00:00:00 --- log: started retro/06.08.24 00:11:29 --- join: Cheery (n=Cheery@a81-197-19-23.elisa-laajakaista.fi) joined #retro 02:12:16 --- join: nighty (n=nighty@tin51-1-82-226-147-104.fbx.proxad.net) joined #retro 03:33:27 --- join: nighty_ (n=nighty@tin51-1-82-226-147-104.fbx.proxad.net) joined #retro 03:43:45 --- quit: nighty (Read error: 110 (Connection timed out)) 03:48:37 --- join: jas2o (n=jas2o@WNPP-p-203-54-32-152.prem.tmns.net.au) joined #retro 03:51:06 --- part: jas2o left #retro 05:31:47 --- quit: virl (Remote closed the connection) 06:13:54 --- join: timlarson_ (n=timlarso@65.116.199.19) joined #retro 06:14:38 --- quit: timlarson_ (Read error: 104 (Connection reset by peer)) 08:31:19 --- quit: Shain (Remote closed the connection) 08:41:23 --- join: Shain (i=steve@c-67-161-56-76.hsd1.ca.comcast.net) joined #retro 09:37:16 --- join: Ray_work (n=Raystm2@199.227.227.26) joined #retro 10:01:57 --- join: timlarson_ (n=timlarso@65.116.199.19) joined #retro 10:08:26 --- join: nighty (n=nighty@tin51-1-82-226-147-104.fbx.proxad.net) joined #retro 10:24:45 --- quit: nighty_ (Read error: 110 (Connection timed out)) 10:33:13 --- join: neceve (n=claudiu@unaffiliated/neceve) joined #retro 11:26:52 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-175-210.pools.arcor-ip.net) joined #retro 11:35:05 --- quit: Snoopy42 (Read error: 145 (Connection timed out)) 11:35:29 --- nick: snoopy_1711 -> Snoopy42 11:57:17 --- quit: Cheery ("Download Gaim: http://gaim.sourceforge.net/") 12:12:31 --- quit: neceve ("Leaving") 12:53:05 --- quit: nighty (Client Quit) 13:12:14 --- join: Quartus__ (n=Quartus_@209.167.5.1) joined #retro 13:56:51 --- quit: timlarson_ ("Leaving") 14:51:44 good evening 14:51:57 Hi crc. 14:52:42 Hi crc. 14:52:54 Hi Quartus. 14:53:03 Hey Ray_work. 14:53:11 Ray! 14:53:14 how was your trip? 14:53:29 Excellent. I'll tell you about it more when I'm home, okay? 14:53:37 ok 14:56:52 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-175-210.pools.arcor-ip.net) joined #retro 14:58:09 --- quit: snoopy_1711 (sterling.freenode.net irc.freenode.net) 14:58:09 --- quit: Snoopy42 (sterling.freenode.net irc.freenode.net) 14:58:54 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-175-210.pools.arcor-ip.net) joined #retro 14:58:54 --- join: Snoopy42 (i=snoopy_1@dslb-084-058-175-210.pools.arcor-ip.net) joined #retro 15:00:30 --- join: snoopy_1611 (i=snoopy_1@dslb-084-058-175-210.pools.arcor-ip.net) joined #retro 15:01:00 --- quit: Snoopy42 (Nick collision from services.) 15:01:12 --- nick: snoopy_1611 -> Snoopy42 15:02:35 --- quit: snoopy_1711 (Read error: 145 (Connection timed out)) 15:06:27 Well, that's neat. I can patch the Standard Forth floating-point emulation package directly into the retroforth interpret loop, and floats work cleanly. 15:07:05 355e 113e f/ f. -> 3.14159292035398230 15:07:23 woah, cool! 15:08:12 _very_ cool 15:08:45 2e fsqrt f. -> 1.41421356237309504 15:08:46 etc. :) 15:11:21 i didn't know that facility existed. 15:11:27 thanks Quartus. 15:11:40 The Standard float stuff? Eckert's work from tinyboot.com. 15:11:49 He was kind enough to patch a bug in fsqrt that I found when testing. 15:12:02 also was not aware of that. will check it out when home. 15:12:28 It lacks the transcendental functions. 15:12:56 I had a joke... 15:12:57 I could throw in some lower-accuracy versions, I have the source. 15:12:59 not really. 15:13:11 That's very cool. 15:17:41 --- quit: Quartus__ ("used jmIrc") 15:17:51 --- quit: Ray_work ("User pushed the X - because it's Xtra, baby") 15:29:34 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #retro 16:52:24 Quartus, that wasn't an intelligent turn. 16:59:52 If you're going to start spraying foul language into a technical channel, you can expect to be banned. This, however, is not an appropriate place to discuss it. If you should wish to discuss it, feel free to /msg me privately. 17:02:12 it was a big failure, when you are so arrogant! you won't destroy my favorite forth! retroforth has a good design in every of it's parts, when you can't see it then, that acknowledges everything I said. 17:04:53 I have considerable respect for retroforth's design, as I've said both here and elsewhere. Its sound construction is the very reason I'm able to build on top of it. 17:08:02 so everything what I said was true. 17:09:50 whatever, that you banned me showed me your ugly face and I don't want to talk to you again for next time. 17:10:33 I'm sorry about this, folks. This is happening in here because it's the only other channel I'm currently in. I'll leave, hopefully he'll wind down. 17:10:36 --- part: Quartus left #retro 17:34:30 * docl is personally thrilled to have such good ans support for retro. it's only being made as an extension, and will probably lead to more improvements 17:38:09 yup 17:40:10 docl: there will be a new rxforth build soon (0i) which makes some significant changes to the rx core wordset and features 17:40:32 neat 17:41:55 will this be a seperate Rx from the one in retro, or will the changes be integrated there as well? 17:42:30 at this point it's separate 17:42:50 I may switch to using it as the basis of Retro 9.3 or 9.4 though 17:43:31 cool 17:43:33 it's already cleaner overall, and it integrates some things from 9.2 (notably prefixes) far more closely 17:43:43 * crc will pastebin his NOTES file 17:43:45 that's good 17:45:31 http://retroforth.net/paste/?id=158 17:45:49 the vector words, like 'is', can be used in definitions in the new core 17:46:18 --- join: Quartus (n=trailer@CPE0001023f6e4f-CM013349902843.cpe.net.cable.rogers.com) joined #retro 17:55:50 crc, let me know if I can assist in integrating the ANS stuff with the new rx core. 17:56:12 I will 17:56:14 well, s/integrate/maintain compatibility/. 17:56:20 :) 17:56:27 some of it will find its way into the core 17:57:04 I am seriously looking at changing >in to be compatible 17:57:06 I'd like to keep the layer working. I don't know if any of your planned changes will destablize it -- nothing leaps out at first glance. 17:57:22 For that you'd have to parse line-by-line. Shouldn't be a huge change. 17:58:08 I think you'd want to add REFILL for that (or equivalent, but that's the Standard name for it). 17:59:32 By the way, I have a disassembler running, and I've patched it to display word names where possible. This may not fit in line with your design goals, but some way of reliably determining the length of each word would let me write a proper SEE. 18:00:09 In an early Quartus Forth, I compiled a sentinel value just after the final code in a word. That could sometimes be tripped up by embedded string literals or other matching code sequences. 18:00:27 I switched over to storing the length in the headers. 18:01:00 I'll look into this as well 18:01:47 http://retroforth.org/projects/darcs/Retrospect/retrospect has some code for doing hex dumps of words; it's not flawless in determining the length of a word though 18:01:47 Ok. I could have ; keep its own list of lengths, but then I'd have to rely on another heuristic for kernel words. 18:02:27 I'll check it out. I've noticed while disassembling that ;; doesn't appear to convert final calls to jumps. 18:03:02 not at present 18:03:20 Ah ok. 18:03:27 I disabled it while I look for a way to handle tail call elimination in a better way 18:04:39 I optimized a few primatives. The unchanged ANS fib bench runs as fast now under rf as it does under gforth-fast. 18:04:46 cool 18:05:28 Oh -- the other benefit to storing lengths is that you can reliably inline any position-independent code without using ret as a sentinel. 18:05:46 I also use the lengths for recursive extraction when turnkeying. 18:07:53 I optimized - + < > dup tuck for the fib, using an ANS 486 postfix assembler. 18:10:16 Were you here when I was hopping up and down excited about integrating a standard floating-point emulation into the interpreter? 18:11:01 Still entertains me a great deal to be able to do float stuff just by plugging these things together like Lego. 18:13:42 yes 18:17:28 Yesterday's excitement was a bignum package. It's probably wrong to have so much fun with bignums, but here we are. :) 18:26:10 Lego, great analogy. 18:26:12 oops 18:26:18 --- nick: nanstm -> Raystm2 18:26:26 --- mode: ChanServ set +o Raystm2 18:35:52 * crc is going to bed; will be here tomorrow 18:35:54 goodnight 18:35:57 cya 18:44:36 * Raystm2 is still reading the sources. 18:44:42 goodnight crc. 18:45:15 I'll do a little more on chess tonight and start the posting habit all over again. 18:45:44 Quartus: this is really great work. 18:45:58 Thanks, Ray. Just about to post a bug-fix. 18:46:53 Okay, please keep me abreast, I'm certainly interested. :) 18:47:35 It's up. 18:47:46 Fun to watch a couple guys that obviously know what they are doing, and bring those, sometimes different, experiances together. 18:48:18 Like watching two old plumbers discuss the best path for the drain. 18:48:24 :) hehe 18:48:24 Ew. 18:48:31 sorry. 18:48:36 :) 18:49:26 I meant that to be very complimentary. 18:49:34 Noted. :) 18:50:57 You pasted? or are you hacking the darcs directly? 18:51:14 Oh. It's in a static place now: http://quartus.net/retro/retro-ans.fs 18:51:37 Ah! okay. Excellent. You plan to host the module? 18:52:02 That's my copy, anyway. crc has been putting versions into the release since .8. 18:52:23 ya. cool. Reminds me to update. 18:52:51 I used to be scared of Retro's fast change time. Now I'm thrilled by it. 18:53:37 It's the very picture of a language that has always been capable, and always in flux. 18:54:32 I'll be keeping that link up to date with changes to the ANS layer. 18:54:32 I suppose a portion of it will ease into a norm. Seem's like there will always be this aire of rebellion, the Forth of Forth. 18:54:58 Very cool. :) good to know. and Thank you, cuz now I plan to use these words. 18:55:26 The layer steps on as little of retro's namespace & functionality as possible, so you can still use retro-y things. 18:55:51 >if and ;then and ;; and :: and vectors and all that. 18:56:07 I think that's just beautiful. The way you did that. 18:56:09 Prefixes are gone, though, except for numeric ones. 18:56:37 I'm gonna miss !name @name. :( 18:57:10 You know it was less me than that it just worked out like that. RetroForth was already a pretty normal Forth. There are only a few names that conflicted, 9 I think. 18:57:30 Hmm no kidding. 18:57:38 That I didn't know. 18:57:39 Ray, there's always variable foo : @foo foo @ ; : !foo foo ! ; if you really miss'em that much. 18:57:45 hehe 18:58:27 no I'd be compelled to always use the shorter and more efficient "foo @", every where it need appear. 18:59:24 Though, I do tend to create those defs in colorforth. Don't know why. Sometimes as interface, I suppose. 18:59:53 Usually no more then 2 or three letters long, like unix commands. 18:59:57 It's faintly possible I could hook prefixes back in. I'll have a look in rx and see where they are applied. 19:00:22 I'm sure it's not that necessary. 19:00:44 The base prefixes are all there. 19:01:02 Those, I like. 19:01:14 The math one's, I need to start using. 19:01:31 The !name @name ones, might make code unreadable. 19:01:55 Yes, they're handy. I expanded the ' one to also do double-cell values, so 'Raystm2 produces 1937009970 5398905 on the stack. 19:02:24 oh no, you did-ent. :) 19:02:38 I love to see letters as numbers on the stack. :) 19:02:42 Heh. 19:02:46 I make chess games out of it. 19:04:03 2variable x 'Raystm2! x 2! x 8 type -> syaR!2mt 19:04:38 The general applicability of that I'm not sure, but it's an extrapolation from 'a 'ab and so on. 19:05:36 hehe :) 19:05:42 that's great! man. 19:07:16 I meant to ask crc about that, his ' used base 255 instead of 256 for some reason. Never notice on single chars, of course. 19:08:05 I see. 20:38:12 * Raystm2 suddenly distracted by Heart on PBS. 22:02:46 ping 23:18:49 --- join: Cheery (n=Cheery@a81-197-19-23.elisa-laajakaista.fi) joined #retro 23:59:59 --- log: ended retro/06.08.24