IRC log started Sun Nov 7 00:00:01 1999 [msg(TUNES)] permlog 1999.1107 -:- SignOff hcf: #TUNES (Leaving) -:- NetSplit: varley.openprojects.net split from asimov.openprojects.net [12:23am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net] -:- Netjoined: varley.openprojects.net asimov.openprojects.net -:- abi [nef@bespin.dhs.org] has joined #tunes -:- Fufie [stig@tunnel-44-24.vpn.uib.no] has joined #tunes -:- lar1 [LarMan@dialup-209.244.105.246.SanJose1.Level3.net] has joined #tunes yer on late So are you Friends are over no im always on this late friends are over so u run for irc? Heh One is installing Linux on his box did u get it fixed on yer box? Ugh, no I have installed 3 times on my server tonight I am going to retreive my data from linux and reinstall debian "So, when you typed in the date, it exploded into a sheet of blue flame and burned the entire admin wing to the ground? Yes, that's a known bug. We'll be fixing it in the next release. Until then, try not to use European date format, and keep an extinguisher handy." :) have u heard of pickOS? No some guy asked about it on comp.os.research. he says it has a hashed filesystem for storing stuff like xml docs 01:10am Hmm, sounds cool I would like to learn more about that reflective fs that water was talking about BRB - I need to reboot (winstupid) -:- SignOff lar1: #TUNES (Leaving) 01:20am -:- SignOff eihrul: #TUNES ([x]chat) -:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) -:- binEng [bineng@e161.ryd.student.liu.se] has joined #tunes -:- NetSplit: adams.openprojects.net split from sterling.openprojects.net [04:32am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [adams.openprojects.net] -:- Netjoined: adams.openprojects.net sterling.openprojects.net -:- Fufie [stig@tunnel-44-24.vpn.uib.no] has joined #tunes -:- binEng [bineng@e161.ryd.student.liu.se] has joined #tunes -:- yoo [yoo@user-37kbas3.dialup.mindspring.com] has joined #Tunes * yoo/#Tunes sings some tunes -:- SignOff yoo: #TUNES (Leaving) -:- Kaufmann [Kaufmann@dial865.infolink.com.br] has joined #tunes Shalom! Is anyone awake? hi I'm reading about the language Escher ? 06:10am Aaaairgh, Nutscrape just crashed '?' what? Are you using the latest version? 4.7, right? Yeah, 4.7 ? = Escher? I've gotta learn about what's out there, ya know.. so what's Escher? It's some kind of functional language. Quite recent, it seems like. Nice semantics? I haven't come so far yet. :) abi: escher? escher is a functional logic programming language at http://www.cs.bris.ac.uk/Research/Declarative/escher.html see. abi knows everything today is my father's birthday Congratulations to him then abi: Do you really know everything? bineng: wish i knew obviously not ;) we're going out to eat as always where is Faré's personal site? never mind we're going to a barbecue place... you've never had a real barbecue until you've had an Uruguayan barbecue 06:20am Uruguayan barbecue? I suppose I won't have that privilege for... forever. the American barbecue tradition is pathetic. It's based on burning random meats on big gas-pumping grills. I mean, they do HAMBURGERS! That's outrageous I guess. Since in Uruguay practically all they have to eat is cow, bovine butchering and cooking is a whole art What should the grill be like, then? It's hard to describe... I'll take a photo and send it to you :) the first error is using gas instead of coal 06:30am wow, life :) kauf: yes, it is rather hideous, sin't it? you don't just toss the meat into the fire; you put it on a support, distant from the fire, so that it'll cook gradually, while you splash water into it to keep the meat soft and make sure that the inside of the meat gets cooked too This is /very/ hard to talk about in English sounds like my method of water-cooking-with-wok Hmm, water? I haven't thought abourt that. Not some kind of grill-mix? I ahve the vegetables, and base (rice) at the bottom of the wok, and put thin-sliced meat at the top combination steaming/broiling/baking Wow, amazing... since last night I received 17 emails, none of them in anyway important, and only two worth reading Anyway, that sounds interesting, Zhi but wtf is a wok? hmm, its a cooking dish think of a round metal shield - like a buckler with a smallish bottom its fairly shallow. asian cooking stir-fry pot interesting cheap and versatiles, you can make stir-fries, or water-stirred dishes soups, stews, etc good for steaming since you can get a steaming pot which is a little larger than the base 06:40am Wow, this /is/ related to CS :) cooking science heh my father almost wound up getting a degree at the Cordon Bleu this was about five years ago So what's up? 06:50am hmm, not a great deal I thought so :) working on an os, but temporarily stopped :) what's /your/ OS like? heh :) ok, no kernel, no processes, single address space, persistenjt virtual memory system, trusted native compiler, non-forgable references. No processes? just threads well, given a single address space we don't have process sandboxes Seems pretty insecure? but beyond that we don't need a notion of a process anyhow so I put it in twice bin: well, that's the point of the trusted native compiler. bin: which compiles a nice safe lisp dialect into machine-code, and prevents interfaces from being violated bin: this of course, relies on the compiler being correct, which is a possibly dangerous assumption Zhivago, that's a nice concept bin: but probably humanly achievable 07:00am kauf: thanks, simplicify is my style :) The same here.. all the complexity here ends up in the compiler which is, naturally where I'm stopped on :) How about memory managment? That becomes pretty complex too hmm, I dont see why. my alisp compiler/vm has pretty much the same idea, but will run as a process on top of arbitrary traditional OS's at first; only later will it be extended to bootability kauf: alisp? alisp is not Scheme, but it's a helluva lot like it, and (although first it will run as a regular process on top of UNIX) I intend to make it bootable, eventually. see? hmm, ok, well, we have a system called mlisp which compiles to native code and can produce linux x86 binaries as well as x86 bootimages its a scheme derivative, but fairly primitive at the moment I'm working on raising the instruction set up to the function level wha'ts alisp like? Zhivago: Persistant virtual memory, sounds hard at least. big: its actually pretty simpel big: becuase we're also journalling it gets a little trickier big: but basically we just use disk as memory, and ram as cache I suppose you want performance too big: and demand page, and sync bin: well, the journalling is for security as well bin: since we always want a clean/consistent image to return to incase of a catestrophic failure. right bin: but it should have acceptable performance, expecially without a kernel layer or ipc to slow things down bin: turns context switches into register save/load's which is also a bonus bin: but essentially this system consistens of a contigious image of this form bin: boot+nonpaging+paging alisp is like Scheme, only even more functional-like - in fact, purely-functional programming is supported and encouraged; the compiler, being reflexive, can be arbitrarily flexible in its interpretation of alisp code. It also supports many logic programming concepts Are you doing the compiler from scratch? bin; the boot loads itself and the non-paging system, does some setup, then tells the non-paging section to go its object system is very similar to Self's, and strongly related to the compiler's inner workings bin: yes, this is all our own work, and all implemented in scheme at the moment, except for the x86 booter which is 1k of asm Self is a nice thing indeed kauf: reflexive or reflective? Zhi, sorry s/reflexive/reflective/ kauf: ahh good :) kauf: what does it produce? Zhi, what do you mean? I mean, what does the compiler output? Zhivago: does the compiler take input in ASCII form or something else? strictly speaking, extended intensive parse-dags kauf: um, I mean, when you compile something, what do you get at the end of the process to run? bin: the reader takes in ascii, the compiler takes the s-expressions that the reader produces Zhi, it's an interactive compiler... you compile functions with it, and apply these functions to desired input to get desired output kauf: um, what doe sit compile _to_? what's the target of the compiler? kauf: i'm familiar with lisp compilers. Zhivago: I'd like to ask you the same question :) 07:10am Zhi, I thought you were :) s/thought/assumed/ kauf: so, doe sit produces native code? or byte-code? or what? bin: which one? Zhivago: What does the compiler produce? bin: x86 native code, gas assembly, or C Zhivago: but in such case, is it possible to access something that was not compiled together with the caller? (I'm not as familiar with this as you two) bin: I'm not sure I undersatnd the question Zhi, it produces parse-dags which are then interpreted - i.e., assembled and executed. Some kind of platform-independent bytecode will eventually also be supported. The usual form of executable-level persistence is simply saving the parse-dag object directly to persistent storage. kauf: ahh kauf: fair enough, I need to use an incremental native compiler for my work. Zhivago: Existing components in the system must be visible and accessible in some way bin: yes bin: they're just objects in the memory-space, which are reachable directly or indirectly though references Hmm, compiled objects..? bin: any kind of objects, including compiled function/method objects bin: strings, code, instances, classes, types, methods, namespaces, pairs, vectors, and so on Can you change something that's compiled? I.e., can you modify functions? you can replace them if you're able to Is the environment a Scheme (or whatever) run-time environment, or is that just for functions? hmm, I'm not sure I understand the question again. the environment is a graph of objects which should work more or less like a scheme environment hm if you mean, 'could you use this system like a normal scheme system' then the answer is 'mostly' 07:20am ok.. how does the internal environment of compiled functions and the system environment fit together? Are they the same, or totally disjoint? zhiv: why did you choose a lisp-1 (ie Scheme) instead of a lisp-2? lisp-1's scale really badly though they can look pretty.. um, you might want to clarify what you mean by environment, since that doesn't make much sense to me fufie: what is a lisp-2? sorry a lisp-2 has separate namespaces for functions and variables a lisp-1 uses the same namespace/lookups fufie: oh, well, that's um, a trivial matter I mean the VM.. the place where objects live, like filesystem or something fufie: I've found no real problem between CL and scheme in that regard fufie: the major problem imho with scheme is the lack of a decent type-system fufie: although if you can pointout some problems with a single space that I've overlooked, I'd be greatful. bin: oh, they all live in memory. zhiv: scheme is the dorian gray of langueages.. young, beautiful and will always remain so.. but useless for practical stuff fufie: standard scheme yes, but I don't intend to use a stancard scheme Fufie, the system being reflective, that becomes a non-issue... if you want to enforce separate namespaces, add another symbol table to the system and write some system rules enforcing that if the type of an object is function, then it goes to symbol table A, otherwise to symbol table B. zhiv: that was my guess too, and that is why I wanted to know if you chose lisp-1 or lisp-2 fufie: well, um, in my experience there's no significant difference fufie: if yor experience is otherwise, I'd like to hear about it kauf: implementation-wise it is not a big issue, but one has to choose at some point Zhivago: Yes, that's the thing I'm interesting in; how does this look, how do programs see and reference objects? bin: ok, there are no programs bin: just a mess of objects that lurk in memory, and refer to each other ie (define (foo list) (list list)) is kindof different in lisp-1 and lisp-2 Zhivago: Is there a distinction between the inside of functions and the outside (i.e. memory)? fufie: yes, I realise that, but um, having been programming in CL for several yeras, and then moving across to scheme, I've found that to be a non-problem Zhivago: you said functions were compiled, not? bin: um, they're all in the one address space, there's no memory protection bin: yes, they're compiled zhiv: for smaller programs and when you think clearly it should be ok.. but the moment you introduce dynamic scope which can be really useful it is easy to run into trouble with lisp-1 Zhivago: Then how does a function know what other objects there are? bin: well, it has references to them Zhivago: gathered at compile-time, or how? fufie: well, you don't usually want dynamic scope in functions, at least in this system that's not so useful. bin: yes, at compile-time, or passed as arguments ic 07:30am * Kaufmann/#tunes is off now bye fufie: hygenic macros tend to be safer to use in those cases, unless I'm talking about something else zhiv: I don't know the particular system, but a "system-lisp" should imho have dynamic scope as it can reduce side-effects for the rest of the system cya fufie: ok, just to make sure we're talking about the same thing, can you give an example? later kaufmann zhiv: hygienic macros is not important here.. that is a separate discussion -:- SignOff Kaufmann: #TUNES (I don't remember, I don't recall, I've got no memory of anything at all...) zhiv: a really banal example: assume *debugging* is a dynamic variable (ie defvar in CL).. I find this useful: (let ((*debugging* t)) (test-function-which-can-be-a-whole-subsystem))) fufie: ok, so we're looking at specials it also makes threading a lot easier at least in one of my systems fufie: good fufie: hmm, I'll have to think about that one, since I haven't been using specials since CL, mostly because I've had no use for them. 07:40am there is a gabriel/pitman article about lisp-1 vs lisp-2 from 88 should be interesting, I'll keep an eye out. Issues of Separation in Function Cells and Value Cells. url? just found bib(tex) entries to it but have read it and I think I have a copy at my office (I'm at home now) hmm, ok, well, it it's convenient, you could email to brian@designix.com.au I'll try to remember if I find an electronic version if it isn't easy, don't worry, I'll keep an eye out 07:50am -:- AlonzoTG [Alonzo@216-164-131-170.s170.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes -:- eihrul [lee@usr5-ppp05.lvdi.net] has joined #tunes -:- SignOff AlonzoTG: #TUNES (Read error to AlonzoTG[216-164-131-170.s170.tnt2.lnhva.md.dialup.rcn.com]: Connection reset by peer) * Fufie/#tunes yawns agreed.. -:- hcf [nef@me-portland-us126.javanet.com] has joined #tunes 10:10am -:- SignOff binEng: #TUNES (zzz..) -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us126.javanet.com]) -:- hcf [nef@me-portland-us126.javanet.com] has joined #tunes -:- T5G [root@t1o68p71.telia.com] has joined #tunes -:- lar1 [lar1@1Cust109.tnt3.sfo3.da.uu.net] has joined #tunes Hey This is lar1 speaking live from Debian! (I got unlazyed and reinstalled :) ) * T5G/#tunes is away, auto-gone [tv/lp] potato? Fufie: Slink. I am updating my packages now Fufie: The whole reason it broke was trying to update slin -> potato :) hehe.. have a potato at work.. * lar1/#tunes hopes that this time he won't hose his libc 12:20pm -:- newbie [root@235-CORU-X13.libre.retevision.es] has joined #tunes -:- newbie is now known as BOrg -:- SignOff BOrg: #TUNES (Changing server...) yeah.. going potato requires updating almost everything due to glibc2.1.. though the CMUCL in potato seems newer 12:30pm -:- Downix [down@d-gnaps-81.ici.net] has joined #tunes hey-lo all Hey What is ldso? ld.so is how UNIX links shared libraries to objects saves HD space Hmm, libc6 is dependent on it yes unless you want everything with static libs I want to make sure I select it this time ;) ok Does potato use glibc or libc or is there a difference? glibc is just GNU's version of libc6 Ah xdm is the thing that makes me log in with X, right? potato uses glibc2.1, slink uses glibc2.07 I think xdm is the glossy X-based login yes Fufie: Ok... thanks last time stupid xdm was installed by deafult and my X server was misconfigured and bah! No more xdm! :) 12:40pm xdm is usually not needed xdm is for wimps -:- NetSplit: varley.openprojects.net split from sterling.openprojects.net [12:47pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net] 12:50pm hey all Can someone point me to OS kernels I could look at? go to the review subproject pages there's a whole slew of OSes there looking through there but I mean any good ones to review? hmm, i'll check :) looking for recomendations here the'res L4-Linux (Linux running on L4 microkernel).... Linux is a pain to look over L4 is sloppy though I have L4 tho and no good implementations really * Downix/#tunes agrees you could wait a year for mine :) Core is supposedly releasing Clementine as well I quote "this fall" -:- ChanServ has changed the topic on channel #tunes to: www.tunes.org sure... which he supposedly already has some posix services on I want it now tho why do you need it now? :) to study you could check out EROS maybe they just GPL'd it recently What is newt? I am not going to be using any of them, so licence isn't important you also have the MIT's Xok I just want to see 01:00pm I have Xok Xok is more like what I want, but not quite what are you looking for? that's a convo that covered what I want sounds kinda like what i'm trying to write :) slightly different list of requirements than a normal kernel and kinda what AlonaoTG is writing but that is what I want, spelled out his is more complicated than mine :) I just want it to work but first I need a kernel to base everything else on there's also Nemesis Nemesis? yes -:- SignOff lar1: #TUNES (Leaving) almost every OS design I see tho is too large for my needs Nemesis might be nice though :) 512K is a cramped space, but it's enough for my needs well... 512k for the whole OS? that's not umm, *cough*, that possible :) unless you want some real slow-downs..... Amiga has no slow-downs well, trade-offs i should say and it was on a 128K ROM trade-offs I can deal with and the kernel needs memory to store objects in beyond just its code and tables * Downix/#tunes nods so long as you allow that :) its possible can easily fit just the code and tables in 512k I just need enough of an OS in that 512K to boot a program either the rest of the OS or a single program 01:10pm well, microkernels are usually no larger than 100k so you'll have lots of room to spare yes but I don't need things in the kernel that are in a typical microkernel since it'll all be ROM'd what about bugs in the kernel itself? :) put things like file systems, etc all as libraries Keep the kernel as small as possible to minimze bugs less LOC == less bus to repair just wish my kernel was ready at this point :) have you coded anything? i have code... its all old can I see? lets take this to a message :) 01:20pm !ChanServ:*! lilo used GETPASS on channel #freebsd -:- SignOff Fufie: #TUNES (Leaving) http://biz.yahoo.com/prnews/991105/ca_pc_worl_1.html oops 02:00pm -:- Fufie [stig@tunnel-44-24.vpn.uib.no] has joined #tunes sawmill crashed on me again :( 02:10pm -:- Kaufmann [Kaufmann@dial145.infolink.com.br] has joined #tunes Shalom! hi What's up? stress-testing my system with 11k mails in xml-format.. still fast otherwise little happens which system? 02:30pm my distributed hypermedia-system ah 02:40pm -:- SignOff Kaufmann: #TUNES (Read error to Kaufmann[dial145.infolink.com.br]: Connection reset by peer) man, cvs is annoying can someone tell me what the hell I'm doing wrong? [root@survivor123 /root]# export CVSROOT=pserver:anoncvs@cvs.eros-os.org:/home/anoncvs/cvs [root@survivor123 /root]# cvs login cvs login: can only use pserver method with `login' command cvs [login aborted]: CVSROOT: pserver:anoncvs@cvs.eros-os.org:/home/anoncvs/cvs any clue? try putting the : in front of pserver 03:20pm -:- SignOff Downix: #TUNES (BitchX: made with real honey.) -:- lar1 [lar1@1Cust109.tnt3.sfo3.da.uu.net] has joined #tunes Hey -:- lar1 [lar1@1Cust109.tnt3.sfo3.da.uu.net] has left #tunes [] -:- lar1 [lar1@1Cust109.tnt3.sfo3.da.uu.net] has joined #tunes -:- SignOff lar1: #TUNES (Leaving) -:- air [brand@p0wer.qzx.com] has joined #tunes 03:40pm -:- SignOff Zhivago: #TUNES (Ping timeout for Zhivago[th.merddin.com.au]) !ChanServ:*! ^lilo used GETPASS on channel #freebsd -:- Zhivago [brian@th.merddin.com.au] has joined #tunes -:- Fare_ [fare@s206.paris-112.cybercable.fr] has joined #Tunes -:- Fare_ [fare@s206.paris-112.cybercable.fr] has left #Tunes [] !ChanServ:*! lilo used GETPASS on channel #lunarix -:- lar1 [lar1@1Cust109.tnt3.sfo3.da.uu.net] has joined #tunes Hey hey 04:30pm It takes a long time to get 100+ Mb on a 56k modem :( -:- SignOff lar1: #TUNES (Read error to lar1[1Cust109.tnt3.sfo3.da.uu.net]: No route to host) 04:40pm -:- lar1 [lar1@2Cust88.tnt20.sfo3.da.uu.net] has joined #tunes -:- SignOff lar1: #TUNES (Leaving) -:- NetSplit: lackey.openprojects.net split from devlin.openprojects.net [05:20pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [lackey.openprojects.net] -:- Netjoined: lackey.openprojects.net devlin.openprojects.net -:- Zhivago [brian@th.merddin.com.au] has joined #tunes * Fufie/#tunes yawns.. -:- lar1 [lar1@sdn-ar-014casfrMP184.dialsprint.net] has joined #tunes -:- SignOff lar1: #TUNES (Leaving) -:- hcf [nef@me-portland-us1036.javanet.com] has joined #tunes -:- SignOff Fufie: #TUNES (sleep) -:- abi [nef@bespin.dhs.org] has joined #tunes -:- root [root@dialup-209.245.138.28.SanJose1.Level3.net] has joined #tunes Hey -:- SignOff root: #TUNES (Leaving) -:- lar1 [lar1@dialup-209.245.138.28.SanJose1.Level3.net] has joined #tunes That was stupid :) 06:40pm -:- SignOff lar1: #TUNES (Leaving) -:- hcf has changed the topic on channel #tunes to: Free Reflective Computing System || brain abi: OpenC++ is C++ plus a MOP at http://www.hlla.is.tsukuba.ac.jp/~chiba/openc++.html 08:30pm abi: DoPL is Dictionary of Programming Languages at http://cgibin.erols.com/ziring/cgi-bin/cep/cep.pl 09:40pm -:- td [x@1Cust113.tnt3.wilmington.nc.da.uu.net] has joined #tunes -:- beholder [beholder@ppp-029.m2-1.sub.ican.net] has joined #tunes -:- beholder [beholder@ppp-029.m2-1.sub.ican.net] has left #tunes [] -:- hcf_ [nef@me-portland-us203.javanet.com] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- hcf_ is now known as hcf -:- _ruiner_ [DIY@ppp245.wi.centurytel.net] has joined #tunes -:- SignOff _ruiner_: #TUNES (destroy what destroys you) -:- SignOff Fare: #TUNES (Ping timeout for Fare[quatramaran.ens.fr]) -:- SignOff eihrul: #TUNES ([x]chat) -:- SignOff td: #TUNES (td has no reason) [msg(TUNES)] newlog 1999.1108 IRC log ended Mon Nov 8 00:00:04 1999