00:00:00 --- log: started forth/04.07.22 00:14:04 --- join: Klaw (~anonymous@ip68-228-92-218.oc.oc.cox.net) joined #forth 00:19:58 --- quit: Fractal (Read error: 110 (Connection timed out)) 01:10:34 it has no Yahoo-like directory, sucks ! 01:17:10 Hi 01:17:24 Make one, serg ;) 01:30:13 it must be designed from beginning 01:30:36 * Serg_penguin is writing tree site engine w/ pages, in SQL, all site in one table 01:49:51 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 01:50:30 Hi 02:00:40 i gonna change my matchbox-cam - it has broken line in 7-th shot at same place ;( 02:34:35 --- quit: Herkamire ("bed") 02:37:07 Serg_penguin: :/ 02:37:21 Maybe you can expect that from very cheap cameras. ;) 03:24:13 --- join: aum (~aum@port-204-54-210.fastadsl.net.nz) joined #forth 03:24:21 anyway, it has a year of warranty, and i'll use it to full extent 03:24:55 --- join: crc (crc@0-1pool176-2.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 03:53:41 --- join: arke (arke@melrose-251-251.flexabit.net) joined #forth 03:54:38 HI 03:54:48 iH 03:54:56 Moo..ooM 03:55:22 Hello ... olleH 03:55:30 Good morning arke 03:57:31 good morning charles childers 03:58:56 Are you going to start that again? 03:59:51 What did you expect, a sane discussion? :) 03:59:59 Yes 04:00:00 For once 04:00:02 :-) 04:00:14 Optimist... 04:00:21 I always am 04:00:52 * crc is attempting to get the new kernel working... 04:01:25 :) 04:02:06 Which kernel? 04:02:33 The new kernel for RetroForth/Native 04:02:45 Ah, OK. 04:02:57 A lot of nice stuff - once I get it working 04:04:37 Multitasking, better keyboard driver, improved HDD driver... 04:05:18 * Robert thinks of Chuck's HDD "driver". 04:05:29 Mine 04:05:40 Mine's in assembly. ~100 lines 04:05:51 Loads/saves one sector at a time, using PIO 04:05:55 Very reliable though 04:06:03 OK, any special features? 04:06:33 Not yet 04:06:44 I'm working on quite a bit though: 04:06:59 Persistant storage (to first track on hard drive) 04:07:05 --- quit: lalalim_ (Remote closed the connection) 04:07:10 --- join: lalalim (~lalalim@p508AB336.dip.t-dialin.net) joined #forth 04:07:17 Networking (TCP/IP, PPP) via NIC 04:07:27 VGA/SVGA graphics 04:07:31 crc: coooooool 04:07:38 crc: OH OH OH OH OH OH OH I DO GRAPHICS!!!! 04:08:14 crc: Nice.. Well, I'm going out now, see you later. 04:08:20 Ok 04:08:28 If you manage to get SVGA working properly I'm interested. :) 04:08:56 Oj 04:08:59 Ok 04:09:11 * crc is tired - I just got home from work 04:13:21 arke: can you do a nice 320x240 256 color screen for the boot logo? 04:13:40 I'm not much good with graphics :-( 04:14:03 Assuming that all goes well, I'll have this up and running by the weekend 04:14:40 I have to strip out a few pieces of 'C' code that Tom used in the kernel before I can fully test it 04:14:53 heh 04:15:00 crc: I can start once I'm back homke. 04:15:08 crc: but yes, I can do graphics. 04:15:15 crc: will you let me cheat? 04:15:16 cool 04:15:20 How? 04:15:41 let me perform vm86 calls? 04:16:08 The BIOS functions? 04:16:13 and let me put a pre-pm hook, for init 04:16:22 yes, but only the mode change ones. 04:16:30 Go right ahead 04:16:32 putpixel etc. I can do myself :) 04:16:33 cool 04:16:39 but 04:16:42 I'm not opposed to that 04:16:45 i know nothing of vm86 etc. 04:16:54 so you'll have to help me cheat ;) 04:17:08 well, i meant that i don't know how to do it etc. 04:17:29 i need to be able to claim the real mode 0xa000 segment 04:17:36 The current boot sector changes the mode before switching to PM 04:17:39 Sure 04:17:43 cool. 04:17:44 The memory layout is linear 04:18:00 cool. 04:18:04 is vm86 hard? 04:18:08 Yes 04:18:12 But it's doable 04:18:21 I've done a kernel using it before... 04:18:28 Ok. 04:18:32 Though I prefer to use pure PM code :-) 04:18:47 would you prefer graphics code in asm or forth? 04:18:50 or both? :) 04:19:10 Either one is fine, if you document it :-) 04:19:28 (Just basic comments) 04:19:46 eww comment ;) 04:20:10 Just something like: "This draws XXXX at X,Y" or something 04:20:42 I hate totally uncommented code, but I don't like detailed, line by line comments either 04:21:23 is stack comment enough? 04:21:49 Yes 04:21:57 my words are all less than 80 bytes, including stack comment and name 04:22:02 Cool 04:22:08 so I would rather do 04:22:20 Most of mine are (except for the assembly ones) 04:23:16 : putpixel ( x y c - ) push fbaddr fbseg @ swap pop l! ; 04:23:26 is that commented enough? :) 04:23:30 Yes 04:23:44 col 04:23:46 cool* 04:23:53 its pretty much self-documenting 04:24:05 something you can't really do with most languaghes other than forth 04:24:08 I have a word named 'frame' that points to the start of the video RAM 04:24:10 I agree 04:24:19 That's one of the things I *love* about Forth 04:25:01 :) 04:25:02 Though I left out the stack comments in the current codebase -- they're in the Manual :-) 04:25:08 eww! 04:25:09 :) 04:25:46 No, since most of the binary size is actual Forth code that's compiled on startup, I wanted to keep it as clean as possible 04:26:10 Besides, do you really need stack comments for things like 'swap', 'dup', and so on? :-) 04:26:52 yes 04:26:57 i always do stack comments 04:26:57 Why? 04:27:06 That's fine 04:27:07 ITS A FORMALITY :) 04:27:16 RetroForth is informal 04:27:27 Most of the documentation is in the manual 04:27:37 (Which hasn't been released yet) 04:27:52 I don't think anyone other than me really understands the codebase at the moment 04:28:23 (Tom hasn't touched it since about 6.10, which was a long time ago) 04:28:46 And I didn't even share the code for 7.0 until a week or two ago :-) 04:29:16 The final codebase for 7.0 won't be released to the world until August 1 04:29:30 Though it's a lot nicer than the 6.xx releases were 04:29:42 :) 04:29:48 I just have to get the native version up to speed 04:30:06 It hasn't recieved any real updates since 6.03 :-( 04:30:13 ;) 04:30:34 I tried dropping the 7.0 codebase onto the old kernel and it totally f***ed up 04:30:39 have you ever considered metacompiling retroforth? 04:30:41 Everything broke horribly 04:30:42 Yes 04:30:51 That's a longer term goal (8.0 maybe) 04:31:06 I've been writing an assembler, target compiler, and so on 04:31:15 It's a slow process though 04:31:56 i was thinking 04:31:59 if you metacompile 04:32:11 and keep osdep code separate 04:32:32 then the different versions are likely to stay up to date even if you work on only one 04:32:38 The OS dependent code is mostly separate already 04:32:52 The only routines that depend on a specific OS are KEY, EMIT, and BYE 04:33:22 Which is why it can be released for new OSes so quickly 04:33:35 Going from Native to Linux took a month or so 04:33:46 Then from Linux to Windows was about a month and a half 04:33:53 FreeBSD took two weeks 04:34:04 BeOS was two days 04:34:09 uum 04:34:15 you forgot an important word 04:34:18 What? 04:34:22 key? 04:34:47 What's the difference between KEY? and KEY 04:35:58 KEY ( - c , block until a key is available and return it ) 04:36:13 KEY? ( - f , checks to see if a key is available ) 04:37:27 osdep words, i think, are 04:38:07 EMIT KEY KEY? AT CUR@ BYE and the block words 04:38:27 --- join: crc_ (crc@0-1pool176-2.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 04:38:38 Under the hosted versions, it Under the hosted versions, it blocks until ENTER is pressed. SubsequentUnder the hosted versions, it blocks until ENTER is pressed. Subsequent calls to KEY fetch from the buffer calls to KEY fetch from the bufferblocks until ENTER is pressed. Subsequent calls to KEY fetch from the buffer 04:38:56 --- quit: crc (Read error: 104 (Connection reset by peer)) 04:39:12 Under the hosted versions, it blocks until ENTER is pressed. Subsequent calls to KEY fetch from the buffer 04:39:22 eww 04:39:38 getchar semantics 04:39:39 --- nick: crc_ -> crc 04:39:41 eww 04:39:55 It's always worked for me 04:40:11 :-) 04:40:30 KEY? could be added to the native version easily though 04:40:48 I'm not sure how to handle it under Linux/FreeBSD/Windows/BeOS using syscalls 04:41:20 under unix, you can just set stdin to not buffer 04:41:41 How? 04:41:44 you can even do that with ANSI C 04:41:51 I haven't found a way to do that cleanly yet 04:41:56 theres a function called setbuf 04:42:05 in ANIS C 04:42:20 although thats C 04:42:32 but look in glibc for setbuf 04:42:34 That'll work for the BeOS port, which uses LIBC. I still have to think about Linux, FreeBSD, and Windows 04:43:03 the unices can do it with syscalls 04:43:28 What syscalls? 04:43:31 with windows too, but itll be a bit different 04:43:38 not sure, ioctl probably 04:43:48 * crc hates ioctl 04:44:00 What's wrong with just having key? 04:44:12 I like KEY :-) 04:45:15 Granted, writing games is harder... 04:45:27 But I have GameForth for that 04:45:50 bleh 04:45:59 getchar emantics are the worst 04:46:21 Why? 04:47:19 gotta go 04:47:23 email me and ask 04:47:23 ok 04:47:25 bye 04:47:28 bye 04:47:47 --- quit: arke ("Lost terminal") 04:48:17 crc: GameForth ? 04:48:41 GameForth is an extended version of RetroForth designed for programming games 04:50:23 Not released to the public yet 04:50:30 Windows only 04:50:56 Being developed at the request of a Forth programmer interested in using RetroForth instead of Win32Forth :-) 04:53:59 It's a low priority project, but is kind of fun 04:54:05 Nonstandard design :-) 04:54:22 * Serg_penguin aiming the smoking hole of /dev/null at Windows 04:55:50 I don't like Windows, but most of RetroForth's users seem to use it! 04:56:09 Linux is pretty well served by IsForth and gForth 04:56:20 There's a growing number of FreeBSD users 04:56:44 And it'll be the only alternative to gForth on BeOS 04:57:06 Eventually I'll be switching GameForth to use libSDL, then it'll run under the other OSes as well 04:57:36 Personally, I'm hoping to stop using any 'traditional' OS within a couple of years 04:57:55 RetroForth/Native is the foundation for my planned OS :-) 05:01:45 I wonder if I should switch to using LILO or GRUB to boot RetroForth/Native 7.0 05:03:15 * Serg_penguin dreams of UNIX in Forth, but has not enough skill to implement one 05:04:30 Tom is working on a simple UNIX-like kernel that has RetroForth built into the kernel 05:05:00 (Actually his work is what I'm using as the basis for the RetroForth/Native 7.0 kernel) 05:09:17 Personally, I want a very non-UNIX environment to use/program for 05:11:31 * crc is sick of traditional OSes 05:23:45 I'm not 05:31:37 Why not? 05:37:55 There are few OSes that don't draw heavily from the Unix design 05:38:15 Oberon System 3 comes to mind 05:45:13 --- quit: lalalim (Read error: 110 (Connection timed out)) 05:45:31 --- join: lalalim (~lalalim@p508ABDEB.dip.t-dialin.net) joined #forth 05:51:09 --- quit: Serg_penguin () 05:59:23 --- join: tucknip (pickroll@dialup-4.153.70.2.Dial1.Atlanta1.Level3.net) joined #forth 06:01:13 Hello tucknip 06:01:19 hi 06:06:02 * crc is continuing to hack RetroForth/Native's new kernel 06:06:38 what's retroforth's claim to fame? 06:07:19 * Robert returns 06:08:04 It's small, portable (across the various x86 OSes), and easy to learn, modify, and use 06:08:06 wb Robert 06:08:14 * aum downloads, makes... 06:08:22 is it x86 only? 06:08:25 Yes 06:08:29 shame 06:08:32 Why? 06:08:37 alienates mac users 06:08:43 True 06:08:49 But I don't have a Mac 06:08:54 and mac users outnumber linux users 06:08:57 When I get one, I'll do a Mac version 06:09:19 So? 06:09:28 I *do not have a mac* 06:09:39 I *can not develop/test for a platform I don't own* 06:09:39 not that i'm too worried - i'm an x86 linux user myself 06:09:42 :-) 06:09:51 Eventually there will be a mac version 06:10:05 Eventually being a long-term goal 06:10:08 is it easy or hard to add C primitives? 06:10:15 It's pretty easy 06:10:29 Look at the RetroForth 7.0-beta4 codebase for 'generic.asm' 06:10:31 are there hooks already, or do i need to write assembler shims? 06:10:41 You have to write assembler shims 06:10:52 crc: are you the developer? 06:10:55 Yes 06:11:10 would you consider adding a framework to ease implementation of c prims? 06:11:14 Yes 06:11:21 Once 7.0 is finished 06:11:22 aum: Among coders? 06:11:35 Robert: yes 06:11:35 (re: mac users outnumbering Linux/x86 users) 06:11:46 Robert: you might have a point 06:11:56 * aum tends to think linux coders way outnumber mac coders 06:12:12 They probably do, because x86 hardware is cheaper 06:12:16 wtf is 'fasm'? 06:12:17 And since it's a stand-alone system, all x86 coders can use it. Right? 06:12:24 FASM is the flat assembler 06:12:31 http://www.flatassembler.net 06:12:34 can't see fasm on debian feeds 06:12:42 It's not on them 06:12:53 Sadly... 06:12:58 nasm not suitable? 06:13:02 * crc also maintains the BeOS port of FASM 06:13:08 You can use NASM with a little work 06:13:23 What are the advantages of FASM? 06:13:23 The 6.30 release can be built with either NASM or FASM 06:13:33 1) Written in assembly language 06:13:38 2) Well supported + documented 06:13:40 I'm a NASM user now, but I'm getting a little annoyed because of its lack of macro features. 06:13:52 3) No libc dependencies (except on the BeOS port) 06:14:00 4) Good macro features :-) 06:14:04 Cool. :) 06:14:10 5) Fewer bugs than NASM 06:14:13 6) Less ' 06:14:20 But NASM has excellent documentation, is widely ported, and I already bothered learning it... 06:14:21 6) Less 'red tape' to deal with 06:14:32 Does it support different output formats? 16-bit code? 06:14:42 FASM is VERY close to NASM's syntax 06:14:43 Yes 06:14:50 16 bit or 32 bit code 06:15:16 flat memory (.com), PE, ELF (executable), ELF (relocatable), COFF, etc file formats 06:15:39 --- part: tucknip left #forth 06:15:50 No external linkers needed for plain assembly projects under Linux/Windows/DOS (unlike NASM) 06:16:20 There are a few directive changes: (NASM -> FASM) 06:16:24 RESB -> RB 06:16:28 RESW -> RW 06:16:32 INCBIN -> FILE 06:16:42 And so on 06:16:47 It's pretty straight forward 06:17:33 it's pretty quiet - doesn't say much 06:17:43 What RetroForth? 06:17:50 RetroForth is as silent as possible 06:17:57 Stays out of the way :-) 06:18:16 .s doesn't work 06:18:24 In which version? 06:18:35 linux-console-fasm 06:18:43 6.30? 06:18:44 crc: Nice :) Does it use any advanced syscalls? (asking because of portability reasons) 06:19:03 Not that I'm aware of 06:19:15 'words' gives incomplete output 06:19:19 --- part: segher left #forth 06:20:29 What's the last word listed by 'words' ? 06:21:11 ( 06:21:14 .s works here (under Windows, but the code for .s is totally portable) 06:21:17 before that, \ 06:21:30 That's the right ending 06:21:48 There's also a macro and forth vocabularies 06:21:54 Try 'forth words' 06:21:57 'macro words' 06:22:01 'domain words' 06:22:12 (Domain was dropped in 7.0) 06:22:30 6.3k binary ?!? 06:22:34 Yup 06:22:40 holy shit 06:22:42 Version 7.0 is <4k binary 06:22:53 you should do a version for ARM and similar embedded cpus 06:22:56 Is that the compiler? 06:23:12 That's the compiler and interpreter and basic words 06:23:15 OK. 06:23:28 A good chunk of that is raw ascii source that's compiled on startup 06:23:35 So it *could be smaller* 06:23:37 :-) 06:23:39 fat.f is a whole 1811 bytes - man that's bloated!! :P 06:23:39 I think my 16-bit Forth's corresponding binary is around 3kB. 06:23:50 That's pretty amazing with Forth.... 06:23:53 I plan to do an ARM version one day too 06:23:54 Useful but still small. 06:23:55 Yup 06:24:10 4k is pretty small for a 32-bit Forth :-) 06:24:22 Yes, but not extreme. 06:24:33 i'd love a forth-powered PXT GPRS cellphone 06:24:34 Still, small enough to fit everywhere. 06:24:39 I agree 06:25:50 here I was, feeling good about stripping a forth binary down to 144k 06:25:58 That's one of the things I like about RetroForth: It runs under Linux, FreeBSD, Windows, and BeOS (and should be easily ported to other OSes) 06:26:08 144k? What takes all THAt space? 06:26:12 aum: Your forth probably does a lot more than RF :-) 06:26:18 Is it written in C? 06:26:20 crc: slightly 06:26:23 crc: yes, in C 06:26:32 * aum has written a plugin framework for FICL 06:26:41 Ahh, FICL is pretty nice 06:26:49 * crc has used FICL in the past 06:27:03 * aum has coded plugins for GUI, sockets, even embedding a python vm 06:27:06 Wow 06:27:10 That's impressive 06:27:29 possible to implement methods of python classes in Forth 06:27:34 cool 06:28:01 manipulates python objects faster than the python bytecode interpreter 06:28:18 I've never done much with python 06:28:26 python r0x0r5 06:28:33 Once I discovered Forth, I haven't used much else 06:28:36 but python has a price - slowness and bloat 06:28:54 I have written a simple Lisp interpreter (in C) though... 06:29:03 tcn likes python 06:29:14 what i like about python is its casual attitude - it takes care of types in the background, and lets you think in the abstract without being pulled up every 5 seconds by low-level red tape 06:29:28 and, the 'everything is an object' paradigm 06:29:37 * crc hates ' 06:29:44 * crc hates 'objects' 06:30:00 crc: i thought you meant that you hate the ' word 06:30:10 No 06:30:26 I use ' sometimes 06:30:44 i use ['] more often than ' 06:30:48 : ' 32 parse find ; 06:30:53 What's ['] ? 06:31:05 crc: it's the immediate version of ' 06:31:12 In RetroForth that'd be: 06:31:24 macro : ' 32 parse find ; forth 06:31:33 Defining ' as a compiler macro :-) 06:32:06 (The compiler searches the macro vocabulary first, and handles the words in it as if they were IMMEDIATE) 06:32:23 In RF 6.30 'parse' is called 'word' 06:32:27 crc: try this in retroforth: 06:32:37 : fib dup 3 < if drop 1 else 1- dup recurse swap 1- recurse + then ; 06:32:56 Why? 06:33:00 1) No ELSE 06:33:05 what??! 06:33:06 2) No RECURSE 06:33:18 To calculate FIB I do: 06:33:29 : .1 >r over + swap r> repeat .1 ; 06:33:29 : fib 0 1 rot .1 swap drop ; 06:33:40 (That example is for RF 7.0) 06:34:03 I've never liked ELSE 06:34:20 crc: i take it your main background is assembler? 06:34:23 My method is faster than the recursive method 06:34:50 I learned to program in BASIC, then went through Pascal, C, Oberon, and finally assembly + Forth 06:34:56 With some use of LISP 06:35:08 i gather your heart is in the bit-bashing side of things 06:35:34 I try to do things in a simple, practical way 06:35:46 Recursion is complex, so I try to avoid it 06:35:56 recursion can be very wasteful 06:36:03 Yes it can 06:36:25 It can be useful at times though 06:36:35 retroforth would be a good OS for a pen-computer 06:36:43 Like a Palm? 06:36:46 or a wristwatch computer 06:36:53 crc: no - a computer in a pen 06:37:03 I'd love to see that :-) 06:38:35 The main use of recursive FIB is for benchmarking? 06:40:08 I can't seem to find any problem with '.s' 06:40:35 crc: .s don't work on my debian box 06:40:48 What does it do? 06:40:52 nothing 06:40:59 What's on the stack? 06:41:19 ahh, tried again with linux-console - it's ok 06:41:27 Odd 06:41:33 I'll look into that 06:41:36 k 06:41:54 I don't think I included '.s' in RetroForth 7 06:42:36 I never actually used it 06:44:59 anyway, way past bed time 06:45:01 g'nite 06:45:04 heh, my FIB finds FIB(45) in 0.012 seconds :-) 06:45:04 --- quit: aum () 06:45:06 g'night 06:45:14 I should go to be too 06:45:33 Good night all! 06:45:37 --- quit: crc ("Time for work...") 08:36:49 --- join: wossname (wossname@HSE-MTL-ppp61837.qc.sympatico.ca) joined #forth 08:57:49 --- join: Serg[GPRS] (~z@193.201.231.126) joined #forth 08:58:57 hi ! 08:59:41 --- quit: Serg[GPRS] (Client Quit) 09:12:13 --- join: Fractal (jah@selling.kernels.to.linus.torvalds.at.hcsw.org) joined #forth 09:23:51 Hi Fractal 09:24:08 How jah doing? 09:24:46 --- join: tgunr (~davec@vsat-148-63-4-106.c001.g4.mrt.starband.net) joined #forth 09:30:45