoite> and so on, I don't know how many languages you speak ;)

[05:20:59] <kokito> de nada meianoite :)

[05:21:18] <kokito> just three meianoite :P

[05:21:36] <[Katisu]> congrats mianoite and _hugo

[05:21:43] <bonefish> kokito: Hi!

[05:21:50] <kokito> hey bonefish!

[05:21:51] <meianoite> fair enough, I do manage to communicate in English... mostly.

[05:22:20] <meianoite> [Katisu], thanks

[05:22:25] <bonefish> Congrats to our accepted students! :-)

[05:22:27] <kokito> bonefish, we are all celebrating the GSoC announcement

[05:22:43] <meianoite> looking forward to being useful

[05:22:50] <meianoite> kokito I already submitted the news to OSNews

[05:22:51] <bonefish> kokito: Yeah, where's the booze? ;-)

[05:22:58] <meianoite> (even though I tend not to like them. =P)

[05:23:14] <kokito> hidden under the table bonefish :)

[05:23:50] <bonefish> kokito: Not here. *sigh* ;-)

[05:24:01] <meianoite> BGA used to have a TON of beer at his place

[05:24:26] <bonefish> I suppose we should all visit him then. ;-)

[05:24:45] <kokito> meianoite, I submitted a news to the haiku-os.org website. hopefully the admins will publish it at some point (maybe after some editing).

[05:25:09] <meianoite> but he already have drinked it all

[05:26:53] <meianoite> aye. I definitely must work on my English skills.

[05:27:08] <meianoite> *but he's already drinked it all

[05:27:09] <meianoite> better.

[05:27:22] <bonefish> drinked -> drunk(en) ?

[05:28:12] <meianoite> http://wsu.edu/~brians/errors/drank.html

[05:28:48] <meianoite> have drunk, I guess

[05:29:02] <meianoite> =P

[05:35:41] *** eNGIMa|thiele has quit IRC

[05:37:19] <Grackle> What do the little map markers on the gsoc page mean? (looking at this http://code.google.com/soc/haiku/about.html )

[05:37:34] <meianoite> it's where the students are from

[05:38:47] <Grackle> ...but they don't lead anywhere.

[05:38:52] *** Sloar has joined #haiku

[05:38:56] <Grackle> Is that just because they haven't put information in yet?

[05:38:58] <meianoite> seems not to work on Firefox, btw

[05:39:07] <Grackle> Oh, lame.

[05:39:08] <meianoite> nor on IE

[05:39:18] <_hugo> hi guys. thanks. congratulations to you too meianoite

[05:39:19] *** eNGIMa|thiele has joined #haiku

[05:39:31] <kokito> Grackle, I believe they point to a geographic location on google maps

[05:39:58] <meianoite> btw, geist, you were wrong

[05:40:08] <meianoite> the current code IS O(n)

[05:40:09] <meianoite> =P

[05:40:28] <bonefish> Good night!

[05:40:34] <_hugo> see you Ingo

[05:40:43] *** bonefish has quit IRC

[05:40:50] <Grackle> kokito, do they work for you?

[05:41:36] <kokito> actually, no Grackle ;)

[05:41:50] <Grackle> heh ok

[05:42:08] <kokito> Grackle, I should have said "they are supposed to..." :)

[05:42:11] <Grackle> I don't really need them to work, I just wondered what they were since they didn't seem to do anything.

[05:42:14] <Grackle> hehe ok

[05:42:29] * kokito is a bad communicator :P

[05:49:57] *** guma has quit IRC

[06:06:06] *** Kernel86 has quit IRC

[06:11:57] *** tkcvcv has joined #haiku

[06:12:05] <tkcvcv> how is the hardware support on haiku for hard drives?

[06:12:19] <tkcvcv> i.e. if I have a new sata HD that works fine under linux, can I assume I will get full performance out of it on haiku?

[06:14:52] <meianoite> KDE got 40 slots, NetBSD got 6, GCC got 8, FreeBSD got 25, PHP got 7 slots

[06:14:59] <meianoite> we're bigger than NetBSD! than PHP!

[06:15:10] <meianoite> and 1/5 as big as KDE

[06:15:11] <meianoite> we so rock :D

[06:15:13] <kokito> tkcvcv, I think there is limited SATA support in Haiku, but I don't know the specifics.

[06:15:39] <meianoite> and we're not even production ready yet! yeeey :D :D

[06:17:13] <Grackle> yay :)

[06:17:54] <meianoite> we rule so hard it's not even funny

[06:18:00] <tkcvcv> is this for google summer of code?

[06:18:05] <tkcvcv> how many did haiku get?

[06:18:07] <meianoite> tkcvcv, yep!

[06:18:08] <tkcvcv> where are the projects posted?

[06:18:14] <meianoite> EIGHT slots, dude!! EIGHT!

[06:18:39] <Grackle> http://code.google.com/soc/haiku/about.html haiku is right here

[06:18:39] <meianoite> http://code.google.com/soc/haiku/about.html

[06:18:46] <meianoite> Grackle beat me to it

[06:20:41] <tkcvcv> what does red ballon w/ black dot in it mean?

[06:21:24] <Grackle> It's a google map marker, it's supposed to take you to the student's location, but it doesn't work.

[06:21:49] <tkcvcv> ah; thanks

[06:21:49] *** tkcvcv has left #haiku

[06:22:10] <Grackle> It's almost like I'm smart, but I'm just repeating things.

[06:24:18] <Grackle> Plan 9 got 13 slots? Hm.

[06:24:38] <meianoite> Grackle you gotta respect the Bell Labs people

[06:24:51] <meianoite> and Plan 9 is quality code afterall

[06:25:04] <meianoite> not to mention how network transparency does interest Google. a LOT.

[06:25:07] <Grackle> I just hear so little about it, so I was kinda surprised.

[06:25:32] <meianoite> the key is: would Google benefit from it?

[06:25:41] <meianoite> in the case of Plan 9? HELL yes.

[06:25:56] *** MrSunshine has quit IRC

[06:26:17] <meianoite> at least I can see how they could

[06:26:25] <meianoite> but, consider this

[06:26:32] <meianoite> we're not production ready yet!

[06:26:41] <meianoite> they have been for almost a decade now

[06:27:02] <Grackle> Yeah, this is so exxciting. I only found out about Haiku a couple weeks ago, but already so much has happened in that time.

[06:27:05] <meianoite> so we're in great shape, all things considered

[06:27:16] <meianoite> but things are not done yet on the site

[06:27:26] <kokito> I have met one of the Plan 9 guys at a couple of conferences in Japan, and he was very fond of BeOS

[06:27:27] <meianoite> Mozilla Inc for example has NO slots ATM

[06:27:35] <meianoite> and I just can't believe they actually got none

[06:27:37] <Grackle> I definately like the BeOS approach to things.

[06:28:10] <meianoite> Grackle don't we all? :)

[06:28:21] <Grackle> I guess that's why we're here.

[06:28:30] <meianoite> heh ;)

[06:28:59] <_hugo> meianoite: how old are you, if you don't mind me asking

[06:29:05] <meianoite> 24

[06:29:34] <meianoite> I believe you don't mind telling me your age as well...?

[06:29:39] <_hugo> not at all, i'm 22

[06:29:43] <meianoite> I even suspect you're younger than me

[06:29:46] <meianoite> there.

[06:29:47] <meianoite> :)

[06:29:55] <_hugo> why would you suspect that?

[06:30:05] <meianoite> your sheer energy

[06:30:10] <Grackle> haha

[06:30:15] <geist> that's about how old axel is too, IIRC

[06:30:17] <_hugo> that was unexpected

[06:30:18] <meianoite> just look at those commit logs ;)

[06:30:33] <meianoite> welcome, geist

[06:30:44] *** Kernel86 has joined #Haiku

[06:30:48] <meianoite> btw, geist

[06:30:56] <meianoite> do you mind taking a look here?

[06:31:05] <meianoite> http://svn.berlios.de/viewcvs/haiku/haiku/trunk/src/system/kernel/scheduler.cpp?rev=20427&view=markup

[06:31:21] <meianoite> especially scheduler_enqueue_in_run_queue(struct thread *thread)

[06:31:25] <geist> hang on, gimme a minute

[06:31:28] <geist> in the middle of something

[06:31:34] <meianoite> and scheduler_reschedule(void)

[06:31:38] <meianoite> sure

[06:31:48] * JonathanThompson hopes it isn't in the middle of two slices of bread at a Cannibal's Convention

[06:32:09] <geist> no, adding a new job to my backup server

[06:32:39] <meianoite> take your time

[06:32:46] <JonathanThompson> Just got an interesting error message from ssh: "ssh-rand-helper produced insufficient data"

[06:33:15] <meianoite> I'm just surprised to see that what you told me last night no longer holds, seems someone just threw away the O(1) code... if it ever got to be there in the first place.

[06:33:58] <JonathanThompson> Well, it IS in a rapid state of flux.

[06:34:19] <meianoite> JonathanThompson, would you EVER discard O(1) code and replace it with O(n) code?!!

[06:34:32] <JonathanThompson> Yes, in certain situations :)

[06:34:58] <meianoite> low overhead O(1) code implied ;)

[06:34:59] <JonathanThompson> As in, when O(1) code requires something that's not feasible in the condition it's used in, or... it doesn't work 100% reliably.

[06:35:17] <meianoite> well, that's not "O(1)" code. that's BROKEN code.

[06:35:26] <kokito> LOL

[06:35:27] <meianoite> one that fucks up in constant time ;)

[06:35:29] <JonathanThompson> Broken code can be O(1) :)

[06:35:51] <JonathanThompson> What's worse is when it is a subtle break.

[06:35:59] <meianoite> that's just like people who like to speed in the middle of the city

[06:36:12] <meianoite> what the hell are they trying to do, get to the red sign first?!

[06:36:13] <JonathanThompson> You don't? :)

[06:36:32] <meianoite> no, I like them green. so I keep a civilized pace. :)

[06:36:34] * JonathanThompson notes meianoite has clearly never driven on Telegraph Road in the Detroit area

[06:36:47] <JonathanThompson> (AKA US 24)

[06:36:47] <geist> meianoite: ugh, you're right

[06:36:54] <geist> someone did fuck up the run queues

[06:36:55] * meianoite never driven abroad

[06:37:02] <geist> damnit, this is a piece of crap

[06:37:14] <meianoite> it's too bad BerliOS doesn't have SVN blame like newos.org has

[06:37:28] <meianoite> geist, it seems that I'll HAVE something to do afterall :D

[06:37:33] <meianoite> L O L O L O L

[06:37:41] <geist> but anyway, it's easy to fix

[06:37:48] <geist> that just some data structure thing. no biggie

[06:37:53] <meianoite> yeah, we both know so

[06:37:54] <meianoite> btw

[06:38:13] <JonathanThompson> I grew up in the Detroit area, and that's one very well-implemented set of lights and such for getting through a dense population area at 45-50 mph (legal speed limit on cruise control) or about 70 mph (the illegal but used higher harmonic speed).

[06:38:14] <geist> actually that whole linked list needs to be replaced with a doubly linked list

[06:38:19] <meianoite> could you *please* explain why you're limiting your bitmat to 32 bits?

[06:38:21] <geist> my guess is it's stupid in the sem code as well

[06:38:31] <geist> hmm? bitmat?

[06:38:34] <meianoite> actually, I'd not do double linked lists... but circular lists, I would.

[06:38:39] <meianoite> the LK code

[06:38:46] <meianoite> 32 priority levels

[06:38:49] <geist> so it can use clz

[06:38:52] <meianoite> sure

[06:38:53] <meianoite> but

[06:38:54] <Grackle> Heh JonathanThompson, I just got a warning for doing 70 when I should have been doing 50,

[06:39:06] <JonathanThompson> Where at, Grackle?

[06:39:07] <Grackle> I live out in rural maine though.

[06:39:22] <geist> meianoite: you know lk is designed to run on like 50mhz arms, with 64KB of ram?

[06:39:23] <meianoite> geist, consider this

[06:39:29] <meianoite> 1100000 =  96 - 128

[06:39:30] <meianoite> 1000000 =  64 -  95

[06:39:30] <meianoite> 0100000 =  32 -  63

[06:39:30] <meianoite> 0000000 =   0 -  31

[06:39:30] <meianoite> so

[06:39:30] <meianoite> we could have 4 integers

[06:39:32] <meianoite> and do prio >> 5

[06:39:36] <JonathanThompson> Ah, you see, you do 70 MPH on Telegraph road in the Detroit area (and in Detroit proper) duriing rush hour, along with everyone else :)

[06:39:50] <geist> heck, there's not enough ram to run more tha 5 or 6 threads at a time

[06:39:50] <meianoite> and choose the bitmap that way

[06:40:03] <meianoite> well, THAT explains it.

[06:40:07] <meianoite> but newos has similar code

[06:40:10] <geist> lk == little kernel

[06:40:12] <JonathanThompson> geist, my next machine won't have less than 4 gigs actual RAM :)

[06:40:15] <Grackle> Yeah, I was just by myself there, an easy target.

[06:40:29] <geist> i wrote it in a few days for a friend's project

[06:40:32] <meianoite> (unless I misread something completely)

[06:40:48] <geist> first at foremost it's supposed to be little, can't use more than 8k or so of footprint

[06:40:55] <geist> I'll deal with adding features later

[06:41:04] <JonathanThompson> Oh, and if you actually get on the interstates in Michigan, 80 mph is the minimum you'll be doing - in the slow lane :)

[06:41:19] <meianoite> geist I still can't help but be flabbergasted by how COOL it is that your latest design so closely reminds my own

[06:41:20] <geist> in fact, it really doesn't need 32 queues, but it's no biggy because of the clz instruction on the arm. single cycle to determine the first used queue

[06:41:50] <Grackle> haha geez

[06:42:00] <geist> probably only needs 4 priority levels or so

[06:42:07] <JonathanThompson> The Seattle area is COMPLETELY different.

[06:42:10] <geist> and actually I think it's configurable (as long as it's less than 32)

[06:42:25] <meianoite> I can see that.

[06:42:52] <meianoite> still

[06:42:55] <JonathanThompson> I've been frustrated getting onto I-90 from Mercer Island and getting stuck behind some idiot doing 35 mph without blinkers on, with no traffic issues, and a little moisture.

[06:43:05] <geist> anyway, yeah as i look at scheduler.cpp in haiku

[06:43:08] <meianoite> do you approve the usage of 4 bitmaps as I described?

[06:43:15] <geist> it's really screwed up. someone messed it up

[06:43:22] <meianoite> (beats doing 2 levels of priority queues)

[06:43:22] <geist> for one thing it's not part of thread.c

[06:43:36] <geist> so it can't inline a bunch of those thread queue stuff

[06:43:49] <geist> though that being said it was derived from an earlier version of the newos stuff

[06:44:11] <geist> and of course newos isn't exactly incredibly optimal in this area, but it's mostly a matter of just optimizing the data structures

[06:44:13] <meianoite> geist I'm under strong impression that it has NEVER used your O(1) code.

[06:44:19] <meianoite> I remember seeing similar code in 2004

[06:44:22] <meianoite> 3 years, dude

[06:44:36] <geist> haiku mostly sat still kernel wise until relatively recently

[06:44:37] <meianoite> so if the fuckup indeed happened, it has been this way for a LONG time

[06:44:55] <geist> it did. I think it was part of the initial 'oooh neat we have a kernel lets start fookling with it'

[06:44:56] <_hugo> meianoite: for some reason there is a project for a O(1) scheduler. in order to improve it.

[06:45:01] <geist> that frankly, we're still fixing bugs from

[06:45:09] <meianoite> _hugo you missed the discussion last night

[06:45:12] <_hugo> meianoite: i did not

[06:45:28] <JonathanThompson> I've  been reading the discussion on the list, and now this, and I find myself wondering if Haiku will have a notably better scheduler than BeOS and Zeta.

[06:45:30] <meianoite> geist stated the code on the kernel WAS O(1) already

[06:45:44] <geist> meianoite: anyway, the first thing you should probably do is make the queue pointer for threads double linked

[06:45:47] <_hugo> geist assumed the code that was there was the same as a particular newos' version

[06:45:59] <geist> yeah I mostly did

[06:46:25] <meianoite> well... I can't say I'm not happy with that. it means I'll get to code :D

[06:46:28] <meianoite> real code

[06:46:35] <_hugo> which is fine :-) i think it is early to go blaming on someone fucking the code. instead, concentrate on improving it

[06:46:36] <meianoite> not just some makeup code

[06:46:37] <meianoite> REAL code ;)

[06:46:49] * JonathanThompson notes tonight as the night meianoite stops producing faux code :)

[06:46:58] *** mmadia has joined #haiku

[06:47:13] * JonathanThompson waves a dead chicken at mmadia in welcome

[06:47:17] <meianoite> _hugo there's no one to blame, I'm only puzzled at the fact that such an ugly duckling sat there untouched

[06:47:18] <geist> i do know to be fair that the unified double linked list in newos was quite late

[06:47:19] <geist> probably 2005 or so

[06:47:32] <geist> and I had gone back and changed most linked lists to that

[06:47:36] <geist> including the scheduler queues

[06:47:37] <mmadia> mmmm... tasty.

[06:47:45] <_hugo> meianoite: well, im puzzled how haiku works considering the amount of people that worked on it in their free time over these past years

[06:47:54] <_hugo> lots of stuff gets rushed in order to work.

[06:47:59] <_hugo> working first, optimizations later

[06:48:15] <JonathanThompson> Organized mayhem, that's what it is, _hugo :)

[06:48:22] <meianoite> geist I'm still wondering if DLL (as in double-linked-lists) are really necessary

[06:48:27] <JonathanThompson> (semi-organized)

[06:48:28] <_hugo> JonathanThompson: that's opensource i think

[06:48:31] <geist> for scheduler stuff, absolutely

[06:48:32] <meianoite> why not a SLL with lazy removal?

[06:48:43] <geist> remember, the thread queue list node is recycled for lots of stuff

[06:48:54] <geist> the general idea is the thread belongs to at most one queue of something

[06:49:09] <geist> so if it's not in the scheduler list, it's either running (no queue) or blocked on something (on a semaphores queue)

[06:49:10] <meianoite> right

[06:49:16] 

[06:49:22] <geist> and the same next pointer is used for both

[06:49:30] <meianoite> now I lost you.

[06:49:41] <geist> and in the case of semaphores, it's much more likely that you'll get pulled out of the middle of a list

[06:50:11] <meianoite> the CSLL I used worked like this

[06:50:27] <geist> yeah, so it's generally pretty important to use a doubly linked list

[06:50:38] <geist> i can't believe someone replaced my code with this junk

[06:50:40] <meianoite> array[prio] -> last thread -> current thread -> other threads -> last thread

[06:50:50] <meianoite> so I indirectly pointer at the current thread

[06:50:56] <meianoite> removal is trivial then

[06:51:01] <meianoite> so is addition

[06:51:09] <geist> i have no idea what that means

[06:51:15] <meianoite> it becomes tail by adding to the front of the last, then rotating

[06:51:20] <meianoite> twice

[06:51:36] <geist> that sounds totally tied to that scheduler algorithm

[06:51:49] <meianoite> yep, it is.

[06:51:50] <geist> just stuff it in a DLL

[06:52:00] <geist> pop it when you run, push it ack when you're done

[06:52:06] <geist> it's just a little data structure

[06:52:17] <meianoite> to be honest, I didn't even pop it out

[06:52:32] <meianoite> unless I removed it from the ready q.

[06:52:34] <geist> and frankly it's a wast eof time cycle counting the linked list when you're about to go spend thousands of cycles context switching

[06:53:07] <JonathanThompson> Ah, yes, the 2% optimization that takes 80% of the time :)

[06:53:21] <meianoite> I can't see where I was counting there, but anyway. I see your point.

[06:53:25] <JonathanThompson> (to optimize in creation stage)

[06:53:39] *** kr1stof has joined #haiku

[06:53:42] <meianoite> IF you ever happen to need shaving even more memory, consider that CSLL with that extra indirection

[06:53:55] <geist> i really dont understand it at all

[06:54:13] <JonathanThompson> CSLL?  Circularly Singly-Linked List?

[06:54:30] <meianoite> consider a head node to a circular singly-linked list

[06:54:39] <meianoite> that's what I meant by array[prio]

[06:54:48] <meianoite> the array of priority queues

[06:54:54] <meianoite> ok so far?

[06:55:09] <meianoite> JonathanThompson correct

[06:55:10] <geist> okay

[06:55:25] <geist> i know what the data structure is I'm just not sure what 'other threads' 'last thread', etc mean in that context

[06:55:41] <meianoite> other threads === ellipsis ;)

[06:55:42] <JonathanThompson> So, 32 CSLL's where the current pointer is referenced in that array?

[06:55:58] <meianoite> JonathanThompson actually, 120.

[06:56:06] <JonathanThompson> Details, details :)

[06:56:11] <meianoite> geist, so

[06:57:02] <meianoite> last thread = the last thread that was scheduled on a given prio level. not blocked, still ready to run, preempted by blowing the quantum or just yielding

[06:57:53] <meianoite> current thread = the thread that will be run in case that prio level is selected. the "first" ready thread on that prio level.

[06:58:06] <meianoite> other threads = well, the other threads on that circular list

[06:58:20] <geist> how is that not just implicit with the location in the list, or is that actually what you mean?

[06:58:21] <meianoite> last thread, again = closing the circular list

[06:58:36] <meianoite> what I mean is that if you do

[06:58:57] <meianoite> list_head -> current thread -> other threads -> last thread -> current thread

[06:59:03] <meianoite> it's hard to remove current_thread

[06:59:25] <meianoite> in case it gets unscheduled due to blocking

[06:59:40] <geist> well, okay yeah

[06:59:45] <meianoite> but if you use the last thread to indirect into current_thread, it's dead easy

[06:59:46] <geist> but i would have never done it that way in the first place

[06:59:55] <meianoite> if you needed to shave bytes... ;)

[07:00:01] <geist> you dont

[07:00:09] <meianoite> true

[07:00:16] <meianoite> but I didn't design it for 4096 threads

[07:00:19] <meianoite> but for millions of threads

[07:00:29] <JonathanThompson> Beyond possibly ensuring you don't have cache line sharing issues, a few bytes won't likely mean much in practice.

[07:00:30] <meianoite> then I realized that's still a couple years away from happening ;))

[07:01:01] <geist> but anyawy, what I was saying was this queue is reused for other things the thread can belong to

[07:01:04] <JonathanThompson> The transitions between kernel and user space will eat up a lot of that in overhead anyway.

[07:01:07] <geist> such as a primitive it's blocked on (in this case semaphores)

[07:01:10] <meianoite> (actually Ingo beat me with the cluestick)

[07:01:19] <geist> and in those cases it's frequent to pull threads out of the middle of a queue

[07:01:27] <geist> and therefore having doubly linked lists makes that trivial

[07:01:32] <meianoite> sure.

[07:01:50] * JonathanThompson throws out the thought of using TST's instead, just to make life more interesting :P

[07:01:59] <meianoite> I'm still having a hard time understanding why a thread would be pulled out the middle of a wait queue...

[07:02:10] <meianoite> TST?!

[07:02:14] <JonathanThompson> Because it no longer needs to wait :)

[07:02:15] <geist> sem_acquire(blahblah, TIMEOUT)

[07:02:26] <geist> which actually happens relatively often

[07:02:43] <meianoite> uh... ok.

[07:02:50] <geist> actually unless it changed in haiku, sleep() was implemented as waiting on a semaphore with a timeout

[07:02:54] <meianoite> the case where the thread gives up waiting on the sem

[07:02:57] * JonathanThompson wonders how much is obvious when he's employing geek humor

[07:03:13] <geist> meianoite: right. basically a timer will come along and remove the thread from the sempahore

[07:03:20] <meianoite> JonathanThompson I still don't know what TST means, so I could't laugh ;)

[07:03:27] <JonathanThompson> Ternary Search Tree :)

[07:03:46] <meianoite> ok... I need to brush up my geek humour :)

[07:03:55] <geist> also, getting a signal in kernel space may force you to get pulled from a semaphore

[07:04:07] <meianoite> geist ok ok ok I got it now ;)

[07:04:11] <JonathanThompson> <-- has special form of internal geek humor: Rosetta Stones cost extra

[07:04:28] <geist> in lk, I abstracted all of that into a generic wait_queue structure

[07:04:39] <geist> that all the locking primitives use. actuallt works quite well

[07:04:47] <meianoite> I'd still implement it with a clock structure instead of DLL, but that's just me

[07:05:02] <meianoite> (and in this case the CSLL with indirection still works just fine)

[07:05:11] <geist> you have to search the list

[07:05:24] <JonathanThompson> That's what I was thinking, geist.

[07:05:34] <meianoite> semaphores don't tend to become this crowded, do they?

[07:05:39] <geist> they can

[07:05:42] <JonathanThompson> What if you have 100 threads or so that need to be searched to remove one out of the queue?

[07:06:08] <geist> but it's all microoptimization at this point

[07:06:15] <meianoite> JonathanThompson, in this case you don't search. you don't use a queue. you use a better structured... structure. to handle it.

[07:06:19] <geist> it seems to make sense to use the most obvious choice, and then optimize later

[07:06:33] <geist> which is very much the mentality of newos

[07:06:38] * JonathanThompson notes meianoite is at a loss 4 WORDS (8 bytes)

[07:06:42] * _hugo supports geist

[07:06:46] <meianoite> that's actually the very core of O(1) algorithms

[07:06:55] <meianoite> not searching

[07:06:56] <geist> some of the bigger misfires on haiku (like the haiku scheduler) is because of premature optimization

[07:07:12] <meianoite> but structuring it in a way the desirable data item is handily available

[07:07:40] <geist> anyway, anything else you want to talk about? I frankly dont care about the low level data structure

[07:07:47] <geist> provided it's not some really dumbass idea :)

[07:08:06] <JonathanThompson> Use a red-black tree :)

[07:08:09] <JonathanThompson> :P

[07:08:10] <meianoite> not really atm

[07:08:22] <geist> i think the biggest thing right now is to get the queues per cpu

[07:08:25] <geist> and deal with all the fallout of that

[07:08:26] <meianoite> just liked to point you to that fuckup on Haiku's scheduler

[07:08:37] <geist> yeah, absolutely. I looked in the svn logs

[07:08:44] <geist> it was done by someone named 'lillo' back in 2002

[07:08:48] <JonathanThompson> Who gets a dart up their rear? :)

[07:08:50] <meianoite> that was Angelo Mottola

[07:09:02] <meianoite> and I really have *no* idea why he did that

[07:09:09] <geist> to be fair he may have actually had furhter optimizations

[07:09:25] <geist> he obviously was tryign to factor out the queue search that newos does (and is quite stupid about)

[07:09:46] <meianoite> geist if you don't mind I'm definitely adopting LK's idea

[07:09:47] <geist> but he didn't take it that one step farther

[07:09:55] <meianoite> just because it was my idea too ;)

[07:10:00] <geist> oh sure

[07:10:02] *** stargater has joined #haiku

[07:10:12] <stargater> moin

[07:10:14] <meianoite> and it happens to be how Linux does it as well

[07:10:15] <geist> just make some slightly more generic bitsearch taht works across multiple words and you're okay

[07:10:30] <geist> actually you might want to double check that there isn't a compiler intrinsic for that

[07:10:36] <geist> there's obviously a __builtin_clz

[07:10:41] <meianoite> more generic than bitmap[prio>>5]?

[07:10:53] <geist> i'm thinking there's a __builtin_ffs that does the search across multiple words

[07:11:24] <meianoite> ARM even likes that idiom a lot

[07:11:50] <geist> at the previous company i wrote an OS for we were on arm7 and had to hand code it

[07:11:59] <meianoite> especially considering bitmap is 4 int32 long

[07:12:00] <geist> but you can actually do a bit search in 7 cycles on arm7, so it wasn't so bad

[07:12:15] <meianoite> (or would be)

[07:12:17] <geist> on arm9 it's a single cycle with clz

[07:12:28] <geist> nice thing is __builtin_clz picks the proper one for you

[07:12:33] <geist> and I'm thinking there's something on x86 for that

[07:12:38] <meianoite> cool.

[07:13:05] <geist> the only slight bummer about the bitmap approach is you have to track how much stuff is in the queue so you can tell if the queue went empty (and thus need to set the bit zero)

[07:13:26] <geist> I forgot what I did for lk. I think I may have done a is_list_empty(), which for that CDLL just compare prev to next

[07:14:03] <geist> oh, also for the per-cpu stuff, look at the cpu structure

[07:14:12] <geist> there's already a per-cpu structure, I forget where it's declared

[07:14:19] <geist> it's designed for you to stuff stuff like this in

[07:14:20] <meianoite> here

[07:14:29] <geist> and it's cacheline aligned, etc

[07:14:45] <geist> and you can look it up very quickly. get_cpu_struct() or something like that

[07:14:54] <geist> it is a single pointer dereference to get it

[07:15:00] <meianoite> __builtin_clz(run_queue_bitmap[prio>>5])

[07:15:09] <meianoite> that wouldn't hurt, would it?

[07:15:18] <meianoite> especially given ARM

[07:15:21] <geist> umm, what?

[07:15:24] <geist> that doen't even compile

[07:15:31] <meianoite> ok.

[07:15:38] <geist> __builtin_clz takes a single uint32

[07:15:49] <meianoite> consider run_queue_bitmap ain't int32 anymore

[07:15:54] <meianoite> but int32[4]

[07:16:03] <geist> right, but like I said it takes a uint32

[07:16:13] <geist> i was thinking there was a variant that searches a run of bitmaps

[07:16:22] <meianoite> as soon as you index the array, what do you get?

[07:16:26] <meianoite> an uint32.

[07:16:27] <geist> ie, __builtin_foo(uint32_t *, size_t len);

[07:16:44] *** mmadia has quit IRC

[07:17:07] <meianoite> with up to 128 priority levels, >>5 gets the 2 bits that would select the correct bitmap

[07:17:10] <meianoite> 0, 1, 2 or 3

[07:17:13] <geist> and for that matter a routine that just does 4 ifs and 4 clzs is probably fine

[07:17:34] <geist> oh if you like

[07:17:39] <geist> again, that's just microoptimization stuff

[07:17:49] <meianoite> but it works ;)

[07:17:50] <geist> just get it working first

[07:17:55] <meianoite> and relieves the limitation of 32 priority levels

[07:17:59] <meianoite> which we must, anyway

[07:18:07] <geist> obviously

[07:18:14] <meianoite> I like my solution. ;)

[07:18:30] <meianoite> I even suspect that becomes just a couple of instructions on the arm

[07:18:36] <meianoite> compared to a single clz

[07:18:39] <meianoite> IF so

[07:18:47] <geist> I'm actually not sure what you're talking about now

[07:18:56] <geist> i'm really failing to follow you here

[07:19:02] <meianoite> what parameters do clz have?

[07:19:07] <meianoite> does the barrel shifter work there?

[07:19:18] <geist> are you talking about preselecting which word to look at with a 2 bit bitmap?

[07:19:22] <meianoite> can it do indirect addressing?

[07:19:38] <meianoite> either preselecting or just using it, depending on what clz can take

[07:19:47] <meianoite> (I didn't read the opcode format yet)

[07:20:05] <meianoite> just the semantics, "count leading zeroes"

[07:20:11] <geist> for arm you're probably better off doing them all in parallel, then selecting which one you want

[07:20:15] <geist> to hide the load stalls

[07:20:52] <meianoite> there's a SINGLE load unless I'm very mistaken

[07:20:59] <geist> ?

[07:21:02] <geist> for 4 words?

[07:21:19] <meianoite> no, not to find the highest priority

[07:21:30] <geist> see, again I think we've gotten off track from each other

[07:21:31] <meianoite> but to either write or clear the bit

[07:21:42] <geist> oh wow, I was totally not talking about that at all

[07:21:45] <geist> that's trivial

[07:22:00] <geist> you can do that in about 4 instructions

[07:22:21] <geist> and it'd be dominated by the load/store stall that you're inevitably going to have to do anyway

[07:22:34] <geist> so being clever with the barrel shifter or so wont help you too much

[07:22:45] <meianoite> probably true...

[07:23:21] <geist> i was talking about taking 4 words of bitmap and selecting the first one

[07:23:25] <meianoite> can't the arm do some trickery like the save register stuff where you specify a range of registers to save on the stack?

[07:23:37] <geist> actually true. yeah

[07:23:45] <meianoite> so you can do multiple load of a 4 word bitmap on 4 registers?

[07:23:57] <geist> sure, you'd do that for the clz stuff probabaly

[07:24:04] <meianoite> so, a single load

[07:24:13] <meianoite> 4 conditional CLZs

[07:24:21] <geist> like I was saying, you'd probably just load all 4 words and then start clzing while interleaving code to decide which one to use

[07:24:24] <meianoite> (I'd guess)

[07:24:29] <geist> yeah, probably

[07:24:42] *** gotaku has joined #haiku

[07:24:50] <meianoite> note that I'm in no shape or form an ARM expert

[07:24:53] <geist> for setting or clearing one bit in that bitfield, you're better off just deciding which word, loading it, setting/clearing the bit, and then storing it

[07:25:00] <meianoite> I only know it's got some cool tricks

[07:25:10] <geist> i'd say I'm pretty close to being an expert ARM programmer

[07:25:34] <geist> the really fun tricks come when you stuff 4 bits of some reult into the flags register of CPSR

[07:25:43] <geist> and then use conditional stuff

[07:25:49] <meianoite> great, you might give me a hand whenever I decide to venture into the embedded world... I got a Symbian 9 phone ;)

[07:26:11] <geist> http://newos.org/WebSVN/filedetails.php?repname=extern&path=%2Flk%2Ftrunk%2Fapp%2Fstringtests%2Fmymemcpy.S&rev=0&sc=0

[07:26:20] <geist> see .L_not16bytealigned

[07:26:28] <geist> that's a trick I picled up a while back

[07:26:41] <geist> you can stuff stuff into the flags register and then conditionally do stuff

[07:26:55] <geist> in this case it's to conditionally move 1-15 bytes to get something aligned

[07:26:56] <meianoite> and there's this very nifty Genesis emulator for Symbian 9 originally written for RISC OS

[07:27:04] <meianoite> and then ported to the GP32x

[07:27:24] <geist> one of my other projects was to write a full ARM emulator

[07:27:25] <geist> http://newos.org/WebSVN/listing.php?repname=extern&path=%2Farmemu%2Ftrunk%2F&rev=0&sc=0

[07:27:26] <meianoite> and then to Symbian, both UIQ and S60

[07:27:35] <geist> and it runs pretty darn fast. on my machine it emulate about a 70mhz arm

[07:27:52] <meianoite> interpretative one?

[07:27:55] <meianoite> static rec?

[07:27:56] <geist> right

[07:27:57] <meianoite> dynarec?

[07:28:13] <geist> it's about as fast as you can get while still interpreting

[07:28:34] <geist> any further speed increases will have to be from dynamic code gen

[07:29:02] <geist> getting it down to 10:1 or so cycle wise is pretty tough to do while still intepreting

[07:29:26] <meianoite> http://www.gtoal.com/sbt/

[07:29:33] <meianoite> I wonder if you're familiar with this paper

[07:29:41] <geist> but of course like most of my projects I did it to learn

[07:29:54] <geist> in that I like to make all the mistakes and reinvent the wheel, it teaches you more

[07:30:39] <meianoite> heh, that's another similarity we have then. we design similar schedulers, and we like fucking stuff up and then fixing them ;D

[07:31:53] <meianoite> that switch(PC) trick is just so cool.

[07:32:04] <meianoite> you let the compiler do all the boring optimization for you

[07:32:32] <geist> works if the code is small

[07:32:53] <geist> actually the most annoying thing about emulating ARM is the fact that you can screw with PC

[07:33:17] <meianoite> I'd wager it works most of the time, as long as the compiler doesn't run out of memory

[07:33:21] <geist> so I tended to have to segment opcodes into stuff that writes to PC and stuff that doesn't

[07:34:08] <geist> that paper is relatively interesting, though it relies on a bunch of assumptions

[07:34:13] <meianoite> you can always break that big switch it into multiple levels if it gets too big

[07:34:23] <geist> btw, if you want to talk about generic osdev stuff, we should probably take this somewhere else

[07:34:27] <geist> like #osdev

[07:34:59] <geist> there are more folks there that may be itnerested as well

[07:35:02] <meianoite> well, that paper is not directed to the ARM per se, but smaller and repetitive code like console games. I just thought you'd be interested

[07:35:08] <meianoite> good to know

[07:36:19] <meianoite> anyway, I'm not into generic osdev stuff. I have *zero* experience with generic osdev stuff, to be honest. I just happened to design a half-decent scheduler as a OS class assignment... ;)

[07:37:01] <geist> aww damn, kurt vonnegut died

[07:37:24] <meianoite> I just hope that's not Mambo Kurt

[07:37:53] <meianoite> aye, not even close

[07:38:17] <meianoite> well, he was 85

[07:38:25] <meianoite> lived a nice life, it seems

[07:38:38] <geist> actually, no he had a pretty screwed up life

[07:39:12] <meianoite> I'm trying to keep the cheerful mood, nevermind my ignorance ;)

[07:39:29] <geist> anyway, gonna go do something else for a bit

[07:39:37] <meianoite> and I'm off to bed

[07:39:45] <meianoite> g'night, folks

[07:40:10] <meianoite> g'night, geist, _hugo, kokito, JonathanThompson, Grackle

[07:40:28] *** meianoite has left #haiku

[07:43:26] <stargater> wow i = http://www.blachford.info/computer/begeistert012/S_IthmarCircular.jpg

[07:43:57] <gotaku> ?

[07:49:23] <Grackle> http://www.blachford.info/computer/begeistert012/S_IMG_3467.jpg haha awesome, one of the first things I did with BeOS was watch futurama episodes

[07:49:59] <Grackle> The first time I installed it on physical hardware, that is.

[07:54:16] *** kr1stof has left #haiku

[08:08:02] *** NeonLicht has quit IRC

[08:08:04] *** NeonLicht has joined #haiku

[08:19:40] *** gotaku has quit IRC

[08:38:13] *** umccullough_lap has joined #haiku

[08:41:58] *** mazon is now known as Mazon

[08:43:31] <umccullough_lap> main screen turn on

[08:46:02] <Grackle> we get signal

[08:50:20] *** mmu_man has joined #haiku

[08:50:20] *** ChanServ sets mode: +o mmu_man

[08:50:55] <mmu_man> plop

[08:51:05] <mmu_man> aw, looks I'll be having work to do this summer :))

[08:55:01] *** mmu_man changes topic to "Looks like there are 8 projects accepted for SoC... http://code.google.com/soc/haiku/about.html | Haiku Tech Talk video http://video.google.com/videoplay?docid=236331448076587879 - photos http://haiku-os.org/index.php?q=gallery&g2_itemId=1288 | http://haiku-os.org | IRC logs: http://echelog.matzon.dk/?haiku | http://cia.navi.cx/stats/project/OpenBeOS | http://svn.berlios.de/viewcvs/haiku/"

[09:03:34] *** umccullough_lap has quit IRC

[09:06:36] *** Teknomancer has joined #haiku

[09:07:25] *** tombhadAC has quit IRC

[09:09:57] *** PulkoMandy has joined #haiku

[09:15:21] *** silverpower has quit IRC

[09:16:46] <absabs> yayaya

[09:17:22] <absabs> my application is accepted

[09:17:35] <_hugo> congratulations

[09:17:40] <absabs> :)

[09:17:49] <absabs> thanks

[09:18:01] <absabs> I will do my best from now on

[09:18:04] <absabs> :P

[09:18:12] <_hugo> absabs: what was your application?

[09:18:21] <absabs> Firewire stack

[09:18:22] <_hugo> which even

[09:18:27] <_hugo> i see

[09:19:34] <absabs> aha

[09:19:45] *** silverpower has joined #haiku

[09:20:00] <absabs> _hugo:Your application is alse accepted

[09:20:09] <_hugo> absabs: indeed

[09:20:11] <absabs> congratulations

[09:20:15] <_hugo> thank you

[09:20:21] <absabs> :)

[09:20:38] <absabs> congratulations to our eight student

[09:20:44] <absabs> :)

[09:23:19] <kokito> hello absabs. welcome!

[09:24:00] <absabs> hello

[09:24:14] <absabs> Thanks Haiku

[09:25:11] <absabs> Contribute to haiku has always been my dream

[09:25:41] <kokito> well, absabs, your dream has now become true. :P

[09:26:12] <absabs> thanks

[09:26:13] <absabs> hehe

[09:26:18] <absabs> :)

[09:28:49] <kokito> absabs, just curious: where are you from?

[09:29:29] <absabs> I come from China

[09:29:36] <kokito> cool

[09:29:42] <absabs> hehe

[09:29:45] <kokito> where in China?

[09:30:07] *** silverpower_ has joined #haiku

[09:30:38] <absabs> Yang Zhong city, JiangSu provience

[09:30:42] *** silverpower has quit IRC

[09:30:56] *** silverpower_ is now known as silverpower

[09:31:11] <Teknomancer> i would have applied if GsoC was a year ago :(

[09:31:11] <_hugo> absabs: are you China right now?

[09:31:12] <Teknomancer> no loner a student

[09:31:13] <Teknomancer> pffft

[09:31:20] <Teknomancer> longer=longer

[09:31:21] <absabs> yes

[09:31:22] <_hugo> in China that is

[09:31:26] <Teknomancer> aargh

[09:31:48] <_hugo> interesting. i hope i can visit China in the future

[09:32:10] <absabs> The city is an island on ChangJiang River

[09:32:22] <absabs> _hugo:Welcome!!

[09:32:49] <absabs> only smaller than ChongMing island of ShangHai

[09:33:48] <_hugo> absabs: what are you studying?

[09:33:55] <absabs> CS

[09:34:25] <kokito> absabs, what region of China is Yang Zhong located at?

[09:34:46] *** slaine_ has joined #haiku

[09:35:11] <kokito> ah! close to Shanghai and Nanjing, right?

[09:35:16] <absabs> yes

[09:35:33] <absabs> close to Shanghai

[09:35:57] <kokito> 感謝

[09:36:04] <absabs> 不客气

[09:36:17] <kokito> :)

[09:36:26] <absabs> kokito:你会说中文

[09:36:27] * _hugo crashes

[09:36:35] <absabs> :)

[09:36:51] <kokito> no, absabs. 日本語 :)

[09:37:03] *** Ingenu has quit IRC

[09:37:06] 

[09:37:09] <Teknomancer> :)

[09:37:13] <kokito> LOL

[09:37:15] *** Ingenu has joined #Haiku

[09:37:20] <Teknomancer> hi Ingenu

[09:37:31] <absabs> :)

[09:37:52] <absabs> kokito:where are you from

[09:38:21] <kokito> 中文不理解 :)

[09:38:30] <kokito> absabs, I am from Argentina.

[09:38:36] <Ingenu> hi

[09:39:03] <absabs> a beautiful country

[09:39:25] <absabs> kokito:你中文说得很棒

[09:39:54] <kokito> ok, absabs, that was too difficult for me to understand :)

[09:40:03] <absabs> en,en

[09:40:22] <Grackle> heh all I see are question marks

[09:40:30] * Grackle goes to sleep now

[09:40:31] <Grackle> night all

[09:40:38] <kokito> night Grackle

[09:40:45] <absabs> kokito:your Chinese speaking is very good

[09:40:46] <absabs> night

[09:41:20] <kokito> absabs, 大感謝 :P

[09:41:38] <absabs> kokito:太客气了:)

[09:41:53] <kokito> absabs, actually, I cannot speak a word of Chinese.

[09:42:31] <kokito> ok folks, time to get some sleep here. nice chatting with you absabs. will see you online.

[09:42:47] <_hugo> see you kokito

[09:42:57] <kokito> night _hugo

[09:43:07] <absabs> kokito:nice

[09:43:12] *** kokito has quit IRC

[09:44:10] *** emitrax has joined #haiku

[09:47:40] <emitrax> good morning :D

[09:48:42] <absabs> good morning

[09:49:04] *** silverpower has quit IRC

[09:49:45] *** silverpower has joined #haiku

[09:54:36] *** eNGIMa has joined #haiku

[10:02:22] *** Jixt has joined #haiku

[10:02:44] <Jixt> hi

[10:03:51] <Jixt> Is there someone who can tell me more about the USB Stack of Haiku? I can't find any status on the haiku homepage

[10:05:18] <_hugo> Jixt: what do you wish to know about? usually the source is the best reference

[10:06:33] <Jixt> _hugo, I want to know if the bulktransfer protocol is fully implemented and if the ehci,uhci and ehci are fully implemented...

[10:06:58] <Jixt> *ehci,uhci,ohci

[10:07:41] <_hugo> im not sure if they are fully implemented, but there is code for uhci, ohci and ehci

[10:08:01] <_hugo> like i said, the best reference is the source tree.

[10:08:22] <_hugo> or the mailing list, usually that works too

[10:08:36] <Jixt> ok, I will try the mailing-list

[10:15:14] *** wkornewald has joined #haiku

[10:15:23] <wkornewald> hi everyone

[10:15:41] <Ingenu> lo

[10:15:53] <wkornewald> mmu_man: I hope you aren't unhappy about having to mentor a student, now :)

[10:16:12] <_hugo> hey wkornewald

[10:16:21] <wkornewald> hi guys

[10:16:58] <wkornewald> it's really cool that Google assigned an 8th student to us

[10:17:11] <_hugo> yep

[10:17:29] <Jixt> an extra one?

[10:17:49] <wkornewald> I don't know if my nagging after midnight helped :) there was a chat late in the night/morning with Google...

[10:18:33] <wkornewald> Jixt: we only expected 7, but then mmu_man (Francois Revol) joined as a mentor in the last second (actually, it was already too late :)

[10:19:05] <Jixt> nice ;-)

[10:19:35] <absabs> very nice

[10:22:23] <wkornewald> sigh...I'm having a hard time with the network preflet design. I'd want to have all networks/connections shared between all locations (profiles) and allow for overriding certain preferences (depending on the network type), but I don't really know if this added complexity is worth it or if it's sufficient to offer a function for simply copying an existing network from another location into the current one

[10:22:58] <CIA-17> hugosantos * r20661 /haiku/trunk/src/add-ons/kernel/network/protocols/udp/udp.cpp: made UDP agnostic to the underlying network protocol (preparing some work into the future).

[10:23:19] <_hugo> wkornewald: thats a good question

[10:23:23] <wkornewald> many networks will probably appear in multiple locations

[10:23:24] <_hugo> question/point

[10:23:43] <_hugo> in fact i think the override capability makes a lot of sense

[10:23:52] <wkornewald> e.g., you'll want your VPN connection everywhere

[10:24:20] <wkornewald> _hugo: it would just be: "Use Global Preferences" or "Use Custom Preferences"

[10:24:51] <_hugo> well, in terms of how showing this information to the user, you could have the common preferences (global), and something like "in this location, i'll use XXX as YYY"

[10:25:23] <wkornewald> the custom prefs for ethernet would be fully customizable, but PPP would only allow for changing the number and selecting "Wait For Dial Tone"

[10:25:37] <_hugo> so its clear what is a global setting and is a custom setting in the location

[10:25:47] <_hugo> *and what is

[10:26:36] <wkornewald> I don't think there should be the possibility to override individual settings, but just a whole group (often just a subset, like in case of PPP)

[10:26:59] <wkornewald> otherwise the interface gets very complicated

[10:27:06] <_hugo> what is a group?

[10:27:13] <wkornewald> a group of settings

[10:27:18] <_hugo> right, but who defines that

[10:27:23] <wkornewald> we :)

[10:27:31] <_hugo> eheh, yeah, but i mean in the interface

[10:27:50] <_hugo> is it the ethernet settings for instance

[10:27:57] <wkornewald> it's pre-defined and depends on the interface type.

[10:28:00] <_hugo> that says what "groups" there are?

[10:28:01] <_hugo> ok

[10:28:25] <wkornewald> I was thinking of a list of networks on the left and the settings on the right.

[10:29:07] <_hugo> i would put "Common settings" in the list of networks, but separated and probably on the top

[10:29:42] <_hugo> or Global settings

[10:29:45] <wkornewald> the right side then has a marker with a checkbox at the top "- [ ] Use Custom Preferences ------------" and below there is a marker "- Global Preferences -------"

[10:29:46] <_hugo> whatever name is better

[10:29:59] <_hugo> sure, but where do you configure the global settings?

[10:30:21] <wkornewald> below the global marker

[10:30:42] <_hugo> in the same view then?

[10:30:54] <_hugo> im not sure its clear that changing preferences there will affect all locations

[10:30:55] <wkornewald> yes. the problem is that if a user creates a new interface in a non-Global location he will have to configure some of the global settings, first

[10:32:12] *** wildur has joined #Haiku

[10:32:24] <_hugo> anyway, im sure you have a lot more thought into this than i have :-)

[10:32:31] <_hugo> im just throwing ideas

[10:32:38] <_hugo> you have put..

[10:32:45] <_hugo> need to eat real food, instead of words

[10:32:48] <wkornewald> I'm just not sure if it'S worth the complexity

[10:32:59] <wkornewald> ok

[10:33:54] <_hugo> wkornewald: my university VPN for instance, uses the exact same preferences in terms of credentials and routes/etc, but it changes the VPN server depending from where im connecting from

[10:34:09] <_hugo> depending on :-) bad day today, eheh

[10:34:10] *** Teknomancer has quit IRC

[10:34:19] <wkornewald> why is that?

[10:34:32] <wkornewald> can't they just offer one single server?

[10:34:56] <_hugo> they use different servers for some reason

[10:35:50] <_hugo> well, in fact, the problem is that inside the campus, wireless communication is very limited. and they use a different domain name for the vpn server inside the campus and outside

[10:36:11] <wkornewald> why do you think that listing both custom and global preferences above each other wouldn't work well?

[10:36:15] <_hugo> (i would say they are too lazy to configure the same zone with different settings depending on incoming interface)

[10:36:58] <wkornewald> (where the custom prefs aren't shown when you use the global ones for that location)

[10:37:06] <_hugo> wkornewald: i think it would be better to show the global preferences uneditable. and when you select "custom preferences in this location"

[10:37:09] <_hugo> then you would be able to edit

[10:37:20] <_hugo> then you keep the view simple

[10:37:52] <_hugo> possibly there could be a button that would jump to the view that enables you to edit the global preferences

[10:37:59] <_hugo> but this should reflect your current "location"

[10:38:02] <wkornewald> but how will you handle the case when you create a new PPP interface from within a custom location? if not all prefs are visible...

[10:38:13] <_hugo> i.e. no longer the location you are editing, but "global preferences" or something

[10:38:57] <_hugo> wkornewald: if an interface is specific to the location

[10:39:03] <_hugo> there is no global preferences, only custom ones

[10:39:21] <wkornewald> I wanted to have all networks shared, actually

[10:39:54] *** procton_ has quit IRC

[10:40:03] <_hugo> you could have a flatted view that shows all interfaces, and this location-specific interface could show up too, but as "Dialup at Home"

[10:40:07] <_hugo> as an example

[10:40:12] <_hugo> flattened :-)

[10:40:27] <wkornewald> let's see if from this point of view: sharing preferences only makes sense if you often change the base preferences and all other locations need to update accordingly. otherwise, you can just create copies of existing interfaces. the only problem is that you have to manually copy all new interfaces to all locations

[10:40:37] <_hugo> right

[10:40:46] <wkornewald> so, does this case happen, often?

[10:40:51] <_hugo> yes

[10:40:59] <wkornewald> example?

[10:41:15] <_hugo> Wireless interface, usually uses DHCP to obtain an IP, but depending on the network

[10:41:24] <_hugo> might require additional auth.

[10:41:55] <wkornewald> but wouldn't those be different networks?

[10:42:28] <stargater> wow cool 8

[10:42:31] <_hugo> the specific ones, yes. but the general case is just a general case

[10:42:32] <stargater> i love 8

[10:42:48] <wkornewald> what do you mean with specific and general?

[10:43:04] <wkornewald> I mean, those networks will have different SSIDs, anyway?

[10:43:09] <_hugo> wkornewald: by general i mean default. the default settings are used when you attach to a new network

[10:43:13] <_hugo> yep

[10:43:28] <wkornewald> then, those are different networks (I distinguish at a network, not interface level)

[10:43:41] <_hugo> you cant expect a user to configure its connection for each new wireless network (different SSID) it connects to

[10:44:04] <wkornewald> but that's what we do in every OS, ATM. you need a new password every time

[10:44:11] <_hugo> by default you should be able to select the network to attach to and it would just work

[10:44:21] <_hugo> you just need to configure when additional information is required

[10:44:41] <_hugo> you might not require a password

[10:44:45] <_hugo> it could be an open network

[10:44:57] <mmu_man> damn I won't be allowed to take vacation this summer :(

[10:45:02] <mmu_man> :D

[10:45:07] * mmu_man pets wkornewald

[10:45:10] <wkornewald> I think we can expect that every network has a different configuration

[10:45:16] <wkornewald> hi mmu_man :)

[10:45:31] <wkornewald> and for open networks there is no configuration necessary

[10:46:10] <wkornewald> so, the user could just click "Connect" and that's it. what was the problem, agai? :)

[10:46:21] <_hugo> wkornewald: well, go ahead :-) i think an iterative development model is better anyway. if your approach has shortcomings im sure people will identify them

[10:46:24] <mmu_man> is there any time table for that SoC ?

[10:46:41] <wkornewald> http://code.google.com/support/bin/answer.py?answer=60325&topic=10729

[10:46:51] <_hugo> wkornewald: no problem at all, should just press Connect, yes

[10:46:58] * mmu_man lazy today :))

[10:47:21] <wkornewald> _hugo: but what is wrong with that, then?

[10:47:26] <mmu_man> March 12: Mentoring organization application deadline

[10:47:27] <_hugo> wkornewald: nothing :-)

[10:47:28] <wkornewald> you can have different configurations depending on the SSID

[10:47:34] <mmu_man> so I wasn't overdue, was I ?

[10:47:37] <stargater> ahh jixt have post in the mailing list , nice :-)

[10:47:44] <wkornewald> mmu_man: actually, you were :)

[10:48:08] <wkornewald> _hugo: I think my suggestion is the same as yours :)

[10:48:15] <_hugo> wkornewald: it probably is :-)

[10:48:20] <mmu_man> ohn march

[10:48:36] <wkornewald> march?

[10:48:40] <mmu_man> nevermind

[10:49:20] <wkornewald> _hugo, so is there any case where you change the base prefs and all others need to update accordingly?

[10:49:49] <_hugo> wkornewald: for the common user that is probably rare

[10:50:10] <_hugo> wkornewald: as long as the default configuration is sane

[10:51:03] <wkornewald> in that case, the default configuration has no special meaning, though

[10:51:37] <_hugo> wkornewald: default in the sense that if you get to a new location, auto-configuration (if the network so it permits) should work

[10:51:45] <_hugo> without additional interaction with the user

[10:52:00] <wkornewald> it would just be: all hard interfaces are automatically added to all locations (but with their own preferences), but the custom ones (PPP, Mobile IPv6, VPN, etc.) are not shared, at all, but you can import them from other locations

[10:52:23] <wkornewald> yes, auto-configuration should always be the default behavior

[10:52:53] <_hugo> ok, i still think there should be some sort of global settings

[10:52:56] <_hugo> for stuff like VPN

[10:53:03] <_hugo> and specially Mobile IP

[10:53:16] <wkornewald> you can manually import that stuff, though

[10:53:25] <_hugo> by keeping the configuration specific to a location, you are reducing mobility

[10:53:39] <_hugo> i want to move between locations seamlessly

[10:53:47] <raph_ael> hello

[10:53:49] <wkornewald> we can detect if something is configured the same in multiple locations

[10:53:50] <_hugo> and continue to use my VPN connection and/or Mobile IP

[10:54:06] <_hugo> well, we could

[10:54:11] <mmu_man> idealy everyone woudl have a GPS and use it to set teh location :))

[10:54:30] <wkornewald> it adds implementation complexity, but from the end-user POV it's the same

[10:54:37] <_hugo> wkornewald: but you are trying to go around the fact that those settings are in fact global

[10:54:48] <wkornewald> yes, that's right

[10:54:51] <_hugo> the whole point of VPN and mobile IP and is to establish an overlay so your location is not relevant

[10:55:34] <wkornewald> that's why I wanted to have globally shared networks (and no location-specific ones, at all)

[10:55:55] <_hugo> but location settings are relevant as well

[10:56:06] <_hugo> and not only in terms of interfaces

[10:56:12] <_hugo> but in the overall system

[10:56:25] <wkornewald> yes, that's what the override feature is for

[10:56:32] <_hugo> for instance, at "Work" i might need a proxy to connect to the http sites

[10:56:37] <_hugo> while at "Home" i might not require it

[10:56:49] <_hugo> well, but the override feature should have a wider application

[10:57:04] <_hugo> it should even be presented in some way it could be usable by applications

[10:57:16] <_hugo> (im stretching over R1)

[10:57:31] <wkornewald> what kind of wider application?

[10:57:55] <_hugo> wkornewald: i have a very specific example, and i even submited a bug to Adium a long time ago

[10:57:56] <_hugo> related to this

[10:58:08] <_hugo> while at home, i wanted to connect to MSN directly

[10:58:19] <_hugo> while in the university i wanted it to use a specific proxy

[10:58:36] <wkornewald> well, all this would do is that the VPN connection which you normally only use for your university locations would also be available at home where it doesn't make any sense, but it probably doesn't disturb, either. the advantage is that you can create a new university location and have VPN directly available

[10:59:08] <wkornewald> _hugo: yes, that can be done, easily. just choose the location and say "Use Custom Preferences"

[10:59:18] <_hugo> what preferences? where do i set this?

[10:59:30] <_hugo> i think i should do this in the application, not in the network preferences

[10:59:30] <wkornewald> in your browser, for example

[10:59:37] <wkornewald> exactly

[10:59:44] <_hugo> so we should have a location data store

[10:59:59] <wkornewald> the printer preferences would allow for setting the default printer based on the location

[11:00:03] <wkornewald> ,too

[11:00:09] <_hugo> there are even more examples. while at home i might want to connect automatically to my network share

[11:00:36] <wkornewald> this can be done, but that's not the current problem

[11:01:01] <_hugo> well, im trying to prove that the location issue has a wider range

[11:01:17] <wkornewald> I already thought of that. there is nothing that stops you from expanding its use

[11:01:33] <_hugo> why not build the network preferences over this instead?

[11:01:37] <wkornewald> you just offer locations and the apps support prefs per location or not

[11:01:51] <wkornewald> how would you do that?

[11:02:10] <wkornewald> IOW, what would this change regarding the network preferences application?

[11:02:17] <_hugo> you would watch for specific preferences in the location store regarding the current location

[11:02:36] <wkornewald> are you talking about the UI or the implementation

[11:02:37] <wkornewald> ?

[11:02:46] <_hugo> well, in terms of the preferences application, how the information is presented its a bit irrelevant

[11:02:47] <_hugo> :-)

[11:02:55] <_hugo> im mostly thinking about implementation details, sorry

[11:03:05] <wkornewald> I don't care about the implementation, ATM (or never :)

[11:03:13] <wkornewald> the UI must be usable

[11:03:28] <wkornewald> easy to use and encourage even non-techies to setup their home network

[11:03:32] <_hugo> right, but the UI must also reflect the underlying capabilities. (without bloat)

[11:03:53] <_hugo> and network settings should not be confined to the "network preferences" preflet

[11:04:00] <wkornewald> and the locations feature must not be too cumbersome, so it actually gets used instead of ignored (because you have to setup all interfaces anew for every new location)

[11:04:03] <_hugo> they should be able to be set in-place

[11:04:06] <_hugo> that is, in the browser, etc

[11:04:57] <wkornewald> the networks preferences only allow for configuring the interfaces

[11:05:04] <wkornewald> nothin more

[11:05:33] <wkornewald> what other apps do is thus more or less decoupled (unless they allow for creating new locations)

[11:05:40] <_hugo> wkornewald: interfaces is something that is a bit low level. while most times we have a relationship of 1 to 1 in terms of devices and connections, this is something that will eventually change

[11:06:07] <wkornewald> what do you mean? (btw, the preferences would be on a network level, not interface, most of the time)

[11:06:17] <_hugo> (10:04:58 AM) wkornewald: the networks preferences only allow for configuring the interfaces

[11:06:42] <wkornewald> I tried to be specific. this doesn't imply much about the UI

[11:06:45] <_hugo> what i mean is that you may use different interfaces to establish a connection to the same network

[11:06:59] <wkornewald> yes, indeed

[11:08:01] <wkornewald> but I think this could be a little bit too complicated. I mean, do you really want to first create a network and then attach the interfaces to it?

[11:08:25] <stargater> question, how can in haiku a atribute:Icon watch , so then i can have a file  with 000111000011100111 ?

[11:08:27] <wkornewald> (from the UI)

[11:08:45] <_hugo> wkornewald: sometimes you should just provide the network credentials and then let the stack do its work. interface configuration is on auto

[11:09:47] <wkornewald> different interfaces require different authentication methods. for example, I can connect to my router via WiFi (requires only password), Ethernet (no login), and VPN (requires username and password)

[11:09:51] <_hugo> note that you would have a network, and no specific interface assigned to it

[11:10:21] <_hugo> wkornewald: that is your home network

[11:10:34] <wkornewald> yes, but I couldn't comfortable configure it

[11:10:57] <_hugo> what will happen in a couple years is that you an AP that authenticates you through EAP to an AAA server deep in the network that understands your credentials. and you may do so in very different APs

[11:10:58] <wkornewald> I mean, I don't want to setup a network just to hold the interfaces in it

[11:11:01] <_hugo> from different operators

[11:11:37] <_hugo> is that you can connect to an AP..

[11:11:40] <wkornewald> switching from WiFi to Ethernet is automatic, anyway. only VPN wouldn't work easily, but then I could create a "location" that has this

[11:12:01] <_hugo> well, a VPN is not a location, it is a network

[11:12:06] <_hugo> also

[11:12:20] <_hugo> how do you specify in the configuration that the VPN connection should be the same

[11:12:25] <wkornewald> yes, but it could also be understood as "Home Network"

[11:13:05] <wkornewald> I don't want to have to specify that. I just want to create separate networks. there will also be a "Mobile IPv6" network type which allows for doing what you said

[11:13:10] <_hugo> wkornewald: this is dependency issue here. a Location is something that arises from your current connection. a Network is something you connect to

[11:13:15] <_hugo> there is :-)

[11:13:18] <_hugo> i really need to rest

[11:14:02] <wkornewald> _hugo: isn't Mobile IPv6 what you are talking about?

[11:14:03] <_hugo> interesting discussion, but i need to take a break and eat something :-)

[11:14:14] <_hugo> wkornewald: VPN connections also apply

[11:14:27] <_hugo> wkornewald: Mobile IP is too dynamic

[11:14:52] <wkornewald> OK, let's eat something (me, too ;) and get back in a few minutes

[11:15:03] <_hugo> :-) brb

[11:15:40] <mmu_man> hmm

[11:15:49] <mmu_man>  [revol@Zeta /Data/haiku/trunk]# HAIKU_INSTALL_DIR=/haiku jam -q install-haiku

[11:15:50] <mmu_man>  Unknown suffix on <src!add-ons!accelerants!intel_extreme>accelerants - see UserObject rule in Jamfile(5).

[11:15:51] <mmu_man> f*

[11:15:52] <mmu_man> still broken

[11:16:02] <mmu_man> (yes I did rebuild jam just now)

[11:18:52] *** GreyGhost has joined #haiku

[11:31:20] <mmu_man> hmm

[11:31:39] <mmu_man> guess the cirrus Jamfile wasn't changed correctly

[11:31:53] <mmu_man> revert accelerant/Jamfile should do it

[11:33:56] *** absabs has quit IRC

[11:34:05] <mmu_man> ahh :)

[11:36:32] *** Sloar has quit IRC

[11:54:50] *** axeld has joined #Haiku

[11:54:51] *** ChanServ sets mode: +o axeld

[11:55:17] <stargater> hi axeld

[11:55:29] <JonathanThompson> Greetings, axeld.

[11:55:42] <axeld> Hi there

[11:55:49] <JonathanThompson> Seems someone in the US is forgetting that you're a bunch of timezones away :)

[11:55:51] <mmu_man> plop

[11:56:05] * JonathanThompson should be in bed, but he's an insomniac

[11:56:22] *** Ingenu has quit IRC

[11:57:02] <_hugo> hey axeld

[11:57:02] <axeld> JonathanThompson: who would that be? :-)

[11:57:05] *** Ingenu has joined #Haiku

[11:57:06] <axeld> Hi _hugo!

[11:57:17] <JonathanThompson> The guy working on the scheduler.

[11:57:35] <_hugo> JonathanThompson: he's actually in brazil, but about the same timezone

[11:57:40] <Ingenu> hey JonathanThompson, know what a static variable inside a class function mean ?

[11:57:46] <Ingenu> (or rather what it does)

[11:58:01] <JonathanThompson> Ah, ok, you can't make any sane guesses about a name not being in the US, due to diversity :P

[11:58:08] <JonathanThompson> Yes, Ingenu.

[11:58:26] <JonathanThompson> Has identical semantics as one outside of the class member.

[11:58:52] <JonathanThompson> Though you need to initialize it first within that member function.

[11:58:57] <Ingenu> so it's a global var ?

[11:59:04] <JonathanThompson> No, a localized global :P

[11:59:15] <Ingenu> reduced scope

[11:59:17] <wkornewald> re. hi everyone

[11:59:21] <JonathanThompson> Everyone that uses the method can use it.

[11:59:21] <wkornewald> _hugo: continue?

[11:59:22] <axeld> JonathanThompson: he's from Brazil, actually :-)

[11:59:33] <Ingenu> ah

[11:59:34] <JonathanThompson> I'll try not to fault him for it, axeld :P

[11:59:58] <_hugo> wkornewald: actually i was preparing to leave a bit, i'm sorry. pretty tired, been up all night. :-)

[12:00:11] <axeld> JonathanThompson: it's only 4 hours away usually; even though it's 5 now because of DST

[12:00:19] <JonathanThompson> Perhaps I should have had a clue he wasn't US-native: his english/spelling while I've seen it here and on the list is too good :P

[12:00:20] <axeld> _hugo: sleep well :)

[12:00:26] *** emitrax_ has joined #haiku

[12:00:34] <wkornewald> ok. then, I'll go cycling. seep well

[12:00:41] <wkornewald> sleep, even :)

[12:00:44] <JonathanThompson> :)

[12:00:53] * JonathanThompson imagines _hugo seeping

[12:01:15] <_hugo> well, im tired, but don't actually feel like sleeping. i would rather fix this one TCP issue

[12:01:16] <_hugo> eheh

[12:01:26] <wkornewald> hi emitrax_

[12:01:31] <JonathanThompson> That's when you should walk away, _hugo :)

[12:01:52] <JonathanThompson> The sooner you get started, the more likely you are to take longer, if you start when tired.

[12:02:01] <wkornewald> yep, I those bug fixing sessions never let me sleep well :(

[12:02:04] <_hugo> JonathanThompson: my coding cycles aren't very usual i would say

[12:02:24] <JonathanThompson> Are you one of those Bill Gates types for that?

[12:02:40] <_hugo> im not sure, how would you describe a Bill Gates type

[12:02:44] <_hugo> besides the uglyness

[12:02:55] <Ingenu> and evilness

[12:02:57] <JonathanThompson> Go for a couple days straight, eating junk, take a long nap, repeat.

[12:03:31] <_hugo> JonathanThompson: im not that constant. i can take a long sleep session, or multiple 2h naps

[12:03:36] <JonathanThompson> I suspect his wife doesn't make that as easy to get away with anymore, assuming he still codes himself.

[12:04:17] <emitrax_> hi wkornewald

[12:04:26] <JonathanThompson> One day about 11-12 years ago when I was a bit younger and didn't have as many health problems, I remember doing a tag-team effort with someone at work, that ended up being a 26 hour work day.

[12:04:49] <_hugo> :-)

[12:05:11] <JonathanThompson> Since I was on the clock at the time, that gave my supervisor a new challenge: figuring out how to get that entered into the computer correctly :P

[12:05:30] <_hugo> after ~24-28h i dont feel like sleeping anymore. but then you get too tired and fall asleep just by sitting in the couch

[12:05:37] <_hugo> but anyway

[12:06:10] <_hugo> i think ill be up a bit more

[12:06:13] <JonathanThompson> It was weird, getting to watch the other guy's facial hair grow in front of my eyes :P

[12:06:16] <_hugo> wkornewald: still regarding the perflet

[12:06:31] <stargater> cu

[12:06:38] *** stargater has quit IRC

[12:06:59] <_hugo> wkornewald: i was wondering whether you were making a distinction between Location and Network

[12:07:53] *** emitrax has quit IRC

[12:08:16] <wkornewald> locations are the same as network profiles

[12:08:32] <wkornewald> they are containers for a set of overridden preferences

[12:08:37] <wkornewald> for multiple networks

[12:08:53] <wkornewald> a network is nearly the same as an interface

[12:09:06] <_hugo> right, i don't agree with that

[12:09:31] <_hugo> a Network is a service which you attach to via an interface (not a specific one)

[12:09:55] <wkornewald> every network is part of a location. I thought of possibly making every network part of every location and only allowing overriding

[12:09:59] <_hugo> and this interface may not even be remain the same during the connection's lifetime

[12:10:13] <_hugo> s/be//

[12:10:25] <wkornewald> _hugo: for which network types can this happen?

[12:10:34] <_hugo> wkornewald: all kind of overlay networks

[12:10:46] <wkornewald> VPN won't care because it uses whatever internet connection is there (ethernet, WiFi, etc.)

[12:10:57] <wkornewald> Mobile IP won't, either

[12:11:08] <wkornewald> and I don't intend to stick WiFi on top of ethernet

[12:11:11] <wkornewald> :)

[12:11:19] *** absabs has joined #haiku

[12:11:44] <wkornewald> do you have an example where this wouldn't work?

[12:11:49] <_hugo> i was thinking that one may want to prevent a VPN connection to be established over a specific internet, and that would be a network preference

[12:12:00] <_hugo> specific interface, sorry

[12:13:05] <wkornewald> why would one want to do that? are there any computers that have multiple internet connections at the same time where you'd want to choose a specific interface?

[12:13:14] <_hugo> wkornewald: yes

[12:13:17] <_hugo> there are plenty

[12:13:34] <wkornewald> which ones?

[12:13:34] <_hugo> ethernet/wifi, wifi/3g, wifi/gprs, etc

[12:13:39] <_hugo> all kinds of multi homing

[12:14:11] <wkornewald> but if you already have a WiFi connection why should you also connect via 3g and pay money? :)

[12:14:31] <_hugo> wkornewald: who decides this?

[12:14:48] <_hugo> that wifi is "better" than 3g

[12:15:08] <wkornewald> well, there are things that just don't make any sense to do. people can sit in their care backwards and drive forwards, but that doesn't make sense and car makers shouldn't build seats that can be rotated

[12:15:52] <wkornewald> if WiFi doesn't make sense then it should be disabled

[12:15:53] <_hugo> wkornewald: i think its pretty reasonable to expect a multi-homed device, like a computer with both a wireless interface a and 3g card both connected to the internet

[12:17:04] <_hugo> saying "the user should disable WiFi" doesn't quite follow the trends. but i don't know what you are targetting :-)

[12:17:29] <JonathanThompson> For traveling purposes, perhaps something new needs to be defined:  a network route configuration that describes the transitions a user may make, and the settings of that route, where "route" is the set of servers between point A and point B and how it all fits together.

[12:18:10] <JonathanThompson> For example, perhaps if there's more than one server along a traveling route, you don't want to sign onto some particular server for whatever reason, and prefer to use others.

[12:18:11] <wkornewald> I think this stuff shoud just work. _hugo: what we can do is this: allow for reordering the networks in the list such that some of them get higher precedence

[12:18:32] <_hugo> wkornewald: that is a good possibility

[12:18:40] <_hugo> wkornewald: the UI must reflect this though

[12:18:43] <wkornewald> (reordering via drag-n-drop, so those who really need it can use it)

[12:19:12] <wkornewald> we can add a simple note ("Note: You can change the priority of interfaces by dragging them up or down.")

[12:19:23] <_hugo> wkornewald: this sounds reasonable to me

[12:20:14] <wkornewald> ok

[12:20:59] *** Sapote has joined #haiku

[12:21:08] <wkornewald> then, the question remains whether you should be able to selectively override preferences or only be allowed to change a subset of those settings or always override the full settings (the alternative in all cases is to use the global settings)

[12:21:56] <_hugo> wkornewald: in terms of interfaces i think its reasonable to copy the profile for new locations

[12:22:16] <_hugo> but i still think there is network configuration that doesn't fit this "interface configuration" model

[12:22:36] <_hugo> specifically VPNs and Mobile IP

[12:22:47] <wkornewald> I think that the most common case could be that VPNs are shared, yes you said it :)

[12:22:58] <_hugo> ok then

[12:23:04] <wkornewald> PPP should probably also be shared

[12:23:33] <_hugo> wkornewald: not so sure

[12:23:36] <wkornewald> the problem is that, depending on the location, you have to override a subset of those settings. e.g., the phone number could change

[12:23:42] <_hugo> exactly

[12:23:53] <wkornewald> in a hotel you need a dial prefix

[12:23:54] <_hugo> i think PPP is a location-dependent configuration

[12:23:59] <wkornewald> you might not even get a dial tone

[12:24:04] <_hugo> i.e. its completely new

[12:24:29] <_hugo> is/has

[12:24:34] <wkornewald> I just don't want that people enter their username and password for every location. the only things that change are a few phone line settings

[12:25:01] *** Jixt has quit IRC

[12:25:12] <wkornewald> VPN is similar. probably only the server changes

[12:25:15] <_hugo> wkornewald: that doesn't seem a normal case to me. if you are using the same ISP (same credentials) you will probably always dial to the same number

[12:25:17] *** Jixt has joined #haiku

[12:25:27] <_hugo> ah, i see what you mean

[12:25:38] <_hugo> you could override parts of the information in the location

[12:26:04] <_hugo> or we consider that credentials are not network configuration and are stored elsewhere

[12:26:15] <wkornewald> yes, exactly. for example, in my "Hotel" location I would override the phone number such that it dials a prefix ("0") and doesn't wait for a dial tone (because that's not available in that hotel)

[12:26:23] <_hugo> so in fact the user/pass is shared, somewhere else, but the profile to dial is new

[12:26:45] <wkornewald> that's what I intended with the Global vs Custom configuration

[12:26:48] <JonathanThompson> If you're using dialup, there's actually a good chance you'd want to dial a more local number, due to line quality and cost of dialing.

[12:27:02] <JonathanThompson> You *could* do the "dial out" method, but that's a more expensive way to work.

[12:27:15] <JonathanThompson> Unless you happen to have unlimited long distance in your plan.

[12:27:27] <wkornewald> JonathanThompson: are you talking about different ISPs or the same ISP?

[12:27:32] <axeld> wkornewald: I've missed the beginning of this conversation, but it should be perfectly fine to have WiFi and ethernet configured at the same time

[12:27:34] <JonathanThompson> Same ISP, even.

[12:27:59] <JonathanThompson> This is true in the US: I don't know about other countries and how that all works.

[12:28:11] <wkornewald> axeld: yes, of course. but if both are connected to the Internet one of them should get higher precedence, at least

[12:28:46] <wkornewald> JonathanThompson: do ISPs in the US offer local numbers?

[12:28:51] <axeld> wkornewald: yes, but the networking stack could figure that out

[12:29:00] <JonathanThompson> Most often that's all they offer for dialing into, wkornewald.

[12:29:08] <wkornewald> axeld: oh, how would that work?

[12:29:18] <JonathanThompson> And that's all people that don't have unlimited long distance plans would reasonably expect to do.

[12:29:20] <_hugo> axeld: the network stack can guess, but get it wrong

[12:29:33] <JonathanThompson> Even with unlimited long distance, line quality tends to suffer if you call long distance.

[12:29:34] <wkornewald> JonathanThompson: I mean, do they offer a local number, a "distant" number and a "foreign country" number?

[12:29:39] <_hugo> axeld: you may seem to be connected to the internet because you were configured via DHCP, but in fact not be

[12:29:42] <axeld> wkornewald: since the recent changes, a driver is supposed to provide information about the line speed and quality

[12:29:49] <_hugo> and if you give preference to ethernet even then

[12:29:58] <JonathanThompson> It isn't 100% consistent with all ISPs.

[12:30:05] <_hugo> you'll get locked out

[12:30:13] <JonathanThompson> A "distant" number is anything outside of your regular calling area.

[12:30:27] <JonathanThompson> And that tends to be $ pretty quickly, in addition to lower line quality.

[12:30:28] <_hugo> so i think we should be able to manually specify the preference of interfaces

[12:30:30] *** GreyGhost has quit IRC

[12:30:35] <axeld> _hugo: true, but can't the routing mechanism be smart enough to try a different interface if one doesn't work out?

[12:30:45] <JonathanThompson> So, that's the reality of dialup in the US.

[12:31:00] <wkornewald> JonathanThompson: ok, thanks for the info. as long as the login information doesn't change this should be possible to do with overrides

[12:31:05] <_hugo> axeld: well

[12:31:15] <JonathanThompson> I thought it'd be useful for me to bring that up :)

[12:31:22] <_hugo> axeld: that is more the work of a router than the host

[12:31:31] <_hugo> besides, that would only work for TCP

[12:31:32] <wkornewald> axeld: I think that if it can be automated it should be done

[12:31:33] <axeld> _hugo: since not all drivers can provide IFF_LINK info, it would be nice if the stack would have means to to something like that

[12:31:39] <JonathanThompson> After all, a lot of countries don't have nearly as many telephone area codes or the same laws/rules for fees.

[12:31:59] <axeld> wkornewald: yes, but if it's too much work, it shouldn't be done for R1 :-)

[12:32:02] *** JamesB192 has joined #haiku

[12:32:09] <_hugo> and im talking about deciding whether a destination is reachable via a specific interface, not link status

[12:32:14] <JonathanThompson> Whenever an ISP does have a single distance number, that's inevitably costly.

[12:32:43] <wkornewald> ok, then, I think we don't even need to offer precedence prefs in the preflet

[12:33:00] <axeld> wkornewald: maybe for R1

[12:33:10] <_hugo> axeld: it would be highly complex. its hard to decide whether a host isnt there because we have no way to reach it or if it is down

[12:33:50] <_hugo> and as you have no way to know, you can tag the default route as "unreachable" or something, you would need to try each destination

[12:33:51] <axeld> _hugo: true enough, but then the other route would fail as well and equal the score

[12:34:06] <_hugo> *you cant tag

[12:34:43] <wkornewald> axeld: I don't think we should introduce such a feature and later remove it, again

[12:34:43] <_hugo> axeld: this could make sense to do per route_info, not route_entry

[12:35:08] * JonathanThompson goes to bed, after imparting useful US-related wisdom into the people that are working on dialup networking for Haiku

[12:35:09] <wkornewald> axeld: it's just an edge case, anyway

[12:35:24] <_hugo> i still think we should be able to change interface preference manually

[12:35:25] <axeld> _hugo: I was thinking about having a default route per interface with a score computed by line speed and quality

[12:35:57] <_hugo> axeld: sure, that makes sense. but the better default route could really not provide connectivity to the internet

[12:35:58] <axeld> wkornewald/_hugo: as a manual intervention this could be useful, yes

[12:36:03] <_hugo> the best

[12:36:21] * mmu_man pets UML use-case

[12:36:28] <mmu_man> hmm long since I touched that thing

[12:36:37] <axeld> _hugo: exactly, and that's why successful connections vs. failed connections would influence the score as well

[12:37:06] *** kb7sqi has quit IRC

[12:37:13] <wkornewald> axeld, _hugo: how should this be displayed in the preflet? I think this is such an edge case that if we can get it to work automatically in most cases we don't need to offer it via the UI

[12:37:17] <_hugo> axeld: i don't think a failed connection via a default route should influence it's metric

[12:37:29] <_hugo> wkornewald: drag and drop sounded good

[12:37:30] <axeld> _hugo: why not?

[12:37:53] <_hugo> axeld: because you have no way to know about the reason why that specific connection failed

[12:38:00] <_hugo> so this could result in a lot of false negatives

[12:38:11] <wkornewald> _hugo: drag-n-drop is manual. if it's automatic you need to tell the system somehow

[12:38:12] <axeld> wkornewald: I agree it's not very important for the UI, but maybe the stack should support this feature when you manually tweak the configuration

[12:38:21] <_hugo> wkornewald: no need to tell the system about automatic

[12:38:38] <wkornewald> and how does the system know it should be automatic?

[12:38:39] <axeld> _hugo: but false negatives aren't a problem as they would affect all routes

[12:38:46] <_hugo> wkornewald: it does, default behaviour

[12:39:07] <wkornewald> and what if you drag something around and then want the automatic behavior, again?

[12:39:27] <_hugo> axeld: imagine the following case. i have two ethernet interfaces, one is connected to my campus, and the other to the internet

[12:39:39] <_hugo> i cannot connect to the campus hosts via the internet one, and vice versa

[12:39:51] <wkornewald> but then you don't have a default route, do you?

[12:39:54] <_hugo> i keep trying to connect to the internet, ill lose connectivity to the campus

[12:40:11] <_hugo> because ill keep decreasing the campus' route metric until it is unusable

[12:40:27] <_hugo> you do, both have default routes

[12:40:30] <_hugo> because the campus network is big

[12:40:35] <axeld> _hugo: only if you're using the default route for the campus

[12:40:41] <_hugo> of course

[12:40:48] <axeld> _hugo: correct would be to have a subnet route

[12:40:53] <_hugo> my campus has a network > than /24

[12:40:56] <_hugo> axeld: not really

[12:41:03] <axeld> _hugo: ie. I would consider this a problem of your configuration

[12:41:09] <axeld> _hugo: how so?

[12:41:10] <_hugo> in fact, my campus has a lot of different networks

[12:41:22] <_hugo> i cant even have a single CIDR that includes all of them

[12:41:29] <_hugo> without including masks i dont own

[12:41:30] <axeld> _hugo: then create several routes for them

[12:41:33] <_hugo> axeld: with DHCP?

[12:41:34] <wkornewald> your campus really has no user-friendly network with VPN servers that change depending on your location :)

[12:41:48] <_hugo> this is a real world example

[12:41:54] <_hugo> this is not a broken setup

[12:42:12] <wkornewald> I guess this problem wouldn't exist with IPv6?

[12:42:19] <axeld> _hugo: how is that supposed to work? The stack could never know which interface to use

[12:42:32] <axeld> _hugo: it would need to introduce per connection routes

[12:42:40] <_hugo> axeld: it would use the default one, and half the connections (campus or internet) wouldnt work

[12:42:46] <wkornewald> _hugo: axeld is right. by manually configuring the routes you don't get the desired effect, either

[12:42:55] <_hugo> but if you change the metric, you change this behaviour

[12:43:02] <_hugo> axeld: exactly

[12:43:03] <_hugo> thats why i said

[12:43:10] <_hugo> we can do it in route_infos

[12:43:16] <_hugo> but not in route_entries

[12:43:24] <_hugo> well, a new route_info

[12:43:28] <_hugo> with full destination

[12:43:51] <wkornewald> _hugo: sorry. I meant: by manually moving around networks in the preflet you won't be able to create correct routes

[12:44:10] <_hugo> wkornewald: you dont create routes, by moving them you change their cost

[12:44:13] <_hugo> cost/metric

[12:44:18] <wkornewald> or that

[12:44:19] <wkornewald> :)

[12:44:23] <wkornewald> still won't work

[12:44:28] <_hugo> wkornewald: why not?

[12:44:53] <wkornewald> well, it would work if you create routes on the command-line or in configuration files

[12:45:07] <axeld> _hugo: route_info is supposed to be used in something like TCP - the actual route would change as needed

[12:45:30] <_hugo> axeld: yes, this could be used to change the route if the we are using is failing

[12:45:35] <_hugo> if there is a new compatible one

[12:46:14] <axeld> _hugo: but the metric is part of the route, not the info - the info is just a reference to a route giving a target address

[12:46:15] <_hugo> wkornewald: in my view, the perflet should message the net server with the new preference, it knows how it configured the interfaces, and change the costs

[12:46:22] <wkornewald> I just don't want that our preflet is overloaded with complex prefs for edge-cases. I want that even my uncle can setup his home network instead of calling me because he doesn't understand all those prefs

[12:46:33] <_hugo> axeld: i know, we just need a bit more info in route_info

[12:47:23] <_hugo> axeld: this could be used for UDP as well, we could cache the destination in the UdpEndpoint

[12:47:53] <axeld> _hugo: I don't see how this could work

[12:48:07] <_hugo> axeld: using route_info for this or what i said about UDP?

[12:48:31] <axeld> _hugo: ah, now I get closer: you mean the stack should actually know about all route_infos as well, right?

[12:48:40] <_hugo> yes

[12:48:47] *** rcjsuen has joined #haiku

[12:49:03] <_hugo> both TCP and the underlying protocols should feed information into the stack so it tries to update route_info

[12:49:22] <_hugo> TCP could say "too many un-acked packets", and the underlying could say "ICMP dst unreach"

[12:49:34] <_hugo> this would hint the stack to try and switch the route_info to use another route

[12:50:14] <axeld> _hugo: right now, the route_info implementation is still a bit in the air, but it could be done like that

[12:50:48] <_hugo> i think this could work really well, specially if the network signals us with ICMP dst unreach

[12:50:50] <axeld> _hugo: but still, the stack should be smart enough to actually switch preference of a route using this information

[12:51:30] <axeld> _hugo: also, it's still problematic for anything else than TCP, as it would be on the client's burden to resend the data

[12:51:33] <_hugo> well, its a bit more complex route selection

[12:52:06] <_hugo> axeld: UDP is unreliable, packets could be lost in the network anyway

[12:52:12] <_hugo> the app should always try to resend or ignore

[12:52:19] <axeld> indeed

[12:52:43] <wkornewald> does this affect the preflet in any way?

[12:52:49] <_hugo> wkornewald: no :-)

[12:53:16] <axeld> _hugo: okay, you've convinced me - go ahead and implement it ;-))

[12:53:21] <wkornewald> I'm just wondering what would happen if you had to setup your campus network.

[12:53:30] <_hugo> wkornewald: regarding the perflet, you could have dragging the interface and around and a "reset" button

[12:53:35] <JonathanThompson> Oh, I should be sleeping, but other than a facial explosion, I thought I should mention at least one more thing about dialup in the US that's important.

[12:53:36] <_hugo> axeld: eheh, bug fixing first!

[12:54:06] <wkornewald> _hugo: how would that work with your campus network (>24)?

[12:54:09] <JonathanThompson> Make sure you have at least one alternate number for any given location for the same network, because not everyone can have success calling into the same number, for various reasons.

[12:54:12] <_hugo> wkornewald: it wouldnt

[12:54:31] <_hugo> i mean, it would only change all the connections either to one ethernet

[12:54:32] <_hugo> or the other

[12:54:55] <JonathanThompson> if(busy||NotConnectingSuccessfully) TryAlternateNumbers();

[12:54:56] <wkornewald> yep. how are the students configuring their network, then?

[12:54:57] <_hugo> but still there are other simpler cases that could benefict from you being able to specify which interface should be used

[12:55:27] <_hugo> wkornewald: this was in my research lab. i added a lot of routes manually

[12:55:41] <wkornewald> but the other students don't have to cope with that?

[12:55:58] <wkornewald> they just plug in and surf?

[12:55:59] <_hugo> no because the default route they get gives them access outside

[12:56:09] <wkornewald> but they can't get inside?

[12:56:36] <_hugo> well, my research lab network was not really my campus network, it was separate, thus the issue

[12:56:53] <_hugo> so you could think of the campus network as the internet, and my research lab network as the other one

[12:57:24] <wkornewald> ok. BTW, can DHCP give you a non-default route?

[12:57:25] <_hugo> anyway, what i was discussing with axel could work for other things as well

[12:57:40] <_hugo> wkornewald: DHCP only specifies the router you should use

[12:58:03] <_hugo> besides the subnet mask

[12:58:25] <wkornewald> and wouldn't automatic route choosing be better in most cases than manually via drag-n-drop? or would the automatic stuff still work, even if you have drag-n-drop?

[12:58:33] <_hugo> it would still work

[12:59:07] <wkornewald> ok, so I could reorder the networks *incorrectly*, but in the end it would still work correctly?

[12:59:35] <_hugo> well, it could.

[12:59:42] <axeld> wkornewald: just with a slow start :)

[12:59:45] <_hugo> yeah

[13:00:04] <wkornewald> ok...

[13:00:13] <axeld> That's why I think the stack could be smart enough to reorder internally

[13:00:25] <_hugo> i think the default behaviour

[13:00:35] <_hugo> should be what you mentioned axeld

[13:00:51] <wkornewald> axeld: wouldn't that be done even if I set the precedence manually?

[13:00:58] <_hugo> setting the metric to be a product of the media "quality"

[13:01:44] <_hugo> wkornewald: as i see it, the stack would set metrics itself as long as the metric was not overriden

[13:02:05] <axeld> yes, but that can vary as well, ie. you might have two different WiFi cards connected to two different networks - and the signal strength can vary between them :-))

[13:02:23] <_hugo> yes

[13:02:33] <_hugo> or you can remove the cable from your ethernet card

[13:03:38] <wkornewald> so, I think the preflet shouldn't allow for changing the order

[13:03:43] <_hugo> eheh

[13:04:07] <_hugo> wkornewald: being able to change the order does not add any clutter to the interface

[13:04:10] <wkornewald> this is really an edge case and those who need it can edit the configuration files

[13:04:22] <_hugo> ugh :-)

[13:04:25] <wkornewald> it adds a note and a button for resetting the order

[13:04:31] <_hugo> so dont add a note

[13:04:45] <wkornewald> and how do people know what that reset order button is for? :)

[13:05:01] <_hugo> or if people try to move the order, since you dont want users to frack it up, you could have a BAlert asking for confirmation

[13:05:19] <wkornewald> or, maybe a checkbox "[x] Choose Network Priorities Automatically"

[13:05:27] <_hugo> wkornewald: the reset button would only be enabled when they were moved

[13:05:37] <_hugo> that could work too

[13:05:46] <_hugo> drag and drop would not work as long as that was on

[13:06:10] <wkornewald> hmm, maybe a button is better, only shown when dragged....

[13:06:40] <_hugo> the checkbox looked good as well

[13:07:07] <_hugo> [x] Determine Network Preference Automatically

[13:07:26] <_hugo> and the list order would reflect the current order

[13:08:09] <wkornewald> but I doubt that this feature is useful for more than a handful of people

[13:08:42] <_hugo> wkornewald: only a handful of people will go to the network preferences anyway

[13:09:09] <wkornewald> I don't think so. many people will have to configure their PPPoE or WiFi connection

[13:09:10] <_hugo> all the rest should either be connected or get a "Please enter the password to network X" dialog

[13:09:29] <_hugo> a connection to WiFi should not go through the network preferences

[13:09:38] <_hugo> it is too complicated

[13:09:52] <wkornewald> WiFi is too complicated?

[13:10:04] <_hugo> no, going to the preferences just to connect

[13:10:13] <_hugo> there should be a deskbar replicant, part of an application which also does scanning

[13:10:29] <_hugo> where you go an select "Connect to this network" from a menu

[13:10:40] <_hugo> it would figure out whether you need a password or not

[13:10:45] <_hugo> and ask one if so

[13:10:49] <_hugo> that would be it

[13:11:02] <wkornewald> the preflet should still allow for changing the location prefs

[13:11:07] <wkornewald> for WiFi

[13:11:13] <_hugo> sure. but that is advanced usage

[13:11:24] <_hugo> thats why i think you should be able to move interfaces around

[13:11:33] <wkornewald> yep. but we still have a few users with PPPoE

[13:12:13] <_hugo> PPPoE should be similiar to the WiFi case. All simple "Connects" should be there

[13:12:49] <wkornewald> you have to create it somehow...I'm wondering if we could split the preflet into two parts

[13:12:57] <_hugo> in my opinion you should only go to the network preferences when you need to specifically define stuff such as WEP vs WPA

[13:13:00] <wkornewald> replicant and preflet

[13:13:07] <_hugo> please do :-)

[13:13:14] <_hugo> we already have NetworkStatus

[13:13:20] <_hugo> but it should handle configuration as well

[13:13:27] <_hugo> in my opinion of course

[13:14:21] <wkornewald> I'll try to work something out. but I still have the question whether it's better to share all network prefs with the global location (and thus with all locations)

[13:15:03] <axeld> I think it's good to have all the configuration in a single place - but maybe that can be combined somehow

[13:15:42] <axeld> Ie. the NetworkStatus app could launch the prefs starting with the right tab

[13:17:28] <wkornewald> axeld: yes, I had thought of something like that, too...but OTOH maybe the replicant could offer all the basic configuration and the preflet could offer the non-trivial stuff, only

[13:17:43] <_hugo> im not sure where the logic would be, but i think the replicant should detect new interfaces/networks and let the user connect to them in place

[13:17:47] <_hugo> instead of launching the preferences

[13:18:03] <wkornewald> IMHO, in any case, there should only be one way to do something

[13:19:01] <_hugo> i think there should be simple ways to do simple tasks :-

[13:19:02] <_hugo> :-)

[13:20:16] <wkornewald> _hugo: I don't think that the replicant should automatically show all available networks. that's what the NetworkManager does and IMHO it sucks because you're helpless if the context menu doesn't reload. it should open a new window. only basic functions (select known network, disconnect, ...) should be available. the rest should be done via a separate window

[13:20:30] <wkornewald> (NetworkManager under GNOME)

[13:21:15] <_hugo> wkornewald: hm

[13:21:31] <_hugo> so you've identified what is wrong with networkmanager

[13:21:34] <_hugo> we can do better :-)

[13:22:18] <_hugo> but i still think showing the networks in the context menu is a good idea

[13:22:23] <_hugo> and we can update it dynamically

[13:22:31] <_hugo> and even show signal bars

[13:22:32] <wkornewald> well, it should continually reload and collect available networks

[13:22:39] <_hugo> of course

[13:22:49] <_hugo> most gnome stuff is very blocking, thats why networkmanager sucks in that regard

[13:22:56] <wkornewald> but I don't know if that's desired (battery usage?)

[13:23:23] <wkornewald> and can a WiFi card collect networks all the time and be connected, at the same time?

[13:23:37] <wkornewald> AFAIK, it can't listen on multiple channels at the same time

[13:23:59] <axeld> wkornewald: Mac OS X also show the available networks in the popup menu, and I always found it quite handy

[13:24:21] <wkornewald> ok, if this can work well enough it should be done, then.

[13:24:26] <_hugo> wkornewald: scanning is done by sporadically changing channel

[13:24:45] <wkornewald> yep. so, this shouldn't be done all the time in the background, right?

[13:24:53] <wkornewald> otherwise you lose packets

[13:24:56] <mmu_man> btw, there is an app that does that already and much more for for laptop

[13:25:08] <_hugo> not all the time, but you can do it eventually

[13:25:08] <mmu_man> used to be hosted on beclan

[13:25:23] <mmu_man> can't remember the name

[13:25:39] <wkornewald> axeld: does OS X only lookup the networks when you click the replicant?

[13:25:46] <wkornewald> or all the time?

[13:26:51] <_hugo> wkornewald: i've used OS X and cant answer that. there is a delay accessing that context menu though

[13:26:59] <_hugo> but its still fast to hop all the required channels

[13:27:11] <_hugo> *too fast o

[13:27:14] <wkornewald> does it grow after you've clicked?

[13:27:32] <_hugo> wkornewald: it does not

[13:27:39] <_hugo> nor does it show signal

[13:27:42] <wkornewald> (i.e. detect more networks)

[13:29:03] <wkornewald> hmm, even the ethernet prefs could move to the replicant

[13:29:36] <_hugo> basic ethernet preference is connect/disconnect

[13:29:37] <wkornewald> do we need a preflet, at all? :)

[13:29:39] <_hugo> always auto configure

[13:29:51] <_hugo> yes. if you want to configure a static address

[13:29:54] <_hugo> for instance

[13:30:12] <wkornewald> the status info (which IP is configured, which router, etc.) could contain all that :) but it's probably too overloaded

[13:31:37] <Ingenu> boneyard

[13:31:42] <_hugo> the status window could have a "Configure" button that would launch the perflet with the interface selected

[13:32:34] <wkornewald> _hugo: what I meant was that since you have to display "IP Address" anyway, this could be a text field that can be changed. you could also select auto config and then those text fields would be replaced with just text

[13:32:59] <axeld> wkornewald: the NetworkStatus is *optional* so yet, we need a preferences app

[13:33:12] <wkornewald> axeld: who said it's optional? :)

[13:33:20] <_hugo> i really think it should not be optional

[13:33:21] <_hugo> eheh

[13:33:25] * wkornewald is just throwing around ideas

[13:33:36] <_hugo> i mean, it is, but should be started by default

[13:33:40] <axeld> wkornewald: it is; it's your own call to run it

[13:34:06] <wkornewald> axeld: it should be ran by default, IMHO. it's important enough

[13:34:16] <axeld> wkornewald: why would it be important?

[13:34:37] <wkornewald> it could display connection errors. or the connection status of your WiFi network

[13:34:55] <axeld> wkornewald: that doesn't make it important

[13:35:13] <wkornewald> well, WiFi networks don't connect immediately

[13:35:17] <axeld> wkornewald: it's useful when you have WiFi, yes, but it's almost useless on a desktop computer - nor should it be mandatory

[13:35:20] <wkornewald> you have to wait a moment

[13:35:34] <axeld> wkornewald: you mean I should think before I talk? ;-)

[13:35:45] <wkornewald> hehe :)

[13:37:51] <wkornewald> hmm, I don't think we can move all the network prefs into the replicant, so the replicant will probably not be needed...but it should be shown when you have a WiFi card or when there is a connection error/misconfiguration

[13:38:21] <wkornewald> maybe we just need an option in the preflet for that

[13:38:37] <_hugo> yes, it should be used to talk back to the user

[13:40:15] <wkornewald> probably we don't need the option...just show always when there are certain network types. but then, a laptop user who doesn't use WiFi might always see that replicant....

[13:40:17] <axeld> Indeed, the net_server could start it when the need arises

[13:40:51] <axeld> wkornewald: I wouldn't do it like this - only start it when there is a need to have it

[13:40:54] <wkornewald> for preferences, I think I like the idea that the replicant opens the preflet with the right network pre-selected

[13:40:59] <axeld> wkornewald: otherwise, let the user decide

[13:41:23] <wkornewald> axeld: doesn't a user with a PPP connection always need that preflet?

[13:41:38] <axeld> wkornewald: also, don't forget NetworkStatus could sit as a replicant at your desktop if the user wants it that way

[13:41:58] <wkornewald> ouch :)

[13:42:09] <axeld> wkornewald: it looks kinda neat :-)

[13:42:16] <axeld> (as it can be any size)

[13:42:17] <wkornewald> do you have a screenshot?

[13:42:25] <axeld> No, but I can make one

[13:42:26] <wkornewald> is it just an icon or what?

[13:42:53] <wkornewald> what does it look like on the desktop, then?

[13:43:03] *** myob has joined #haiku

[13:43:05] <_hugo> like a replicant :-)

[13:43:06] <axeld> wait a sec :)

[13:43:13] <axeld> _hugo: exactly :-))

[13:43:20] <_hugo> eheh

[13:43:21] <wkornewald> hehe :)

[13:43:54] <wkornewald> ok, so: the replicant only offers connecting, disconnecting, status. for prefs the preflet is consulted

[13:45:05] <myob> argh, I hate cli irc clients...

[13:45:50] <axeld> wkornewald: http://www.pinc-software.de/images/Haiku-NetworkStatus.png

[13:46:15] <wkornewald> neat, but I hope you don't want this feature to be configurable via the preflet :)

[13:46:31] <_hugo> wkornewald: this is standard BeOS, just drag the replicant to the desktop

[13:46:32] <wkornewald> how does one install the preflet?

[13:46:36] <wkornewald> replicant!

[13:46:42] <wkornewald> wow... :)

[13:47:04] <wkornewald> does the app open a window? I thought it only installs a replicant

[13:47:19] <axeld> wkornewald: I kept "Show Replicants" turned on so that you can see

[13:47:20] <wkornewald> or can I drag replicants from Deskbar?

[13:47:26] <myob> damn, I need to play with haiku again...

[13:47:40] <axeld> wkornewald: when you start it, you can decide wether it installs itself to the Deskbar or not

[13:47:41] <myob> been about six months since I had a PC it booted on

[13:47:49] <wkornewald> ok

[13:48:01] <axeld> wkornewald: exactly like ProcessController

[13:48:21] <wkornewald> good. then let's please move on with the next open question for the preflet:

[13:48:48] <wkornewald> should all networks be avilable on all locations (but Global prefs can be disabled/overridden)

[13:49:13] <wkornewald> the advantage is that you can easily create new networks from any location and have them avilable in the other locations (important for VPN)

[13:49:36] <wkornewald> the disadvantage is that the preflet will show more networks than necessary, in the list

[13:50:10] <_hugo> i wouldnt say it should more networks that necessary

[13:50:16] <_hugo> should show

[13:50:18] <wkornewald> but I don't think that anyone will have so many connections that this gets too annoying. also, when disabled the networks won't show up in the replicant

[13:50:33] <mmu_man> hmm damn what was that app from david reid for laptop stuff

[13:51:04] <wkornewald> _hugo: but then you can't easily share stuff

[13:51:11] <myob> mmu_man david reid? I know petterhj had one that was hosted on beclan (which is reids...)

[13:51:31] <mmu_man> well, maybe, I nkow it's on beclan yes

[13:51:36] <mmu_man> can't remember the name

[13:51:44] <mmu_man> handled apm, wifi and other stuff

[13:52:02] <myob> beportable

[13:52:08] <myob> dunno if it was ever released

[13:52:09] <mmu_man> right, that's it :)

[13:52:14] <wkornewald> _hugo: and this could be pretty annoying for VPN and PPP

[13:52:17] <myob> to date nobody has a 'show me local networks' thinking

[13:52:20] <myob> thingy*

[13:52:24] <mmu_man> it's not mentioned on the public part of beclan

[13:52:24] <_hugo> wkornewald: well, in my pov, there would be a base config (global) with groups (like you stated) override by locations

[13:52:31] <mmu_man> and it seems my pass doesn't work anymore

[13:52:32] <myob> although patrick's drivers support it

[13:52:43] <wkornewald> _hugo: groups? :)

[13:52:50] <_hugo> wkornewald: groups of settings

[13:53:02] <wkornewald> ok, let's call them by their name ;)

[13:53:11] <_hugo> what is their name? :-)

[13:53:18] <wkornewald> groups of preferences :)

[13:53:20] <_hugo> eheh

[13:53:22] <_hugo> ok then

[13:53:25] <wkornewald> or sets of prefernces :)

[13:53:35] <wkornewald> I just forgot the other two words whenI said "groups" :)

[13:53:45] <myob> need to make michael lotz do that wifi stuff for haiku I guess ;)

[13:54:24] <wkornewald> _hugo: isn't that what I suggested, then?

[13:54:41] <_hugo> wkornewald: i guess. but i hope that solves your problem of whatever is best

[13:54:53] <wkornewald> _hugo: if I create a new VPN network in my "Hotel" location it will also be available in the "Global" location and thus everywhere

[13:55:01] <_hugo> i just think you should be able to have connection completely embedded in locations, i.e. no global presence

[13:55:08] <_hugo> connections

[13:55:12] <_hugo> argh, damn typing

[13:55:18] <wkornewald> hehe

[13:55:33] <wkornewald> _hugo: how do you share settings then?

[13:56:01] <_hugo> wkornewald: in that particular case you dont want to share

[13:56:09] <wkornewald> you could do this by "importing" networks from other locations, but that should only create a duplicate because otherwise you get ugly dependencies

[13:56:10] <_hugo> the connection is specific to the location

[13:56:23] <_hugo> let me put it this way

[13:56:38] <_hugo> a connection is a set of groups of preferences

[13:56:59] <_hugo> you could have this set in the global preferences, and have groups overridden by locations

[13:57:10] <_hugo> or you could have the whole set contained in a location

[13:57:24] <_hugo> when you are in a particular location, you instantiate the complete set and create the connection

[13:57:31] <_hugo> so it doesnt matter whether it is global or not

[13:58:19] <wkornewald> hmm...maybe that's indeed better...then, you'd have to manually go to "Global" and create it there. the "Add..." function could also show a list of location-private networks

[13:59:17] <_hugo> sure, if for some reason you want to make it global. but i think most people will add the connection to its proper location from the start

[13:59:59] <wkornewald> something like "Add -> Make Global -> Hotel -> Work (VPN)" ugh...deep hierarchies...

[14:00:19] <_hugo> i was thinking more of:

[14:00:26] <wkornewald> maybe a checkbox is better :)

[14:00:29] <_hugo> select "Hotel"

[14:00:35] <_hugo> whether on the left, top, or whatever

[14:00:38] <_hugo> that affects what you see

[14:00:43] <_hugo> then press on Add

[14:00:46] <_hugo> this Add is Hotel specific

[14:00:52] <_hugo> and then pick VPN

[14:01:16] <_hugo> instead of Hotel you could have picked "Global preferences" or "All locations"

[14:01:31] <wkornewald> yeah, my use-case was for making an existing location-private network global

[14:02:33] <_hugo> ah ok :-)

[14:04:20] <wkornewald> hmm, so, how should this overriding work, then?

[14:05:36] <wkornewald> 1) show custom and global above each other (- [ ] Use Custom Preferences --------   ......    - [ ] Global Preferences -----------)

[14:05:55] <wkornewald> I don't think this is good.... :-/

[14:05:58] <_hugo> i would say only show one group

[14:06:05] <_hugo> with a checkbox below

[14:06:07] <_hugo> stating

[14:06:12] <_hugo> "Override these values for this location"

[14:06:19] <_hugo> which would enable the fields

[14:08:01] <wkornewald> 2) put a box around groups of prefs and have a drop-down "[Use Global Preferences \/]" (or "Custom")

[14:08:46] <_hugo> that works too. but i feel that a checkbox is more user friendly. usually stuff that is visible (vs. hidden) are a better choice

[14:08:56] <wkornewald> also, there would be a checkbox "[ ] Make Globally Available"

[14:09:15] <wkornewald> or "[ ] Available In All Locations"

[14:09:24] <_hugo> ah, dont think that one is good

[14:09:40] <_hugo> i think its better to specialize than to generalize

[14:10:35] <wkornewald> the checkbox would just move the interface to "Global"

[14:10:47] <wkornewald> in case you made a mistake

[14:11:12] <wkornewald> and added it to a custom location

[14:11:16] <_hugo> moving the user around without specifying so its also not good for usability. people get lost

[14:11:34] <wkornewald> we wouldn't move him around

[14:12:37] <wkornewald> hmm, maybe the moving around can be done with "Add..." and "Remove" (where remove asks whether to remove from all locations or keep the entry in the current one)

[14:12:43] *** BuildFactory has quit IRC

[14:12:51] <wkornewald> nah...

[14:12:54] <_hugo> :-)

[14:13:47] <wkornewald> well, how do you remove it from Global then? :)

[14:14:12] <_hugo> select, delete

[14:14:33] <wkornewald> and lose all prefs?

[14:14:35] <_hugo> "this configuration is also used in the following locations, do you really wish to remove?"

[14:14:37] *** mmu_man has quit IRC

[14:15:10] <wkornewald> I mean, what if you made it global, but then want to keep it only in one location

[14:15:19] <_hugo> drag it to the location

[14:15:33] *** BuildFactory has joined #haiku

[14:15:35] <wkornewald> locations aren't visible like that

[14:15:50] <wkornewald> it will probably only be a drop-down

[14:16:06] <wkornewald> they shouldn't take up a lot of space (since that feature isn't used by everyone)

[14:16:23] <_hugo> so right click and "Move", or a Move button

[14:16:24] <wkornewald> (and even then, they don't need much space)

[14:16:57] <_hugo> Move is not as clear as the drag though

[14:17:42] <wkornewald> I don't like the move thingy...

[14:17:47] <_hugo> :-)

[14:17:55] <wkornewald> :)

[14:18:01] <_hugo> im sure you'll figure out something better

[14:18:46] <wkornewald> maybe...maybe I can just add a button "Remove From Global"...

[14:19:26] <wkornewald> _hugo: didn't you want to sleep?

[14:19:27] <wkornewald> :)

[14:19:35] <_hugo> i was thinking the same

[14:19:40] <_hugo> but im writing some code

[14:20:25] <wkornewald> get some sleep. if you work too much you burn out. but don't work too little, either ;)

[14:20:47] <_hugo> eheh

[14:20:59] <_hugo> im well enough to finish this bit, ill rest after

[14:23:39] <wkornewald> ok, I'll go. bye. good night

[14:23:49] <_hugo> see you wkornewald :-)

[14:23:50] <_hugo> have fun

[14:24:01] <wkornewald> thanks. up the hills, some cycling :)

[14:24:06] *** wkornewald has quit IRC

[14:28:14] <CIA-17> jackburton * r20662 /haiku/trunk/src/servers/app/Desktop.cpp:

[14:28:14] <CIA-17> Menu windows are no longer the preferred keyboard events target (for

[14:28:14] <CIA-17> now, at least). This feature wasn't used yet anyway, and turning it

[14:28:14] <CIA-17> off cures bug #1152.

[14:31:25] *** rcjsuen has quit IRC

[14:40:09] *** StylusEater_Work has joined #haiku

[14:52:01] *** emitrax__ has joined #haiku

[14:59:50] *** emitrax_ has quit IRC

[15:01:21] *** kr1stof has joined #haiku

[15:09:03] <Sapote> Hi to all! I live in Argentine. I have used beos from version 5, at the moment I use linux for several years, come following the project haiku and is incredible since the last times have progressed. Congratulations!

[15:10:54] <Sapote> my natural language is spanish, for contact to me in IM: guillermobenite at jabber dot org

[15:11:09] <Sapote> or in #haiku-es

[15:15:11] <CIA-17> jackburton * r20663 /haiku/trunk/src/kits/opengl/GLView.cpp:

[15:15:11] <CIA-17> sizeof(B_PAGE_SIZE) != B_PAGE_SIZE. Fixed bug 1151. Sorry for the silly

[15:15:11] <CIA-17> typo. Also uses calloc() so that the direct_buffer_info struct is

[15:15:11] <CIA-17> initialized to 0.

[15:17:35] <Jixt> is there someone known to the haiku code?

[15:17:52] <_hugo> Jixt: to a bit of it

[15:18:02] <Jixt> is there a way to get the TRACE log messages into the syslog?

[15:18:19] *** GreyGhost has joined #haiku

[15:18:25] <Jixt> some option in the kernel settings file?

[15:19:48] <_hugo> it is possible, yes

[15:20:12] <Jixt> do you also know how to do it?

[15:20:34] <CIA-17> axeld * r20664 /haiku/trunk/src/add-ons/translators/jpeg/ (JPEGTranslator.cpp be_jerror.cpp):

[15:20:34] <CIA-17> * EXIF data is now also parsed in case there was no ioExtension message passed in;

[15:20:34] <CIA-17>  the orientation was only honoured in case there was a message before.

[15:20:34] <CIA-17> * Disabled error alerts; they are more annoying than helpful.

[15:20:34] <CIA-17> * Bumped version to 1.2.0.

[15:20:49] <_hugo> you need to edit the kernel configuration and set "syslog_debug_output true"

[15:21:19] <_hugo> you can else edit directly in the source if are you building your image

[15:21:41] <Jixt> yes, I did that already. And I only see the TRACE_ERROR messages :-S

[15:22:15] <_hugo> Jixt: most of the TRACE() calls are #ifdef'ed in each file. what messages do you want to see?

[15:23:28] <Jixt> I want to  see the messages from the usb stack. I've got it running, but I like to see all the messages when a usb device is connected

[15:23:45] <Jixt> example -> http://svn.berlios.de/viewcvs/haiku/haiku/trunk/src/add-ons/kernel/busses/usb/uhci.cpp?view=markup

[15:24:20] <myob> gah I can't figure out how to read abck the buffer in irssi...

[15:24:52] <_hugo> Jixt: DEBUG must be set

[15:25:11] *** guma has joined #haiku

[15:25:32] <Jixt> ok, in which file?

[15:36:29] *** emitrax_ has joined #haiku

[15:40:18] <Jixt> I should build the usb stack in debugging mode

[15:42:44] *** petterhj has joined #haiku

[15:43:05] *** guma has quit IRC

[15:44:51] *** emitrax__ has quit IRC

[15:50:25] <Ingenu> I'm working on officially shitty C++ code (found no less than 4 goto)

[15:52:48] *** guildencrantz has joined #haiku

[15:56:11] *** Sloar has joined #haiku

[16:05:15] *** jevin has joined #haiku

[16:05:18] <StylusEater_Work> goto's???? ouch

[16:05:39] <Ingenu> yeah don't tell me

[16:05:54] <Ingenu> game dev studio more proggies than artists

[16:06:00] *** slaine__ has joined #haiku

[16:06:01] <Ingenu> and most coders suck

[16:12:10] <Sapote> goto!?

[16:12:10] <Sapote> :D

[16:12:59] <Ingenu> yeah :(

[16:13:26] <Ingenu> so get Haiku done people so I can rejoyce about something

[16:14:40] <jevin> didnt you hear? i finished it last night

[16:14:45] <jevin> its done. finished.

[16:15:08] <Ingenu> so release it now ;p

[16:15:24] * jevin runs

[16:15:52] *** slaine__ has quit IRC

[16:16:23] *** slaine__ has joined #haiku

[16:16:43] *** GreyGhost has left #haiku

[16:17:10] *** slaine_ has quit IRC

[16:17:47] *** guma has joined #haiku

[16:20:37] *** Mazon is now known as mazon

[16:30:41] *** absabs has quit IRC

[16:35:33] *** Jixt has quit IRC

[16:38:52] *** MrSunshine has joined #haiku

[16:41:14] *** kr1stof has quit IRC

[16:41:37] *** slaine__ has quit IRC

[16:42:05] *** slaine_ has joined #haiku

[16:44:00] *** rcjsuen has joined #haiku

[17:01:42] *** raph_ael has quit IRC

[17:03:15] *** raph_ael has joined #haiku

[17:04:48] *** TTRanger has quit IRC

[17:08:01] *** arikato has quit IRC

[17:08:35] <DeadYak> wow, 8 students? nice

[17:09:21] <Ingenu> yep

[17:09:32] *** arikato has joined #haiku

[17:11:09] *** Jixt has joined #haiku

[17:13:04] <DeadYak> Ingenu: incidentally, goto has its place

[17:13:13] <DeadYak> Dijkstra did more damage with that paper than possibly any one else in computing history

[17:13:23] <myob> wonder can they do 8 times as much as the last student :P

[17:13:31] <DeadYak> myob: what last student?

[17:13:54] <myob> andrew g

[17:14:17] <DeadYak> *blink*

[17:14:20] <DeadYak> drawing a blank....

[17:15:58] <myob> last year, networking, did buggerall

[17:16:10] <myob> got paid by haiku directly

[17:16:13] <DeadYak> ah.

[17:16:21] <DeadYak> forgot about that

[17:18:42] <jevin> haiku paid him to do nothing? :-?

[17:20:55] *** jevin has quit IRC

[17:26:31] *** slaine_ has quit IRC

[17:28:30] *** slaine_ has joined #haiku

[17:30:45] *** emitrax__ has joined #haiku

[17:32:37] *** Anxiety has joined #haiku

[17:32:54] *** slaine__ has joined #haiku

[17:33:24] *** nutela has joined #haiku

[17:33:53] <nutela> Hi guys first moment on #haiku

[17:34:09] <nutela> and it's... quiet (everybody must be coding)

[17:34:53] <nutela> i wanted to ask, is anybody using R5 on Core Duo mobo?

[17:35:06] <Ingenu> tried, didn't work

[17:35:15] <Anxiety> Good question, I'd like to know that too

[17:35:24] <Ingenu> ASUS P5B-E Plus & E6600 w/ 2GB

[17:35:37] <Ingenu> locked somewhere

[17:35:38] <Anxiety> My first guess would be no for SATA reasons

[17:36:07] <Ingenu> CD-ROM are still PATA, why would SATA crash the system ?

[17:36:38] <Ingenu> but it might be at the disk icon it froze, may check it when I get back from work later

[17:37:14] <Anxiety> not crash, but be unable to load... or did werent you just not able to load the install cd?

[17:37:15] *** Jixt has quit IRC

[17:37:58] <Ingenu> install CD did freeze at some icon

[17:38:02] <Ingenu> I don't remember which

[17:38:19] <Ingenu> now I realize I'm all SATA anyway so I wouldn't have been able to install it ^^

[17:38:27] <Ingenu> (at least it would have ran as a LiveCD)

[17:38:46] <Anxiety> The bigger question, for me anyways... does anyone have Haiku running on a Core 2 Duo?

[17:39:03] <Ingenu> get me a bootable Haiku CD and I'll try

[17:39:07] *** emitrax_ has quit IRC

[17:39:13] <DeadYak> if you hit space immediately when the CD starts you can set some options that make it display what it's doing on the console

[17:39:27] <DeadYak> which should tell you exactly why it's halting

[17:39:40] <Ingenu> for BeOS5 ?

[17:39:41] <DeadYak> lack of SATA should actually show you a KDL indicating it wasn't able to find a bootable device, not a hard freeze

[17:39:42] <DeadYak> yes.

[17:39:45] <Ingenu> remember which option it is ?

[17:39:56] <DeadYak> trying to remember the layout of bootman's menu, it's been a while

[17:40:06] <Ingenu> for me too ^^

[17:40:09] <DeadYak> there's a debug output option, but I seem to remember the name being a bit counter-intuitive

[17:40:17] <Anxiety> http://home.gna.org/pingwinek/index.html ?

[17:40:24] <Ingenu> at worst I'll check them all ;)

[17:41:19] <Ingenu> nice it works now (that website)

[17:41:27] <Ingenu> ok will get it at home and give it a try

[17:41:37] <Ingenu> does Haiku handle SATA ?

[17:41:44] <DeadYak> sorry I can't do any better right now but I'm at work, no BeOS CD here

[17:41:55] <DeadYak> fairly sure it's under safe mode options though

[17:42:06] <DeadYak> and I believe Haiku has an SATA driver, yes

[17:42:57] <Ingenu> np DeadYak I've been using BeOS as my main OS during a few years, should be able to find it by myself

[17:42:57] <DeadYak> note however....

[17:42:57] <DeadYak> the console output is nowhere near as verbose as serial is

[17:43:03] <Ingenu> (or rather : I did use BeOS)

[17:43:06] <DeadYak> so enabling serial debug output and using a serial cable to another PC is arguably more useful

[17:43:13] <DeadYak> same :P

[17:43:22] <nutela> so does R5 work on Core (2) Duo

[17:43:32] <DeadYak> nutela: probably not without some patches

[17:43:33] <myob> seriously doubt it

[17:43:36] <nutela> I mean on mobo

[17:43:38] <Ingenu> I don't have a serial cable

[17:43:41] <myob> won't boot on my Pendium-D

[17:43:45] <nutela> Hi MYOB!

[17:44:21] <[Beta]> myob, (from above) ... hugo has probably done 2x as much already

[17:44:21] <DeadYak> bear in mind R5 is like... 7 years old at this point

[17:44:21] <nutela> dang

[17:44:21] <nutela> I want to upgrade my P3

[17:44:21] <myob> zeta dows, works fine - except its parellel ATA support is bloody useless

[17:44:21] <myob> [Beta]: I know... Axel MkII

[17:44:22] <Anxiety> I got Ubuntu installed last time, I'm going to see about using that to develop for Haiku

[17:44:22] <Anxiety> all my previous efforts to get some BeOS on my system werent successful

[17:44:22] <Ingenu> nutela not booting normaly

[17:44:34] <nutela> I tried doing some music on windows with sonar but I just don't like it

[17:44:52] <nutela> aw zeta is crap, I don't want to use it

[17:45:03] <Ingenu> I suppose there's no drivers for most my hardware anyway ^^

[17:45:10] <Ingenu> XFi, GF8...

[17:45:13] <myob> well, I don't want to use it either - I can barely rip a CD and never a DVD reliably on it

[17:45:46] <nutela> is it possible to use the old R5 or Dan0 drivers on zeta?

[17:45:53] <nutela> I tried but zeta wouldn't boot

[17:45:58] <myob> not without killing my SATA...

[17:46:16] <nutela> do those new core duo mobo's support old IDE?

[17:46:30] <myob> my "possibly the fastest BeOS box in the world" doesn't like anything older than 2006 or so - zeta 1.2.1 is teh oldest it'll boot

[17:46:56] <Anxiety> Mine only sort of does

[17:47:04] <nutela> what are it's specs?

[17:47:11] <Anxiety> it has a seperate chip to handle it, its not 'part' of the motherboard anymore

[17:47:35] <nutela> I thought sequitur had bugs on zeta it didn't have in R5

[17:47:36] <myob> 2x2.8, 2GB RAM, 6800 Ultra, 2x250GB SATA

[17:47:41] <DeadYak> as in IDE on a card?

[17:47:48] <myob> its one of the fastest supported boxes anyway ;)_

[17:47:55] <nutela> What DeadYak?

[17:47:57] <myob> as in 2D, overlay, etc not VESA

[17:48:03] <Anxiety> no, like a seperate bios chip or something... kind of like most built-in raid controllers

[17:48:07] <DeadYak> never mind, misunderstood

[17:48:08] <DeadYak> ah.

[17:48:15] <nutela> I mean what does it offer only SATA those new core duo mobo's?

[17:48:23] <DeadYak> myob: my athlon64 3200+ / X800 ran zeta pretty nicely too

[17:48:27] *** MikeW has joined #haiku

[17:48:30] <Anxiety> no, Ive got IDE still

[17:48:39] <Anxiety> 4 ide, 8 sata

[17:48:56] <myob> DeadYak:  ran not runs though isn't it :P

[17:48:57] <Anxiety> actually, let me double check that

[17:49:27] *** slaine_ has quit IRC

[17:49:31] <nutela> i am waiting on Euan too compile the radeon driver on R5 so I hope I can finally use DVI and widescreen resolutions for my flatpanel

[17:49:36] <myob> DeadYak: CPU is probably somewhat equal, the graphics I think you've got me beat - think...

[17:49:57] *** jiuda_D`arkness has joined #haiku

[17:50:11] <Anxiety> http://www.newegg.com/Product/Product.asp?Item=N82E16813131030

[17:50:44] <nutela> you have R5 on that Anxiety>

[17:50:45] <nutela> ?

[17:50:58] <Anxiety> no, unfortunately

[17:51:07] <Anxiety> then again, I havent tried lately

[17:51:08] <nutela> ok so that doesn't work

[17:51:11] <jiuda_D`arkness> hi all

[17:51:21] <myob> R5 runs on s.f.a these days

[17:51:28] <Anxiety> I was going to give it another shot when BeMax or what ever its called got out of beta for the newest version

[17:51:30] <nutela> I must ask Eric Hackborn to release source to sequitur

[17:51:39] <nutela> sfa?

[17:51:53] <myob> I must partition my laptop and install haiku on it, actually - it sees the IDE, R5 doesn't, zeta doesn't

[17:51:57] <nutela> and find those pesky little buggies

[17:51:58] <myob> sweet fuck all, nutela

[17:52:12] <jiuda_D`arkness> emitrax__ are you ok?

[17:52:22] *** kokito has joined #haiku

[17:52:23] <emitrax__> jiuda_D`arkness: hi

[17:52:36] <emitrax__> jiuda_D`arkness: I left you a pm on itbug

[17:52:44] <emitrax__> jiuda_D`arkness: did you get it ?

[17:52:45] <nutela> myob ill look it up on wikipedia

[17:53:15] <jiuda_D`arkness> yes

[17:53:58] <myob> doubt wikipedia has it, vaguely obscure acronym that it is

[17:54:39] <nutela> harhar : http://www.thepatri0t.net/2006/12/26/do-not-do-not-buy-asus-p5b-deluxe/

[17:55:24] <nutela> http://en.wikipedia.org/wiki/Sweet_fuck_all

[17:55:28] <Anxiety> err 400

[17:55:30] <nutela> wtf

[17:55:39] <nutela> 400?

[17:55:42] <nutela> what's that

[17:56:03] <Anxiety> invalid request

[17:56:35] <jiuda_D`arkness> emitrax__ we speak in private?

[17:56:42] <MikeW> not sweet fanny adams?

[17:56:51] *** tqh has joined #haiku

[17:56:59] <nutela> 4xx Client Error

[17:56:59] <nutela> The request contains bad syntax or cannot be fulfilled.

[17:56:59] <nutela> 400 Bad Request

[17:57:02] *** MikeW has quit IRC

[17:57:21] <jiuda_D`arkness> emitrax__ italian is better XD

[17:57:39] <nutela> so anyone running R5 on Core (2)?

[17:57:49] <CIA-17> darkwyrm * r20665 /haiku/trunk/src/add-ons/kernel/drivers/midi/usb_midi/Jamfile: Enable build of usb_midi on R5 using same Jamfile techniques as usb_hid

[17:57:50] <nutela> tqh you have a Echo Mia?

[17:58:22] *** JamesB192 has quit IRC

[17:58:26] <tqh> nutela no, at least not that I'm aware of :)

[17:58:38] <nutela> oh

[17:58:43] <nutela> err :-P

[17:58:44] <myob> right, hometime... yay

[17:58:47] * DeadYak pets tqh

[17:59:02] <nutela> myob : who has Echo Mia MIDI you know?

[17:59:07] <myob> nobody

[17:59:08] *** myob has quit IRC

[17:59:16] <nutela> I think somebody from bezilla had

[17:59:25] <nutela> I'll look up haiku cia

[17:59:30] <emitrax__> jiuda_D`arkness: query

[17:59:48] <emitrax__> jiuda_D`arkness: ma li leggi i miei messaggi ?

[17:59:58] <tqh> can't think of anyone that has mentioned it.

[18:00:10] <jiuda_D`arkness> emitrax__ sto provando a chiamarti ma non mi risp

[18:00:29] <emitrax__> jiuda_D`arkness: su jabber?

[18:00:53] <emitrax__> jiuda_D`arkness: parliamo in privato che non mi sembra il caso di parlare in italiano in canale

[18:00:54] <nutela> I think on bedrivers

[18:01:17] <jiuda_D`arkness> yes

[18:01:18] <kokito> time for a #haiku-it channel? ;)

[18:01:22] <nutela> ah tigerdog has

[18:01:37] <emitrax__> kokito: if you wanna join us :)

[18:01:40] <DeadYak> nutela: that card's supported for playback only

[18:01:43] <DeadYak> kokito!

[18:01:46] <emitrax__> kokito: sorry about that

[18:01:47] <jiuda_D`arkness> wait that i trying to remember my gmail password XD

[18:01:58] <netster403> *yahn*

[18:01:59] <netster403> : idle     : 571 hours 9 mins 55 secs (signon: Mon Mar 19 14:51:55 2007)

[18:01:59] <emitrax__> omg :D

[18:02:01] <kokito> not a problem emitrax__ ; just an idea.

[18:02:05] * tqh thinks Swedish is easier to understand :)

[18:02:21] *** mazon is now known as Mazon

[18:02:24] <DeadYak> "Please note: as of this writing, echo cards 24-bit cards are incompatible with the ASUS A8VDeluxe motherboard at least through BIOS rev. 1013. Not a BeOS issue, per se, but important to know."

[18:02:26] <nutela> XeD is here>

[18:02:27] <nutela> ?

[18:02:28] <kokito> emitrax__, io no parlo italiano

[18:02:36] <kokito> DeadYak!

[18:02:39] <emitrax__> kokito: I know! :D

[18:02:59] *** netster403 has quit IRC

[18:03:06] <DeadYak> Korli wrote that driver if I'm not mistaken

[18:03:12] <nutela> yes DeadYak

[18:03:31] <nutela> only playback? where is that written?

[18:03:44] <DeadYak> nutela: the hardware compatibility matrix comments

[18:03:51] <nutela> oh no I should have bought the Gina 3g :-(

[18:04:08] <nutela> isn't that old DeadYak?

[18:04:10] <DeadYak> nutela: that might've changed, looking if the driver src says anything

[18:04:22] <nutela> DeadYak: is this private talking?

[18:04:46] <DeadYak> nope

[18:04:54] <DeadYak> google for the BeOS Hardware Compatibility Matrix

[18:04:57] <nutela> isn;t that old info DeadYak?

[18:05:13] <jiuda_D`arkness> emitrax__ on the server irc.azzurra.org i have a chan called #beos

[18:05:34] <DeadYak> nutela: that's why I'm checking the driver

[18:05:43] <nutela> yeah well check this: http://cia.vc/stats/author/korli

[18:07:18] * DeadYak nods

[18:07:20] <nutela> Firmware to support new revision 4 Mia MIDI hardware

[18:07:31] * nutela test

[18:07:55] * nutela thinks beshare "Ah..."

[18:07:56] <DeadYak> cool

[18:07:58] <tqh> ah I suspect tigerdog has a lot of cool stuff :)

[18:08:23] <nutela> 24 bit 96kHz!

[18:08:24] *** tqh has quit IRC

[18:08:35] *** tqh has joined #haiku

[18:08:38] <kokito> tqh!

[18:08:47] *** TTRanger has joined #Haiku

[18:09:00] <DeadYak> TTRanger: do you use any Echo cards by any chance?

[18:09:15] <TTRanger> DeadYak Not presently, no, but I'm open to the idea of it.

[18:09:27] <DeadYak> ah

[18:09:28] <DeadYak> just wondered :)

[18:09:32] <TTRanger> If we can get a driver that will support their current cards...

[18:09:32] <DeadYak> what audio hardware are you using for TT then?

[18:09:34] <nutela> I'll be telling you soon

[18:09:47] <tqh> kokito Hi :)

[18:09:48] <TTRanger> nutela What will you be telling us?  :-)

[18:09:48] <nutela> yeah TTRanger

[18:10:02] *** slaine__ has quit IRC

[18:10:05] <nutela> how the Echo Mia Midi works with R5 and Phos

[18:10:13] *** Lelldorin1 has joined #haiku

[18:10:22] <TTRanger> nutela Are you testing it?

[18:10:57] *** miqlas has joined #haiku

[18:10:58] <miqlas> re!

[18:11:15] <nutela> : TTRanger not yet

[18:11:31] <nutela> Whereis Atrus ;-)?

[18:11:34] <TTRanger> Does anyone know how the media kit is coming along in Haiku?

[18:11:45] <nutela> how does one do private messaging?

[18:12:00] <nutela> ah

[18:13:32] <TTRanger> "/msg nameofperson"

[18:13:35] <TTRanger> minus the quotes

[18:13:47] <Ingenu> heading home now

[18:14:06] *** Ingenu has quit IRC

[18:15:01] <nutela> Thanks TTRanger

[18:15:10] <TTRanger> np

[18:15:19] <nutela> I forgot how it was so Beshare copied IRC rules .. nice

[18:15:55] <nutela> too bad zeta never left beta :-P

[18:16:08] <DeadYak> it does in some ways, yes, in others you can get very confused

[18:16:27] *** jevin has joined #haiku

[18:16:27] <TTRanger> nutela Actually from 1.2 on, it's a pretty darned nice OS.

[18:16:37] <[Beta]> I wish highlighting was context based.

[18:16:40] <TTRanger> Nothing very Beta about how it acts, behaves.

[18:16:50] <DeadYak> [Beta]: context based how?

[18:17:25] <[Beta]> so it knew if people were talking to me, or about unfinished/modern web apps

[18:17:38] *** GreyGhost has joined #haiku

[18:17:46] <TTRanger> wb Grackle

[18:17:47] <TTRanger> oops

[18:17:48] <TTRanger> hehe

[18:17:52] <TTRanger> wb GreyGhost

[18:18:00] <DeadYak> [Beta]: your syntax highlighter doesn't include the [] in your nick?

[18:18:03] <GreyGhost> TTRanger ,thanks you :)

[18:18:14] <TTRanger> Does anyone know how the media kit is coming along in Haiku?

[18:18:18] <[Beta]> heh, and maybe if autocomplete assumed the last to on-join was the one you wanted..

[18:18:36] <[Beta]> DeadYak, i'm only [Beta] on freenode & efnet, the rest have Beta

[18:18:37] <GreyGhost> TTRanger , no idea.. i'm new here ;)

[18:18:48] <DeadYak> [Beta]: oh.

[18:23:35] <CIA-17> axeld * r20666 /haiku/trunk/src/apps/networkstatus/NetworkStatusView.cpp:

[18:23:35] <CIA-17> Fixed two bugs:

[18:23:35] <CIA-17> * FrameReceived() updated the bitmaps, but did not invalidate the view, causing

[18:23:35] <CIA-17>  B_FULL_UPDATE_ON_RESIZE to show an outdated icon.

[18:23:35] <CIA-17> * fInDeskbar was not initialized in case the archive constructor was used.

[18:24:49] *** Lelldorin1 has quit IRC

[18:29:30] *** Jin has joined #Haiku

[18:32:47] *** Fanskapet has quit IRC

[18:35:49] *** tombhadAC has joined #haiku

[18:35:55] *** gotaku has joined #haiku

[18:39:26] <nutela> ok i'm going later

[18:40:28] *** jiuda_D`arkness has quit IRC

[18:46:18] *** mmadia has joined #haiku

[18:50:23] *** tqh has quit IRC

[18:50:45] *** Jin has quit IRC

[18:50:48] *** kr1stof has joined #haiku

[18:56:09] *** Pulko_Mandy has joined #haiku

[18:56:26] *** tqh2 has joined #haiku

[19:04:51] *** Euver has joined #haiku

[19:08:33] *** tqh2 has quit IRC

[19:10:35] *** Ingenu has joined #Haiku

[19:11:51] *** PulkoMandy has quit IRC

[19:12:38] *** mmadia has quit IRC

[19:13:17] *** mmu_man has joined #haiku

[19:13:17] *** ChanServ sets mode: +o mmu_man

[19:14:06] * DeadYak plops mmu_man

[19:15:12] <mmu_man> plop

[19:17:01] *** MikeW has joined #haiku

[19:18:15] *** emitrax__ has quit IRC

[19:19:24] *** tqh has joined #haiku

[19:20:10] *** mmadia has joined #haiku

[19:21:50] *** jevin has quit IRC

[19:24:03] *** Lelldorin1 has joined #haiku

[19:24:14] *** Lelldorin1 has quit IRC

[19:34:08] <[Beta]> is there any reason the svn mail script uses hugo's email rather than "%user% at BerliOS"...

[19:35:11] *** tqh has quit IRC

[19:35:30] *** _Lucretia has joined #haiku

[19:36:40] *** gotaku has quit IRC

[19:38:22] *** _Lucretia_ has quit IRC

[19:38:35] *** MrRagga has joined #haiku

[19:40:52] <kokito> hi mmadia

[19:40:59] <kokito> and mmu_man

[19:43:14] *** Witukind has joined #haiku

[19:43:26] <mmadia> ....  one of these days i should switch my nick to madiam  : P

[19:44:29] <DeadYak> or madman :P

[19:44:31] * DeadYak ducks

[19:44:50] *** GreyGhost has left #haiku

[19:46:20] <mmadia> mad i am : )

[19:46:45] * DeadYak hides the mad trout

[19:47:06] <mmu_man> re

[19:47:14] *** Witukind has quit IRC

[19:47:56] *** miqlas has quit IRC

[19:50:03] *** _konrad has joined #haiku

[19:55:31] *** Pulko_Mandy has quit IRC

[19:56:13] *** Pulko_Mandy has joined #haiku

[19:56:14] *** dr_evil has joined #haiku

[19:56:22] <dr_evil> hi axeld

[19:56:35] *** ChanServ sets mode: +o dr_evil

[19:58:13] <dr_evil> I updated my blog http://haikudev.blogspot.com/

[19:59:41] <_konrad> dr_evil nice

[19:59:44] <mmu_man> hmm :)

[20:00:13] <_konrad> Is it possible to see in which areas the SOC students will work on (tasks)?

[20:00:32] <mmu_man> yes

[20:00:37] <dr_evil> follow the link luke

[20:00:38] <_konrad> mmu_man link=

[20:00:40] <dr_evil> err, _konrad

[20:00:41] <_konrad> ?

[20:00:49] <mmu_man> http://code.google.com/soc/haiku/about.html

[20:00:50] <mmu_man> :)

[20:01:01] <_konrad> Thank you

[20:01:25] <dr_evil> or http://haiku-os.org/news/2007-04-11/its_official_eight_students_to_code_for_haiku_at_gsoc_2007

[20:02:38] <_konrad> A guy at work got intrested in Haiku. Well a few are actually.

[20:05:04] <_konrad> Well gotta go. Bye folks!

[20:05:05] *** _konrad has quit IRC

[20:07:29] *** rgb has joined #haiku

[20:11:13] <CIA-17> korli * r20667 /haiku/trunk/headers/private/storage/DiskDeviceVisitor.h: fixed typo

[20:13:01] <[Beta]> dr_evil, very good news! thank you.

[20:15:30] *** MYOB has joined #haiku

[20:17:32] <CIA-17> korli * r20668 /haiku/trunk/headers/private/net/NetBufferUtilities.h: gcc4 is picky with in-template symbols, they must be qualified

[20:19:09] *** SiCuTDeUx has joined #haiku

[20:19:10] <SiCuTDeUx> :D

[20:19:16] <SiCuTDeUx> hi!

[20:20:51] *** korli has joined #haiku

[20:20:51] *** ChanServ sets mode: +o korli

[20:21:06] <korli> evening

[20:22:11] *** Sapote has quit IRC

[20:22:26] <mmu_man> plop

[20:23:03] <korli> mmu_man: hmm caching :)

[20:23:16] <mmu_man> hmm yummy :))

[20:23:53] <MYOB> can Haiku be built from under Zeta or not?

[20:24:13] <mmu_man> works here

[20:24:43] <mmu_man> I'm using the latest gcc from Oliver in a custom devkit

[20:24:44] <MYOB> ditto

[20:24:45] <MYOB> need it for ffmpeg et al...

[20:24:50] <mmu_man> yep

[20:24:52] <SiCuTDeUx> where is Oliver Dorantes?

[20:24:59] <SiCuTDeUx> long time no see!

[20:25:04] <mmu_man> I even have my local .profile giev me a red warning when it's not enabled :)

[20:25:06] <MYOB> mentioning gcc.... VLC beyond 0.8.6 is a no-go for BeOS

[20:25:15] <mmu_man> what's now ?

[20:25:24] <mmu_man> btw, ffmpeg still has those PRI* macros out

[20:25:32] <mmu_man> but I'm tired to push that thing

[20:25:33] <SiCuTDeUx> MYOB: why?

[20:25:40] <MYOB> I tried compiling it and its close to 100% GCC2.x unfriendly

[20:25:43] <mmu_man> tired to make them understand this is a port *to* BeOS

[20:25:47] <mmu_man> not a port of BeOs to ffmpeg

[20:26:08] <MYOB> I last got the code to compile in December - well, most of it. The BeOS module was fucked due to header changes

[20:26:13] <MYOB> its got worse since

[20:26:43] <MYOB> we'll get 0.8.6b - which is likely to be the most stable VLC ever - but not 0.9x

[20:27:15] <MYOB> that and they appear to be moving to git :(

[20:27:24] <mmu_man> MYOB http://revolf.free.fr/z/haiku-r1-gcc2-x86.zip

[20:27:30] <mmu_man> MYOB http://revolf.free.fr/z/zeta-r1-gcc2-x86-oliver.zip

[20:27:35] <mmu_man> must install both

[20:27:42] <MYOB> mmu_man thanks,. already have all the gubbins set up - I think

[20:27:47] <mmu_man> but as yours works already no reason to use it

[20:28:03] <mmu_man> only got some "redefined" warnings sometimes

[20:28:06] <MYOB> I have an R5 enviroment as well - VLC release builds have been done on it for a while ;)

[20:28:15] <SiCuTDeUx> thats the worst part of using an old compiler...

[20:28:16] <SiCuTDeUx> sux

[20:28:18] <korli> MYOB: seems like the Firefox situation

[20:28:23] <mmu_man> http://revolf.free.fr/beos/beos-r5-gcc2-x86.zip :)

[20:28:25] <MYOB> korli far worse than, though

[20:28:27] *** stargater has joined #haiku

[20:28:35] <MYOB> threading is broken in VLC for BeOS

[20:28:39] <MYOB> networking gone for net_server

[20:28:40] <MYOB> etc etc etc etc

[20:28:50] <stargater> hi

[20:28:57] <SiCuTDeUx> omg...

[20:29:01] * SiCuTDeUx panics!

[20:29:04] <MYOB> plus I guarantee you only me and probably mmu_man could get 0.8.6b to build either

[20:29:28] <korli> axeld: here ?

[20:29:29] <MYOB> or moooooo or David Reid if they hadn't left us...

[20:29:37] <mmu_man> god don't ask me, took me a day to get vlc to build once

[20:30:09] <MYOB> mmu_man a day is good going, it takes me 8 hours to build a build enviroment from clean here

[20:30:12] <MYOB> on my TurboBoxen

[20:30:20] <DeadYak> ouch

[20:30:48] <MYOB> that and the room hits about 45c...

[20:31:01] <MYOB> effectively VLC is running in a linux emulator ;)

[20:31:04] *** MauriceK has joined #haiku

[20:31:05] <SiCuTDeUx> i thinks it's to use new gcc

[20:31:10] <SiCuTDeUx> *think

[20:31:17] <SiCuTDeUx> a newer version...

[20:31:21] <stargater> hi MauriceK

[20:31:28] <MYOB> I can't imagine how painful it'd be with gcc4

[20:31:47] <SiCuTDeUx> MYOB: why?

[20:31:50] <MYOB> as GCC2 is a way faster compiler (to build with)

[20:32:20] <korli> MYOB: why do you think we're still stuck with glibc 2.2.5 ? :)

[20:32:41] <MYOB> korli compiler age also? or something speed related?

[20:32:41] <MauriceK> hey stargater

[20:33:17] *** Lelldorin1 has joined #haiku

[20:33:26] <DeadYak> gcc4 is incredibly slow.

[20:33:33] <DeadYak> at compiling anyways

[20:33:42] <DeadYak> admittedly it's also far more correct but...

[20:33:45] <mmu_man> so

[20:33:46] <MauriceK> yes but compiling is not runtime ;)

[20:33:49] <korli> MYOB: gcc2 isn't supported officially for later versions

[20:33:50] <MYOB> DeadYak as I know all too well from my previous-but-one job...

[20:33:51] <mmu_man> TV by DSL is still broken...

[20:33:52] <mmu_man> 1 week

[20:33:53] <MYOB> korli ah

[20:34:00] <mmu_man> I'll be getting really angry

[20:34:01] <DeadYak> they still have yet to learn how to make readable template error messages :/

[20:34:03] <mmu_man> anyway

[20:34:15] <MYOB> now all I have to work with 'compiler' wise is VisualStudio 6....

[20:34:15] <mmu_man> let's watch SGA 101 again instead :)

[20:34:40] * MauriceK agrees do DeadYak's template statement

[20:35:12] <korli> mmu_man: you have a low rate DSL maybe ?

[20:35:21] <mmu_man> no

[20:35:34] <mmu_man> using multiposte cause I don't have a TV set

[20:35:41] <mmu_man> and *some* (few) channels actually work

[20:35:46] <mmu_man> for a while before locking up

[20:35:46] <MYOB> I regularly wake up thinking I'm so lucky I don't work for a DSL company anymore...

[20:35:57] <mmu_man> :)

[20:36:19] <MYOB> DSL is a fatally flawed technology

[20:36:22] *** Zaranthos has joined #Haiku

[20:36:29] <MYOB> its only a stop gap before every house gets cable or fibre

[20:36:30] <mmu_man> FTTH \o/

[20:36:36] <mmu_man> not here for years I guess /o\

[20:36:37] <CIA-17> korli * r20669 /haiku/trunk/headers/posix/stddef.h: use __builtin_offsetof for gcc4 and superior, it avoids a gcc4 broken build with template use of offsetof()

[20:36:46] <MYOB> I'll got for cable consdering ~80% of houses in Ireland have it...

[20:36:56] <mmu_man> we got cable in my town

[20:37:02] <mmu_man> I actually used it long ago

[20:37:10] <mmu_man> but the company got bought countless times

[20:37:15] <mmu_man> (used to be part of the historical operator)

[20:37:16] <MYOB> same here

[20:37:24] <MYOB> Dunnes->CMI->Chorus->UPC

[20:37:26] <mmu_man> and they switch upload quotas on and off every 6 months or so

[20:37:28] <mmu_man> :(-

[20:37:50] <MYOB> some people went Phoenix->RTE->Telecom->CMI->Chorus->UPC....

[20:38:15] <SiCuTDeUx> Out of Order

[20:38:15] <SiCuTDeUx> The MSN network has issues. There're some errors, please connect again later. Thank you

[20:38:25] <SiCuTDeUx> omg!

[20:38:40] <dr_evil> same here SiCuTDeUx

[20:39:16] <mmu_man> same with Miranda here

[20:39:26] <SiCuTDeUx> ebuddy.como here

[20:39:41] <SiCuTDeUx> and msn messenger live on windows too

[20:40:03] <Pulko_Mandy> works for me now :) (gaim)

[20:40:45] *** procton has joined #haiku

[20:40:54] <dr_evil> MSN Messenger 7.0 (Build 7.0.0816)

[20:41:33] <dr_evil> works now, too

[20:41:42] <dr_evil> but everyone else seems offline

[20:42:17] *** Begasus has joined #haiku

[20:42:23] <MYOB> nearly all my contacts are online...

[20:42:40] <dr_evil> well, going to meet some of then IRL in abot two hours anyway

[20:47:43] <MikeW> passport was down today dr_evil

[20:48:05] *** Lelldorin1 has joined #haiku

[20:48:08] <MYOB> MikeW musta happened while I was in Liffey Valley

[20:48:08] <MYOB> was up as I left work and up when i got home...

[20:48:29] <MYOB> Communicator goes down quicker and more often than DaaT's panties so I can't say its been stable, though...

[20:48:32] <MikeW> well, passport *signin* has been down since about 5:45 GMT according to the microsoft employee that said it

[20:50:05] *** jpd has joined #haiku

[20:50:08] <MYOB> just signed in (again)

[20:50:13] <MYOB> you aren't online, everyone else is...

[20:50:42] *** Begasus has quit IRC

[20:51:01] *** Begasus has joined #haiku

[20:51:15] *** Begasus has quit IRC

[20:51:16] <MikeW> MYOB: I can't sign in yet. The service isn't fully up

[20:51:32] <MYOB> nah, it just doesn't like you

[20:51:40] <MikeW> I see some people from the NL can't sign in either, but I'd guess it depends on which server one is on, rather than geographic location

[20:51:51] *** Begasus has joined #haiku

[20:52:36] <MikeW> there we go

[20:52:46] <MikeW> another spot of vBulletin 2 hacking

[20:52:50] <kr1stof> I cannot sign in here since this morning

[20:53:17] <MikeW> I guess I can't be quite content with software. I just have to mod everything to the gills and make it impossible for myself to upgrade :p

[20:55:55] *** rgb has quit IRC

[20:58:35] *** Lelldorin1 has quit IRC

[20:59:50] *** jevin has joined #haiku

[21:05:49] *** Begasus has quit IRC

[21:07:21] <dr_evil> ping axeld

[21:07:39] *** dsg has joined #haiku

[21:09:14] *** MikeW has quit IRC

[21:09:43] *** MikeW has joined #haiku

[21:12:04] *** DaaT has joined #haiku

[21:12:08] <SiCuTDeUx> kokito: i have a new dog! BTW

[21:12:22] <SiCuTDeUx> kokito: want to guess his name xD?

[21:12:44] * kokito wonders

[21:13:29] * MikeW pokes MYOB

[21:13:56] <SiCuTDeUx> kokito: kokito it is!

[21:13:57] <SiCuTDeUx> xD

[21:14:10] <kokito> LOL

[21:14:13] <SiCuTDeUx> my mother couldn't find a better one!

[21:14:21] <dsg> Can I attach GDB to the haiku kernel easily?

[21:14:24] <SiCuTDeUx> xD so...

[21:14:38] *** MikeW has quit IRC

[21:15:11] <dsg> haiku gets a general protection fault when enabling interrupts under VirtualBox it seems.

[21:16:31] <korli> dsg I don't think so

[21:17:44] *** jevin has quit IRC

[21:18:34] <DeadYak> dsg: Haiku has its own kernel debugger

[21:18:44] <DeadYak> dsg: sounds like you want to be looking at serial debug

[21:19:20] <dsg> DeadYak: Yeah, the kernel dumped me into a debugger, but can I start the debugger immediately and set breakpoints?

[21:19:59] <dsg> And, are there any docs on setting up serial debug?

[21:20:12] *** jevin has joined #haiku

[21:20:46] *** rgb has joined #haiku

[21:20:47] <CIA-17> korli * r20670 /haiku/trunk/src/kits/game/ (PushGameSound.cpp SimpleGameSound.cpp):

[21:20:47] <CIA-17> improve codestyle

[21:20:47] <CIA-17> fix bad init in BPushGameSound

[21:20:56] <DeadYak> looking...I don't think you can generally do what you're describing with a kernel though

[21:20:59] <DeadYak> in terms of breakpoints

[21:21:40] <DeadYak> breakpoints generally rely on certain functionality that more or less requires a running kernel :)

[21:21:53] <dsg> Well, I've done it with a simple kernel (my own small kernel), but I'm not sure where to start with haiku :)

[21:22:06] <dsg> I used bochs then

[21:22:44] *** infamy has joined #haiku

[21:23:13] <infamy> is there a reason why in the buildtools there are test suites for 8bit micros?!? (68hc11)

[21:23:20] <DeadYak> gah where is it...I swear there used to be a doc on setting it up

[21:24:25] <Grackle> infamy: It's probably just part of the standard build tools setup, and not something that we use for Haiku.

[21:25:03] <infamy> ah okay... is there any point in someone removing all of that "fluff" from the buildtools?

[21:25:16] *** korli has quit IRC

[21:25:29] <Grackle> I dunno much about them. That's just a guess, anyway.

[21:26:03] <infamy> well it looks like compiler testsuites

[21:26:25] *** Euver has quit IRC

[21:26:27] *** SiCuTDeUx has quit IRC

[21:26:38] <DeadYak> dsg: hit space immediately at startup

[21:26:41] <infamy> i dont see them getting used... but bah why waste time removing them i guess.... i just saw them and went hmm that is odd.. why are those there..

[21:26:45] <DeadYak> dsg: you should be able to enable serial there

[21:27:00] <DeadYak> infamy: you never know what part of the configure scripts, etc. might need those

[21:27:14] <dsg> DeadYak: Ok, thanks.

[21:28:14] *** korli has joined #haiku

[21:28:14] *** ChanServ sets mode: +o korli

[21:28:21] *** kr1stof_ has joined #haiku

[21:29:43] <dr_evil> you might want to create a haiku->8bit cross compiler later

[21:31:09] <mmu_man> Haiku for C64 ? :)

[21:31:31] <dr_evil> haiku vor atmel development

[21:31:38] <mmu_man> ah :))

[21:31:57] * mmu_man was picturing Deskbar in PETSCII

[21:32:13] <DeadYak> atmel?

[21:32:23] <DeadYak> mmu_man: let's rewrite haiku in c64 basic :)

[21:32:24] <Grackle> Atmel AVR microcontrollers.

[21:32:30] <DeadYak> ahh.

[21:32:34] <Grackle> Kinda like PIC devices.

[21:32:40] * DeadYak nods

[21:32:52] <mmu_man> DeadYak tried Contiki ?

[21:33:03] <DeadYak> mmu_man: not yet

[21:33:23] <Grackle> I wanted to try Contiki on something, but never got around to it. It looks neat.

[21:33:41] <infamy> uip (same author as contiki) is really nice

[21:33:58] *** Kernel86 has quit IRC

[21:34:03] *** Kernel86 has joined #Haiku

[21:34:07] <infamy> works well for a ultrasmall stack just one real limitation... one packet in flight

[21:34:10] <dr_evil> i need to leave for today, bye

[21:34:11] <MYOB> infamy I can't help but think of carry on movies when I see your name here...

[21:34:19] * DaaT pets mmu_man

[21:34:30] <infamy> MYOB: carry on movies?!?!

[21:34:41] <MYOB> infamy never heard of them?

[21:34:48] <infamy> MYOB: nope

[21:34:51] <MYOB> or indeed, the, erm, 'infamous' quote?

[21:35:11] *** kr1stof has quit IRC

[21:35:12] <infamy> nope...

[21:35:17] *** dr_evil has quit IRC

[21:35:21] * infamy is now curious and needs to investigate

[21:35:21] <MYOB> oh well, they're a series of camp british movies; one famous quote being Kenneth Williams screaming "Infamy! Infamy! They've all got it in for me!"....

[21:35:33] <infamy> LOL

[21:35:37] <[Beta]> heh

[21:35:45] <infamy> oh okay i need to find that and use that as my boot up sound

[21:35:59] <MYOB> [Beta] knows what I mean

[21:36:00] * mmu_man pets DaaT

[21:36:09] *** jpd has quit IRC

[21:36:30] <MYOB> when you consider this series's other "big" quote is "ooh matron!" and most famous moment is Barbara Windsor's bra flying off....

[21:37:15] <mmu_man> oh didn't even remember taking those http://revolf.free.fr/beos/BeGeistert-16/bgcam/ :)

[21:39:22] <MYOB> korli looks like a guy I used to work with...

[21:40:07] *** Sarcas has joined #haiku

[21:40:24] <stargater> mmu_man, hehe http://revolf.free.fr/beos/BeGeistert-16/bgcam/frame_0018.jpg  you have took me

[21:42:12] <stargater> http://revolf.free.fr/beos/BeGeistert-16/bgcam/frame_0047.jpg jan and me for yab

[21:42:57] <MYOB> still got my haircut then? :P

[21:43:28] <stargater> no my haircut

[21:43:44] <stargater> ore yours

[21:43:47] <stargater> :-)

[21:44:35] <MYOB> since 1998 ;)

[21:45:40] *** M199 has quit IRC

[21:45:46] *** Euver has joined #haiku

[21:48:48] *** Master199 has joined #haiku

[21:51:56] *** Euver has quit IRC

[21:51:58] *** Master199 has quit IRC

[21:53:52] *** judgen has joined #haiku

[21:54:17] <judgen> lol 80 people here now? this place sure has grown

[21:55:17] <judgen> btw, is tcp/ip working now?

[21:55:56] <_hugo> judgen: yes

[21:56:00] <judgen> fully?

[21:56:04] <_hugo> judgen: define fully

[21:56:30] <judgen> send and recieves packages with less crases than net_server in r4 =P

[21:56:40] <_hugo> judgen: definitely

[21:56:46] <judgen> nice

[21:57:05] *** emitrax has joined #haiku

[21:57:27] <judgen> i installed xubuntu yesterday... xgl/beryl sure is nice but a good designed 2d enviroment is alot more productive

[21:58:03] <DeadYak> I have yet to see compiz/beryl do anything I'd call productively useful as opposed to pointless eye candy

[21:58:07] <judgen> hehe r4 net_server sure was crappy

[21:58:26] <judgen> r4.5 was miles ahead of r4 on my old sytems

[21:58:33] <_hugo> judgen: well, with Haiku you have the possibility to contribute yourself

[21:58:45] <_hugo> so stuff gets less crappy over time

[21:58:55] *** BuildFactory has quit IRC

[21:59:28] <judgen> yeah i know, ive been following this project since openbeos was formed. quite an fast progress i must say

[22:00:05] <Pulko_Mandy> _hugo: does not always works like this... look at linux... :)

[22:00:09] *** BuildFactory has joined #haiku

[22:00:20] <judgen> heard the haiku project got more people from gsoc than expected

[22:00:31] <DeadYak> 8 to be exact

[22:00:33] <judgen> Pulko_Mandy: there is nothing wrong with the linux kernel

[22:00:38] <[Beta]> Can we talk about how good Haiku is, and not how bad we believe Linux to be :P

[22:01:05] <emitrax> :)

[22:01:44] <judgen> linux, bsd, haiku, windows and all the others serves a purpose. So i guess there is no point in bashing one or another about it

[22:02:09] <judgen> the question is rather "whos" purpose =)

[22:02:34] <MYOB> aaaaahhhh its still checking out...

[22:02:44] <judgen> do we have an weather client for haiku?

[22:02:52] <judgen> pineapple servers are down

[22:02:58] <MYOB> thats like 35,000 dollars from google if theey all complete, isn't it?

[22:03:03] <MYOB> (in total)

[22:03:34] *** MrRagga has quit IRC

[22:03:41] <judgen> nice

[22:03:41] *** jpd has joined #haiku

[22:03:42] <judgen> i like google

[22:04:08] *** BuildFactory has quit IRC

[22:04:21] <judgen> is the interns familiar with BeOS/haiku?

[22:04:21] *** BuildFactory has joined #haiku

[22:04:38] <MYOB> well _hugo appears to be + I've seen at least one of the names before

[22:04:47] <Grackle> Heh DeadYak, re: compiz/beryl, I agree. The two forces are joining (have joined?), and I imagine their collective output will be much more sane and usable.

[22:05:07] <[Beta]> For those that aren't, they've got a good 6 weeks to learn.

[22:05:18] <Grackle> 3D accelerated desktops can be very pretty and usable (take mac osx for example, or maybe vista too).

[22:05:53] <MYOB> neither is anywhere near as "3D" as the Linux stuff

[22:05:55] <Grackle> Composited desktops, rather.

[22:05:57] <Grackle> Yeah.

[22:06:02] <[Beta]> MYOB, the $4000 for Haiku will be nice.

[22:06:03] <judgen> Grackle: i dont agree that the vista desktop 3d features brings any advantage in usability or productivity increase

[22:06:09] <MYOB> and I find OS X neither pretty nor usable,

[22:06:18] <Grackle> The 3D stuff I don't really care about, the 2D compositing is nice though.

[22:06:21] <MYOB> its counterintuitive, it gets in your way, and it looks like a designjers wet dream...

[22:06:46] <Grackle> Oh, I like OS X a lot. I can get a lot of work done there. There is stuff I would like to change, though.

[22:06:51] <judgen> hot corners (no 3d needed) would be nice for haiku though

[22:07:34] <judgen> Grackle:  in osx i would like to make the windows minimize when i double click the titlebar just as i can with every other os i use =)

[22:08:51] <Grackle> Ah. In OSX, I do things like this: I open apps from the dock, then navigate between them with the alt-tab app switcher, and make extensive use of the cmd-h (hide) function.

[22:09:24] <Grackle> I don't use macs very often anymore, but I used to do everything on one.

[22:09:28] <MYOB> oddly enough, we have a ctrl-tab function too...

[22:09:56] <Grackle> I don't think it's as elegant, but it's open source, so that can all be changed.

[22:10:20] <Grackle> And it really needs to be alt-tab. Ctrl-tab is just awkward.

[22:10:25] <judgen> Grackle: i think its just my way of using the mouse that interfeers with me liking the "proper ways" i want buttons to the left and only two of them, close and mazimize, and when i double click titlebar it should minimize. I just love it =)

[22:10:27] <MYOB> how is switching through apps elegant or not?

[22:10:34] <MYOB> you don't see anything unless you hold down ctrl...

[22:10:40] <Grackle> Yup judgen, whatever works for you

[22:10:46] <MYOB> and you can change it back to the Windows way by changing in the Menu preferences

[22:11:05] <MYOB> I've seen the Twitcher window about twice, never intentionally...

[22:11:16] <Grackle> Oh, does windows way = shows up immediately?

[22:11:25] <MYOB> no, I meant Alt key

[22:11:30] <Grackle> oh

[22:11:39] <judgen> Grackle: the BeOS Task switcher is very elegant. alt+tab and arrow keys chooses apps, sub-windows and even subdialogues within the apps

[22:11:41] <MYOB> I've no idea what way Windows tabs through windows

[22:11:54] <Ingenu> what's called windows key in BeOS already ?

[22:12:01] <MYOB> and its hold down tab not ctrl, my mistake

[22:12:02] <MYOB> Ingenu Option

[22:12:06] <Ingenu> I remember there's command

[22:12:06] <MYOB> same as on a Mac

[22:12:10] <Ingenu> so ALT would be COMMAND ?

[22:12:16] <MYOB> no, Alt is Alt

[22:12:17] <Grackle> I like it to show up immediately. Even if it's only there for a fraction of a second, it gives me an idea of how the stack is arranged, so I can alt-tabtabtab through the apps easily.

[22:12:28] <Ingenu> so control is command ?

[22:12:29] <MYOB> x86 systems only have 3 modifies

[22:12:35] <MYOB> no, control is control...

[22:12:36] <judgen> Grackle: tried the BeOS task switcher?

[22:12:37] <Ingenu> I remember there's a command somewhere

[22:12:51] <Grackle> Nope judgen.

[22:13:05] <judgen> Grackle: try it, you wont be disappointed in the way it works.

[22:13:06] <MYOB> Grackle well, the thing is 95% of people come from Windows where the task list doesn't show unless you hold down tab for quite some time

[22:13:16] <MYOB> and another 1% of us come from the BeOS world where its the exact same

[22:13:34] <MYOB> hitting ctrl-tab changes to the 'next' app in the stack, just like on Eveyrthing Thats Not A Mac

[22:13:36] <DeadYak> windows key is option I believe

[22:13:39] <Grackle> My XP system shows the switcher immediately..

[22:13:57] <MYOB> must have modified it, because mine don't

[22:14:03] <judgen> Grackle: mine either

[22:14:20] <MYOB> and about the only "modification" I have is making Windows look like Windows 2000 and not a crayola hell

[22:14:28] <Grackle> Weird. I don't recall ever using windows and finding that the switcher doesn't come up immediately.

[22:14:32] <judgen> MYOB: maybe microsoft sold us the crippled down european version? =P

[22:14:38] <Grackle> Let me go give it a try, maybe I just don't notice.

[22:14:39] <Grackle> brb

[22:15:13] <MYOB> or him the wonky US version :P

[22:15:17] <judgen> the best windows ever made must be NT4

[22:15:20] <Ingenu> alt tab always brought up the panel immediatly on NT

[22:15:22] <judgen> MYOB:  hehe

[22:15:38] <MYOB> judgen 2000 surely

[22:15:46] <MYOB> NT4 is fustrating for stuff like not having a devices listing

[22:15:57] <MYOB> 2000 is equally solid once you get to SP4

[22:16:01] <judgen> MYOB: if they had just added usb2.0 and directx support to nt4 i would still be using it

[22:16:02] <Grackle> Ah, control-tab on a windows machine (XP at least) switches immediately. alt-tab brings up the switcher until you let go.

[22:16:16] <MYOB> judgen well, it hasn't got USB-anything...

[22:16:30] <MYOB> not even BeOS 4.5 keyboard/mouse USB....

[22:16:36] <Grackle> I like having the switcher appear immediately and disappear after I let go.

[22:16:44] <judgen> MYOB: there is usb 1.1 and directx6/7 for nt4 but its not officially supported

[22:17:07] <Ingenu> control tab switch window in multi window app (don't remember the name)

[22:17:13] <MYOB> judgen its closer to 'illegally stolen from 2K betas' than 'unsupported'

[22:17:24] <judgen> MYOB: R5 had quite good usb mouse support for its time

[22:17:35] <MYOB> if 2K had decent wifi support I'd have kept using it instead of XP...

[22:17:36] <judgen> MYOB: Bah same thing =P hehe

[22:17:38] <Ingenu> herm ctrl tab does nothing at all in Vista

[22:17:49] <Ingenu> (besides switching between an app windows)

[22:17:52] <MYOB> judgen I know, but R4.5 only supported kdb/mice, which was two things more than NT4 ;)

[22:17:54] <Grackle> Ingenu, oh, I didn't notice that they were windows of the same program. Makes sense though.

[22:18:06] <judgen> MYOB: i have used belkin and d-link devices for wireless in 2k for many years, never any problems

[22:18:27] <MYOB> judgen yes, but its exceptionally clunky

[22:18:35] <MYOB> and new devices = no 2k drivers or applets

[22:18:36] <Ingenu> I remember the days I was very much into the BeOS, was buying my hardware given drivers

[22:18:49] <Grackle> But I find that having to wait for it to come up is kind of frustrating. I end up just switching through quickly until I find what I want.

[22:18:55] <Ingenu> now I just buy whatever I like and it works... (ok using windows :p)

[22:19:04] <judgen> MYOB: all drivers except wia stuff can be made to work in 2k by just editing the ident string in the xp driver

[22:19:23] <judgen> wia stuff dont work since 2k uses twain instead of wia

[22:19:42] <MYOB> Grackle well, this is a case of the beos way of settings - best possible default for most people is not having it come up staight away, and to have it editable is waaaaaay too KDEish...

[22:20:00] *** rgb has quit IRC

[22:20:14] <Ingenu> in BeOS you need wait for the panel to show up ? and w/ Haiku too ?

[22:20:32] <MYOB> Ingenu yes

[22:20:34] <judgen> and now for the most important question of the night for me "do we have slidy tabs now?" haha

[22:20:35] <MYOB> its part of Deskbar

[22:20:40] <MYOB> judgen yes...

[22:20:44] <judgen> yay

[22:20:51] <MYOB> mmu_man did them...

[22:21:05] <judgen> i smell of old alcohol and menthol cigarettes

[22:21:08] <Grackle> Why is it the best possible default? What disadvantage is there to having it come up and disappear quickly? We're using fast computers, it's not like the redraw hurts anything.

[22:21:13] <judgen> i should take a shower

[22:21:30] <Grackle> haha yay slidey tabs

[22:21:31] <procton> Grackle, I don't.

[22:21:36] <MYOB> Grackle because you're the first person I've heard complaining in ten years

[22:21:37] <Ingenu> didn't try haiku on my C2D

[22:21:48] <MYOB> hence its the best possible default for virtually everyone else

[22:22:02] <judgen> maybe not for my dead grandma =P

[22:22:14] <MYOB> and because having it appear and dissapear when all you're doing is flicking between two apps = needless distraction

[22:22:16] <kokito> judgen, BWeather @ http://www.bebits.com/app/1749 . don't know if it works though...

[22:22:21] <MYOB> something OS X is damn good at actually....

[22:22:25] <mmu_man> I did what ?

[22:22:27] <Grackle> Just because they like the current setup doesn't mean they wouldn't like a different one.

[22:22:30] *** Witukind has joined #haiku

[22:22:33] <MYOB> mmu_man slidy tabs... I think?

[22:22:44] <Ingenu> crap alt tab is sent to windows, so can't test it on haiku in qemu

[22:22:45] <mmu_man> ah, right

[22:22:46] <MYOB> Grackle thats using a straw man, though

[22:22:50] <mmu_man> and stippi fixed them

[22:22:59] <judgen> coolio

[22:23:03] <Ingenu> anyone know whether Haiku will run on my Core 2 Duo ?

[22:23:09] <judgen> Ingenu: probably

[22:23:14] <MYOB> Ingenu kernel loads on my dualcore here

[22:23:17] <MYOB> so I guess so

[22:23:20] <MYOB> sees two procs too ;)

[22:23:27] <judgen> Ingenu: it wors fine on my 2x opteron 275

[22:23:32] <judgen> 4 cores total

[22:23:40] <Grackle> MYOB, I suppose, but saying that "nobody complains, therefore it is best" seems a bit logically insecure as well.

[22:23:45] <judgen> but i havent tested latest build though

[22:23:50] <MYOB> Grackle well, its the only way to judge in this case

[22:24:03] <MYOB> feel free to cut the delay out of your own copy of the Deskbar source, though...

[22:24:36] <judgen> Grackle: ever tried beryl? that is not the prefered way for users to set settings for common usage. Better with fewer options and better defaults

[22:24:41] <Grackle> I probably will. I think it's at least worth considering in the usability forum or something.

[22:24:53] <Grackle> No kidding, judgen.

[22:25:18] <Grackle> I use KDE and I like it, but I would agree with anyone who says that the options are too many and too confusing.

[22:25:25] <MYOB> Grackle I seriously doubt many others will support it, due to the visual distraction it causes

[22:25:35] <MYOB> its like the fucking "collapse to dock" thingy in OS X

[22:25:41] <Ingenu> still no installer ?

[22:25:46] <judgen> have all voted down and moded down the windows apps posted on bebits?

[22:25:52] <Ingenu> would BeOS 5 installer work for Haiku ?

[22:26:08] <Grackle> MYOB, do you think people would let it be a compile-time option?

[22:26:22] <Grackle> That would keep the setting out of the preferences interface.

[22:26:29] <MYOB> Grackle well, considering theres a total of 0 options like that in Haiku right now...

[22:26:32] <MYOB> no would be a guess

[22:26:41] <MYOB> Ingenu theorethically yes, in reality I really doubt t

[22:27:00] <Grackle> Really? I don't see why not. Anyone who doesn't want it/doesn't care could just ignore it.

[22:27:01] <judgen> Ingenu: you can install haiku from a beos live cd though

[22:27:15] <Grackle> I guess I don't really mind, it can't be that difficult to change it every time.

[22:27:19] <MYOB> Grackle adding needless complexity

[22:27:25] <MYOB> just like adding a setting for it does

[22:27:34] <MYOB> for, as far as we know - you and nobody else ;)

[22:27:35] <Ingenu> mmhh I'll give it a try

[22:27:40] <judgen> Grackle: also you have the right to make a distro of your own and set it up in what ever way you like afaik

[22:27:52] <MYOB> except you can't use the Haiku name...

[22:27:57] <Ingenu> mmu_man, can you receive private chat messages ?

[22:28:14] <judgen> MYOB: ofcourse not

[22:28:17] * Grackle calls it Maiku or something unoriginal like that

[22:28:37] <DaaT> call it Myku

[22:29:01] <Grackle> Well whatever, thanks for the rational discussion, even if I can't have it my way.

[22:29:13] <MYOB> woah, better make sure not to make any more analogies to DaaTs underwear now...

[22:29:29] <judgen> ill make my own distro and call it bukkake =P

[22:29:38] <Grackle> o.o

[22:29:43] <MYOB> judgen motto: "coming at you from all sides" ?

[22:29:45] *** ekdahl has joined #haiku

[22:29:49] <judgen> the themes will be highly obscene

[22:29:57] <judgen> MYOB: exactly

[22:30:01] <MYOB> like Ubuntu when it started

[22:30:05] <MYOB> Linuxxx

[22:30:53] <judgen> and ill only hire japaneese strippers to code for me

[22:31:02] <Ingenu> gonna try haiku

[22:31:08] <MYOB> why japanese only? racist ;P

[22:31:11] <Ingenu> herm

[22:31:15] <Ingenu> is there GF8 support ?

[22:31:15] <judgen> MYOB: they are pretty =)

[22:31:43] <MYOB> Ingenu the nVidia driver hasn't been touched since Rudolf left, so no...

[22:31:49] <Ingenu> :(

[22:31:50] <MYOB> I'm on the fastest supported nVidia still

[22:32:00] <judgen> MYOB: how else can i make an authentic bukkake distro if the worplace isnt in on it with both heart and soul

[22:32:22] <judgen> MYOB: i have a voodoo5 =)

[22:32:22] <CIA-17> korli * r20671 /haiku/trunk/src/kits/game/PushGameSound.cpp: fixed the wrong test for real

[22:32:23] <mmu_man> Ingenu yes :p

[22:32:26] <MYOB> judgen from what I've been told, plenty of English people are in to it stilll.

[22:32:31] <Grackle> Is there anything planned for a browser solution for Haiku?

[22:32:48] <mmu_man> Ingenu now that's an idea

[22:32:50] <MYOB> Grackle nothing official...

[22:32:51] <judgen> MYOB: isnt english girls know the be ugoes....

[22:32:55] <mmu_man> for all those who can't contribute by coding...

[22:33:02] <MYOB> I would suspect it'll come with Firefox#

[22:33:03] <judgen> better with a good gaelic lass

[22:33:28] <Grackle> Well firefox certainly isn't a very BeOS-y program.

[22:33:29] <judgen> Grackle: firefox is good

[22:33:50] <mmu_man> indeed Moe is a bit... vintage

[22:33:53] <Grackle> FireFox is indeed good. It's massive though.

[22:34:04] <judgen> Grackle: do you know how much manpower it has taken to write firefox? the entire haiku team together havent been close in hours i suspect

[22:34:05] <MYOB> either it or lynx

[22:34:21] <judgen> MYOB: we could port dillo =P

[22:34:26] <Ingenu> anyone here can access haiku donations list ?

[22:34:29] <mmu_man> maybe we'd need an HaikuVirtualGirl (and Boy, must be for everyone, we want Granny to use Haiku too :)

[22:34:31] <MYOB> could glue a http engine to flyingtroll...

[22:34:53] <Grackle> Yes, browsers are huge projects. I just wondered if anyone was thinking about making a native one.

[22:34:54] <MYOB> mmu_man at least one of the Haiku coders would prefer the virtual boy ;)

[22:35:04] <Pulko_Mandy> +++

[22:35:06] *** Pulko_Mandy has quit IRC

[22:35:14] <judgen> i would like a virtual donkey =P

[22:35:17] <Grackle> An existing rendering engine could be used, that would make it a lot easier.

[22:35:25] <Ingenu> gecko ?

[22:35:30] <Grackle> Sure, gecko is good.

[22:35:37] <judgen> khtml i assumed

[22:35:38] <MYOB> gecko needs a lot of work to makle it embeddable

[22:35:43] *** Cowboy_bebop has joined #haiku

[22:35:45] <Cowboy_bebop> Hi!

[22:35:48] <MYOB> ask tqh or fyysik for proof...

[22:35:48] <Grackle> Oh really?

[22:35:48] <Ingenu> hi

[22:35:50] <Grackle> Hi.

[22:35:52] <judgen> howdy cowboy

[22:35:55] <MYOB> and khtml, well, needs porting...

[22:35:57] <Grackle> heh I'll take your word for it

[22:36:10] <Cowboy_bebop> Fine and you?

[22:36:19] <MYOB> if you notice, most alternative OSs that had a native browser, have, well, dumped it

[22:36:30] <judgen> or we could break in to operas hedquarters in bergen and steal their code

[22:36:37] <MYOB> SkyKruzer -> Firefox, etc

[22:36:57] <Grackle> Even KDE's konqueror is dropped in favor of Firefox by a lot of people, and they have a lot of people working on it.

[22:37:01] <Ingenu> I took 'native' as 'well integrated'

[22:37:40] <MYOB> Ingenu well, Firefox has native drag and drop, launches from apps, etc, etc

[22:37:43] <MYOB> just needs to print...

[22:37:44] <judgen> why not build a web broswer that has neither interface or usage.

[22:37:54] <Grackle> Yes, let's steal opera. It would do well.

[22:37:59] <Grackle> haha

[22:38:08] <mmu_man> MYOB it's just a compile-time setting anyway... http://www.gnu.org/fun/jokes/dna.html

[22:38:11] <Cowboy_bebop> Can I have help about compile a linux driver on BeOS?

[22:38:12] <Ingenu> if Opera can run on a Wii it mustn't be too bad

[22:38:15] <judgen> Grackle: we have opera 3.63 allready

[22:38:22] <Grackle> Who needs a browser when you have telnet?

[22:38:22] <judgen> =P

[22:38:27] <Grackle> haha yeah, I noticed.

[22:38:34] <Ingenu> yeah you can send emails with telnet

[22:38:43] <judgen> did anyone but me buy opera 3.63

[22:38:44] <Grackle> And irc, and browse the internets!

[22:39:12] <[Beta]> :o version 2.0 is a topical change. lol

[22:39:21] <Grackle> Sure, it requires an intimate knowledge of each protocol, but with the space and development time savings, how can you go wrong?

[22:39:39] <MYOB> IRC = Vision...

[22:39:55] <judgen> vision is the best irc client i have ever tested so far

[22:40:00] *** Sil2100 has joined #haiku

[22:40:16] *** korli has quit IRC

[22:40:25] <judgen> i dropped baxter faster than you can drop a pen in complete vacum

[22:40:42] <MYOB> [Beta] yeah, and that revision date somewhat implies the Ancient Greeks were only faking it ;)

[22:40:57] <Grackle> I use irssi in gnu screen. It's not the best, but I have grown to like it, and it works well for what I ask of it.

[22:41:04] <mmu_man> MYOB note God doesn't consider it a bug but a purposeful feature :)

[22:41:20] <MYOB> Baxter was cool, around R4

[22:41:30] <MYOB> those background images! good god its like living in a different era

[22:41:39] <MYOB> that and Seth just vanished off the face of the BeOS earth....

[22:41:46] <judgen> on the 8th day god started the computer and sat down to play WoW untill he dominated the whole of azeroth =P

[22:42:34] <judgen> 22:42 here now, maybe is should get to sleap

[22:42:40] <MYOB> I am so intensely glad theres no WoW for BeOS

[22:42:46] <judgen> me too

[22:42:47] <MYOB> I've a feeling we'd "lose" more people...

[22:42:57] <Ingenu> "sleep"

[22:43:01] <judgen> if there was guild wars for BeOS i would probably never get any work done =P

[22:43:04] <MYOB> I used to work in the same building as WoW support

[22:43:09] <MYOB> the "sugar buffalo"

[22:43:13] <Ingenu> been considering doing so since a little while already, too

[22:43:17] <MYOB> they could empty 4 coke machines in 2 hours!

[22:43:22] <judgen> hehe

[22:43:40] <Ingenu> I did only play Eve Online (MMORPG)

[22:43:54] <judgen> eve? is that the cute animé one?

[22:43:55] <MYOB> there was, erm, 3 women

[22:43:58] <MYOB> 170 men

[22:44:00] <Ingenu> took me less than 3 months to quit

[22:44:17] <Ingenu> eve-online.com

[22:44:23] <MYOB> I once played evercrak

[22:44:25] <[Beta]> EVE is an awesome space mmorpg.

[22:44:28] <mmu_man> MYOB are you sure those 3 weren't faked ? ;)

[22:44:29] *** Fanskapet has joined #haiku

[22:44:31] <MYOB> but only because my Vaio came with 3 months free

[22:44:37] <MYOB> mmu_man one had a son who also worked there

[22:44:48] <judgen> i havent stopped playing gw... i am up to almost 4k hours now

[22:44:59] <judgen> hej fanskapet

[22:44:59] <Ingenu> ho

[22:45:06] <Ingenu> played and still play Guild Wars too

[22:45:10] <Ingenu> but it's not the same

[22:45:19] <Ingenu> spent 200hours in the first 6 months after release

[22:45:38] <Ingenu> I'll play it at work

[22:45:47] <judgen> i have gotten tired of gw since they introduced the heroes concept. Made PvE with teammates completley redundant

[22:45:51] <Ingenu> was wondering what to do between 12-14

[22:45:59] <MYOB> I guarantee you she got in to it by coming in to his room, shouting "WHAT THE HELL DO YOU DO ALL DAY! GO OUT AND GET A JOB!", sitting at his PC, and being forced to move, 4 days later....

[22:46:26] *** RedLobster has joined #haiku

[22:46:36] <RedLobster> How did the icon test work again?  0-4 points for each of 7 categories?

[22:46:47] *** jpd has quit IRC

[22:47:26] <judgen> anyone that makes a good full theme "can" be included in haiku i guess, but the default is allready decided i think

[22:47:50] <MYOB> judgen I thought there was going to be no support for icon themeing

[22:47:55] <MYOB> whats there is there - thats the default

[22:47:59] <judgen> MYOB: not in R1 afaik

[22:48:01] <RedLobster> that's not my question.  i'm talking about the icon contest.

[22:48:06] <RedLobster> http://bug-no.petterhj.net/iconcontest/

[22:48:22] <MYOB> RedLobster I wans't answering your question, as I can't remember...

[22:48:43] <RedLobster> :(

[22:48:44] <judgen> MYOB: wasnt there talk about it for R2 or maytbe R3

[22:48:48] <RedLobster> 1138 anyway.

[22:48:57] <MYOB> judgen eww

[22:49:05] <petterhj> RedLobster: I got your mail, and I'm checking it out right now

[22:49:12] <MYOB> themeing is the spawn fo satan - even OSX doesn't do it!

[22:49:20] <judgen> MYOB: there was alot of dreams back when it was still called OpenBeOS =)

[22:49:36] <MYOB> judgen like "finished by Christmas" (2001....)

[22:49:49] <RedLobster> petterhj sweet.

[22:50:02] <judgen> MYOB: i agree, but its good with coloringfor people like me, (colourblind)

[22:50:37] <MYOB> judgen thats what adjusting the colours is for ;)

[22:50:41] <judgen> hehe i dont think you would like my color setup on my boxes

[22:50:42] <MYOB> no need to theme

[22:50:48] <MYOB> screenshot?

[22:50:53] *** pikapika has joined #haiku

[22:51:05] <judgen> MYOB: ill take one from my windows box when i start it

[22:51:08] <RedLobster> it appears that it was 7 criteria, and a 0-4 vote for each.

[22:51:10] <MYOB> a mate of mine uses green/black/blue generally

[22:51:16] <judgen> this one is xfce and recently installed

[22:51:19] <MYOB> drives me infuckingsane to use his PC

[22:51:22] <judgen> hehe

[22:51:28] <judgen> i have alot of tyellow and black

[22:51:31] <Ingenu> BeOS R5.1d0 had skinning

[22:51:32] <RedLobster> petterhj i think i'm right.  because if you look, the highest point total is 28, and it says "average: 4"

[22:51:34] <Ingenu> why wouldn't haiku ?

[22:51:35] <judgen> tyan, yellow and black

[22:51:36] <MYOB> he's not even colourbling

[22:51:39] <MYOB> Ingenu because its EVIL

[22:51:44] <pikapika> hello

[22:51:50] <MYOB> Dano only had it cause BeIA had it - for coporate reasons...

[22:51:52] <RedLobster> beos needing many-user plus game.

[22:52:27] <judgen> MYOB: haiku should have a mandelbrot in the bootloader, like beos =P

[22:52:34] <RedLobster> will the API in haiku eventually drop the B_ in things like B_draw_window, in favor of H_draw_window?

[22:52:45] <judgen> RedLobster: why?

[22:52:50] <Grackle> I imagine a lot of people would hate the theme I use. http://alex.kc0udt.com/upload/data/kdetheme.png

[22:52:55] <RedLobster> well, why would you keep the B?  it's not beos.

[22:53:08] <judgen> RedLobster: the source from older apps can be compiled withouth enourmous amount of change, lets not alter that

[22:53:17] *** BuildFactory has quit IRC

[22:53:21] <petterhj> RedLobster: http://bug-no.petterhj.net/iconcontest/iconset.php?id=1

[22:53:26] <RedLobster> well, you could just use a find/replace on old code.

[22:53:32] <Grackle> But I like it, and it tries to be like BeOS.

[22:53:33] <RedLobster> you could even have a simple utility that does it for you.

[22:53:42] <RedLobster> but it's not beos.  being like beos is one thing.

[22:53:55] <RedLobster> it's haiku.  it needs to be its own thing, and progress.

[22:54:27] <RedLobster> petterhj you rule!

[22:54:28] <judgen> the b could mean alot more than Be =)

[22:54:44] <MYOB> RedLobster thats.... entirely futile

[22:54:45] <petterhj> RedLobster: I know ;)

[22:54:52] <MYOB> the find/replace wouldn't be reliable anyway

[22:54:59] <MYOB> as it'd nuke other suff

[22:55:00] <MYOB> stuff*

[22:55:12] <MYOB> and it would effectively destroy source compatibilty, util or not

[22:55:16] <judgen> b could stand for beer

[22:55:24] <MYOB> consider the aim of the project again

[22:55:27] <judgen> beer_draw_window =)

[22:55:36] <MYOB> then consider what you suggested... do they meet?

[22:55:43] <petterhj> Haiku will break bin. comp. some day, but breaking source comp. will be further into the future

[22:56:07] <petterhj> if breaking it at all

[22:56:12] <petterhj> (my guess)

[22:56:29] <judgen> petterhj: eventually we might have to go 64bit all of us, either we want it or not

[22:56:50] <MYOB> judgen that doesn't "break" more "mildly crack" source compatibility though

[22:56:57] <judgen> MYOB: true

[22:57:19] <RedLobster> petterhj how many voters were there?

[22:57:25] <MYOB> not as much as destroying the entire API

[22:57:26] <judgen> anyone asked yT about their aureal driver?

[22:57:41] <RedLobster> there shouldn't be source compatibilyt.

[22:57:47] <MYOB> judgen you'd have to ask Bernd - thats all thats left

[22:57:50] <petterhj> RedLobster: 2138

[22:57:54] <MYOB> RedLobster once again - look up the aims of the project

[22:57:57] <petterhj> (it says at the result page)

[22:57:59] <RedLobster> haiku needs to be haiku.  source compatibility needs to be virtualized.

[22:58:07] <RedLobster> petterhj awesome!

[22:58:16] <MYOB> what the hell? do you have any idea what you're talking about?

[22:58:26] <RedLobster> and it was conceivably competative, in the sense that people had an incentive to be dishonest and exaggerate.  which is great as a research tool.

[22:58:30] <RedLobster> so amazing that you used Range Voting.

[22:58:40] <RedLobster> is haiku said to be in beta or alpha?

[22:58:45] <RedLobster> as a whole.

[22:58:45] <MYOB> Haiku is going to be Haiku - because Haiku is a BeOS replacement

[22:58:51] <MYOB> hence compatibility is extremely important

[22:58:54] <RedLobster> sure it is.

[22:58:55] <judgen> MYOB: even if we got the code for the aureal cards, if bernd falls in court we would not be able to use it right? or did bernd and the gang remake the aureal driver. Since its soo much better than the crap included in R5.03

[22:58:58] <RedLobster> but i'm saying, virtualize it.

[22:59:15] <MYOB> RedLobster erm, explain that

[22:59:19] <RedLobster> as people start writing code for HAIKU, not beos, older code written for beos becomes less and less important.

[22:59:21] <MYOB> as what you're saying makes no sense at all

[22:59:32] <MYOB> theres no such thing as "virtualising" source compatiblity

[22:59:43] <RedLobster> MYOB Haiku_draw_window is the proper API, but B_DRAW_WINDOW is dynamically replaced when compiling.

[22:59:45] <MYOB> and its COMPLETELY AND UTTERLY POINTLESS

[22:59:55] <RedLobster> yes, using beos conventions is completely useless and utterly pointless.

[23:00:04] <MYOB> no, its not

[23:00:13] *** SiCuTDeUx has joined #haiku

[23:00:16] <MYOB> it would be if this was a project to write a totally different OS

[23:00:17] <MYOB> its not

[23:00:17] <RedLobster> you're not making beos.  you're making haiku.

[23:00:22] <RedLobster> sure it is.

[23:00:25] <MYOB> yes, and what is Haiku?

[23:00:30] <RedLobster> it is to begin with a haiku clone, and expand from there.

[23:00:33] <RedLobster> heard of glass elevator?

[23:00:35] <MYOB> do you have any fucking idea what you're gibbering about?

[23:00:38] <MYOB> clearly not

[23:00:51] <judgen> RedLobster: new to haiku?

[23:00:56] <MYOB> what would changing constant names acheive, other than wholesale destruction of compatibility?

[23:00:57] <RedLobster> Go read about the Glass Elevator project.

[23:00:58] <MYOB> Nothing

[23:01:01] <RedLobster> I used BeOS in 1997.

[23:01:01] <MYOB> absolutely nothing

[23:01:02] <[Beta]> what's glass elevator?

[23:01:09] <RedLobster> lol.  oh dear.

[23:01:14] <Grackle> Geez MYOB, calm down.

[23:01:15] <DeadYak> Glass Elevator is at this point nothing more than discussion.

[23:01:23] <RedLobster> sure.

[23:01:26] <Grackle> I think you're right, but you're awfully harsh.

[23:01:42] <MYOB> Grackle I get harsh with idiots when I'm not being paid to be nice to them

[23:01:51] <judgen> hehe

[23:02:00] <judgen> brittish humor at its best

[23:02:10] <RedLobster> But the point is to make an operating system that replaces beos (source and binary compatible) and then continue where Be left off, no?

[23:02:16] <MYOB> 125 is the starting rate for feigned niceness

[23:02:27] <MYOB> Eh

[23:02:28] <MYOB> No

[23:02:41] <judgen> RedLobster: yes, but also making improovements on the way. And fixing the stuff that Be did wrong

[23:02:42] <Grackle> RedLobster, I see what you're saying, but it's just not necessary.

[23:02:50] <RedLobster> so if Be had kept writing, their next version of BeOS, say version 6, would be a different OS, with different functions, different file systems, perhaps multi-user, etc.

[23:03:01] <RedLobster> judgen which even further supports my point.

[23:03:02] <DeadYak> RedLobster: um...no.

[23:03:03] *** StylusEater_Work has left #haiku

[23:03:06] <MYOB> yes, but would they have changed the names of every constant and function in the API? No

[23:03:08] <RedLobster> DeadYak how so?

[23:03:09] <MYOB> not a fucking chance

[23:03:09] * mmu_man raises the trout

[23:03:19] <DeadYak> RedLobster: in case you haven't noticed Be generally kept compat between releases

[23:03:19] <judgen> RedLobster: your point is that we should do changes, just for changes sake

[23:03:21] <RedLobster> MYOB no, but what's that got to do with anything?

[23:03:22] <MYOB> something tells me you've never worked with software before

[23:03:33] <MYOB> RedLobster becuase thats what you';re fucking suggesting Haiku does!

[23:03:49] <RedLobster> judgen no.  the point is, why would you call something b_draw_window, if you are not b?

[23:03:55] <RedLobster> you are h.

[23:03:57] <MYOB> DeadYak right down to R2 PPC binaries running on R5

[23:04:06] *** jevin has quit IRC

[23:04:07] <mmu_man> RedLobster Because :)

[23:04:13] <MYOB> because thats the function name

[23:04:14] <mmu_man> B_ECAUSE :))

[23:04:20] <MYOB> simple as

[23:04:21] <RedLobster> MYOB yes, in beos.

[23:04:22] <Grackle> Because we are Be compatible, and changing it would be a huge pain in the ass for what accounts for little more than semantics.

[23:04:23] <RedLobster> but this isn't beos.

[23:04:27] <MYOB> and what is Haiku?

[23:04:29] <MYOB> BeOS compatible

[23:04:31] <RedLobster> mmu_man b_cause would have been funnier.

[23:04:36] <MYOB> can you not get the fucking picture here

[23:04:38] <RedLobster> MYOB sure, you can do that by having a wrapper.

[23:04:40] <judgen> because b can be anything. Like Bruno, beer, bleeching, beltbuckle or anything. Let you mind soar. It does not have to mean Be for the programmers using it

[23:04:42] <[Beta]> #define h_ b_ ?

[23:04:46] <RedLobster> so that b_draw_window actually invokes haiku_draw_window

[23:04:53] <petterhj> RedLobster: why make a wrapper just for the sake of it?

[23:04:53] <MYOB> so you can get it by adding needless extra layers?

[23:04:55] <MYOB> why the fuck would you?

[23:05:10] <RedLobster> are most haiku people really this obsessed about the idea of keeping all these weird legacy concepts?

[23:05:19] <MYOB> how is API compatibility a legacy concept??

[23:05:21] <RedLobster> if so, i'm much less optimistic about haiku's future. :(

[23:05:22] <petterhj> RedLobster: yes

[23:05:27] <petterhj> :p

[23:05:28] <[Beta]> its not a legecy concept RedLobster, its the API Haiku has

[23:05:28] <RedLobster> depressing.

[23:05:29] <mmu_man> it's just *names*

[23:05:47] <MYOB> sweet mother of jaysus....

[23:05:56] <MYOB> I'm depressed that people can have such inane ideas

[23:06:01] <RedLobster> it's a legacy concept in that instead of trying to start with a good idea as your basis, and continue it, you are trying to replicate it.  why?!

[23:06:09] <MYOB> effectively you're propsing renaming EVEYRTHING then providing a layer to undo that?

[23:06:14] <MYOB> for NO REASON AT ALL?

[23:06:34] <MYOB> mmu_man got that trout ready?

[23:06:40] <RedLobster> i suggest that the function names have something to do with the OS.

[23:06:49] <RedLobster> instead of the OS which inspired the OS.

[23:06:55] * judgen swings his mighty swordfish over his head

[23:06:55] <MYOB> yeah, and they will - any NEW ones

[23:06:56] *** mmu_man sets mode: +m

[23:07:04] <mmu_man> damn I have a headache

[23:07:08] <mmu_man> some silence please :P

[23:07:08] *** RedLobster has left #haiku

[23:07:27] *** mmu_man sets mode: -m

[23:07:53] <Grackle> That was all rather silly.

[23:08:11] <DeadYak> that's an understatement

[23:08:12] <MYOB> as my last line was before the wall of silence: "however, the functons from BeOS are just that - FROM BEOS"

[23:08:16] <MYOB> but hey, the numpty has gone

[23:08:24] <Grackle> Whee.

[23:08:50] <judgen> well good night MYOB Grackle petterhj DeadYak mmu_man, see you tomorrow maybe

[23:08:57] <judgen> hope you have a nice night

[23:08:58] <DeadYak> night judgen

[23:08:59] <Grackle> http://bug-no.petterhj.net/iconcontest/visualizer/ Haiku is currently using the zumi iconset, right?

[23:09:03] <Grackle> Nighto.

[23:09:12] *** RedLobster has joined #haiku

[23:09:15] <MYOB> that was -bordering- on the Luposian level of pure irritance - but Luposian at least complains about a bug...

[23:09:17] *** RedLobster has left #haiku

[23:09:33] *** judgen has quit IRC

[23:09:34] <DeadYak> yay for asshats

[23:09:37] <petterhj> Grackle: no, the stippi one..

[23:09:40] <mmu_man> Grackle Stippi's I think :)

[23:09:48] <Grackle> oh ok

[23:09:53] <Grackle> They are quite similar, I wasn't sure.

[23:10:02] <Grackle> Stippi looks good.

[23:10:12] <petterhj> yea, I was unsure about it at first, but now i love it!

[23:10:27] <petterhj> Grackle: I'll tell stippi that ;p

[23:10:33] <Grackle> haha

[23:10:34] <[Beta]> lol

[23:11:17] <Ingenu> night people, happy haiku dreams

[23:11:34] <Grackle> cya

[23:12:33] <Grackle> oh tin of pink meat / must we enclue you firmly? / you are not noah.

[23:12:33] *** Ingenu has quit IRC

[23:12:33] <nutela> yes Redlobster was right, why is everything b_ ..?

[23:12:34] <MYOB> just thought of an analogy for that guy

[23:12:42] <MYOB> what does every API call in OS X start with?

[23:12:43] <nutela> LOL

[23:12:45] <MYOB> NS_ .....

[23:12:57] <Grackle> haha

[23:12:58] <nutela> better it had not been MS_ hehe

[23:13:13] <MYOB> is OS X 10.4 NeXTStep? No...

[23:13:33] <nutela> MYOB MYOB

[23:13:42] *** SiCuTDeUx has quit IRC

[23:14:03] *** RedLobster has joined #haiku

[23:14:05] <nutela> how come you get so angry by such a unimportant point of view

[23:14:09] <RedLobster> oh, and a word of advice.  don't shoot the messenger.

[23:14:20] *** nonesuch has joined #Haiku

[23:14:30] <RedLobster> getting hostile at someone for pointing out some very poignant concerns is...silly.

[23:14:32] <MYOB> erm, you werne't the messenger - its your opinion

[23:14:36] <nutela> RedLobster : count how many times that has happened grin

[23:14:46] <MYOB> extremely flawed and completely insane that it is

[23:14:47] <nutela> here we go again

[23:14:51] <DeadYak> how are they poignant?

[23:14:54] <RedLobster> well, you have yet to prove that.

[23:15:01] <DeadYak> they're silly semantic concerns frankly

[23:15:02] <MYOB> RedLobster what is the API prefix in MacOS X?

[23:15:05] <DeadYak> that make no practical difference.

[23:15:06] <MYOB> NS_

[23:15:08] <nutela> you can't prove a definition RedLobster

[23:15:11] <MYOB> is it NeXTStep?

[23:15:23] <MYOB> No, but the API calls are -from- NeXTStep

[23:15:30] <MYOB> just as anything B_ prefixed is from BeOS

[23:15:42] <nutela> That's why OS X is so sloooow

[23:15:50] <nutela> inside it is nextep outside it is OSX

[23:15:52] <RedLobster> what sense will there be in keeping the same function names, when it is _trivial_ to change them?  you think when haiku version 3 is out in 8 years, you're still going to be calling even the new function calls b_this and b_that?  just because it gives you a warm fuzzy feeling?

[23:16:00] <RedLobster> that's absurd.

[23:16:01] <MYOB> its not trival to change them

[23:16:02] <DeadYak> how is it trivial to change them?

[23:16:05] <RedLobster> MYOB yes it is.

[23:16:06] <RedLobster> trivial.

[23:16:07] <nutela> I wonder how Vista looks like inside....

[23:16:08] <MYOB> its extremely UN trivial to change them

[23:16:08] <DeadYak> it breaks every single app for BeOS in existence.

[23:16:13] <[Beta]> but why change them, what is to gain from a change ?

[23:16:21] <MYOB> no, its not. further proof you haven't got a clue about what you're talking about.

[23:16:22] <RedLobster> DeadYak no it doesn't.  you just use a wrapper when you compile them.

[23:16:29] <RedLobster> you're just talking about names.

[23:16:30] <DeadYak> RedLobster = unnecessary overhead.

[23:16:33] <RedLobster> in the source code.

[23:16:35] <MYOB> RedLobster thats still breaking them

[23:16:38] <nutela> OK RedLobster give use the regex how to change them

[23:16:40] <MYOB> and adding large amounts of overhead

[23:16:41] <DeadYak> if you think you can fix it with a bunch of #defines you're kidding yourself

[23:16:42] <RedLobster> it's only overhead for the old apps.

[23:16:54] <RedLobster> you stop writing code in the Be api and start writing code in the Haiku API.

[23:16:58] <MYOB> ...which will comprise the majority of apps

[23:17:08] <MYOB> the "Haiku API" will be the Be API, with new additions

[23:17:16] <RedLobster> so are you guys to tell me that even when you create new functions and classes you will keep the preceding b ?

[23:17:22] <MYOB> no

[23:17:24] <MYOB> where was that said?

[23:17:28] <RedLobster> i'm asking.

[23:17:31] <RedLobster> i didn't say it was said.

[23:17:32] <Grackle> RedLobster, even assuming it is very easy to do, which it isn't, there is absolutely no functional gain.

[23:17:39] <nutela> That's an interesting question

[23:17:42] <RedLobster> oh it's easy to do.

[23:17:43] <MYOB> the exact opposite was said, in fact, if yo check your scrollback buffer...

[23:17:48] <MYOB> no, its not

[23:17:52] <MYOB> how many times do you need to be told that

[23:17:55] <DeadYak> you've written how much code?

[23:18:07] <MYOB> hands up here who actually gets meaningful employment from writing code?

[23:18:07] <RedLobster> plenty.

[23:18:09] <MYOB> .o/

[23:18:10] <Grackle> Yes, if it's easy, then give us the regex.

[23:18:10] <RedLobster> in many languages.

[23:18:17] <RedLobster> lol.

[23:18:21] <DeadYak> I don't believe you.

[23:18:30] <RedLobster> hah.

[23:18:31] <nutela> regex please

[23:18:32] <MYOB> and now, which of that grouping thing its "easy to do"

[23:18:33] <RedLobster> nice come back DeadYak.

[23:18:34] <mmu_man> anyway it's not debatable until R1

[23:18:35] <MYOB> .o.

[23:18:40] <mmu_man> at least

[23:18:50] <RedLobster> this is so silly.

[23:18:58] <DeadYak> RedLobster: then please, show me how to trivially replace expressions on the fly, taking into account their context, etc. where names might be in use in the middle of other names, etc.

[23:19:01] <Grackle> It sounds like a pain in the ass to me, but I am not an employed programmer.

[23:19:02] <nutela> so do new API parts get H_ ?

[23:19:11] <DeadYak> there's a reason nobody does drastic naming changes like that.

[23:19:18] <DeadYak> they are anything but trivial in a large block of code

[23:19:19] <RedLobster> DeadYak any time you see b_draw_window you replace it with haiku_draw_window.  duh.

[23:19:25] <MYOB> yeah, provide us with a layer that will reliably map -ALL- B_ calls to H_ and so on, without affecting anything the coder has named themselves, and which can be done with one header file

[23:19:27] <mmu_man> yeah and ppl will have to look up everytime to know if it's H_ or B_... really nice

[23:19:28] <RedLobster> don't pretend that's complicated or more tricky than it sounds.  it's not.

[23:19:37] <MYOB> RedLobster once again you're showing us you don't know what you're talking about

[23:19:45] <DeadYak> yes it is, try doing search and replace sometime like that

[23:19:46] <MYOB> show us a method that will do that

[23:19:47] <DeadYak> you will break

[23:20:01] <RedLobster> i was in the auidence in mountain view with GLG, and i thought michael seemed to smart.  this is so depressing.

[23:20:09] <MYOB> you'll brreak quite spectacularly

[23:20:10] <RedLobster> good day.

[23:20:13] *** RedLobster has left #haiku

[23:20:16] <Grackle> RedLobster, the way to make changes in open source projects is by writing and submitting code. Just do it. If it works well and people like it, it will stick. You're making no headway here.

[23:20:18] <MYOB> He is smart. You, however, are not

[23:20:21] <Grackle> ...

[23:20:39] <[Beta]> MYOB, heh.

[23:20:41] <[Beta]> Yeah, we need to deal with awkward people better.

[23:20:45] <MYOB> why can I see this being raised over, and over, and over?

[23:20:54] <nutela> What a dick

[23:21:01] <[Beta]> must be in your genes or something

[23:21:04] <MYOB> he's borderline autistic if he's that obsessed with it

[23:21:09] <MYOB> Haiku's own Rain Man...

[23:21:20] <MYOB> *rocks back and forth* must change the Bs.... must change the Bs....

[23:21:23] <Sloar> keep in mind all transcripts are available on haiku's site

[23:21:28] <nutela> haha

[23:21:31] <DaaT> why IS haiku keeping B_?

[23:21:34] * DaaT runs

[23:21:36] <Grackle> Yes, we can be firm without being mean.

[23:21:44] <nutela> LOL

[23:21:50] <nutela> true

[23:22:00] <DeadYak> Grackle: sure, but he wasn't exactly being friendly about it either

[23:22:02] <nutela> put it in the FAQ

[23:22:06] <MYOB> could also tell he's not a coder, of any description

[23:22:17] <nutela> teh FAKING ANSWERED QUESTIONS

[23:22:22] <Grackle> DeadYak: True.

[23:22:51] <Grackle> Frequently Questioned Answers

[23:22:57] <TTRanger> Programmers are such detail people...otherwise why would anyone bother to even debate about something like B_  :-)   Doesn't count for much in the real world, does it?

[23:22:58] <nutela> LOL

[23:23:00] <[Beta]> so, what always bugged me, was the API in US English. Can we change that ? ;))

[23:23:16] <MYOB> yeah, I want SetViewColour()!! :P

[23:23:24] <nutela> LOLLOLLOL

[23:23:28] <MYOB> and I'll scweam and scweam until I'm thick!

[23:23:29] <nutela> ROFL

[23:23:32] <MYOB> (and I can!)

[23:23:36] <MYOB> </just william>

[23:23:57] *** Sil2100 has quit IRC

[23:24:00] <[Beta]> i'm not with him.

[23:24:02] <nutela> oh oh stop it please ROFL

[23:24:03] * kokito will file a big report: modifiy the Haiku API to use the H_ prefix instead of B_

[23:24:05] <Grackle> But it's one character longer, it will cause unnecessary bloat!

[23:24:32] <nutela> huh?

[23:24:45] <Grackle> color vs colour, not H_ vs B_

[23:24:47] <DeadYak> you mean Colour vs Color presumably..

[23:24:48] <DeadYak> yeah

[23:24:56] <Grackle> Just badly timed.

[23:25:44] <MYOB> I forgot how long Haiku takes to build....

[23:25:59] *** oco has joined #haiku

[23:26:02] <Grackle> What are you building it on?

[23:26:05] <Anxiety> just make it one and then include a #define Colour Color :D

[23:26:09] <MYOB> dual 2.8

[23:26:16] <Grackle> oh heh

[23:26:45] <Grackle> I was going to say that it doesn't take too long on my 1.86GHz core solo, but you've got a faster machine.

[23:27:07] <MYOB> what OS, though

[23:27:21] <MYOB> this is a clean build, and its from a BeOS-derivative = slow file IO...

[23:27:22] <Grackle> Linux.

[23:27:25] <Grackle> ah

[23:27:35] <Grackle> I guess that would make a big difference.

[23:27:44] <MYOB> I'm throwing a LOT of machine at this and its still slow...

[23:28:02] <MYOB> that and my tower now sounds like a Pratt & Whitney...

[23:28:08] <Grackle> I imagine that's something that Haiku wants to improve.

[23:28:10] <Grackle> hehe

[23:28:11] <nutela> huh building on beos is slow?

[23:28:24] <[Beta]> exceedingly, nutela

[23:28:41] <MYOB> as is subversion

[23:28:46] <Grackle> Maybe you should set up a linux machine, maybe even a distcc cluster for building.

[23:29:17] *** JamesB192 has joined #haiku

[23:29:22] <nutela> cant you build haiku with haiku?

[23:29:23] *** JamesB192 has quit IRC

[23:29:32] <MYOB> BeOS's file I/O was woeful for the time and its now just unspeakably bad....

[23:30:38] <nonesuch> hey good question nutela i keep forgetting to ask that

[23:30:40] <Grackle> Is zeta any better about that than other BeOS variants?

[23:31:00] <nutela> I'd say its worse

[23:31:07] <MYOB> somewhat worse...

[23:31:14] <nutela> much

[23:31:17] <nutela> well...

[23:31:22] *** ekdahl has left #haiku

[23:31:42] <MYOB> my SATA drive feels no faster than my IDE one did, for instance

[23:32:04] *** JamesB192 has joined #haiku

[23:32:16] <surrounder> gawd I love the fact I run it on slow hardware, it all feels fast on that machine anyways ;)

[23:32:36] <MYOB> ah, the person who makes me think of dodgy Belgian dance music :P

[23:32:48] <surrounder> rofl

[23:32:53] <nutela> Bonzai records hehe

[23:33:23] <MYOB> oh well yer wan that sngs that song is opretty decent looking

[23:34:10] <nutela> What's up with haikunews MYOB

[23:34:11] <nutela> ?

[23:34:19] <nutela> seems pretty dead to me

[23:34:39] <MYOB> apathy#

[23:34:43] *** Cowboy_bebop has quit IRC

[23:35:50] <mmu_man> well let's just hope the file-cache in Haiku will get better than BeOS :)

[23:36:31] * [Beta] pokes MYOB

[23:36:34] <nutela> than BeOS itself

[23:36:34] <nonesuch> soooo...i'm still curious if Haiku is self-building yet

[23:36:46] <nutela> nanotech

[23:37:03] <MYOB> [Beta] supply caffeine and I'll update it more

[23:37:24] <DaaT> pphhhttt

[23:37:25] <[Beta]> does caffeine supply rely on € ?

[23:37:26] <DaaT> excuses

[23:37:43] <MYOB> DaaT you have your sheep for, erm, stimulation...

[23:37:43] <nutela> haha

[23:37:53] <nutela> ROFL

[23:38:13] <nutela> being sharp today right MYOB?

[23:38:19] <DaaT> nope, they're for relaxation

[23:38:22] <DaaT> :P

[23:38:55] *** avkig has quit IRC

[23:39:24] <nutela> night

[23:39:33] <MYOB> riiiiiight

[23:40:05] <MYOB> anyway, I metaphorically left your panties over by a crashed copy of Communicator earlier, you may need to pick them up...

[23:40:17] *** guma has quit IRC

[23:40:30] <DaaT> errrr... my panties? boxers for me, sorry

[23:40:36] <DaaT> you mean DeadYak perhaps?

[23:40:46] <DaaT> I do remember some panties in belgium

[23:40:50] <DaaT> *whistles*

[23:40:52] <MYOB> not unless you were lying to Betsy and Flossy....

[23:41:09] *** axeld has quit IRC

[23:41:25] *** mmu_man has quit IRC

[23:41:30] <DaaT> i never lie to my fluffies

[23:41:35] <[Beta]> hmm

[23:44:04] <MYOB> anyway, need sleep

[23:44:52] <[Beta]> nn

[23:45:31] *** MYOB has quit IRC

[23:48:36] *** oco has quit IRC

[23:49:48] *** petterhj has quit IRC

[23:56:40] *** Fanskapet has quit IRC

[23:56:59] *** Fanskapet has joined #haiku

[23:57:52] *** wildur has quit IRC
top

   April 12, 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 | >