00:00:00 --- log: started forth/04.09.07 00:00:03 i dont have cable or satelite or whatever 01:15:06 --- quit: wossname (Read error: 242 (No route to host)) 01:55:46 wow its all quiet all of a sudden 01:56:03 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 01:56:06 for the last 3 hours anwyways 02:27:35 --- nick: cduce -> futhin 02:29:06 --- nick: futhin -> blahblah 02:29:49 --- quit: wossname (Read error: 110 (Connection timed out)) 02:29:52 --- nick: blahblah -> futhin 02:33:14 --- nick: futhin -> thin 02:48:50 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 03:41:30 --- quit: wossname (Read error: 113 (No route to host)) 03:51:12 --- join: crc (crc@52-pool1.ras11.nynyc-t.alerondial.net) joined #forth 03:53:12 --- quit: ChanServ (Shutting Down) 03:53:28 --- join: ChanServ (ChanServ@services.) joined #forth 03:53:28 --- mode: irc.freenode.net set +o ChanServ 04:06:22 --- quit: ChanServ (Shutting Down) 04:06:38 --- join: ChanServ (ChanServ@services.) joined #forth 04:06:38 --- mode: irc.freenode.net set +o ChanServ 04:07:33 hi crc ! 04:07:55 Hi 04:07:59 do u know any more J2ME mail client ? 04:08:08 No 04:08:10 Mail4ME is _bad_ 04:08:14 :-( 04:08:28 There don't seem to be many free mail clients in Java 04:09:17 it shows setup screen at start, has damn long latency and shows only FROM field from MSG's - then i ask it to read, it constantly says 'network down' 04:22:50 --- quit: ChanServ (Shutting Down) 04:23:05 --- join: ChanServ (ChanServ@services.) joined #forth 04:23:05 --- mode: irc.freenode.net set +o ChanServ 04:34:40 * Serg_penguin is browsing ASCII ART sites 04:37:59 --- join: saon (Ecoder@c-24-129-95-254.se.client2.attbi.com) joined #forth 05:16:31 --- quit: tgunr ("Leaving") 05:16:44 --- join: mur_ (~mur@mgw2.uiah.fi) joined #forth 05:25:21 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth 05:27:10 --- quit: mur (Read error: 110 (Connection timed out)) 05:57:39 * crc hates 16-bit assembly 06:15:52 --- quit: arke (Read error: 110 (Connection timed out)) 06:21:49 --- quit: crc ("Time for bed... Goodnight!") 06:24:38 --- quit: tgunr ("Leaving") 06:27:52 --- join: arke (~Chris@wbar8.lax1-4-11-100-111.dsl-verizon.net) joined #forth 06:28:27 it would really be nice if there was an SDL binding for ficl 06:28:47 its not hard to do, its just busy work 06:29:50 --- join: I440r_ (~mark4@216-110-82-59.gen.twtelecom.net) joined #forth 06:31:56 --- join: mur (~mur@smtp.uiah.fi) joined #forth 06:33:36 --- quit: I440r_ (Client Quit) 06:36:02 --- join: LuckyPhil (~phowlett@CPE-141-168-97-228.qld.bigpond.net.au) joined #forth 06:38:51 --- quit: mur_ (Read error: 60 (Operation timed out)) 06:43:52 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 06:43:52 --- mode: ChanServ set +o Herkamire 06:51:01 --- quit: Serg_penguin () 06:58:10 * arke is away: school --- anybody want to ... uuum ... create ficl SDL bindings? :) 07:01:24 --- part: LuckyPhil left #forth 07:49:22 --- join: mur_ (~mur@uiah.fi) joined #forth 07:57:27 --- join: arke-school (apache@11.198.216.81.dre.siw.siwnet.net) joined #forth 07:57:46 hiya 07:57:50 who's awake? :) 07:57:59 --- quit: mur (Read error: 110 (Connection timed out)) 07:59:34 #define AWAKE 08:00:12 yay 08:00:22 madwork: whats up? 08:00:55 Just workin'. Writing a bunch of Forthy scripts. 08:02:36 --- join: I440r_ (~mark4@216-110-82-203.gen.twtelecom.net) joined #forth 08:05:36 madwork: i get to write a ficl-SDL binding 08:05:46 nobody else has done that 08:05:49 :/ 08:05:57 what a pain 08:08:27 I'm not sure how FICL's user interface is. 08:11:29 its not bad, the only problem is that its horribly documented 08:11:38 you cant get _anything_ done unless you scan ficl.h 08:11:54 but on top of that, it provides an ANS forth 08:12:01 and you can add C functions 08:12:47 Yep. 08:13:10 JAVA 08:13:11 I think Forthy would be easier to interface to SDL. 08:13:31 But at this point, it's even more horribly documented. ;) 08:13:38 well, SDL has lots of functions etc. 08:13:39 :) 08:13:40 hehe 08:13:56 the guy who did ficl came in here once or twice, i think i annoyed him and he never came bacl lol 08:13:57 oopts 08:14:10 Sure. Forthy is all about exporting the C functions to the VM. 08:15:12 in ficl, you have to write wrapper functios 08:15:53 Yes, this is the case for most scripting systems. Mine follows a method similar to Lua's. 08:16:10 void _f_SDL_Init(ficlVm *vm) { ficlVmPush(vm, (ficlCell) SDL_Init(ficlVmPop(vm).i)); } 08:16:41 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 08:16:46 ficlDictionarySetPrimitive(dict, "SDL_Init", _f_SDL_Init, 0); 08:16:49 hi wossname 08:17:13 hi arke 08:18:44 static void word_fsub(FSYSTEM *sys) 08:18:44 { 08:18:44 double result; 08:18:44 08:18:44 result=fs_get_float(sys, -2)-fs_get_float(sys, -1); 08:18:46 fs_remove(sys, 2); 08:18:47 fs_push_float(sys, result); 08:18:49 } 08:18:54 fs_register_func(sys, "f-", word_fsub, FS_DEFAULT); 08:19:40 :/ 08:20:17 Same kinda deal, really. 08:24:42 yeah 08:30:38 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 08:30:45 oh well. just busy work 08:30:49 hi onetom 08:34:52 --- quit: onetom_ (Read error: 110 (Connection timed out)) 08:48:45 --- join: crc (crc@53-pool2.ras11.nynyc-t.alerondial.net) joined #forth 08:49:35 hi crc 08:51:15 Hi arke 08:51:37 * crc is exploring the mysteries of 16-bit x86 assembly language 08:51:55 I haven't used 16-bit assembly for anything other than a boot sector in years 08:51:57 hehe 08:52:06 ask me for help if you need it :) 08:52:10 16-bit is fun 08:52:13 Help! 08:52:21 * crc is doing a 16-bit version of RetroForth 08:52:26 crc: whats teh problem? :) 08:52:29 hehe 08:52:30 it doesn't work 08:52:39 I *think* it's PARSE or QUERY 08:52:42 But I'm not sure 08:53:05 /msg me 08:53:07 :) 08:53:09 l 08:53:11 ok 08:55:18 --- join: randolm (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 08:57:12 --- quit: mur_ (Remote closed the connection) 08:57:47 --- join: mur (~mur@kyberias.uiah.fi) joined #forth 08:57:51 hi randolm 08:58:31 hoy 08:58:43 :) 09:02:31 --- quit: arke-school ("be back in a few hours") 09:10:43 --- quit: wossname (Read error: 110 (Connection timed out)) 09:19:56 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 09:25:00 --- join: FlamingRain_ (Ecoder@c-24-129-95-254.se.client2.attbi.com) joined #forth 09:25:13 --- quit: FlamingRain_ (Read error: 104 (Connection reset by peer)) 09:30:35 --- quit: crc (Client Quit) 09:35:28 morning 09:36:48 --- quit: I440r (Nick collision from services.) 09:36:50 --- nick: I440r_ -> I440r 09:36:53 --- mode: ChanServ set +o I440r 09:36:59 --- join: I440r_ (~mark4@216-110-82-59.gen.twtelecom.net) joined #forth 09:44:51 --- nick: saon -> saon|other_compu 09:50:00 --- join: mur_ (~mur@uiah.fi) joined #forth 09:57:31 --- nick: saon|other_compu -> saon 09:58:36 good morning kc5tja 10:01:05 --- quit: mur (Read error: 110 (Connection timed out)) 10:01:41 --- join: Murrlin (murr@dialup-207-218-213-125.ev1.net) joined #forth 10:03:17 --- quit: Robert ("MONEY") 10:03:25 --- quit: Murrlin (Client Quit) 10:05:53 --- join: t35t0r (~xsc@cookeville-24-158-168-215.midtn.chartertn.net) joined #forth 10:06:02 " if the linux kernel was coded in forth my compiler would compile it in 3 seconds FLAT on an amd k6-3/550" heh can i quote you on that? 10:06:13 i'm gonna email it to torvalds 10:06:25 t35t0r, ive already talked to linus. he KNOWS forth 10:06:35 so why doesn't he make everyone use forth? 10:06:44 but he doesnt like languages without datatypes 10:06:54 no datatypes!!! 10:06:59 so its' like perl1 10:07:00 datatypes is the compilers way of playing mommy and holding the coders hand 10:07:06 err no way 10:07:13 perl has no datatypes 10:07:17 "typing is a crutch for poor programmers" 10:07:19 -- chuck moore 10:07:26 not explicit ones 10:07:33 any 10:08:01 "im sorry but im NOT going to allow you to store that CHARACTER in that memory cell because its reserved for BYTE info" 10:08:02 bleh 10:08:04 lol 10:08:20 well maybe the compile process is ridiculously fast as you stated, but how's the performance? 10:08:42 download the compiler and build it and see for yourself 10:08:59 try ./isforth -fload src/examples/window.f 10:10:05 datatypes are helpers for poor programmers? heh 10:10:11 crutches 10:10:18 same thing 10:10:22 ok lol 10:10:33 its the compiler being mommy 10:10:40 i'll just have ridiculously large variable names 10:10:41 and protecting the little children from themselves 10:11:01 intVar1, stringVar2 ? 10:11:35 anyways i'll go look for a win forth compiler 10:11:40 i don't have a nix computer 10:11:49 lol thats like verbose hungarian notation ??? :) 10:11:54 oh 10:12:00 well there are other forths for windows :) 10:12:12 but they arent as good as isforth *snicker* 10:12:23 i'm anti-variables 10:12:36 yea me too 10:12:36 fight the war against variables! 10:12:45 that's why i do most evertyhign in perl but its slow 10:12:50 really slow 10:14:39 why perl? use forth.. :) 10:14:42 t35t0r: I don't use windows, but winforth (and bigforth) run under windows 10:15:02 yea i will check out forth 10:15:02 gforth too if you have the cgi-win or whatever it's called 10:15:11 yea i have cygwin 10:15:28 cygwin, that's it 10:15:35 i just keep win around for matlab/maple/pspice/labview 10:16:15 isforth wont run in windows 10:16:20 not even in cygwin 10:16:38 yea cygwin only runs pcode 10:16:50 same as win binaries 10:16:59 with cygwin.dll layer 10:18:42 i wonder have some of the fft libraries and other number crunching libraries been converted to forth and if so how are the benchmarks? 10:19:23 google for iforth or marcel hendrics - if anyones done that he has :) 10:30:55 forth doesn't do so well in the shootout 10:33:07 because they used crap forths 10:33:30 i bet iforth would be at the VERY top of that list on most benchmarks 10:33:38 pcode? 10:34:35 Forth will never be as fast as raw C in performance. Most Forth implementations do not keep anywhere enough state information to enable processor-specific optimizations. 10:34:53 does iforth ? 10:36:08 iForth and bigForth both keep colon definitions in a DTC-like representation, which allows them to do high-level code optimizations (see the paper on RAFTS, by Ertl and Paysan). This is why I said *most* Forth compilers. 10:36:31 However, even true, native-code Forth compilers can't match the performance of C because, well, C is typed. 10:36:54 There are many optimizations you can do to boost execution speed if you have type information. 10:37:04 brb 10:38:13 --- join: mur (~mur@smtp.uiah.fi) joined #forth 10:40:13 the processors in our computers are designed for C 10:40:23 they don't really work the same way forth does 10:40:45 so optomizing forth to the extent that gcc optomizes C is a big job, and not always feasable 10:41:37 It is feasible. 10:41:39 ML proves that. 10:42:22 With Forth, it is trivially possible to perform what's called "type inferencing," which lets you deduce/infer the proper types the program deals with without having to explicitly declare or cast types all over the place. 10:42:56 But there has to exist a root-set of objects which are explicitly given types so that the inferencer doesn't wander off into space (and if it does, then you get a type error). 10:43:27 The thing is, in order for a type inferencer to work, you cannot support variable-length arguments (which, true, Chuck doesn't like at all either -- nonetheless, they're damn handy to have). 10:43:54 kc5 you mean things like foo ( ... ) ? 10:44:04 variable NUMBER of parameters ? 10:44:05 Now, the number of arguments a word takes can also be inferred (since it's a kind of type information) 10:44:30 I440r: Your example is meaningless. What doyou mean "things like foo ( ... ) ?" 10:44:50 ... is used in c to say an unknown number of other parameters 10:45:01 look at printf 10:45:06 Yes. 10:45:11 Forth fully supports varargs. 10:45:19 It also supports something C doesn't -- varresults. 10:45:25 yup :) 10:45:25 Consider FIND in most Forths. 10:45:48 ( n1 --- n2 true | false ) 10:45:51 All that goes away if you want type inferencing, static type checking, and therefore, the performance optimizations that such things enables. 10:46:01 returns n2 true or it returns a JUST a false 10:46:06 thats not for find tho lol 10:46:27 that would be ( a1 --- [ a2 t | 1 ] | false ) 10:46:45 ?? 10:47:12 FIND ( naddr -- xt 1 | xt -1 | 0 ) 10:47:26 if find finds what your searching for it returns the CFA of the word plus a flag which will be either true or 1 depending on weather or not the word is immediate 10:47:33 my notation says exactly what yours does 10:48:00 No; if you flatten it out, it reads a2 t | 1 | false. 10:48:01 [ maybe these ] | this 10:48:24 it will return everything between the [ ] or the other singele item 10:48:30 I think you probably meant: [ a2 [ t | 1 ] ] | false ] 10:48:33 its a standard notation 10:48:43 So standard I've never seen it. 10:48:44 ugh. why wrap everything in [ ] 10:48:52 kmc5 is simply splitting hairs 10:48:57 kc5 even 10:49:10 Because [a2 t | 1] is no different from ( xt t | f ) 10:49:22 If you're going to follow a syntactical rule, follow the rule consistently. 10:49:23 who cares about optimizing forth, its already blazing fast ;) 10:49:30 correct but thers ANOTERH OR OPERATION THERE 10:49:43 you dont want ( --- a | b | c ) 10:50:10 I440r: Yes, there is another OR operation. That is precisely the source of the error: AND binds higher than OR. 10:50:17 And binds to the left, not the right. 10:50:23 err no it doesnt 10:50:28 thers no presidence at all 10:50:31 OK, whatever. You keep on dreaming. 10:50:42 thin: That's a provable fallacy. 10:50:44 --- quit: mur_ (Connection timed out) 10:50:46 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 10:51:00 ( a1 --- [ if we return an address we also return one or other flat ] otherwise we reutn just a flag 10:51:14 kc5tja: well i was talking more in the sense of a forth program vs a c program if both are large nontrivial programs 10:51:18 your simply being arguementative 10:51:29 my notation is correct 10:51:37 I440r: Says who? 10:51:41 I don't see how type checking helps optomization 10:52:02 kc5 who says its wrong 10:52:03 you ? 10:52:04 fine 10:52:22 Herkamire: Type checking imposes guarantees on how values are used. You can't add two pointers, for example. That enables the optimizer to make assumptions about how code will be used 10:52:33 sounds logical to me 10:52:45 so type checking has to be done at compiler time 10:52:50 I440r: says me and the rest of the computer programming world, that's who. C'mon man. Get a clue for a change, and stop being different for the sake of being different, then claiming we're so damn argumentative. 10:53:34 hahaha 10:53:42 kc5tja: I don't see how that would help. are you trying to use special purpose registers or something? 10:54:11 I440r: can you code in x86 asm? 10:54:30 Herkamire: no 10:55:05 Herkamire: If you can guarantee invariants about how data is going to be used statically at compile-time, like I said, you can make assumptions about how to re-organize code. 10:55:08 I don't see how it helps to know that two things won't be added together 10:55:24 yeah, I guess 10:55:41 it's just that as far as I'm concerned they are all integers. 10:56:00 But pointers are not integers. 10:56:06 They can be represented as such inside the computer. 10:56:16 But from a logical perspective, they're references. 10:56:22 the only thing I can think of that that sort of information would be useful, is for special registers. eg r0 on ppc behaves differently for a few instructions. 10:56:40 you can subtract pointers 10:57:34 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 10:58:24 You can subtract pointers only if (1) they point to the same kinds of objects, and (2) they both point inside an array of said objects. If they don't, the result is completely and utterly meaningless. And even then, you still need to divide by the size of the object to get a delta index. (I can't conceive of any other reason why you'd want to subtract pointers otherwise.) 10:59:10 generally to see how much is between them 10:59:28 Look, I'm neither arguing for or against pointers. I'm just saying that there exists a set of optimizations which are enabled by the knowledge of types. 10:59:30 eg subtract the bigining of the dictionary from the end to get the size 10:59:44 s/pointers/types/ 11:00:14 I'm sure you're right in general. 11:00:19 t35t0r, yes 11:00:23 The biggest advantage to types is knowing how big they are at compile-time, so you can pre-allocate space on the stack frame, or whatnot. 11:00:46 assumptions (correct ones) tend to lead to optimizations. 11:00:57 unfortunately, I've had way to many experiences where typing got in my way 11:01:13 herkamire: yeah? how did they get in the way 11:01:21 I'm happy to have a programming environment that has no concept of a type-error 11:01:23 Reading Baker's document on lazy allocation, for example, where *ALL* program objects are allocated on the stack, but moved to the (garbage collected) heap only when absolutely necessary is a brilliant example. It's just not possible to do in Forth because, in Forth, structures are just arbitrary bags of cells. There's no innate way to determine the size of a structure. 11:02:00 couldn't you determine the size of a structure by looking at the forth code? :) 11:02:12 yeah, I can see it would be hard to optomize forth because it would be hard to tell what stuff is 11:02:24 thin: That is type inferencing. You're still statically typed even if you don't explicitly give a type. 11:03:37 I want to point out that ML uses type inferencing to also perform memory usage inferencing too! As a result, ML-compiled programs run very fast, and actually have only moderate memory requirements, since the overwhelming majority of space allocated never needs to be garbage collected all that often. 11:04:10 But then, ML performs something like 24 passes over the compiled source code too. Compilation is very slow, but the results are outstanding. 11:04:46 do you happen to be familiar with smalltalk too? 11:04:46 ml sounds pretty cool 11:04:51 i wonder how smalltalk & ml compare 11:05:03 thin: They don't. 11:05:07 i still haven't got the hang of ;s in ML 11:05:14 I had a friend telling me how wonderful ml was, and I read about it. seemed pretty cool. but I couldn't get it to run 11:05:18 Smalltalk is like Forth -- typeless, or more accurately, dynamically typed. 11:05:25 ML is very rigid, strongly typed, etc. 11:05:36 if there's an open source ml that would run under ppc linux I'd like to hear about it. 11:05:39 Smalltalk is object oriented, ML is functional, etc. 11:06:01 Herkamire: There isn't? 11:06:09 there might be now. 11:06:17 I didn't have luck finding one a couple years ago 11:06:20 I thought stuff like SML/NJ runs everywhere. 11:06:29 or I found one and it didn't work... I forget 11:06:39 http://caml.inria.fr/ -- I suspect that if you'll find it, it'll be here. 11:06:41 I'm pretty sure I tried that 11:06:57 (sml/nj 11:07:09 Personally I'm using OCaml, but I don't know about PPC support... 11:07:46 Google for PowerPC site:caml.inria.fr -- the results indicates that PPC is supported. 11:08:06 i tried ocaml, but my inability to understand where to put the ;s stumped me slightly... heh 11:08:18 Topaz: :D 11:08:32 i did write a semi-functional IRC server in it though 11:08:36 a strange first-project, admittedly 11:08:44 Hehe. 11:08:55 but my trying to keep the compiler happy descended into randomly putting ;s in until it stopped complaining 11:09:07 Good method. :) 11:09:36 most OCaml tutorials seem to subconciously assume a familiarity with ML 11:09:52 (ML programmers probably can't understand how people can possibly find it confusing ;) 11:14:33 --- quit: randolm (Read error: 60 (Operation timed out)) 11:15:10 Forth programmers can't find any reason why it's confusing to others either. 11:16:26 It isn't. 11:17:19 :D 11:17:38 C is confusing :P 11:17:46 Nahhh 11:17:47 c is write only :) 11:17:48 "why is C so stupid??" :P 11:18:32 i feel bad for people that learn the imperative paradigm before learning forth or lisp 11:18:33 Hey, I'll be right back. I have to talk to the landlords about renewing my lease. 11:22:10 ugh 11:22:23 can someone read man tzfile and tell me if THEY understand it ? 11:22:54 haha 11:22:55 it took like two reads of man terminfo to understand that enough to start parsing terminfo files... i dont understand how to parse timezone files at all 11:22:59 --- quit: warpzero ("Tried to warn you about Chino and Daddy Gee, but I can't seem to get to you through the U.S. Mail.") 11:24:46 where can i get the source file for time ? 11:24:54 cuz that man page is next to useless to me lol 11:27:27 still don't understand what exactly you are trying to do 11:27:35 are you looking for header files 11:28:05 --- quit: mur (Read error: 60 (Operation timed out)) 11:28:25 --- join: mur (~mur@mgw2.uiah.fi) joined #forth 11:30:38 cool. caml-light installed fine. 11:30:51 now I need to learn a bit of ml again :) 11:31:20 well, I'm back to werk. bbl 11:36:55 no. im trying to figure out how to interpret /etc/localtime 11:36:56 --- quit: arke (Read error: 54 (Connection reset by peer)) 11:37:08 so i can get the local offset from gmt 11:37:41 --- join: arke (~Chris@wbar8.lax1-4-11-100-111.dsl-verizon.net) joined #forth 11:42:15 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 11:43:30 I440r: maybe you should look up the program which creates those files 11:43:35 i forget its name 11:43:46 look up its code 11:44:02 but its in one of the man pages 11:44:32 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 11:46:03 I440r: man tzfile 11:52:25 --- quit: saon ("Leaving") 11:52:38 --- quit: t35t0r ("ircN 7.27 + 7.0 for mIRC (2002/01/10 00.00)") 11:56:31 --- join: warpzero (~warpzero@mi018.dn178.umontana.edu) joined #forth 12:09:02 kc5tja, the xp ebook kept mentioning CRC, is that cyclic redundancy code ? 12:09:53 thin, yes 12:10:27 slava, what does it mean? 12:13:02 thin: No. It's something meaning "Code ???? Card". I don't know what ??? means though. I forgot. It's on the C2 Wiki somewhere though. 12:13:36 here i go, confusing people again :) 12:13:43 Anyway, I have to get to cleaning the house and all, so I will be afk for most of today. 12:15:30 now, the fun challenge of installing Gentoo/MIPS on a SGI O2 12:15:45 "class-responsibilty-collaboration" apparently 12:19:33 hmm cypress hill rocks 12:34:52 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 12:39:07 --- quit: tathi (Client Quit) 12:43:36 --- join: mur_ (~mur@uiah.fi) joined #forth 12:55:23 --- quit: mur (Read error: 110 (Connection timed out)) 13:00:25 --- join: saon (Ecoder@c-24-129-95-254.se.client2.attbi.com) joined #forth 13:00:44 hi saon 13:01:28 hello slava 13:07:34 good morning slava 13:07:43 and saon 13:09:59 man tzfile seems readable to me 13:10:22 subversion is the highest ideal! 13:10:30 subversion is nice 13:10:34 ok. how do i extract the local offset from gmt ? 13:10:37 i dont get it 13:10:45 it's sometimes quite annoying to install though 13:11:09 i did find a bug in the fbsd version of the man page tho :) 13:12:42 I440r: it's not that simple 13:12:47 there's daylight savings time 13:13:02 the offset depends on what time of year it is 13:14:24 ok. so - that man page isnt very helpfull is it 13:15:35 :/ 13:17:46 so i guess i better go dig out the sources to localtime() 13:17:47 ugh 13:19:07 --- quit: saon ("Leaving") 13:29:49 ok - i think ive got a better chance of becoming the POPE than i do over ever understanding THAT mess 13:32:05 dammit i need to write a proper assembler instead of this ad-hoc mess 13:32:14 i have like 5 words for outputting different forms of MOV 13:32:17 so do i 13:32:22 and one of them is broken 13:32:26 i'm not sure which :) 13:34:37 I440r, do you know of any tool that i can run my program under, and then disassemble arbitrary memory regions that i specify 13:34:40 could gdb do this 13:34:57 yes it can! 13:35:04 gdb wont disassemble arbitary addresses - unless they just happen to belong to your process 13:35:10 --- quit: mur_ (Remote closed the connection) 13:35:24 under windows i would say use softice :) 13:35:33 that will disassemble ANY address in memory 13:35:42 --- join: mur (~mur@smtp.uiah.fi) joined #forth 13:41:28 I440r, look at my bizarro assembler syntax: 13:41:29 ( cell -- ) ECX [I]>R 13:41:29 DATASTACK EAX [I]>R 13:41:29 ECX EAX R>[R] 13:41:29 4 DATASTACK I+[I] 13:45:11 --- join: arke|school (apache@11.198.216.81.dre.siw.siwnet.net) joined #forth 13:45:16 hi 13:45:17 :) 13:45:22 hi arke 13:45:41 arke, can you recommend a very very simple x86 assembler in forth i can study? 13:46:28 pygmyforth 13:46:29 :) 13:46:34 that's 16 bit 13:47:44 yep 13:47:55 i need 32 bit :) 13:47:56 --- quit: arke|school (Client Quit) 13:48:53 very very simple x86 assembler AND 32 bit? oooh i'd like to see one :P 13:49:09 thin, well, you can just wait until i finish mine :) 13:49:13 i was writing one :) 13:49:25 but it's in c, oh no 13:49:27 thin, just be warned, i'm not interested in following nasm OR gas syntax ;) it will be a postfix assembler 13:50:06 cool 13:50:11 screw backwards compatibility 13:50:20 as long as its SUPER SIMPLE 13:50:27 other wise i won't respect you ;) 13:50:41 i hate all bloat, and i only respect forthers that follow forth philosophy 13:50:44 it will only do the opcodes that i need 13:50:49 for my STC 13:51:47 --- quit: qFox ("meh ircii's netsplit detection appearantly sux ^^") 13:54:42 great 13:54:58 perfect! 13:57:48 thin, also i'm not sure if it will bring a benefit, but i want to investigate inlining certain words 13:58:21 another idea i had is if you have >r ... r>, instead of moving from ds to cs and back, use a spare register 13:58:39 but of course >r >r ... r> r> etc might need to move to the cs, since you'll run out of registers:) 14:13:09 that sounds more like forth implementation details 14:13:15 not strictly assembler 14:13:24 yes i know 14:13:32 the compiler will be a thin layer around the assembler 14:15:36 --- quit: warpzero (Excess Flood) 14:16:24 --- join: warpzero (~warpzero@mi018.dn178.umontana.edu) joined #forth 14:22:22 simple and x86 don't mix ;) 14:23:16 x86 is a magical architecture 14:24:16 its confusing how encodings are diffeerent for MOV's involving EAX 14:25:05 it's retarded that you have the same mnemonic for radically different operations 14:25:13 yes 14:25:18 there's like 10000 different forms of MOV 14:25:27 LEA is worse actually 14:25:57 isn't it great! 14:26:07 Hmmm... No. 14:26:15 the main reason for that is because of the whole EAX thing 14:26:32 i.e. the whole 32 bit thing 14:26:48 special treatment of EAX is a holdover from 8080 which just had one accumilator register 14:31:52 --- join: randolm (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 14:32:42 --- quit: wossname (Read error: 104 (Connection reset by peer)) 14:33:05 --- join: mur_ (~mur@mgw2.uiah.fi) joined #forth 14:38:19 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 14:39:35 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth 14:39:39 --- quit: randolm (Read error: 104 (Connection reset by peer)) 14:45:38 --- quit: mur (Read error: 110 (Connection timed out)) 14:49:02 --- quit: wossname (Read error: 60 (Operation timed out)) 15:17:10 --- join: mur (~mur@uiah.fi) joined #forth 15:20:10 --- quit: mur_ (Read error: 60 (Operation timed out)) 15:22:35 dlava ive been procrastinating writing an assembler for isforth for 3 years lol 15:23:48 i440r: because you haven't solved the problem of keeping it to nasm style without preprocessing right? 15:24:04 thin no. 15:24:40 i can do delayed processing. i.e. dont interpret a mneumonic till the next one is seen. 15:24:47 I440r, i might have to use far jumps always, here is why 15:24:50 i.e. each mneumonic causes the previous one to be assembled 15:24:51 I440r, when i compile the start of a branch 15:24:58 i can do nasm style without any processing at all 15:25:00 mwahahaha! 15:25:02 I440r, i don't yet 'know' the length of the branches 15:25:17 I440r, so i have to reserve some space for a fixup when i finish compiling the branch 15:25:23 I440r, and do a fixup with the right addr later 15:25:24 I440r, get me? 15:25:26 slava exactly. you usually do something lile "compile branch here 0, 15:25:38 :) 15:25:40 here is on the stack and later on in your code you do here swap ! 15:29:49 i just realised something - spammers mine usenet for email addresses - why not post articles in all groups with like 40 million invalid email addreses 15:30:42 make en waste time mailing 40 million invalid addressses 15:34:11 --- part: thin left #forth 15:35:58 --- quit: warpzero (Read error: 110 (Connection timed out)) 15:38:40 --- join: warpzero (~warpzero@mi018.dn178.umontana.edu) joined #forth 15:53:23 --- join: TheBlueWizard (TheBlueWiz@63.250.24.233) joined #forth 15:53:23 --- mode: ChanServ set +o TheBlueWizard 15:55:48 hi TheBlueWizard 16:01:07 arke hiya 16:01:36 --- quit: warpzero (Excess Flood) 16:04:47 --- join: jdavidboyd (~user@104-170.35-65.tampabay.rr.com) joined #forth 16:06:06 --- quit: jdavidboyd (Remote closed the connection) 16:06:09 --- join: warpzero (~warpzero@mi018.dn178.umontana.edu) joined #forth 16:06:09 --- quit: warpzero (Excess Flood) 16:06:41 TheBlueWizard, got any Forth projects? 16:07:02 --- join: warpzero (~warpzero@mi018.dn178.umontana.edu) joined #forth 16:08:03 arke: had one, but put it on back burner (for so long :) 16:08:29 * arke is back (gone 04:30:31) 16:08:33 TheBlueWizard, what was it? 16:12:01