00:00:00 --- log: started forth/04.07.26 00:01:39 grrrrrrr 00:01:45 sometimes linux really pisses me off 00:02:01 How has it offended you? 00:02:42 what's the point of having protected memory and all that if you let programs take over the keyboard? 00:03:31 twice now I've had to resort to closing my terminal window 00:03:44 What program(s)? 00:03:47 ^C ^Z and ^S had no effect 00:03:49 herkforth 00:03:54 :-) 00:04:03 herk - you generally kill the process from another terminal. 00:04:22 I made a silly change which echoed a bunch of low numbers (1-15) to the terminal 00:04:38 this made it go absolutely haywire, and refused to respond to the keyboard 00:04:57 herkforth only traps SIGILL and SIGSEGV 00:05:28 ayrnieu: I did that. first time it had no visible effect (even though ps informed me that herkforth was no longer running) 00:06:13 * crc has run into similar problems with RetroForth during some experiments 00:06:17 --- quit: kuvos (Read error: 110 (Connection timed out)) 00:06:29 seccond time it visibly quit, and showed a prompt, but then started spitting out garbage on the screen continuously. after waiting 20 secconds or so (in which I tried ^C ^Z and ^S to no avail), I just closed my terminal window 00:06:34 Linux has very crappy handling of TTY's in my opinion 00:07:03 um, unix in general. 00:07:49 it didn't technically crash my terminal, but it effectively did 00:07:54 maybe you should develop within a second herkforth, that filters your development-herkforth =) 00:07:56 programs should not be able to #*@&$^@# up your terminal 00:08:09 well, what terminal emulator did you use? 00:08:14 rxvt 00:08:34 I do not believe that rxvt was malfunctioning 00:08:42 I'm sure it was doing what it is supposed to 00:08:52 I've seen the same sort of crap with xterm 00:08:58 and in the raw terminal 00:10:13 Herkamire: I ran 'svn up' and 'make', and I'm getting an error 00:10:14 crc@herkamire herkforth $ ls blocks/118 00:10:14 ls: blocks/118: No such file or directory 00:10:14 crc@herkamire herkforth $ error in blocks/118 while processing "Shift-Q" before byte 8: Word not founderror in blocks/118 while processing "Shift-Q" before byte 8: Word not found 00:10:36 oops. one sec 00:10:51 * crc doesn't see block 118 either 00:11:21 fixed 00:11:39 I added the shadow block, but forgot to commit the source block... 00:11:45 Ok 00:11:53 Thanks for fixing it :-) 00:11:59 np 00:12:24 crc: BTW, if you run convert-dict you are likely to get some of the problems I just described :) 00:14:55 update again if you want to use block 118 00:15:44 you caught me in the middle commiting stuff, and discovering what I missed. 00:16:20 hm, what assembler does herkforth need? GNU assembler 1.38 gets confused by -mregnames 00:17:53 gnu 00:18:01 ayrnieu: you on ppc? 00:18:25 yes. 00:18:36 MacOSX, specifically. What version of gas do you compile with, then? 00:18:53 GNU assembler 2.14.90.0.2 20030515 00:19:14 a good major version number newer than mine. OK. 00:19:37 -mregnames allows you to use r0 r4 etc to specify registers 00:20:54 I suppose for compatibility I could just define r0 - r8 myself 00:21:50 heh, looks like I just use r0, r3, r4, r5 and r6 00:22:43 humph. broke my footrest the rest of the way 00:24:32 --- quit: tgunr (Read error: 104 (Connection reset by peer)) 00:25:00 ayrnieu: it won't work under Mac OS X 00:25:21 herk - good to know, I guess. 00:25:21 unless you have some way of executing ELF files 00:26:18 --- join: tgunr (~davec@vsat-148-63-4-106.c001.g4.mrt.starband.net) joined #forth 00:36:05 I'm coding a RetroForth source block with various conditional/flow control words from Herkforth :-) 00:36:54 crc: :) that's great :) 00:38:08 I think so. I've found that I like >IF =IF and 0; better than the traditional Forth words for those tasks 00:38:23 :) 00:38:24 me too 00:39:04 I hope to eliminate the traditional words when I do RF 8.0, but that's a ways off yet... 00:39:16 cool :) 00:39:24 I have to get the current users of RetroForth accustomed to them first 00:39:53 :) 00:39:56 yeah 00:40:40 I suppose that's a good thing about not having anybody seriously using herkforth... nobody gets upset about major changes 00:40:46 Right 00:41:17 (I'm in the middle of changing how word names and such are stored in the dictionary) 00:41:28 I have 10-20 people who I know actually use RetroForth, and probably more that I don't know about 00:41:39 Ooh, dictionary changes :-) 00:42:00 so it's counted with the first byte instead of just the string padded with zeros 00:42:15 not sure I like the byte though 00:42:32 I see 00:45:25 I wrote code to alter the dictionary 00:46:20 but running it leaves herkforth unusable, because the editors dictionary searching does not expect a count byte 00:49:48 not sure I want this change 00:49:55 I'll sleep on it 00:50:05 or at least sleap anyway 00:50:48 Ok 00:51:41 * crc just finished coding+testing =if if <>if 0if >a @a !a a 00:51:42 :-) 00:52:48 nice 00:53:18 The words like <; are going to take longer though... 00:54:00 really? why? 00:54:17 Actually, maybe not 00:54:56 they were pretty easy on ppc. I just had to figure how many bytes were compiled by ; 00:56:50 ,<; ,= 20 _bnl ,; ; 00:56:54 : ,<; ,= 20 _bnl ,; ; 00:57:45 ( ,= compiles code to compair the top two items on the stack and drom them. it's used in ; etc.) 00:59:05 20 _bnl compiles a conditional branch forward 20 bytes (bnl == Branch Not Less than) 00:59:35 anyway, I'm off to bed 00:59:52 I'm so glad to hear you are benefiting from herkforth 01:00:05 Good night 01:00:13 --- quit: Herkamire ("bed") 01:07:39 * crc wonders why this is not working 01:18:07 --- quit: cmeme ("Client terminated by server") 01:18:22 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth 01:26:55 --- quit: tgunr (Read error: 54 (Connection reset by peer)) 01:28:36 --- join: tgunr (~davec@vsat-148-63-4-106.c001.g4.mrt.starband.net) joined #forth 02:46:09 --- quit: crc ("Time for bed... Goodnight all!") 03:04:57 --- join: crc (crc@0-1pool88-38.nas48.philadelphia1.pa.us.da.qwest.net) joined #forth 03:15:12 --- quit: crc (Client Quit) 03:16:07 --- quit: tgunr (Excess Flood) 03:17:55 --- join: tgunr (~davec@vsat-148-63-4-106.c001.g4.mrt.starband.net) joined #forth 03:27:25 --- join: crc (crc@0-1pool176-18.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 04:21:16 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 05:46:44 --- join: lalalim_ (~lalalim@p508AB667.dip.t-dialin.net) joined #forth 05:59:29 --- quit: lalalim (Read error: 110 (Connection timed out)) 05:59:45 --- join: crc_ (crc@0-1pool176-55.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 06:12:17 --- quit: crc_ (Read error: 54 (Connection reset by peer)) 06:16:56 --- quit: I440r (Read error: 104 (Connection reset by peer)) 06:17:35 --- quit: crc (Read error: 110 (Connection timed out)) 06:18:42 --- nick: lalalim_ -> lalalim 07:12:07 --- join: I440r (~mark4@216-110-82-203.gen.twtelecom.net) joined #forth 07:21:34 --- part: lalalim left #forth 07:36:59 --- join: ASau (~root@217.16.31.100) joined #forth 07:37:05 Dobry vecer! 07:45:07 --- join: warpzero (~warpzero@mi126.dn180.umontana.edu) joined #forth 07:51:32 Guten Abend, warpzero! 07:51:42 --- join: KptnKrill (~kyle@pool-68-239-36-100.bos.east.verizon.net) joined #forth 07:51:58 God kvaell, KptnKrill! 07:52:06 sup 07:52:14 God kvaell to you too :) 07:52:15 Hi ASau 07:55:47 what does god kvaell mean? 08:00:16 Dobryjj vecher. 08:03:43 Privet 08:03:54 God kväll, KptnKrill 08:04:21 :) 08:04:23 God kvall, Robert. 08:04:41 Er sverige? :) 08:04:57 ? 08:05:06 I'm so confused right now... 08:10:29 --- join: jdrake (~jdrake@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 08:10:38 --- part: jdrake left #forth 08:23:44 Do not do this. 08:24:14 It's the reflection of our internationalism. 08:35:26 --- quit: fridge_ ("Changing server") 08:36:09 --- join: fridge (~fridge@dsl-203-113-229-162.NSW.netspace.net.au) joined #forth 09:28:02 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 09:44:38 --- nick: KptnKrill -> KrillAWAY 09:45:40 Mur, does Eesti belong to Finno-Ugric? 09:45:50 yes 09:46:21 I'm trying to recall, which languages are Baltic. 09:46:34 Lithuanian and Latvian, it looks. 09:59:03 hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi hi 10:00:12 Oh look, it's that annoying spammy guy again :P 10:00:53 Yes, let's run 10:01:06 * tathi runs and hides 10:01:45 OMFG!!!!! 10:02:04 i just found out my friend was hospitalized 2 weeks ago 10:02:06 fuck 10:02:08 aarghghg 10:02:09 brbb 10:03:12 --- quit: tathi ("mowing...") 10:07:37 --- quit: tgunr (Excess Flood) 10:07:54 Ouch 10:08:06 Any idea why? 10:09:30 --- join: tgunr (~davec@vsat-148-63-4-106.c001.g4.mrt.starband.net) joined #forth 10:09:32 no :( 10:12:58 seems that hes alright tho 10:14:20 ls 10:14:32 Ha! 10:14:38 :) 10:14:47 The same interface everywhere! 10:15:21 xlock -mode julia <-- nice. :) 10:15:32 heh 10:15:37 Same goes for the "strange" mode. 10:15:38 julia = sister? :) 10:15:48 :P 10:15:50 No, Julia = old man. 10:15:57 oh :) 10:16:07 heh 10:16:12 julia is a german female name 10:16:16 lol 10:16:18 Last name 10:16:21 For a frenchie 10:16:34 And I also think of the female name... 10:16:38 It's "Julija" in Russian. 10:16:42 :) 10:16:46 * Robert has known a Julia too much 10:17:10 aah 10:17:13 :) 10:17:27 Because Russian is much older language than English, it retains "j" for "i". 10:17:47 :) 10:18:09 Modern languages use "j" for "zh". 10:18:37 :) 10:19:11 Quite opposite the situation is in Altaic family. 10:19:26 ASau our language master :) 10:19:54 Malqar is older than Kazakh, but it has "zhurt" vs. "jurt." 10:19:59 how interesting 10:20:29 indeed 10:21:26 The most curious thing is how "j" sound transforms to "zh" and vice versa. 10:21:54 --- join: jdrake (~jdrake@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 10:21:55 Try to do this, but don't break your tongue. 10:22:20 Dobryjj vecher, jdrake! 10:22:32 terve jdrake! :) 10:22:50 ASau, yo, can I serenade you with my 'poetry'? 10:23:10 Only in Forth. 10:23:30 : SEE MY POINT OF VIEW ; 10:24:07 ASau, http://rafb.net/paste/results/QJKF6V86.html <-- imagine writing that in forth, and tell me it would look better 10:24:52 jdrake: ocaml? :) 10:25:07 arke, bless its soul 10:25:21 look? 10:26:30 What's it meant to do? 10:26:38 one sec 10:27:28 http://homepage.mac.com/wackysteak/Test.png 10:27:35 a star field 10:27:49 I don't have X. 10:27:53 i am trying to look at different methods of generating star fields 10:27:55 X? 10:28:35 Text only. 10:28:39 my next goal is to create a sort of spiral galaxy 10:28:59 ASau, ok, then you probably can't appreciate the beauty, the output of coordinates isn't much like art 10:29:34 well i have to run 10:29:37 later 10:31:28 bye jdrake 10:33:33 Where's ianp? 10:33:51 I'd wanted to consult on IRC client for Emacs. 10:36:09 --- nick: TomasuDlrrp -> Tomasu 10:36:11 jdrake, your code is equivalent to this: 10:36:11 * Tomasu is back (gone 14:40:17) 10:36:19 : star random random point ; 10:36:31 : field #stars times star repeat ; 10:37:39 If you wish to print coordinates first, you have this change: 10:37:46 : print . . ; 10:38:00 : 2random random random ; 10:38:22 : star 2random 2dup print point ; 10:41:25 insomnia can be brilliant sometimes 10:42:02 i just finished and debugged my abstract event framework 10:44:24 :) 10:44:30 question for linuxers 10:44:38 how do i remount / to rw? 10:46:54 nevermind 10:47:31 madgarden, you here now? 10:48:24 --- quit: jdrake (Read error: 110 (Connection timed out)) 10:50:39 mount -o rw /dev/XXX / 10:51:13 mount -o rw / 10:51:24 Maybe. 10:51:37 aack. I forgot to log off last night. Sorry. 10:52:18 anyone have madgarden's little c like script interpreter? or remember its name? I might have it but my $HOME is a mess :o 10:52:27 kc5tja: nobody minds :) 10:52:46 kc5tja: going to work now? 10:53:04 I'm actually off work today, but I will be going to aikido later today. 10:57:08 :) 10:59:44 --- join: jdrake (~jdrake@CPE00045afdd0e8-CM014410113717.cpe.net.cable.rogers.com) joined #forth 11:01:35 wb jdrake 11:01:42 terve arke 11:01:47 yo 11:02:09 does anyone know of a source of info on generating a spiral galaxy? 11:03:24 Nope. Although I suspect it would basically involve random placement of stars with random masses around a significantly larger, central mass (e.g., SIGNIFICANTLY larger, as in, supermassive blackhole massive. :)) 11:03:31 And then apply a spin to it. 11:03:59 kc5tja, that is beyond me 11:05:34 I assume this is for a simulation? If so, what is the simulation for? 11:05:56 this is just for star placement in a static universe. Space game, turn based 11:06:04 Ahh 11:06:31 got to go 11:06:33 lunch :-) 11:06:36 Heh :0 11:06:38 :) even 11:06:39 enjoy. 11:08:34 hello everyone 11:08:43 re warpzero 11:09:08 sup kc5 11:09:12 wheres my kestrel 11:09:28 You know where. 11:09:41 The MISC processor is no more; patent violation. 11:09:45 Now using TTA. 11:09:51 So I guess you got your wish after all. :-) 11:09:52 seriously? 11:09:57 awesome POSSUM 11:10:11 did you get a cease and desist order? 11:10:32 Nope. But someone told me of the sue-fest that Patriot Scientific was going on, and I decided to do some research into the matter. 11:10:49 Sure enough, Chuck and PTSC has valid patents covering the MISC. They won't expire until 2018 at the latest. 11:10:50 ah 11:11:14 TTA is my lover 11:11:36 I like TTAs too, but they have voracious memory bandwidth requirements. :( 11:11:58 can i help you design it 11:12:12 I actually already have the register set designed. 11:12:20 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 11:12:22 anybody on win9x/ME or DOS? 11:12:32 The assembler for it is kind of finished, but I am going to write an emulator for it too. 11:12:56 eh 11:13:03 well im sad i haven't been around 11:13:08 but i got kicked out my house 11:13:15 and my job wont pay me 11:13:33 Performance of the system isn't going to be nearly as good as with the MISC though. :( Due to the voraciousness of its instruction fetch bandwidth, load/store operations will now take 2 cycles, and some forms of branches may take up to 3 cycles. 11:14:02 why don't you just call up chuck 11:14:08 and ask him if you can make a MISC 11:14:38 Chuck has been made aware of my desire to use MISC, and he forwarded my request to his IP manager. I was told, "Contact me any time," and I tried ... three times. 11:14:42 kc5tja: why not make more buses, if possible? 11:14:48 I got no phone calls, no return e-mails, nothing. 11:14:55 kc5tja: :( 11:15:00 arke: It's intrinsic in the nature of Forth. 11:15:08 Forth is just plain not at all parallelizable. 11:15:27 i dont see why memory requirements go up 11:15:28 but sure 11:15:36 Memory requirements go up because: 11:15:44 1) Only one instruction is packed in a single 32-bit word, 11:15:56 2) Instructions are now fetched every freaking clock cycle instead of, say, every five clock cycles. 11:16:11 one forth instruction, that is, not one CPU instruction, right? 11:16:20 Yes, sorry. I meant one Forth primitive. 11:16:28 :) 11:16:36 And in some cases, since I have a dual-bus architecture inside the CPU, some primitives take two CPU instruction words. 11:17:43 well whatever 11:17:58 i dont believe that its that bad but 11:18:05 i haven't exactly tried to do this 11:18:47 I am making this TTA specifically for a stack machine. I have 31 CPU registers (with room for more if I decide to add on later), which closely mimicks the original MISC design. 11:19:28 why don't you design it for a TTA machine 11:19:35 There are three instruction formats: CALL (31-bit effective address, direct-loaded into the PC; no XORing), D=S, and D=imm7, where imm7 is a 7-bit, sign extended immediate value. 11:20:17 Because then there is zero advantage to using it over a more traditional MISC. 11:20:19 Err 11:20:20 RISC 11:20:32 except that its TTA 11:20:34 and not RISC 11:20:42 Except that it can't be used. 11:20:50 Forth has no capacity for instruction-level parallelism. 11:21:18 thats what optimizing compilers are for :D 11:21:20 The results of one operation depends always on the results of the immediately preceding instruction. 11:21:31 kc5tja, you may be interested in works of Chapyzhenko. 11:21:43 ...might... 11:21:46 Dude, if I'm going to go through that kind of trouble, I'll make it run Oberon instead of Forth. 11:21:55 no i know 11:22:04 just leave it open 11:22:10 so that i can use the TTA directly okay 11:22:21 Everything will be LGPL. 11:22:43 kc5tja: the compiler should be able to do parralelization relatively easy 11:22:44 The whole point of this computer is to be *entirely* LGPL. NO hidden secrets. 11:22:59 kc5tja: ...i think 11:23:06 IIRC, Andrejj Chapyzhenko tries to combine TTA with stack. 11:23:10 kc5tja: :) 11:23:16 ASau: I'm alread aware of his efforts. 11:24:02 kc5tja: well why not make a really lame hyperthreading-like thing on the cpu 11:24:04 kc5tja: only thing is making compiler primitiv-aware. which shoudlnt be hard. 11:24:18 warpzero: because he has space/constraints 11:24:24 arke: I *openly* invite you to try. :D 11:24:24 AFAICR, he finds instr. level parallelism in this approach. 11:24:28 i think you could do it really easilly 11:24:39 warpzero: he only has so much room to do it 11:24:40 --- quit: jdrake (Connection timed out) 11:24:49 arke: you just said that 11:24:52 ASau: He has multiple concurrent instruction streams -- e.g., the equivalent of overlapping four colon-definitions on top of each other. 11:24:52 kc5tja: no problem 11:24:59 warpzero: ....because hes on an FPGA 11:25:55 Maybe this was Tarasov's. 11:25:56 kc5tja: well, i guess you could maybe make two cores on one FPGA.... 11:26:47 Someone has expressed idea, that operations on d- and r-stacks can be done in parallel. 11:26:50 Doable, but overkill for my needs. 11:27:11 well i vote for you just 11:27:15 ASau: Sure, because they don't overlap (except when transferring data between them). But what does this get you? 11:27:17 walking over to chucks house 11:27:18 Nothing. 11:27:25 kc5tja: well, i still think that compiler shouldnt be too hard 11:27:27 OK, you can "return" at the same time as performing the final addition or whatever. 11:27:29 arke, it may happen you better to have on-chip RAM. 11:27:52 :/ 11:28:01 kc5tja, saving call-return time. 11:28:07 Which is negligable. 11:28:36 My TTA can already do that anyway. 11:28:54 When you program short definitions, this may become visible. 11:29:04 kc5tja: OVER is only one CPU instruction! :) 11:29:25 kc5tja: oh, nevrmind 11:29:38 kc5tja: i say that you should keep trying to get in touch with chuck or whoever is managing his IP but 11:29:46 kc5tja: but DUP is only one, and DROP too 11:29:52 because obviously this is kinda lame 11:30:27 warpzero: Like I said. Three times with no response. I'm not going to waste my time. 11:30:43 kc5tja: acc_o := dst_t for DROP, dst_t := acc_o for DUP 11:31:09 kc5tja: wait so you have three instruction formats? 11:31:11 arke: Yes. 11:31:22 why not have like 11:31:25 one instruction format 11:31:30 Because code explodes. 11:31:35 99% of Forth code is CALLs. 11:31:38 kc5tja: kaboom 11:31:43 :) 11:32:22 Using a single instruction format means a single call will then take a MINIMUM of one instruction to load the PC-with-auto-push register, then one cell for holding the target address. 11:32:25 kc5tja: eh 11:32:34 kc5tja: OVER = 3 CPU cycles? :( 11:32:36 Hence, a single call goes from 32-bits to 64-bits. That is unacceptable to me. 11:32:55 arke: OVER: DT_T := DS_O 11:33:00 err, wait. 11:33:23 arke: OVER: DT_T := ACC_T , ACC_T := DS_O 11:33:52 kc5tja: you totally need a misc cpu for this 11:33:52 whatd the difference between dt and ds? 11:33:56 :C 11:34:24 DT_T is the top of data stack *trigger*. DT_O is the same register, but *no trigger*. DS_O is the *second* top of stack. 11:34:39 oh, you have a second trigger 11:34:42 er, register 11:34:43 cool 11:34:45 * Tomasu is away: AMG 11:34:48 that makes life much easier 11:34:50 No, no second trigger. Just a register. 11:35:11 thats what i meant 11:35:12 :) 11:35:15 And note that the Forth VM keeps its top of stack in the accumulator, so DT_* is really the Forth's second top of stack, and DS_* is the third. 11:35:26 kc5tja: can you add a population count instruction for me 11:35:55 kc5tja: :) 11:36:00 That's awfully application specific, I think. 11:36:06 its not 11:36:10 i end up wanting one 11:36:13 ALL THE DAMN TIME 11:36:15 in C 11:36:25 kc5tja: do you have temporary registers? 11:36:26 That can't be done in a single cycle anyway. It'd have to be implemented as a step. 11:36:32 arke: No. 11:36:43 how so? 11:36:45 warpzero: I've never had a need for such an operation. 11:36:55 populaation count? 11:36:58 i always need one 11:37:02 number of 1 bits in a word 11:37:03 For what? 11:37:13 lets see 11:37:19 ooh, cool 11:37:20 i wanted one for this CA stuff i was doing 11:37:26 CA? 11:37:30 and i wanted one for some encryption stuff 11:37:33 california? :) 11:37:38 and i wanted one for some compression stuff 11:37:39 but 11:38:04 cortificate authority? 11:38:09 certificate, even 11:38:12 Cellular Automata 11:38:15 tathi: HAH! Dvorak typer! :D 11:38:29 kc5tja: I was just wondering if anyone would catch that ;) 11:38:57 :) 11:39:44 anybody running windows 9x/me? 11:39:48 or DOS? 11:39:57 arke: Not I 11:39:58 why 11:40:11 i have dosemu on here 11:40:22 need mscdex.exe and oakcdrom.sys or equivalent 11:40:32 heh 11:40:59 oh, actually, there's a win98 machine here. 11:41:15 tathi: awesome. could you send me those files please? 11:42:42 kc5tja: whats SWAP/ 11:43:00 SWAP: DS_O := DT_O , DT_O := DS_O 11:43:12 wait 11:43:12 * kc5tja sighs 11:43:16 I keep making that mistake. 11:43:16 you can do that? 11:43:28 SWAP: DT_O := ACC_T , ACC_T := DT_O 11:43:29 Yep. 11:43:49 As long as they're in the same physical instruction word, absolutely. 11:43:51 oh, thats right 11:43:56 Since both moves are executed exactly at the same time. 11:43:58 they just have to be in the same word, right? 11:44:05