IRC log started Tue Jan 12 00:00:00 1999 שםש SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) [msg(TUNES)] permlog 1999.0112 שםש SignOff KeLp: #TUNES (sleep) שםש binEng [bineng@dialup55-1-22.swipnet.se] has joined #tunes שםש SignOff binEng: #TUNES (Leaving) שםש ultima [ultima@user-37kbb2r.dialup.mindspring.com] has joined #tunes שםש SignOff ultima: #TUNES (Leaving) שםש Fare [rideaufr@esmeralda.enst.fr] has joined #Tunes שםש binEng [bineng@dialup247-3-8.swipnet.se] has joined #tunes שםש SignOff binEng: #TUNES (Leaving) שםש hcf [nef@escher.sdi.agate.net] has joined #tunes שםש binEng [bineng@dialup243-3-46.swipnet.se] has joined #tunes hoy binEng hi 11:40am sup?~ reading the new hacker's dictionary... hard copy? nope web? yup :) url? anyway my point is http://www.ccil.org/jargon foldoc includes the jargonfile (aka new hacker dict) includes the jargonfile (aka new hacker dict) may be sought in foldoc at http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?query=includes+the+jargonfile+aka+new+hacker+dict+ mhm \foldoc also has the lang list and a bunch of other good poop compiled into it just goto http://wombat.doc.ic.ac.uk/foldoc/ 11:50am שםש SignOff binEng: #TUNES (Ping timeout for binEng[dialup243-3-46.swipnet.se]) שםש ultima [ultima@user-37kb9sm.dialup.mindspring.com] has joined #tunes hihihi who is here <-- 12:50pm שםש _QZ [brand@p0wer.qzx.com] has joined #tunes ky00 zee! <_QZ> hey que tal, _QZ <_QZ> i forgot to save what lilo said to me yesterday about u <_QZ> when i told him u wanted to be banned again :) btw, ppl, lilo is a she <_QZ> really? yep wot? <_QZ> u sure? what did he say <_QZ> he said something like "dont worry i will get rid of that little shit for u" by what iv seen being said in #linpeople day after day when i used hang ther, lilo is female hcf, no, lilo is a he <_QZ> im pretty sure lilo is a he hcf lilo used to pretend to be female oh whatever cus he thought people listened to females but he's really a male i've talked to him on the phone =) n/m נ ultima/#tunes notes _QZ is rather full of shit lately <_QZ> ultima: no im serious <_QZ> he really messages that too <_QZ> me <_QZ> it wasnt those exact owrds ultima: _QZ is always full o' shit i'm sure he does 01:10pm hcf this is becoming more and more evident especially when I took a look at the BRiX source hehehe <_QZ> u have never seen the brix source _QZ: Sure I have <_QZ> no u havent I have it right here <_QZ> i have not handed it out to anyone wot makes u think I havent just the cpuid poop So? rumour has it So is bin still helping with the web page abi: forget so hcf: I forgot so Decompiler dude <_QZ> ultima: hahahha <_QZ> ultima: u fool what does the version say? <_QZ> lite or nothing, version and build _QZ your so sure of yourself. Why dont you send me the complete source?> <_QZ> what binary do u have <_QZ> u think u have the brix source but u cant tell me that <_QZ> if its the one i gave to tril then u have nothing <_QZ> it has half the setup code and no kernel Then send me a BRiX.img that does something other than print the CPU type ok den, send me the sources <_QZ> not even שםש SignOff ultima: #TUNES (Read error to ultima[user-37kb9sm.dialup.mindspring.com]: Connection reset by peer) 01:20pm <_QZ> hcf: u here? yep <_QZ> i found a bug in the lite versions persistence <_QZ> i was using the modified bit in the page table entry to determine if a page of an object had changed <_QZ> but the lite version doesnt use page tables <_QZ> any ideas around that? probly nothing helpful <_QZ> guess the lite version will get dropped then 01:50pm <_QZ> having actions report to the kernel when they modify objects wont work 02:00pm שםש Tril [dem@bespin.ml.org] has joined #TUNES hoy Tril lilo is male, you can get to his home page from the About link on www.openprojects.net. under the anchor Robert Levin hi hcf I used irc too much yesterday, didn't get anything done lilo is female. was male. or am I baffled? btw, gakuk! if he had a sex change, why doesn't her home page say Roberta? err..wrong window hi Fare ask her! <_QZ> hehe <_QZ> === | ^lilo lilo@varley.openprojects.net <_QZ> === | ircname : Rob Levin) (open source, technology, info <_QZ> why does HIS ircname say Rob Levin? Rob[erta]? <_QZ> stop speaking thru the stupid bot sorry. accident hmm it's just another screen window. speaking thru abi requires /msg Hum. I don't dare announce my own article on /. <_QZ> why not? 02:20pm Fare: Have ESR comment it! שםש SignOff Tril: #TUNES (leaving IRC, doing actual work) QZ: good question QZ: because I'm shy? Because I dunno whether or not it's a nice thing to do? Tril: good suggestion 02:30pm שםש SignOff hcf: #TUNES (Ping timeout for hcf[escher.sdi.agate.net]) שםש hcf [nef@escher.sdi.agate.net] has joined #tunes <_QZ> what is yer article about? 02:50pm שםש Tril [dem@bespin.ml.org] has joined #TUNES _QZ it's about Metaprogramming, basically arguing that it is the next step in helping Free software development, but good info on metaprogramming examples too שםש SignOff Tril: #TUNES (Perl hacking) <_QZ> Fare: where is it at? 03:00pm my article? <_QZ> ya see on www.tunes.org the announcement (I also posted an announcement to the tunes list!) criticism as well as petty spelling/grammar/style fixes WELCOME <_QZ> i am not on that list oh which list are you on? Tril: there? <_QZ> tril is gone Tril: can you tell me whose quote is the bespin motd? Does anybody else find irony in the fact that Bill Gates says he's trying to write a stable OS, and Linus claims he's trying to take over the world? שםש Tril [dem@hh154ws31.hh154lab.wwu.edu] has joined #TUNES I'm here... that's from a comment someone posted on slashdot. an Anonymous Coward I think <_QZ> i am on tunes-lll and the dead os-help/ideas hum. Too bad. Tril: should I announce my article on /. ? Fare: would it get posted? do you mean to ask for help fixing grammar? that too I started but then I realized it would take too long and I'd never get to comment on the content have it read, also :) for your first version I thought the section on reflection was very unclear..did you improve it in the revision? Tril: you might just modify the file, and I'd make a patch! <_QZ> tell hcf about it and it will get corrected :) _QZ hcf is here hcf are you here <_QZ> i know Tril: it was not the section with the most improvements, I think. well, maybe you should post it on usenet before slashdot which groups??? 03:10pm <_QZ> damn thats alot of reading comp.programming.?? create a new group comp.programming.meta :) alt.programming.meta QZ: 17 pages, a lot? <_QZ> 81,000 letters <_QZ> hell yes its alot :) hmm, mozilla doesn't work right with wiki <_QZ> Tril: u have mozilla? as in ns5? after I remove the latex, that's 1758 10709 69207 <_QZ> well im gonna read it later _QZ: I'm running Mozilla/win32 from 01-09 nightly snapshot like I said, the linux binaries are only precompiled for libc6 <_QZ> does mozilla still look like ns4? I don't know if it looks like anything. All it has is a Back and Forward button, animation frames are just numbers 0-13, and the only menu options are stuff like "test widgets" and "javascript console" <_QZ> uhh what part of it do you want me to look at? <_QZ> well ns4 looked like ie4 <_QZ> with those stupid sliding panels <_QZ> and the email stuff was way different what sliding panels? oh docking stuff <_QZ> ns4 was a huge leap over ns3 in the interface area <_QZ> u know the back button and tat stuff was sliding <_QZ> u could move em around 03:20pm extra ui stuff is not in this version, it's just to test the layout engine it seems which is nice thers' no email or anything like that שםש cryogenic [cryogenic@128.163.16.33] has joined #TUNES שםש cryogenic [cryogenic@128.163.16.33] has left #TUNES [] anyone know how locking works? <_QZ> locking? file locking on unix? I mean, these functions say how to get locks and release them but it doesn't say what exactly they do do they prevent other programs from reading the file, or just writing? <_QZ> u make a LCK file in /var/lock <_QZ> and anything accessing files should check for LCK files <_QZ> :) what kind of locking is that? I'm talking about the flock() function. <_QZ> when u lock the file it should have a parameter that does read or write locking <_QZ> if not then it will let others read <_QZ> but not write it says nothing about reading or writing. is it both? <_QZ> the nice thing about brix is that file locking isnt needed :) QZ: that's purely a software onvention! not enforced by the kernel flock has lock_sh for sharing hcf: but it doesn't explain what a shared lock actually is!! what's the point of a shared lock if you can just read/write the file without a lock anyway and why is it called an "advisory lock", does that mean it only affects other processes who also call flock()? and if they don't then they can just ignore the lock 03:30pm <_QZ> no, linux locks the inode Ahem. Why bother, really? I'm writing a program, that's why <_QZ> for what? whaddyawanna lock? see also the end of the man page... it needs to ensure no other programs can read the file for a user id database for the projects tool I'll need to use the same locking for the projects db too this is just for adding new users Fare: What, the linux/Documentation/locks.txt? I looked at it, it just tells the history of locking in the kernel, it doesn't explain how to use locking. well, if there is no docs, I have to write a test program. s/is/are/ <_QZ> Tril: u just want to make sure it does add 2 user simultaneously? <_QZ> s/does/doesny/ <_QZ> ugh uhh..adding a user requires rewriting the entire file. I don't want any other programs reading the file at all until the file is rewritten <_QZ> but the only program that would write is the adduser program right? <_QZ> the web adduser No, the other parts need to modify it too. When a project leader adds a user as a member of their project, the project-edit program has to add to the user's record that they are now a member of that project <_QZ> but only YOUR programs access it correct? <_QZ> if so u can use software locking <_QZ> using LCK files yeah it runs as its own user, the files don't get accessed by anyone else well, i'm assuming the flock function in perl is going to do that for me, or something else, I don't care how it's done wiki uses its own locking by crewating a subdirectory and checking to see if it exists, but I think that's lame does kernel locking affect the root user? unix is lame, anyway Fare: go write TUNES, then. 03:40pm I had better use _QZ/wiki method because I need to open/close the file several times in different functions but keep the lock the whole time I meant, don't bother about the ugliness of the locking method -- it will be ugly, anyway! I meant, you should write tunes, anyway :P :( whynot 03:50pm שםש SignOff hcf: #TUNES (Leaving) Tril: have you had a look at core's stuff? I said, it locks on a 486. I don't have pentiums at home to run it on. tcn said it worked on his pentium so I have not been able to run clementine _QZ says it doesn't do anything שםש Tril [dem@hh154ws31.hh154lab.wwu.edu] has left #tunes [] שםש Tril [dem@hh154ws31.hh154lab.wwu.edu] has joined #TUNES <_QZ> it does but it doesnt uh? i heard uh was he??? <_QZ> abi: forget uh _QZ: I forgot uh Fare: I don't know what I'd do with it if it worked. Does it have example modules to try to load? Fare: what did you run it on? 04:00pm a cyrix (I admit I haven't rebooted my laptop for it) it gives you a prompt but there is no useful language after the prompt I guess a forth interpreter would be neat... tcn would be glad to write one for it I think I mean stick his he already wrote on top of it to delete cookies you set them to expire in the past? 04:10pm שםש SignOff Tril: #TUNES (Read error to Tril[hh154ws31.hh154lab.wwu.edu]: Connection reset by peer) שםש Tril [dem@hh154ws31.hh154lab.wwu.edu] has joined #tunes Who's peer, and why does he keep resetting Tril's connection? That's just NT which is crashing NT, crashing? How surprising! I tried to open a menu on the Start menu, but it froze the computer I had to reboot... oh oh, 246 messages in my tunes box... it actually did it again this time but I waited a few minutes and it went away !ChanServ:*! ^lilo used GETPASS on channel #revinc 04:30pm that's waht you get when the Start menu contents are stored on a Netware server NT, Netware? Are these guys kind of masochistic? lemme guess, they also have VMS installed in the basement, and use VM/CMS for accounting? Tril: what are your symptoms w/ clementine? Fare: Detects hundreds or thousands of extra megabytes of memory, then hangs at the first line of text: "Clementine loader 0.1 (XXXX MB memory)" disk light is turned on when it hangs. for tcn it also prints a ".." and sometimes a heart character, and hangs with the light off. But it worked fine on his Pentium. so we decided it was some kind of 486 problem since it failed on all 486s is it normal to be unable to remove cookies without closing the browser? שםש Beholder [beholder@ppp-158.m2-3.sub.ican.net] has joined #tunes B! T! :) How are you tonight tril Fare: is "make config; make dep" really required before "make doc" to update the web page? yup Beholder,OK working on cookies in perl for the subprojects script. though their result shouldn't change often. Fare: what does junk.script update_sweety do? update html/ to sweety, or update ~fare/ from sweety? why didn't it print anything, or do anything, when I ran it in the middle of the day? Tril: because it uses LOG 04:40pm so see in ~/tmp/junk.log Fare: OK I want to be able to update the web page immediately when I make a change sometimes Sounds interesting. I'd like to learn Perl one day I keep ~fare stuff well separate from ~tunes stuff so I run (as tunes) junk.script update_tunes it just takes so long to do it Tril: if you didn't change anything in the main directory, rerunning make config is a NOP. depend? When will the new page be the tunes home page? If you didn't create any new file or modify the dependency script, rerunning make dep is a nop too B: it doesn't look complete yet! :) Beholder: which new page? Tril: NP, is anyone working on the sub pages? what do you mean fare: I think he means the ~bineng Tril: The ones in the nav bar. I don't hink they link yet. Beholder: are you officially a member yet? B: do you have write access to Tunes? B: did you read my comments in http://www2.tunes.org/~bineng ? B: Nope, nobody has worked on it since bE made it as an example. Be: don't *ever* assume anyone is working on anything. Beholder:fill out www2.tunes.org/poll.html to become listed on the members page :) B: if no one openly says something, and no change was made recently, then no one works on it B: besides, you can feel the changes on CVS... Fare: Ok I saw your comments.. Didn't see them before cuz I always went to the tunes.html it looks like tcn has volunteered to summarize the tunes-lll list (he put a note to that effect on the wiki?TODO page) Tril: Signing up. Beholder: you can change the stuff asked for, too, if you want. I'll modify the script to suit beholder: to know when progress happens, just to a cvs -z 9 update once in a while... I don't think Beholder knows CVS yet... Tril: Nope, if you could explain it, I will try ok, what system are you going to run it on? do you have a unix machine closer to you than bespin? ok you don't need unix you can actually use cvs for windows, I believe. might be more useful that way to have your copy on a local machine Hmm... not that I know of.... although I'm working on it 04:50pm ... on getting a Unix machine what, installing linux on your home machine? :) Beh: you mean, you use windows??? Fare: Yes, I have a family who shares the computer... it's hard to get them to use anything but Windows Beh: FIPS and LILO can help -- multiboot on shared harddisk Fare: Even a multiboot tends to screw them up... even if there is a default... :( even if it is a 1 second timed default :( (I like the removeable IDE HD solution, too) LOADLIN can help, too! I installed Linux as a system debugger on my uncle's machine. He has no problem with that 0 second timed default. just press ALT to get the LILO prompt Fare: What OS's do you run? Fare uses Red Hat Red hat only? No windows at all? Linux, of course windows? what's that? I'm trying CVS for windows right now. It seems to be working Tril: Where can I get it? Tril: CVS... not red hat :) I *never* use windows. When someone in my family has it booted on one of the computer where it still is, I just click on the nice penguin icon, and there am I! abi, wincvs is at http://download.cyclic.com/pub/macintosh/maccvs3.0/wincvs.zip Just what use is windows? besides, there's WINE (that I never use, either) 05:00pm I trained my father to use LaTeX Fare: Games mostly. And my Office Apps. Don't like the X fonts anyways... they look terrible and I'm training my little sister to use Scheme (currently, MacGambit) there are *plenty* of X fonts among which to chose. You may now even use TrueType fonts under X There are office Apps for Linux, too (unless you use Vietnamese, like my mom) and Games too, and there's WINE, too. you have no excuse for supporting M$. Fare: Still can't play half-life in linux... :) An OS cannot make a dent in the home market without the eye-candy Fare: Be realistic Wine sucks... It is only purpose is for people to runn MS Office on a linux machine wine is very good it is progressing rapidly I also have no use for windows, but am forced to use it at the computer lab eye candy is for the brain impaired. Beholder: did you get wincvs yet? I'll tell you how to use it.. You are much more productive when you don't have to even *think* about placing icons on your desktop Tril: Still coming in (Mac is EVIL) נ Tril/#tunes ignores fare your strong opinions don't matter, Fare every user should have whatever they want on their computer Fare: Mac rules :) ahh I found a good cookie library to use sounds like an oxymoron :) besides, WM, E, GNOME, KDE, provide for plenty of eye-candy for those who like Fare: stop bashing current software and go read your tunes mail! :) even if you are completely right :) Fare: You live your life at the prompt? The human brain processes visually and conceptually... therefore a (well done) GUI enviroment is better. I believe all programming should eventually be totally visual (at least flow, and layout) Beholder, I guess the current CVS tree is defunct, if we are using a redesigned one now. B: not true, many people think literately (using symbols and words) instead of visually 05:10pm Tril: Yes there is that OTHER half of the brain :) I guess a comprimise is necessary at least, I think they do. I know I am very graphical :) B: http://www.performancecomputing.com/features/9809of1.shtml people *feel* visually. They *think* symbolically. If you want to deceive people, show them images. If you want to help them think, give them text. Why redesign a CVS tree? At least, there should be a public criticism of the current one before you do (if you do) Fare: What is the difference between thinking and feeling? Beh: it's the difference between a german who votes nazi, and one who does not. [Note: by Godwin's law, the conversation should stop] Who is Goodwin, and why must he interupt my conversation? I'm going to have to go now, but I will be back in about 2 hours if anyone is still on then. נ Beholder/#tunes is AFK 05:20pm OK I'm going home שםש SignOff Tril: #TUNES (Read error to Tril[hh154ws31.hh154lab.wwu.edu]: EOF from client) נ Fare/#Tunes is away (sleep) 05:30pm שםש SignOff Beholder: #TUNES (Read error to Beholder[ppp-158.m2-3.sub.ican.net]: Connection reset by peer) שםש ultima [ultima@user-37kbb1c.dialup.mindspring.com] has joined #tunes <_QZ> hola hello kyoo zee נ ultima/#tunes kicks lilo <_QZ> ok den, send me the sources <_QZ> <_QZ> not even _QZ its cus you dont have any written dude u got like a cpuid and a print function <_QZ> ya that must be it ya it must be send me a functional image at least <_QZ> ok let me turn everything off and make an image ;) ? <_QZ> i have this nifty config file full of defines <_QZ> if they arent defined then sections wont get assembled so define everything <_QZ> how about u wait i dont feel like waiting 2 years while you write some code =) <_QZ> sorry then <_QZ> if i have no code then i have nothing to send u <_QZ> :) ok then just send me wot u dont have =) <_QZ> ok Cute. 06:20pm <_QZ> :) Dude send me a legit image <_QZ> that file was 0 bytes, not 1 byte <_QZ> how did bx get 1 byte i dunno 0.0009766kb heh it must use a float or a double and it loses that in precision <_QZ> do u know where the setup/kerenl begins in my image? i assume setup begins after 512 <_QZ> heh wrong then again u got a shitload of nul's in this image <_QZ> thats cuz the fs <_QZ> those are empty tables 2800 ? <_QZ> 2800? yeah thats where the nul's end 2DC6 seems like your main or something a bunch of register twiddling then a deal of calls then u play interrupt juggling and stuf hmm cpu detection drive detection? more calls then hardlock <_QZ> hehe <_QZ> as u can see yer decompiler doesnt help much :) <_QZ> if only u had the real source in front of u well if I spent a few hours cleaning this up i'd have somethin very readable <_QZ> u still wouldnt have crap nope cus BRiX isnt even crap =) anyway 06:30pm <_QZ> all i can say is that u will be impressed after running brix, typing in its console(text,vga gfx or vesa gfx), shutting down, and having the console retain everything u typed the next time u boot <_QZ> as soon as i get it to do other impressive things then u can see it o so it stores the screen memory on the disk? <_QZ> no is it like lisp or something where the 'kernel' is just a core file? <_QZ> it saves the console as text im supposed to be impressed? =) <_QZ> it reprints it in text mode and redraws in gfx mode <_QZ> i was last night I thought you had a working kernel and everything <_QZ> ya can it run any programs like a compiler? <_QZ> but last time i checked a working microkernel didnt do anything impressive so wots this persistance patent you were whining about cus u really cant patent persistance prior art <_QZ> :) So i'm pretty sure you were bullshitting =) 06:40pm <_QZ> maybe i was about the patent but not about what lilo said <_QZ> besides brix is a free OS, and patents are no good if u dont protect them well <_QZ> but u cant protect them if its free thats the problem though <_QZ> free source that is <_QZ> open source if osmeone patented your exact algorithms they could fuck u ober <_QZ> i should use use open instead of fre <_QZ> no Well I'm working on a license its gona kick ass <_QZ> patents wont hold up if existing work used the contents of the patent See, iuts gonna be like GPL but its gonna have clauses like 'If you read this source, you can not patent an idea contained inside of it' yeah but the problem is with sw its VERY hard to prove its existing work unless you have a userbase linux is one thing <_QZ> if u put it on something like sunsite or some other large server then it will hold up <_QZ> since the file is dated true I would get the source notarized <_QZ> i could even put it on my isp <_QZ> since i do not have root access and cant play around with the clock then the dated file will hold up in court ah not bad LYNX is compiled and installed <_QZ> ofcourse it cant be a windows server cuz we all know how easily they could be hacked and.... :) שםש SignOff ultima: #TUNES (Ping timeout for ultima[user-37kbb1c.dialup.mindspring.com]) 06:50pm שםש ultima [ultima@user-38lc7fo.dialup.mindspring.com] has joined #tunes sorry <_QZ> u got any ideas on howto do what the dirty bit in page tables does without using paging? <_QZ> and without the user code telling the kernel that it has modifed soemthing 07:00pm uh use a non-paged page table of course dont store the dirty bit on the page thats how FBSD does it, it keeps a page table in kernel memory that cant be paged out <_QZ> non-paged page table?? yeah then when a page is accessed a fault is generated I'm assuming your looking at the problem of keeping track of which pages are in use, but you dont want to page all of them in to check? <_QZ> no oh hrm, whats your problem then? <_QZ> i use the dirty bit for persistence ok, so you can save which pages are in use? <_QZ> but i have 2 kernels, one with paging and one without Ah <_QZ> i was making one without for speed why not just save the entire memory space? <_QZ> but there is no reason i cant use a single page dir for all apps I mean, the kernel is gonna know whats allocated and whats not <_QZ> the normal kernel swaps page directories when it switches threads nod <_QZ> and that makes it somewhat slower that makes sense <_QZ> but if i used a single table in the lite kernel i can still use the dirty bit and have speed too <_QZ> and i thought i might have to scrap the lite kernel then why dont u? just keep track of which thread each page belongs to <_QZ> oh crap maybe 2-4 byte overhead per page, make big pages <_QZ> u had to keep talking didnt u <_QZ> :) of course I did <_QZ> this stupid lite kernel is nothing but trouble ditch it u dont need 2 kernels <_QZ> and everyone thinks paging is soooooo hard just make options to turn on and off specific parts paging is fucking easy <_QZ> but its the non-paging kernel that is a bitch to make Oh? שםש Tril [dem@xws112.xtrn03.wwu.edu] has joined #tunes <_QZ> i havent run into a single problem with the normal kernel hi tril re ultima 07:10pm <_QZ> ok if i double the size of each page table so the first 4k holds the page table and the second 4k holds a threadID of each page <_QZ> that should work right hmm well I dont liek that howbig are your pages? <_QZ> 4k ah nevermind, I do like that <_QZ> a page table holds 1024 page pointers <_QZ> so an extra 4k per 4meg so, 8k per 4mb? thats pretty good <_QZ> well u have to have 4k per 4meg for paging <_QZ> that adds an extra 4k wow This is kinda weird <_QZ> anyone know the exact time it takes to flush the cache? oh shit <_QZ> eh? how are you flushing it? <_QZ> changing cr3 oh <_QZ> it invalidates every entry hmm no ide idea... nod