IRC log started Mon Aug  9 14:27:43 1999
[msg(tunes)] permlog 1999.0809
* ruiner/#tunes thinks eihrul helped design the "improvements" to C to make C++
<eihrul> no...
<hcf> abi: forget go
<abi> hcf: I forgot go
<eihrul> i just am disgusted by people who seek to attack what they don't understand
<eihrul> i can attack C++
<hcf> abi: go os is at http://www.soi.city.ac.uk/~gel/go.html
<eihrul> but i feel you have no right to
<smkl> c++ is easy to attack because it's so popular
<ElGato> well my first programming language was c++, and it scarred me
<ruiner> yes!
<ruiner> go gato
<eihrul> i programmed in C for 4 years before C++
<eihrul> and C++ was like the promised land after that
<ElGato> i'd rather have learned scheme and asm
<ruiner> I started with qbasic....then vb....then c++....and I hated it....so I stepped down to c
<eihrul> it offered so much of what C was lacking
<eihrul> so maybe it's your progression to the language that's wrong :)
<eihrul> not the language itself
02:30pm
* ruiner/#tunes thinks anybody who wants to learn C++ should definately learn C first
<ElGato> i agree
<eihrul> i agree with that
<eihrul> so many people have very limited understanding of memory addressing
<eihrul> and all the little things that make C useful
<ElGato> i went and learned c after that then asm
<ruiner> learn how to throw a fastball before you learn to throw a curve
<ElGato> *sigh*
<ElGato> if only someone showed me asm first
<eihrul> intel manuals are very good
* ElGato/#tunes has 9 of em :)
-:- _sham [sham@phila-dialup478.nni.com] has joined #tunes
<ElGato> ruiner: i feel sorry for you, learning basic first
<smkl> http://www.cs.cornell.edu/talc/default.html -- asm that is checked so it can't access uninitialized memory
-:- Tril [dem@bespin.dhs.org] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
<ElGato> Tril: is christmas island independenr or does it belong to the US or what?
<smkl> i think it belongs to oz
<ElGato> s/independer/independent
<Tril> i forget, www.cx has all the info
<Tril> www.cx is the DNS for .CX, hosted in UK.
<_sham> elgato: are you considering registering a CX domain?
<ElGato> _sham: uh no, why the hell would i want to do that?
<_sham> well you were thinking of registering a domain name, and CX is a viable choice
02:40pm
<ElGato> i should reg tunes.mx ;)
<_sham> tunes.nu
<ElGato> nu, nu
<ElGato> what's nu?
<smkl> niue 
<_sham> Nieue
<ElGato> hmmm, never heard of it
<_sham> ElGato: how's work on the compiler?
<smkl> it belongs to nz
-:- Tril [dem@bespin.dhs.org] has left #tunes []
<_sham> put your code on bespin
<ElGato> i hate it when you ask that
<ElGato> no wait
<ElGato> i hate nothing
<ElGato> i am annoyed when you ask that
<_sham> ok
<ElGato> i still don't know about tokenizing
<_sham> you're channeling your anger
* _sham/#tunes feels like Dr. Katz
<ElGato> i can't make a token for oh say the number "1234" because there might be a variable named "1234"
<ElGato> what to do
<smkl> get flex
<ElGato> flex flex flex
<ElGato> im sick of hearing about flex
<smkl> hehe
<_sham> flex this, flex that... bah!
<smkl> why would you allow variables called "1234"
<ElGato> cause then the bastard programmer who names his variables after numbers will sart bitching
<smkl> are you going to make everyone use peano's numbers or something ?
<ElGato> peano?
<smkl> an (italian?) mathematician
02:50pm
<ElGato> hmmm
<smkl> you would have zero and then function that returns the next number
<_sham> abi: Peano is an italian mathematician who invented algebra and calculus while drunk one night
<abi> i haven't a clue, _sham
<smkl> peano?
<smkl> abi: peano?
<abi> i don't know, smkl
<_sham> abi: smkl?
<abi> _sham: i haven't a clue
<ElGato> i guess i'll just not let anyone use numbers in their variables
<hcf> smkl: the 'who' needs a \, as in '\who'
<hcf> s/smkl/_sham/
<_sham> abi: smkl is a spam bot who got stranded in #tunes
<abi> wish i knew, _sham
<_sham> abi: smkl is a spam bot \who got stranded in #tunes
<_sham> heh 
<ElGato> ElGato?
<abi> you are annoyed when you ask that
<ElGato> hahaha
<ElGato> hickserv?
<abi> hickserv is, like, wastes his time so effectively or the official Tunes knowledge moocher
<smkl> abi: _sham is a norwegian fisherman!!!
<ElGato> is that a finnish insult?
<smkl> yes
* ElGato/#tunes writes that down
* _sham/#tunes hits ElGato with a book of finnish insults
<ElGato> im gonna go to finland someday
<ElGato> see if i like it, mabye stay
<_sham> start a family in finland?
03:00pm
<ElGato> yeah, i'll steal linus torvalds' wife
<ElGato> no ' after that, pardon
<ElGato> abi: arrow logic
<ElGato> abi: arrow logic?
<abi> no idea, elgato
03:10pm
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf [nef@me-portland-us646.javanet.com] has joined #tunes
<ruiner> later all...
-:- SignOff ruiner: #TUNES (Leaving)
* ElGato/#tunes is away: (Auto-Away after 10 mins) [BX-MsgLog On]
-:- wale6361 [user2471@69pm1.ifu.net] has joined #Tunes
-:- wale6361 [user2471@69pm1.ifu.net] has left #Tunes []
03:20pm
-:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp35.lvdi.net])
-:- eihrul [lee@usr5-ppp45.lvdi.net] has joined #tunes
-:- SignOff ElGato: #TUNES (ElGato has no reason)
-:- liar [brand@p0wer.qzx.com] has joined #tunes
-:- SignOff smkl: #TUNES (sleeping ...)
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us646.javanet.com])
-:- SignOff abi: #TUNES (Ping timeout for abi[bespin.dhs.org])
-:- abi [nef@bespin.dhs.org] has joined #tunes
-:- ultima [ultima@user-37kb9st.dialup.mindspring.com] has joined #Tunes
<liar> has anyone here used the invlpg instruction?
<ultima> wtf?
<ultima> why would you use the invlpg isntruction?
<liar> to invalidate a tlb entry
<liar> the manual isnt clear on what i use for the operand
06:30pm
<eihrul> liar: a simple memory address
<eihrul> ala load effective address
<liar> so i use the first address in the page?
<eihrul> yep
<liar> linear?
06:40pm
<eihrul> shouldn't be
<eihrul> should be virtual
<liar> same thing
<eihrul> gah
<eihrul> linear seems too much like physical :)
<liar> do u have an eflags reference handy?
<liar> nm found it
<liar> ok maybe i didnt find it
<eihrul> you have intel manuals?
<liar> ya
<eihrul> should be in one of them
<liar> i can never find a complete set in it
<eihrul> i've seen one in both the first and third, methinks
-:- Kenn [cpe2@1Cust121.tnt7.tampa.fl.da.uu.net] has joined #Tunes
<Kenn> I'm look for Mr. Right . . . :)
<liar> i will look in the amd manual, they have their shit in order
<Kenn> is Fare here?
-:- SignOff _sham: #TUNES (Ping timeout for _sham[phila-dialup478.nni.com])
<Kenn> under another name?
<eihrul> liar: amd manual is better?
<Kenn> oh
-:- SignOff Kenn: #TUNES (Leaving)
<liar> eihrul: yes it is
<eihrul> will have to check it out
<liar> ok found them in the amd manual
<liar> amd has a more detailed table of contents
<liar> intel only lables chapters 
06:50pm
<liar> whats the best way to handle stuff like sleep timers?
<eihrul> i have some code to do it if you want to look
<eihrul> not the best... 
<liar> how does it do it
<eihrul> one sec... lemme reread this code :)
07:00pm
<eihrul> okay... i do it like thus:
<eihrul> keep a list of all the sleep timers sorted, closest deadline first
<liar> ok got that
<eihrul> then just check if the clock has met the first deadline
-:- _sham [shamino@phila-dialup456.nni.com] has joined #tunes
<liar> thats how i do it now
<eihrul> if it has, then just pull timers off the list until you find a deadline greater than the clock
<eihrul> well that's what i'm doing
<eihrul> but i'm not using absolute time to avoid overflow
<liar> i use a 64bit counter
<eihrul> i simply use the time relative to the first task inserted
<eihrul> that way i don't have to really worry... i can hold up to 32 bits of time no matter what
<_sham> i use an MMX counter
<liar> i take the current counter value and add the time to sleep to that value and place it in the list
<eihrul> avoids using 64 bit counters nicely :)
<eihrul> er rather, i'm keeping two clocks...
<liar> then i can just compare the current counter to the timer value 
<eihrul> one that restarts at 0 every time a timer is inserted onto a perfectly empy list
<eihrul> and the absolute clock
<liar> hmm, i will have to do that
<eihrul> and all timers subsequently inserted
<eihrul> re relative the first clock, rather than the latter
<liar> ok
<eihrul> and to prevent having to compare the clock to the timers
<eihrul> i simply just store a counter down clock till the next timer will go off
<liar> and what do u do if someone tells a thread to sleep for more than 497 days? :)
<eihrul> when that hits 0, i know an a timer daedline has happened
<eihrul> so i could use 64 bit counters
<eihrul> but i still think my solution allows for longer timers without overflow :)
<liar> a counter down clock?
<eihrul> every time a timer is inserted on the list
<eihrul> i find out how many clocks left until the first timer goes off
07:10pm
<eihrul> and i just count that down
<eihrul> it saves me a comparison :)
<liar> but u have to decrement it each clock tick
<eihrul> yep... but that's simplre than an increment and compare
<eihrul> easier to test for zero :)
<liar> why increment?
<eihrul> to know when a timer has gone off
<eihrul> or you could simply store the absolute clock at which the next timer will go off
<eihrul> however you wanna do it
<eihrul> which would probably be an even better solution
<eihrul> so you only have to maintain one timer
<eihrul> er clock :)
<liar> i think i'll do that
<eihrul> i think i'll do that too 
<eihrul> now that i've thunk of it :)
<liar> heh
<liar> hmm maybe the second counter would be better
<liar> well maybe not
07:20pm
<liar> just compare the timer value to the loewr 32bits of the 64bit counter
<eihrul> having only one counter could lead to weird overflow conditions
<liar> no
<eihrul> maybe not on 64 bit counter :)
<eihrul> but who knows... millions of years for now...
<eihrul> we'll be having the y2g problem
<eihrul> and it'll be all your fault
<liar> u can use a 64bit counter for the primary and use 32bit relative counters for the timers and just compare the lower 32bits of the primary to the timer
<liar> i will accept full responsibility for all y2g problems :)
<eihrul> then you'd have to modify your 32 bits 
<eihrul> to be offset by 64timer % 32bits
<eihrul> to do that
<liar> in sleep() do timer=primary_counter&&0x0ffffffff+timer_value
<liar> and ignore the overflow
<liar> then in the scheduler just compare the timer to the lower 32 of primary
<eihrul> so long as you don't check for greater than or less than
<eihrul> and just equality
<liar> u only need equality
<eihrul> yep... just checking :)
-:- SignOff eihrul: #TUNES (Leaving)
-:- eihrul [lee@usr5-ppp45.lvdi.net] has joined #tunes
<eihrul> that sucked
<liar> heh
<eihrul> did an EOF in my irc client or something
<eihrul> and it just locked
<liar> === SIGNOFF #tunes eihrul Leaving
>>> _sham [shamino@phila-dialup456.nni.com] requested VERSION  from #tunes
<_sham> heh
07:30pm
-:- hcf [nef@me-portland-us739.javanet.com] has joined #tunes
<liar> eihrul: u know much about protection?
<eihrul> kinda sorta
<liar> i found out that if cpl=0 then the stack dpl must also be 0
<eihrul> yep
<liar> does ds dpl also need to be 0 to access pages with supervisor level?
<eihrul> thinking... :)
07:40pm
<eihrul> not quite sure, but prob not :)
<eihrul> you can always try and see what happens
<liar> i hope to hell not
<eihrul> if you get a GPF, you'll know we were both wrong
<liar> i would have to change ds for every kernel function and restore it before leaving
-:- SignOff _sham: #TUNES (Ping timeout for _sham[phila-dialup456.nni.com])
<eihrul> yes... you need to do that
<liar> oh crap nevermind i have to do that no matter what
<ultima> Doesn't writing an OS just totally SUCK?
<liar> but if i dont need a dpl0 ds then i can just set it to dpl3 ds and not need to restore
<liar> ultima: yes
<eihrul> but it's the suckyness that makes it fun
<liar> uhh
<liar> no the suckyness is what makes u feel really good AFTER its all done
<liar> the thrill of making a better os is what makes it fun
<eihrul> not necessarily better
<eihrul> it's just a mountain to climb :)
<liar> well if i wasnt making something i thought was better than why the hell would i make it???
<eihrul> why the hell not?
<eihrul> what else do you have to do? :)
<liar> :)
<liar> play games?>
<liar> go play some paintball
<liar> a little bball... i can think of lots of things to do
<eihrul> but programming something, better or not, is still more fun IMHO
<liar> i dont like to program
<eihrul> umm??
<eihrul> are you sick or something?
<liar> i am making brix so i never have to program again but i will be able to make lots of programs
<eihrul> mentally? :)
<liar> thats the power of brix
<eihrul> apparently years of nissan commercials have been lost on you :)
<eihrul> "Enjoy the ride."
<liar> oh i am enjoying the ride
07:50pm
<eihrul> i would hate to have my programs made completely for me :)
<eihrul> simply because i wouldn't get to make them
<liar> eh?
<eihrul> <liar> i am making brix so i never have to program again but i will be able to make lots of programs
<liar> right
<liar> do u know how brix works?
<eihrul> kind of sort of
<liar> its UI is beyond anything u have seen
<eihrul> a few menus? :)
<eihrul> i like my xterms much better, thank you
<liar> no menus
<eihrul> then what constitutes the brix UI?
<liar> u can design how stuff looks
<eihrul> how so?
<liar> u can configure the window
<eihrul> i can do that in X11 :)
<liar> no u cant
<liar> try to get rid of buttons on netscape
<liar> try to add buttons
<eihrul> well, could you possibly be a bit less vague and a bit more descriptive?
<eihrul> about what exactly makes the UI better?
<eihrul> so that i could know why :)
<liar> ok let me give u a little background info
<liar> for a text object u have a show method that will render the text inside a view
<liar> there is also a show document for the text object that will create a window, a view and then call the show method
<liar> the document can hold code and objects inside it and also reference external objects
<liar> so u can edit that text show document to add buttons or other controls to the window
<liar> u can add a find button and link it to a text find method
<liar> u can put that button anywhere u want
<liar> u can also put other documents inside documents
<liar> and u can do all this by dragging stuff around
<hcf> ole ;)
<liar> hcf: not quite
<eihrul> kind of like smalltalk ide in a way :)
<liar> u guys can make all the comparisons u want between brix and smalltalk but brix is a million times faster than smalltalk
<eihrul> i wasn't commenting on the speed
<eihrul> considering brix has JIT compilation
<eihrul> of some sorts
<liar> so either i am a brilliant coder or brix is nothing like smalltalk
<eihrul> i'd assume it's very much so faster
08:00pm
<liar> brix doesnt have jit
<eihrul> that's what your docs said
<liar> i have never used the word JIT
<eihrul> well, compiled
<liar> jit is right before u run it
<liar> brix compiles when it downloads the method and never again unless needed
<eihrul> which is why it's faster than smalltalk
<hcf> ok, enuf w/ the smalltalk, now compare it to something else that will annoy liar ;)
<liar> heh
<eihrul> okay, java!
<liar> noooo
<eihrul> or visual basic!
* eihrul/#tunes cackles.
<liar> nothing like vb
<eihrul> i know... but hcf made me do it
08:10pm
-:- water [water@ppp-ip140-c11.tscnet.net] has joined #tunes
<water> hey
<hcf> water!
<water> hcf!
<eihrul> fire!
<water> what's up, hcf? where are tril and fare, etc?
<hcf> water: uhm, not here for reasons i'm not sure of
<water> k
<water> my laptop is still down for the count
<water> :(
<water> although my hard drive is backed up
<water> any news for Tunes? is fare finally getting this act together?
<Mr_Wrong> Fare officially turned the project over to me, Water.
<hcf> things have been pretty dead lately
<water> whoa
<water> mr_wrong: huh?
08:20pm
-:- Tril [dem@bespin.dhs.org] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
<water> tril!
<Tril> water!
<Tril> did you call me from California?
<water> ?
<water> no
<Tril> ok
<water> i also was too busy to meet with maneesh :(
<water> well, we were both too busy
<Tril> i got a job at NAS the ISP bespin is hosted at
<water> ok
<Tril> now I have a key I can go in and fix server if no one is there :)
<water> tril: got the news from jecel?
<Tril> no
<water> hmm
<Tril> self for linux is out?
<water> not yet
<water> jecel's now the president of a company
<Tril> oh
<Tril> yes, but... is anybody else in the company? :)
<water> he's gave the specs for his new pc
<water> yes, he's got a few workers
<Tril> he's trying to build those Merlin hardware boxes huh
<water> yep.
<water> they'll be very modern, apparently
<Tril> what's the url?
<water> he's going to use a Xilinx FPGA and some flash-rom(?) chips to emulate Self/R or whatever
<water> his website is completely not informative
<water> but here's the url
<water> http://200.210.69.43
08:30pm
<water> the best info was in his email to the self lists
<water> oh yeah, he also will include co-processors to handle i/o and graphics and sound
<water> i'll forward the e-mail to you
<Tril> ok
<water> anyway, my laptop has crashed hard
<water> paper-writing has gone on, but "unplugged"
<Tril> hm
<water> the "reflective arrow logic" paper is getting pretty detailed
<water> although i don't have any good formally-proven results about it
<water> most of my proofs are just lemmas
<water> but then, that makes it simpler to explain
* water/#tunes just learned of a newly published paper on LANL that relates quite well to tunes, it seems
<liar> water: can u forward that email to me?
<water> liar: sure
<water> liar: email again? (?)@qzx.com
08:40pm
<liar> brand
<water> ok. it's on its way
<liar> hmm i have all the standard emails @qzx.com except for god. i need to go make that one right away :)
<water> Tril: how're Fare and yourself?
<Tril> I haven't seen fare for a while
<Tril> I'm ok, work on type junk progressing slowly, as usual
<ultima> If it ain't tril.
<ultima> Hi.
<water> want to meet in seattle for a brainstorming session?
08:50pm
<hcf> water, Tril: icuc, http://www.cs.bham.ac.uk/research/xslam/xslam.html
<liar> hmm jecel is gonna pay apple for firewire
<water> yep
<liar> 640k cache???
<water> i guess so
<liar> i would think 512k would be easier to come by then 640
<liar> and 10ns is kinda slow for cache
<water> it's for an FPGA, though
-:- The_Groove [jam@ldsl30.dnvr.uswest.net] has joined #tunes
<The_Groove> ultima: can you have something for me tomorrow?
<water> liar: the cpu will only run at 100mhz
<Tril> oh, I gues it wont run Win2000, no reason to buy one
<water> hehe
<liar> well it sounds like it will be overclocked to 100mhz :)
<Tril> ok..water are you back for a while?
<water> i have about four weeks, then three weeks out again, then a year and a half in
<Tril> ok
<water> i might be going to a school next week, though
09:00pm
<Tril> hi ultima find that disk? :)
* Tril/#TUNES is away: (E-MAil me if you want to talk, I will not sit around on IRC waiting for it thanks!) [BX-MsgLog Off]
-:- SignOff Tril: #TUNES (Tril has no reason)
09:10pm
-:- SignOff water: #TUNES (Ping timeout for water[ppp-ip140-c11.tscnet.net])
-:- water [water@ppp-ip95-b30.tscnet.net] has joined #tunes
-:- SignOff The_Groove: #TUNES (ircII EPIC4pre2.004-19990718 -- Accept no limitations)
<liar> whats a hula pointer?
<eihrul> a pointer to a hula!
<liar> its some kinda pointing device
<liar> used by notebooks
<eihrul> i only know of the stubby things
<eihrul> the trackballs, and the touch pads, and the serial mice
<eihrul> but not a hula pointer :)
<eihrul> unless the stubby things are hula pointers
<hcf> Hula Point cursor pointer from Usar Systems ...
<hcf> the Hula Point, based on Fujitsu magnetic field detection Hall effect sensor technology,
<liar> and?
<hcf> thats all the page i was looking at had for info, more coming
<water> the Hall effect? cool!
<water> i know what it is, then
<liar> its not that joystick
<water> it's a sort of a puck with an electromagnetic coil that detects motion based on induced voltages
09:30pm
<liar> i have never seen a puck pointer
<water> it's just like a mouse without a ball
<liar> on a magnetic mouse pad?
<water> i think so. i remember using a Hall-efffect mouse with my old CAD workstation
<liar> i know what hall effect is
<hcf> hula iz also a fingertip mouse
<liar> my final project for my Associate degree used em
<hcf> info + pics @ http://www.stealthcomputer.com/peripherals_oem.htm
<water> well, gotta go for now
-:- water [water@ppp-ip95-b30.tscnet.net] has left #tunes []
<liar> water: and yer cad mouse was probably optical
<liar> optical has much better resolution than magnetic
09:40pm
-:- rend [rend@216-119-133-166.ipset13.wt.net] has joined #tunes
<rend> does tunes have a website?
<hcf> rend: see the topic?
<rend> -:- #Tunes  There isn't a topic.
<hcf> oh
-:- hcf has changed the topic on channel #Tunes to: Reflective Computing System <http://www.tunes.org/>
10:00pm
<hcf> see it now?
<rend> what i see now is a broken 2 gig hd.
<rend> a broken pro 100 intel card
<rend> and a broken pentium box
<rend> track 0 bad, disk unusable
<rend> so, what is actually completed on tunes?
-:- SignOff hcf: #TUNES (Leaving)
10:10pm
-:- rend [rend@216-119-133-166.ipset13.wt.net] has left #tunes []
<liar> 9x or nt that is
<liar> oops
<ultima> busted =)
<ultima> hey liar why did you change your nick ?
<liar> cuz
<liar> do u not like it? :)
<ultima> Heh
<ultima> I'm neutral on that ground =)
<liar> i changed it cuz of abi
<liar> abi: nickometer _QZ
<abi> '_QZ' is 23% lame, liar
<liar> abi: nickometer liar
<abi> 'liar' is 0% lame, liar
<ultima> nickmeter ultima
<ultima> nickometer ultima
<abi> 'ultima' is 0% lame, ultima
<ultima> even
<liar> u are 0
<ultima> What is this lameness?
<ultima> nickometer abi
<abi> 'abi' is 0% lame, ultima
<ultima> abi is lame
<ultima> nickometer abi
<abi> 'abi' is 0% lame, ultima
<liar> its the letters
<ultima> More of this stuff...I remember it used to have Karma 
<ultima> nickometer _m00
<abi> '_m00' is 45% lame, ultima
<ultima> nickometer _m00^
<abi> '_m00^' is 61% lame, ultima
<liar> uppercase and other non-alpha chars
<ultima> nickometer _m00^__AaA
<abi> '_m00^__AaA' is 99.83% lame, ultima
<ultima> nickometer _m00^__AaA%
<abi> '_m00^__AaA%' is 99.89% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$
<abi> '_m00^__AaA%@*#&$' is 99.9824% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$;;;
<abi> '_m00^__AaA%@*#&$;;;' is 99.9907% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$;;;*@#&$&*@#$
<abi> '_m00^__AaA%@*#&$;;;*@#&$&*@#$' is 99.9955% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$
<abi> '_m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$' is 99.9962% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$(@*#$@*#&$@#&$#&$@(*#&$(@*#&$(@*#&$(@$*
<abi> '_m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$(@*#$@*#&$@#&$#&$@(*#&$(@*#&$(@*#&$(@$*' is 99.9968% lame, ultima
<ultima> nickometer _m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$(@*#$@*#&$@#&$#&$@(*#&$(@*#&$(@*#&$(@$**!@)#(*UBSDFBSDF9234hU23bn4____23-490i424;4l@#4898sdf9ya89u98`97`*&!~*793287
<abi> '_m00^__AaA%@*#&$;;;*@#&$&*@#$*(@#&$@#&$(@*#$@*#&$@#&$#&$@(*#&$(@*#&$(@*#&$(@$**!@)#(*UBSDFBSDF9234hU23bn4____23-490i424;4l@#4898sdf9ya89u98`97`*&!~*793287' is 99.9978% lame, ultima
<ultima> damn
<ultima> karma
<ultima> karma ultima
<ultima> no more karma command?
<liar> nickometer @!&*#^(&^!@#%(@!^#%@#%^@#&@!%#&%@%&**&%@!%*(@!%#(*&@^!%#&^((&*@!^#*)&@!^#&*^@&JHGJH#GVJG*&#CG@!#GC@!^#TC*^@!#T*C@T!@C#T@!^*#TBC^*@!G#^C@!B#^!@(#^C#@#@!C^#!@(*C&#C!@#&(T@!C#@!*#&(TC!@*&#T&C!@T#*&!@TC#&T!@#*(&CT@!#*&T@!*&#C!@B&*#C!@#
<abi> '@!&*#^(&^!@#%(@!^#%@#%^@#&@!%#&%@%&**&%@!%*(@!%#(*&@^!%#&^((&*@!^#*)&@!^#&*^@&JHGJH#GVJG*&#CG@!#GC@!^#TC*^@!#T*C@T!@C#T@!^*#TBC^*@!G#^C@!B#^!@(#^C#@#@!C^#!@(*C&#C!@#&(T@!C#@!*#&(TC!@*&#T&C!@T#*&!@TC#&T!@#*(&CT@!#*&T@!*&#C!@B&*#C!@#' is off the scale lame, liar
<liar> HA
<ultima> Heh
<ultima> nickometer !@#*(&89y3129uh3(!@#*&!(@&*#(!82739817@#&!@#$ &$)@#$&@$)@&$@*$&@(T&@CT!)@TR&TE@T#)$^TT~$#^~T!^T!$^T$^T~@^#T$)@^T$^&@T$&^@T~#$T@^T~#$^T@^T)~@TT#^T)~@#$(@*#$(@#$_@*~#$(*_@#~*$(@*#&$_*@&#$_#&$^~@#*$@#&$&@^#$&^@#~H(UHE{~@*#&E*~@&#$_(*&~@#*$&@~#Y$U@#H${~@#U${~@#U$*@&#$_~@&#$~+@#&$_~(*@&#*$_(&_~@*#$&@~_(#&$~_@*#&$_&*~H@#HU${~@#*$~@#&$*~(@&#$Y@U#H${@~UI#H$~{*(#$+@&*#*_($&@~(#Y$I~UH#{$OHUI~@{#H$~@#*$&
<abi> '!@#*(&89y3129uh3(!@#*&!(@&*#(!82739817@#&!@#$' is 99.9959% lame, ultima
<ultima> Whoa there
<ultima> wtf
<liar> haha
<liar> u had lowercase
<liar> nickometer *&&%^
<abi> '*&&%^' is 99.932% lame, liar
10:30pm
<ultima> nickometer *
<abi> '*' is 14% lame, ultima
<liar> nickometer *&&%^U
<abi> '*&&%^U' is 99.9321% lame, liar
<ultima> nickometer **
<abi> '**' is 38% lame, ultima
<ultima> nickometer *!
<abi> '*!' is 38% lame, ultima
<ultima> nickometer *!-
<abi> '*!-' is 97.05% lame, ultima
<ultima> nickometer *!-#
<abi> '*!-#' is 99.85% lame, ultima
<ultima> nickometer *!-#^
<abi> '*!-#^' is 99.932% lame, ultima
<ultima> nickometer 1
<abi> '1' is 22% lame, ultima
<ultima> nickometer 1
<abi> '1' is 22% lame, ultima
<ultima> nickometer !
<abi> '!' is 14% lame, ultima
<ultima> nickometer 1234567890`-=~!@#$%^&*()
<abi> '1234567890`-=~!@#$%^&*()' is 99.9934% lame, ultima
<ultima> nickometer 1234567890`-=~!@#$%^&*()QWERTYUIOP{}|[]\;':",./<>?
<abi> '1234567890`-=~!@#$%^&*()QWERTYUIOP{}|[]\;':",./<>' is 99.9967% lame, ultima
<ultima> hmmmmmm
<ultima> I'm just not lame enough, apparently.
<liar> i didnt want to be 23% lame so i changed my nick to queuezee but that was too long and annoying me cuz it made everything i said stick out to the right more so i changed it to liar
<liar> damnit, this stupid persistence thread is making it hard to understand my scheduler
10:40pm
<eihrul> what's wrong with your scheduler? :)
<liar> nothing
<liar> its just my persistence thread
<liar> its not a normal thread
<liar> it has no dpl3 stack cuz its the only thread that runs in cpl0 and it also has not entry in the active thread list
<liar> it is also not in any osp and has no page directory
<abi> okay, liar.
<eihrul> abi: shut up
<liar> forget it
<abi> liar: I forgot it
<eihrul> stupid bot
10:50pm
<ultima> dumb dumb dumb
<eihrul> i vote we kill abi\
<liar> well i guess i have to increase the size of my scheduler by about 50 bytes to keep the persistence thread crap from slowing down normal switches
<eihrul> but we just need to figure out how to bring her alive first
<liar> hey do u know howto save the fpu stack and state?
<ultima> I second that vote.
<eihrul> i haven't gotten there yet :)
<eihrul> try me next week
<liar> heh
<liar> i allow the user code to tell the kernel that it uses the fpu so the scheduler will save/restore fpu for it, and the scheduler even checks to see if the thread uses the fpu so it can save/restore but it is only missing that save/restore code :)
<liar> and i have never used the fpu in my life
11:00pm
<eihrul> there should be ways to just detect if the FPU is used
<liar> how?
<eihrul> not sure yet... but probably some flags set somewhere
<liar> well i like my way
<eihrul> well autodetecting it should be simple :)
<eihrul> there's actually an interrupt i think
<eihrul> for lazy FPU switching
<liar> hey if the user doesnt tell me its gonna use the fpu then im not gonna save/restore it :)
<eihrul> but interrupts can :)
<eihrul> i remember reading that in the intel manual
<liar> uhh
<eihrul> basically... once another task is switched in
<eihrul> it can generate an interrupt before the other task uses the FPU
<liar> i dont use tasks
<eihrul> tasks, threads, whatever you want to call it
<liar> i dont use tss's
<eihrul> but it shouldn't be bound to tss's
<liar> well my way is much simpler than that way
<eihrul> how is it much simpler?
<eihrul> having the CPU telling you it's been used
<eihrul> versus having the user tell you it's been used :)
<liar> the user code calls a kernel function that sets a bit in the thread entry (SIMPLE). the scheduler checks that bit and saves/restores (SIMPLE)
<eihrul> so you can make the lazy fpu interrupt thingy set the bit
<liar> or i can setup some stinkin exception handler that will save/restore it
11:10pm
<liar> anyway until i have the details of the lazy fpu way, it remains very complex and difficult
<liar> OK, brix should be working perfectly now
11:20pm
<liar> UGH, i hate it when i open the same file in 2 windows and edit both :(
<eihrul> haha
<eihrul> i just did that
<liar> eihrul: do u know howto make nasm display an address while assembling the code?
<ultima> I just use one xemacs session =)
<ultima> liar you can make nasm dump a listing but there is no 'interactive display' of any sorts
<liar> damn
<liar> im getting tired of inserting 'jmp short $', assembling, running bochs, getting address, removing the jmp, reassembling and rerunning bochs
<ultima> just dump a listing
<ultima> that should do what you want
<ultima> it shows the addresses of everything
<ultima> nasm -l outfile
<liar> that would take too much time
<ultima> It would take more time than running bochs?
<liar> bochs runs brix quite fast
<liar> it takes about 1/2second to run brix in bochs
<liar> brix is a highly optimized OS so it runs fast on a 2mhz 386 :)
11:30pm
<liar> bochs: panic, exception(): 3rd exception with no resolution
<eihrul> gah... sure, you leave out half the features of more conservative systems
<eihrul> and it's easy to call it optimized :)
<liar> eihrul: what features?
<liar> the brix kernel has more features than any other uk
<ultima> uh oh...
<liar> and its nly 9k
<eihrul> well mine is 3.5k so far :)
<eihrul> but it's half implemented :)
<liar> name a feature u think i have left out
<eihrul> address space switching
<liar> eh?
<ultima> heh
<ultima> my caffiene has run out
* liar/#tunes never had any caffiene to start with
<ultima> I have all of the mov/push/xchg/pop opcodes
<ultima> liar: I did. And now I'm crashing. b00m.
<liar> am just crashing
<liar> s/am/im/
<liar> eihrul: what is that?
* ultima/#Tunes has learned a lot tonight though.
<liar> eihrul: and why would i wanna do it
<ultima> An incredible amount.
<ultima> I've decided the direction I'm going to take UltOS at least =)
<liar> i got an incredible amount done on brix in the last 3 nights
<liar> i got it doing a nice triple exception
<ultima> not bat liar.
<ultima> bad even.
<eihrul> triple exceptions? :)
<eihrul> that's a lot of nesting going on
<eihrul> what happens when you run out of stack?
<liar> eihrul: so why would i wanna switch address space?
<ultima> goodnight all...
<ultima> see ya tomarro
-:- SignOff ultima: #TUNES (Leaving)
<liar> u get a page fault
<liar> and u get killed
<liar> and then the kernel tells the language that it killed whatever method
<eihrul> i'm down to the last few glue functions
<liar> u still havent answered me
<eihrul> different processes
<eihrul> or because it's the masochistic thing to do
<eihrul> you decide
<liar> what
<liar> brix changes address space
11:40pm
<liar> thats why each osp has a pgdir
<eihrul> oh, ok 
<eihrul> that's good enough for me
<liar> i have pervasive multithreading
<liar> i have persistence
<liar> object management
<liar> lets see a normal uk handle that
<liar> do u have event lists?
<liar> and can u hook em to irqs?
<eihrul> i will have most of that but persistance and object management plus a few other goodies :)
<eihrul> when i'm done, of course... 
<eihrul> but right now i have near squat
<liar> what other goodies?
<liar> can yer kernel hook to other kernels?
<liar> didnt think so :)
<eihrul> ummm... but of course
<eihrul> that's not that hard to implement
<eihrul> you simply have to keep track of which tasks are local
<eihrul> which aren't
<eihrul> then you can simply route all messages to non-local tasks to a user level process that manages them
<eihrul> but again, that's all for the future
<eihrul> right now, i've just started :)
<liar> the way brix works it makes everything simple
<liar> all u gotta do is send messages and the kernel and ofs takes care of everything
<eihrul> again... that's not unique to brix
<liar> u dont care or need to know where shit is at
<eihrul> but more to message passing microkernels :)
11:50pm
<liar> no no no
<liar> brix message passing is better than uk message passing
<liar> they arent the same thing
<liar> uk mp is sending a message to a thread or process or whatever
<eihrul> neither does the kernel
<eihrul> nor the sending process
<liar> brix mp is sending a message to an object
<liar> brix sends its messages with fork() and call()
* eihrul/#tunes cackles.
<eihrul> just a rehash of the same 'ol thing (tm)
<liar> fork is async and call is sync
<eihrul> ult: you there?
<eihrul> er wait
<eihrul> he's not apparently :)
<liar> hah
<liar> ugh i need to sleep but want to finish the kernel tonight
<eihrul> trying to write an assembler for the dude's VM
<eihrul> he didn't even wait for me to finish
<liar> huh?
<eihrul> n'm
[msg(TUNES)] newlog 1999.0810
IRC log ended Tue Aug 10 00:00:00 1999