IRC log started Fri Aug 6 00:00:01 1999 [msg(TUNES)] permlog 1999.0806 -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.dhs.org]) -:- SignOff Tril: #TUNES (Ping timeout for Tril[bespin.dhs.org]) -:- Connection closed from clarke.openprojects.net: Success -:- Connecting to port 6667 of server clarke.openprojects.net [refnum 1] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES -:- Your host is clarke.openprojects.net, running version u2.10.05.18.(ipcheck4-5) -:- This server was cobbled together Wed Apr 28 1999 at 13 24:47 EDT -:- clarke.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv -:- [local users on irc(2)] 1% -:- [global users on irc(156)] 45% -:- [invisible users on irc(193)] 55% -:- [ircops on irc(11)] 3% -:- [total users on irc(349)] -:- [unknown connections(0)] -:- [total servers on irc(33)] (avg. 10 users per server) -:- [total channels created(118)] (avg. 2 users per channel) !clarke.openprojects.net Highest connection count: 21 (20 clients) !clarke.openprojects.net Welcome to Open Projects! You are on 2 ca 1(2) ft 14(14) tr. -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.dhs.org " -:- TUNES [tunes@bespin.dhs.org] has joined #tunes -:- [Users(#Tunes:3)] [ TUNES ] [ Plundis ] [ Crimson ] -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 7.837 secs!! -:- Mode change [-s] for user TUNES -:- abi [nef@bespin.dhs.org] has joined #tunes -:- Tril [dem@bespin.dhs.org] has joined #TUNES -:- mode/#Tunes [+o Tril] by ChanServ -:- eihrul [lee@usr5-ppp128.lvdi.net] has joined #tunes -:- hcf [nef@me-portland-us337.javanet.com] has joined #tunes -:- smoke [smoke@13dyn63.delft.casema.net] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp128.lvdi.net]) -:- SignOff smoke: #TUNES (tsjoeke tsjoek) -:- eihrul [lee@usr5-ppp114.lvdi.net] has joined #tunes -:- smkl [sami@MCCLIX.rdyn.saunalahti.fi] has joined #tunes -:- SignOff smkl: #TUNES (Ping timeout for smkl[MCCLIX.rdyn.saunalahti.fi]) -:- SignOff Tril: #TUNES (Tril has no reason) -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES -:- mode/#Tunes [+o Tril] by ChanServ -:- Bill [user9973@N090P008.dipool.highway.telekom.at] has joined #Tunes -:- SignOff Bill: #TUNES (Leaving) -:- smkl [sami@MCCLIX.rdyn.saunalahti.fi] has joined #tunes -:- _sham [sham@phila-dialup618.nni.com] has joined #tunes -:- core [core@core.suntech.fr] has joined #tunes alors moi je dis, bonsoir. hateverway hattay ustjay aidsay <_sham> hahaha pig latin hehe good evening people, you'll adjust for your timezone :) >>> _sham [sham@phila-dialup618.nni.com] requested TIME from #tunes >>> _sham [sham@phila-dialup618.nni.com] requested unknown ctcp READJUST_YOUR_ATOMIC_CLOCK from TUNES * eihrul/#tunes is wondering whether it's worth hashing process id's. what's your alternative? walking a list every time you need one? (or need to get to the process identified by the given id?) well, i was thinking that you could simply cache the last resolved pid and for lots of IPC you usually do several operations to/from the same pid eh ? what good would that do? it would prevent pid lookup after the first time yeah, if the system didn't run concurrent processes, perhaps well, as soon as the process you IPC to will reply and lookup the caller back, it will defeat the point of your cache, and that's just considering two processes, not the usual 2000. every process has it's own little cache well, how do you wait on messages on different ports from different processes at the same time then? :P i don't have ports :) pid's are effectively the ports... i'm keeping it simple pid's, not thread ids? (or do you plan to have threads,) well, still, waiting on multiple pids defeats the point :P no threads this time around :) 01:30pm what does that have to do with pid resolution? wow, openprojects is so that there is almost a server per client. *g* well, unless you maintain a multiple cache of pid's, you'll loose the advantages of the cache if you don't use just one, and if you use a cache of multiple pid's, it's pointless, since it's called hashing, and you better do it from the start. if you want to really keep it simple, you can restrict the number of concurrent processes, and just use an array of pointers, indexed by pid.. i don't like the idea, but for a simple test system, it works. (well, unix is a test system in this respect them, if you will :P) the reaosn i didn't like hashing is because you either have to keep the hash table a fixed size or even worse, GROW it :) <_sham> core: you think it's inefficient to keep a table of processes indexed by PID? _sham: not inefficient, it's actually efficient. just sets an arbitrary limit :) <_sham> core: look up tables = tres good so does hashing almost <_sham> ;) eihrul: well, you can use multilevel hash tables, like a two level one; so you grow it without resizing _sham: i know; it's just imposing limits, of say, 4096, 16384, 65536, whatever, processes. eihrul: sort of like x86 page tables but for hashing any kind of objcets, growing it on demand. but that's a more complicated solution than a simple array with a fixed-for-systems-life limit :) i dunno, the cache just seems cleaner you only have one lookup per transaction after that, it's free eihrul: with an array or a hash, you never have a lookup. eihrul: that's the whole point of both methods. eihrul: message passing is already overkill. you have the troubles of managing limits/frailties of either of those methods why is message passing overkill? eihrul: it is a lot more overhead than direct method invocation i mean; so adding a lookup to it is quite a lot. eihrul: there is no frailties to hashing, it's just more complicated to put together. s/is/are/ well you can't exactly have DMI between processes, can you? eihrul: yes, that's what i'm saying eihrul: you can't, but there is no need to add to the overhead :) eihrul: but considering you're writing a test system, you can pretty much do as you like, i guess :-) i'm just trying to evaluate what my choices are... eihrul: this is a generic problem, not just limited to pid's i know it is eihrul: hiding pointers using indexes of some kind for human readability and security, that is. eihrul: fyi, i use a two-level hash table as a generic solution; this makes for a bit more than 1 million indexes growing the memory usage on demand and very efficiently accessed; but that's just me. how many entries per level? 1024 01:40pm what if you have one stray pid in a second level, then you're using a lot of bytes... the hash is handled by the memory manager, so that it can use the underlying hardware efficiently. the virtual memory manager of clementine that implements management through paging, uses 1024 entries/level on a 4K-pages system. eihrul: yes, you can have one pid preventing an entire second-level area from being freed. eihrul: but new pid's are allocated from those, first. well shouldn't pids generally try to be unique in time? well, clementine uses them more for reference than anything else, as components do not care what process id, thread id, or whatever they're running in; that's low level and irrelevant details :) they use other means, so pid's needn't be as much unique as possible. so the generic implementation of hashing tries to reuse as much as possible, actually. but what about when communicating with other processes/threads? they just call a method in a component; if it happens to run in another process/thread, the system routes it through. if a server goes down, but fails to notify a client, reuse could allow some other process to take the id and continue communication eihrul: again the id isn't relevant; if an used component goes down, the caller is always notified and can revert to another component, possibly using another pid/tid, without problem (well, this is not completely implemented yet, but that's how it works.) okay... can you please actually describe what exactly your component system is/does :) well, i don't want to sound like i'm teaching stuff, that wouldn't be my place at all i'm very modest about my project :) is it a communications channel established between two processes of some sort? but if you want to know - okay - it's made entirely out of independent components that describe an interface, and also describe what they need doesn't matter if it's not your place to teach, i welcome any ideas :) 01:50pm if the imported component for a given functionality that your component needs, is in the same address space for some reason, direct calls will occur; if it's in another process space, it will be encapsulated in messages transparently; if it's on the other side of the world, same. a 'connection' is established when you open the functionality the component provides with the intention to use it. if it's disconnected before you close it on your own (component crashed you are notified. this isn't done yet, but components are going to implement a Notify() call for that purpose. so you're using some kind of IDL to define interfaces? yeah. some kind. not corba-like at all, but you got the main lines right, yes. although, i want to underline, that my compiler doesn't generate _any_ stub code, it just compiles the interface to binary, which is stuck to the component code/data etc, thanks to XCOM. if your component crashes, all its imports are automatically liberated in cascade, too, of course. ah, definitely not something i should be trying to emulate right now :) well, i have a working system that uses this principle, but it took us one year to get there and i'm not trying at all to convince you of anything, you just asked how it works :) i know i did <_sham> i will have to make some major adjustments to XCOM for it to fit my needs i think it's a really neat principle, with a lot of advantages, a lot of which were inspired by Fare :) _sham: feel free! i could just use a fixed size hash table :) _sham: as long as you don't make incompatible changes, please go ahead and do so, and contribute back. that'll be great. seeing as i'm not implementing page swapping, it couldn't hurt to have a limit or two eihrul: right, if you're just toying with a test system, you probably don't want to get into something TOO hairy, if you're going to rewrite it soon. <_sham> core: wouldn't it be proper to create a language that further extends XCOM? _sham: a language..? XCOM is barely a low-level transport specification and you can almost do a direct mapping of unique pids if you allocate correctly from hash table _sham: please describe what you mean, though :) all ideas are very welcome. eihrul: well, if you allocate a table of say, 4096 entries, that's quite decent for a test :) <_sham> well maybe you could make a language that fully exploits the "object" system of XCOM the hash tbale is only going to point to the pcb's... it's not actually going to contain them <_sham> like a real object oriented language, instead of just C objective c! :) <_sham> or is C adequate? _sham: again, XCOM doesn't impose any object system of any kind; the one i'm talking about is specific to Clementine, that happens to use XCOM. but XCOM is just a binary specification, you put what you want in it. _sham: you can compile C++, Java or OCAML to XCOM objects without problem :) core: how comparable is the complexity of making a small xcom loader versus an elf loader? _sham: actually, the generic 'interface' section scheme is here to allow different operating systems to use the same tools and perhaps loaders eihrul: it is a lot simpler. look at retro's ah, i'll look into it but ah there bootloaders that support it? er are or would i have to make one? eihrul: xcom doesn't have two sets of sections for objects and executables, unlike ELF; doesn't drag years of kludges, either :) 02:00pm eihrul: clementine is made so that the loader needn't understand xcom, but it would be very easy to extend GRUB to understand XCOM. If you could do that, you'd be my new best friend, as i don't have time :) <_sham> ah retro (and no, retro's xcom loader is not written in forth :))) i guess i could play with that once i get around to it -:- SignOff Tril: #TUNES (Tril has no reason) eihrul: if you need an XCOM-aware bootloader for your system, and you want to extend grub to do that, it'd be very simple. if you do it, and contribute the changes back, it'd be very nice. i'm sure i could fit that in sometime it's not like i have a job or anything :) eihrul: clementine's full featured XCOM loader is 287 lines of C, comments included; and it's as powerful as ELF. 287 is pretty big :) eihrul: that'd be insanely cool, even if it's simple to do; i don't have time at all right now, outside of clementine and giving my free time to my girl. eihrul: big?! :) that's bigger than my malloc and page allocation implementations combined :) trying to keep the kernel small, that's why i don't want to grow hash tables eihrul: linux's elf loader is 1366 lines of C eihrul: the 287 lines include a lot of comments. like about 6-10 lines per function to describe it. guess that's pretty good then but if you have one monstrous 277 line function, maybe not :) lol no, i'm not a pig. question regarding tss: assuming i map all my process kernel stacks at exactly the same virtual address, is there any reason to reload esp0 in my system tss? eihrul: i guess not. <_sham> core: what's better: using TSSes or doing manual context switching? esp0 per se, no; of course the current stack pointer will have to be. but i shouldn't have to fiddle with the tss at all, right? :) _sham: i use manual context switching so i can just use one TSS per processor in the system. but that's just me :) ok, brb, girl issue :) 02:10pm -:- SignOff _sham: #TUNES (Ping timeout for _sham[phila-dialup618.nni.com]) -:- _sham [sham@phila-dialup089.nni.com] has joined #tunes core: hurry up with that girl issue :P <_sham> eihrul how would task switching with only one TSS work? you simply store all the registers in the kernel stack for each process and load the other various tasks registers manually 02:20pm -:- fare [fare@quatramaran.ens.fr] has joined #Tunes -:- binEng [Anders@dialup93-1-8.swipnet.se] has joined #tunes -:- SignOff fare: #TUNES (Connection reset by pear) -:- SignOff binEng: #TUNES ( <k!14>) * core/#tunes returns anyone still alive? :) yep so i should not have to fiddle with the tss, right? :) oh, you mean, if using manual switching? no. okay, good well, even if i don't maintain TSSes in the GDT, i still allocate one for every process, for something quite useful.. the I/O bitmap :) also... when implementing messaging, is it better to allocate the page entries (messaging window) on demand or reserve some? 03:00pm i'm just going to set the IOPL field on my processes :) IOPL is a little global :) thought it was per thread as it's stored in eflags? well, for your message window.. there's no definitive answer of course. you could make it so that processes can allocate "messageable" memory, that will automatically be page aligned, and that can then be simply remapped into the destination instead of copied .. eihrul: yeah, it is; i mean, I/O-port-wise sure the process can access any io port, but it's just for fun :) and then threads could directly compose the message in the allocated 'messageable' memory.. saves you any copy.. as long as it's mapped read only in the destination, i guess that's fine .. unless it's an async request in which case you'll have to duplicate at least once, i guess. what's the url for xcom, as well? ah, found it eihrul: http://xcom.tunes.org that you're welcome to implement for GRUB or generally tweak to your needs :) -:- liar [brand@p0wer.qzx.com] has joined #tunes 03:10pm hey liar. give image of brix to core hrmm, grub only supports elf right now, eh? 03:20pm eihrul: and a.out too IIRC eihrul: i think freesbee uses a.out support. (or it did till 3.0-RELEASE anyway) (soon freesbee people will stop making fire with a silex.) ah, aout was on the next page :) hehe well, if there are two, there can be three :-) it's kind of sloppy looking load_image seems to take up 5 or 6 pages :) that's projects taken over by gnu, for you :P hehe.. optimise some. but do XCOM first :-) it's not so much optimization but the notion of modular programming seems lost on whoever wrote this :) even procedural programming even programming? :-) well, the multiboot spec blows IMHO. but grub is quite nice feature-wise. hehe... you could say that what about it blows specifically? x86 centric.. they could have made a more generic one.. like a helicopter view of booting :-) .. the physical memory info they provide is limited to one size, what about systems with holes ? (clementine supports them fine) ... and grub looks for a signature in the middle of the code of your system! that's quite dangerous/unpredictable, even with a checksum i don't believe in it. no more than the parameter it provides you to tell you grub booted you. unsafe. yep, that does seem to slightly suck that's just my opinion. it works well as far as i can tell from an user point of view, and it can speed up bootstrapping a new system, a lot. and it's free, so hey :) but where exaclty do you come across such gaps in memory? well, some ISA cards can require a hole from 15 to 16 megs (physical) for their aperture; that's not common but it's planned (called the 15M memory hole in bioses). i _think_ grub handles this specific case (linux does not). and a lot of non-x86 systems have those; such as the amiga, the sgi, etc. memory in different regions. physically speaking. 03:30pm anyone used/seen litestep? heard of it, but that's it liar: friends of mine use it liar: they are pretty good at graphics therefore their desktop looks quite impressive and non windows-ish. <_sham> what are the holes used for, core? core: this documentation up to date? _sham: well, the amiga has different types of RAM, some that is shared with external chips (like PCI memory; but the amiga is an UMA) .. some isn't.. and it's mapped at different physical addresses... so does the sgi.. _sham: on the PC the hole is just to cope with an ISA shortsightedness _sham: it's just a remark because i'm a nitpicker tho. grub does work fine. <_sham> what about things like texture memory on the voodoo3, is that memory mapped? yeah, but it's not "allocatable" memory so you don't care if it's not declared as RAM, you can still remap it into virtual addresses i'm talking about allocatable ram :) ack, hungarian notation! <_sham> so if it isn't declared as RAM, how is it declared? _sham: it's not declared _sham: it's pci memory.. you don't want to use it as ram for your processes _sham: it's just decoded, and can be read and written, but that's it; you only want to use it as a whole, as framebuffer; or texture buffer. core: you have any bootable xcoms i could test with? eihrul: you can produce code that boots and write an XCOM object of them as easily as an elf one :) ah, i'll just port chaos to it then 03:40pm but would be easier if i didn't have to :) <_sham> core: in manual context switching, how does the kernel switch to code of lesser or higher privileges? how can you switch to different privileges manually? <_sham> man, i don't understand how this would work ;\ why would you want to? eihrul: or just make stupid boot code :) sham: the user code will call the supervisor code thru a call/int gate _sham: it doesn't matter; the code will always be interrupted while in PL 0 (timer interrupt, or system call, whatever) _sham: so it doesn't matter if it entered it from PL 3 :) sham: when the supervisor code is finished it will return to the user code <_sham> ahh yes of course you have to use one kernel stack per thread for that to work. which means that if i compute 17 + 5 in brix it consumes 40 MB of kernel stacks :-) *run* :) ha ha ha hehe so when do you give me an image ? :) after i have one to give anyone who doesn't have more than 40 MB of ram is a cheap bastard :) i had to goto a party last night but maybe i will work on it tonight eihrul: that's just for kernel stacks.. i'm not speaking of page tables ;) core: okay, so maybe 120 MB... but who doesn't that much these days? :) brix only uses 4 page tables to map ***ALL*** code into 03:50pm he's got you there... cause it doesn't do much :-) no matter how much code u load it will bever exceed 4 page tables oh yeah, you limited yourself to 4 MB, i forgot :-) 4096 simultaneously loaded methods if only C had templates... i'm just trying to tease liar's pride into proving me wrong sending an image, but.. :P liar: yeah, we talked about that before IIRC iirc? <_sham> core: late one night in #tunes liar started sending me brix source code, probably in a drunken moment of disorientation or something :) liar: if i remember correctly; sorry. and late one night sham starting killing dcc conections like the fool he is :) <_sham> heh _sham: oh, so you're behind windows2000testing.com? :-) <_sham> core: Windows is my personal messiah _sham: so people only get 7 of your commandments before the messiah crashes? <_sham> heh core: due to the weather, of course :) or only get an evaluation version of your message, followed by "your license of microsoft messiah has expired." eihrul: of course :) 04:00pm <_sham> core: if microsoft offered you a job as head programmer on the windows 2k project with a starting salary of 6 figures, would you take it? :) was crack.linuxppc.org cracked? _sham: i don't want to give ideas to french taxes, but i'm not far from that. it's still open, so probably not if i'd cracked that box, i'd make sure noone else cracked it... they have the root passwd on the web page _sham: but to answer your question, if they offered a tenfold improvement over my salary + any conditions i like (let's dream), of course i would, as long as i have time to work on cool projects outside of it (among the conditions, having a life outside m$, being able to leave after a few years, and having limited clock time at work. ) core: which grub to start from, 0.5 or 0.5.92? liar: they gave it away eihrul: latest is best i guess :) gnu grub? it was on alpha.gnu.org, so i'd assume yep. :) <_sham> liar: so do you think windows 2k will be revolutionary in any aspect? _sham: well, objectively - the wild usage of COM+ is revolutionary; system is more and more componentised. sham: HAHAHAHAHA the rest, blargh. more like provoke one there are a few programming ideas that i find neat like how the MMC works, even if it's completely pointless and the implementation is unusable, but it's a good idea. <_sham> hah microsoft attempts to benefit from open source without doing its part :P core: xcom looks very cool so far eihrul: thanks.. it sure can be still improved a lot, it's biased towards clementine, and i need you and other people to unbias it :) 04:10pm i'll be sure to do so as there seems to be much better documentation for it than ELF :) yeah, lack of proper ELF docs was one of the many things that led me to wrap up XCOM :) but ELF does have the cooler name atleast, acronym-wise... i'll give it that <_sham> you should have named it TROLL _sham: TRivial Object Lean Loader? eihrul: bah, i like XCOM too ;) but yeah elf is cool. hmm should i work on brix and get an image out tonight or should i install my dual board in this machine right now? <_sham> so i guess XCOM XCOM would be <_sham> the game X-COM: UFO defense in XCOM format liar: get an image, installing your dual board will fubar your system for approximatively 5 months :) _sham: uh uh.. i remember the game :) oh no, i just gave you another excuse not to show anything about brix for a year :PP well i compiled an smp kernel about a month ago and installed smp linux on my dual board and it worked... wow, i have five whole users on my OPN server. it's getting crowded! uh uh, you say that now ;-) so it shouldnt fubar anything <_sham> OPN? OPN is the Open Projects Net, an IRC network for Open Projects _sham: openprojects <_sham> oh <_sham> heh <_sham> core: /kill a few ;P _sham: hehe _sham: nah, it took me a lot of spamming to get them to connect in the first place :)) <_sham> /msg rand_user MESSAGE FROM THE INTERNET - the internet is currently down. please check back between 1:00 and 7:00 haha core: u have yer own opn irc server? core: give me ops man :) <_sham> haha liar the Ircop i was an ircop on efnet /msg corny26 *OPENPROJECTS SERVER NOTICE* Due to traffic jams, it is recommended that you reconnect to heinlein.openprojects.net; thanks for your cooperation. you may proceed now. liar: yeah, it's been a while. it's not like i advertised for it or anything tho. <_sham> wow i have _a lot_ of thinking i have to do to define my final conception of this object system <_sham> liar: so what, did you abuse your powers on efnet? the forcibly bad thing about object programming is that you start from the principle that you know from the beginning all that your program will do. which is plain wrong usually :) and makes it very hard to fit stuff in later, when you discover some member is private and should be public, etc :-) 04:20pm core: what connection does heinlein have? and why would u name it heinlein? :) efnet opers are castrated, they can't do much beyond trading 12-year-old girls pictures while /killing channels that run those, i guess :) <_sham> irc.qzx.com? liar: E3 ; and you don't know who heinlein is? liar: all OPN servers are named after famous sci-fi writers; robert heinlein is one, he wrote starship troopers, among others. uhh <_sham> core: you should have named it herbert starship troopers SUCKED they don't have a server called herbert? OMG. that's true. damn! <_sham> or perhaps clarke oh well :) they have clarke <_sham> or wells :) they should have wells..? no, they don't. damn #2! :) core: u named yer box after a sucky writer, what a shame :) <_sham> or maybe you could start writing some sci-fi novels, win an award or two and name it after yourself :P liar: hee hee. it's more like i thought up of a scifi writer on the spot when lilo was asking for a name. liar: and the box is actually called orb; it hosts a bunch of things. will be moved to a seperate OSS-projects box by the end of the month. liar: the best in all of that is that heinlein is two rooms away from me (when i'm at work) and i can hit it at 100 mbps :P -:- NetSplit: lackey.openprojects.net split from sterling.openprojects.net [04:24pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [lackey.openprojects.net] -:- Netjoined: lackey.openprojects.net sterling.openprojects.net -:- Crimson [crimson@chaosdev.org] has joined #tunes -:- liar [brand@p0wer.qzx.com] has joined #tunes * Crimson/#tunes is gone. Gone since Wed Aug 4 18:07:00 1999 liar: the best in all of that is that heinlein is two rooms away from me (when i'm at work) and i can hit it at 100 mbps :P i already had it that way, core. abi: the best? core: i don't know <_sham> the best in all of that? the best in all of that is that heinlein is two rooms away from me (when i'm at work) and i can hit it at 100 mbps :P abi: best? core: no idea abi: no, the best in all of that is me. okay, core. <_sham> abi: people? people are people the best in all of that? <_sham> heh somebody said the best in all of that was me. screw you abi :-) silly bot anyway. <_sham> hmm maybe i should port XCOM to GGI -:- SignOff liar: #TUNES (Read error to liar[p0wer.qzx.com]: EOF from client) -:- liar [brand@p0wer.qzx.com] has joined #tunes !drow:*! Lackey's getting a serious set of updates now, folks. Be back as soon as possible. <_sham> the game, that is !core:*! good luck drow :) -:- _sham [sham@phila-dialup089.nni.com] has left #tunes [] _sham: actually GGI runs on clementine as xcom components :) i have bad timing. always :) heh how come u didnt have any ggi in the image u gave me? !spice:*! and now for aour daily quote from Linus: !spice:*! Are you being stupid on purpose, or were you born that way? !core:*! that's from linus? !spice:*! It is :) !core:*! damn, i should have copyrighted it while in kindergarten :) 04:30pm !core:*! i like what he said about GGI more tho :) "it's like killing an ant with an atom bomb" (setuid problems with X being an ant) !core:*! i don't collect linus' sentences, but i remember that one. but anyway. :-) does www.windows2000test.com have a web page? supposedly supposedly what? :) can u connect? it's been rumoured * core/#tunes missed a train nevermind :) ok, it's like 312 trillion degrees and i'm sweating 300 oz of water every minute, but i'm going to try and sleep :) eihrul: don't hesitate to email -> core@suntech.fr 'night :-) cya cya :) -:- SignOff core: #TUNES (zonk) 04:50pm -:- hcf [nef@me-portland-us208.javanet.com] has joined #tunes -:- _sham [sham@phila-dialup328.nni.com] has joined #tunes -:- SignOff smkl: #TUNES (sleeping ...) -:- SignOff _sham: #TUNES (Ping timeout for _sham[phila-dialup328.nni.com]) -:- _sham [sham@phila-dialup031.nni.com] has joined #tunes >>> _sham [sham@phila-dialup031.nni.com] requested USERINFO from #tunes <_sham> hcf i have a question about linking and calls <_sham> can you answer it? :P probly not ask the os-help mlist <_sham> do CALLs use relative or absolute offsets? i know they use relative ones for calls within the program but what about libs, do they use far calls (absolute) or near calls (relative) to call library functions? probly far calls 06:40pm -:- SignOff hcf: #TUNES (Leaving) -:- SignOff _sham: #TUNES (Leaving) -:- ElGato [dave@bespin.dhs.org] has joined #tunes -:- ruiner [nate@ppp242.wi.centuryinter.net] has joined #tunes liar go on roger wilco on bespin 08:50pm * ElGato/#tunes is away: (Auto-Away after 10 mins) [BX-MsgLog On] * ElGato/#tunes is back from the dead. Gone 0 hrs 0 min 16 secs -:- SignOff ElGato: #TUNES (ElGato has no reason) -:- hcf [nef@me-portland-us943.javanet.com] has joined #tunes -:- ElGato [dave@bespin.dhs.org] has joined #tunes * ElGato/#tunes is away: (Auto-Away after 10 mins) [BX-MsgLog On] -:- Closing Link: TUNES[bespin.dhs.org] by clarke.openprojects.net (Ping timeout for TUNES[bespin.dhs.org]) -:- Connection closed from clarke.openprojects.net: Success -:- Use /Server to connect to a server -:- Connecting to port 6667 of server irc.us.openprojects.net [refnum 0] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES (from lewis.openprojects.net) -:- Your host is lewis.openprojects.net, running version u2.10.05.18.(ipcheck4-5) (from lewis.openprojects.net) -:- This server was cobbled together Wed Apr 28 1999 at 12 02:19 EDT(from lewis.openprojects.net) -:- lewis.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv -:- [local users on irc(5)] 1% -:- [global users on irc(171)] 42% -:- [invisible users on irc(241)] 58% -:- [ircops on irc(14)] 3% -:- [total users on irc(412)] -:- [unknown connections(0)] -:- [total servers on irc(31)] (avg. 13 users per server) -:- [total channels created(124)] (avg. 3 users per channel) !lewis.openprojects.net Highest connection count: 11 (9 clients) !lewis.openprojects.net Welcome to Open Projects! You are on 4 ca 1(2) ft 14(14). -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.dhs.org " -:- TUNES [tunes@bespin.dhs.org] has joined #tunes -:- [Users(#Tunes:9)] [ TUNES ] [ ElGato ] [ hcf ] [ ruiner ] [ liar ] [ Crimson ] [ eihrul ] [ abi ] [ Plundis ] -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 7.614 secs!! -:- Mode change [-s] for user TUNES damn wing commander the movie is dumb duh 09:40pm liar: as a rule... most movies are dumb not true MOST not all the entire police academy series was excellent yes, but that's the exception... not the norm drama at its best heh heh....those movies sucke d for every police academy there's 30 bad disney movies, 200 bad love stories, and 500 steven segall action "thrillers" star trek 7-9 were beyond excellent those were horrible all the star trek movies are NO, star trek 7-9 are the best 3 movies made no, they suck NO, YOU SUCK I could make a better movie about my ass -:- SignOff ElGato: #TUNES (Hey! Where'd my controlling terminal go?) ruiner: so in yer movie what happens after u pull yer head from yer ass? :) the borg try to assimilate it, only microsoft beats them to it ah... so that's why everything coming out of microsoft resembles shit in some form or another... yeah they managed to make the coolest part about star trek (the borg) suck in the movie....#8? u know if the borg really existed and tried to assimilate earth they would see m$ as the primary source of technology and assimilate them and then the borg would self destruct after the m$ virus code infected them kinda like how lucas took the cool alien sidekick (chewie) and tried to rework it to sell more toys (jar jar) so i guess thats one good thing about m$ borg fodder, who would have thought! planetary defense against borg-like races they'd see how horrible it is and not use it you mean, it they might however take up some of M$'s anticompetetive practices the borg are one :) 09:50pm qz....do you have networking capabilities in brix yet? how did st8 suck? the magnetic boots were lame ruiner: no, u wanna write them? -:- water [water@ppp-tnt-20.tscnet.net] has joined #tunes how were they suppose to walk on the hull? water! for that kind of money, sure hey waiter and no, I don't want to qz....not my cup 'o tea then why did u ask? I'll think I'll be a jabrone for a while longer and download mp3's for a few more months I was considering if it was really release worthy yet.... hoy water hey hcf * water/#tunes hasn't had his laptop working while at sea... :( i had to connect with my roommate's computer so, my only progress on the papers has yet to be made electronically-available 10:00pm * water/#tunes looks at tunes mail for the last two weeks -:- SignOff ruiner: #TUNES (Leaving) bye all. i'll be back in an hour or so -:- SignOff water: #TUNES (Leaving) 10:10pm -:- billyboof [hatefull@ptnm-sh4-port136.snet.net] has joined #tunes -:- SignOff billyboof: #TUNES (hatefull@antisocial.com) -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES -:- mode/#Tunes [+o Tril] by ChanServ hoy Tril hi 11:10pm i still dont agree with u on reply-to :) ok. run yer own mlists :) btw, if the users of a list really do want reply-to, I will set it. But I try to talk them out of it, first. nah, i will get u to add reply-to :) by not setting reply-to yer making it harder for those of us that arent idiots how? I thought it was the other wqay around cuz we have to manually type the list email address or cut and paste what mailer do you use? I don't know what you are talking about. netscape when i get an email from the list i am expected to reply back to the list let's say you get a message from Paul dufrp@tunes.org, To: tunes-lll@tunes.org that way everyone knows the question has been answered or they get to hear my responce then you hit Reply to All. It should automatically put dufrp and tunes-lll in to the To: or cc: fields. ya but then paul gets dups 11:20pm But he doesn't, because he sets a Reply-to: tunes-lll@tunes.org himself in his mails. when he does, hitting reply to all sends to only the tunes-lll@tunes.org address i think that is an extra step that we must take paul doesnt set reply-to did u read the anti-reply-to page? no if he sets reply-to then i hit reply, not reply-to-all no i didnt read it i dont have time to read crap it has more arguments on it.. which i dont care to make to you because you can just read it. brb * Tril/#TUNES is away: (afk) [BX-MsgLog Off] u cnat just run away from me damnit :) 11:30pm abi: reply-to munging is at http://www.unicom.com/pw/reply-to-harmful.html, http://www.metasystema.org/essays/reply-to-useful.mhtml 11:50pm abi reply-to it has been said that reply-to is http://www.unicom.com/pw/reply-to-harmful.html [msg(TUNES)] newlog 1999.0807 IRC log ended Sat Aug 7 00:00:00 1999