00:00:00 --- log: started forth/14.09.28 00:26:06 --- nick: Tod-Autojoined -> TodPunk 00:34:32 --- join: true-grue (~grue@89-178-248-61.broadband.corbina.ru) joined #forth 01:50:15 --- quit: impomatic (Ping timeout: 272 seconds) 02:42:29 --- join: impomatic_ (~digital_w@87.113.116.210) joined #forth 03:59:37 --- join: proteusguy (~proteusgu@ppp-110-168-230-47.revip5.asianet.co.th) joined #forth 04:14:31 --- quit: proteusguy (Remote host closed the connection) 05:51:00 --- join: impomatic (~chatzilla@87.113.116.210) joined #forth 06:53:37 --- join: Zarutian (~zarutian@168-110-22-46.fiber.hringdu.is) joined #forth 08:02:36 --- join: zlrth (~user@c-71-206-216-100.hsd1.pa.comcast.net) joined #forth 08:48:51 --- join: proteusguy (~proteusgu@ppp-110-168-230-34.revip5.asianet.co.th) joined #forth 09:04:06 --- join: _spt_ (~Jaat@unaffiliated/-spt-/x-5624824) joined #forth 09:16:34 --- join: _spt2_ (~Jaat@host-92-12-208-7.as43234.net) joined #forth 09:18:19 --- quit: _spt_ (Ping timeout: 245 seconds) 09:18:58 --- quit: Zarutian (Quit: Zarutian) 09:35:52 --- join: xyh (~xieyuheng@2001:250:3002:5550:6ea1:cc0f:bcb2:b187) joined #forth 09:58:14 --- join: Zarutian (~zarutian@168-110-22-46.fiber.hringdu.is) joined #forth 10:09:16 --- nick: MrM0bius -> MrMobius 11:40:33 --- quit: xyh (Ping timeout: 260 seconds) 12:16:30 --- join: xyh (~xieyuheng@2001:250:3002:5550:6ea1:cc0f:bcb2:b187) joined #forth 12:32:57 --- quit: zlrth (Ping timeout: 244 seconds) 12:44:46 --- join: aranhoide (~smuxi@130.Red-81-32-187.dynamicIP.rima-tde.net) joined #forth 13:51:02 --- join: _spt3_ (~Jaat@host-92-12-208-7.as43234.net) joined #forth 13:54:58 --- quit: _spt2_ (Ping timeout: 260 seconds) 14:10:45 --- quit: proteusguy (Remote host closed the connection) 14:45:21 --- quit: true-grue (Read error: Connection reset by peer) 15:40:29 --- quit: _spt3_ (Quit: irc- et) 17:23:06 --- quit: xyh (Ping timeout: 260 seconds) 18:08:39 --- join: DocPlatypus (~skquinn@c-50-162-13-180.hsd1.tx.comcast.net) joined #forth 18:09:12 [1032172.190735] ptrace of non-child pid 7264 was attempted by: gdb (pid 7272) 18:09:40 this is apparently why see doesn't work in GNU Forth on amd64, at least on Ubuntu. 18:10:15 I feel really stupid, because I should have seen this when I first had problems about three years ago. and I am just now noticing today. 18:11:44 supposedly amd64 disassembly was fixed in 0.7.1. I got this same result in 0.7.3, and now 0.7.9_20140402 18:11:45 --- join: xyh (~xieyuheng@121.33.190.134) joined #forth 18:18:28 xyh: are you a GNU Forth user by chance? 18:34:59 DocPlatypus: you have to give xyh some context. 18:35:21 oops 18:35:44 would have explained had I gotten a positive answer... 18:36:53 I'm wondering now if this bug is not peculiar to old Ubuntu 18:37:06 what does see in GNU Forth do? use gdb to inspect memory? 18:37:13 apparently yes 18:37:26 *if* the word is code 18:37:41 does see spit out x86 intel assembly? 18:37:42 if the word is written in Forth it works fine 18:37:59 yes, it is supposed to if the word is code. however on amd64 it hangs until I control-\ 18:38:16 and then: s" pkill gdb" system s" reset" system 18:38:17 humm... 18:38:42 which, btw, is a real bear to type in when gdb is eating one out of every few keystrokes 18:39:18 Zarutian: do you have access to an amd64 box you can try to reproduce this on? 18:40:40 DocPlatypus: nope, I do not have such access sorry. 18:41:14 only other box I can try it on is running a version of Ubuntu at least as ancient 18:41:50 well, there's an amd64 box with OpenBSD on it 18:42:17 but I no longer care if GNU Forth fully works on OpenBSD or not because soon I won't be running it 18:43:06 DocPlatypus: I do not use GNU Forth, but I will view the channel log for you questions :) 18:43:46 DocPlatypus: humm can you hexdump out see ? 18:44:13 Zarutian: when see fails it does give me a hexdump, I'm not sure what good that's supposed to be 18:44:26 see itself is written in Forth 18:44:33 alright! 18:45:06 the problem actually occurs in name-see 18:45:08 if you can pastebin that code somewhere or point to the source where it is accessible via web 18:45:10 one of the words see calls 18:45:17 ok 18:45:29 I see! (/pun not intented) 18:45:51 name-see does what? print out the name of the word? 18:46:22 does GNU Forth have a single combined name and code dictionary or split ones like eForth? 18:47:11 http://tny.cz/65ad307a 18:47:20 I think there are separate wordlists 18:49:09 oky that big minus constant is a bit suspictious 18:49:36 that's just the way see spits it out. it's probably something else in the proper source 18:50:01 most likely and huge hexedemical constant 18:50:48 it's actually a constant called alias-mask 18:50:50 DocPlatypus: I use archlinux on x86-64, how can I test the bug for you ? 18:51:19 xyh: not familiar with Arch, but see if it has a GNU Forth package versioned 0.7.1 or later 18:51:39 gforth-0.7.1 (or 0.7.2 or 0.7.3 or even 0.7.9_xxxxxxxx) 18:51:59 anything later than 0.7.0 as the bug is supposedly fixed in 0.7.1 18:52:20 DocPlatypus: perhaps 0x8000000000000000 ? 18:52:49 Zarutian: yes 18:53:05 okay catching up, seeing your alias-mask comment 18:53:07 the problem isn't there though. it's after that 18:53:20 hey, I was curious 18:53:28 version 0.7.3-1 18:54:24 it may be under xt-see 18:54:29 DocPlatypus: somewhere in the first outermost IF THEN clause or? 18:54:54 ok 18:55:34 I have found it... see calls parse-name, find-name, and name-see. name-see calls xt-see. xt-see calls seecode. 18:55:39 among others 18:55:59 alright. It might be in seecode? 18:56:24 and then seecode calls discode 18:56:51 and discode fork-exec into gdb to disamble? 18:56:56 http://tny.cz/cd55e744 18:57:12 apparently yes. and somehow gdb isn't being launched as a child of gforth 18:57:54 xt-see is very ugly which is why I'm skipping it. it's a very long mess of nested if/else/then 18:58:48 might be worth patching idsasm-gdb to just hexdump the code instead of relieing on gdb 18:59:22 * Zarutian is more used to running Forth on bare metal or in specialized VMs. 18:59:28 yeah 18:59:46 eventually I will be making pinball machines that run Forth on bare metal if all goes according to plan 19:00:46 which will eventually involve writing my own Forth engine, and I don't know the details on what I'll have available to me yet 19:01:34 * Zarutian really should continue specifing the J1 as boolean combinational and sequentual virtual circuitry for the SMPC stuff he is working on. 19:03:17 I think the gdb thing is just an ugly hack, intended only to "get us by" until a proper amd64 and/or 386 disassembler is written 19:03:43 (that SMPC only gives me binary wires and two type of gates: XOR (that doesnt have netlatency) and AND (that does have netlatency) 19:04:18 it is better getting by on hexdumps imnsho 19:05:27 yeah 19:05:41 I'd actually be just as okay with "+ is code" like another Forth did 19:06:02 GNU is lazy to ues gdb, the so called "philosophy" of ``using little programs to do their job well'' is just because they can not use C to write a big complex program as one 19:06:05 or show me a Forth equivalent 19:06:26 * Zarutian is curious about the pastebin DocPlatypus used tny.cz. 19:06:38 I just happened upon it when looking for pastebins 19:06:52 or shall I say pastebin.com alternatives 19:07:24 I just find it interesting that they pay for pastes. And in bitcoin no less. 19:07:35 oh 19:07:39 that, I didn't know about. 19:08:10 I just stumbled upon that while looking at your pastes 19:08:21 and what's SMPC? 19:08:47 Secure Multi Party Computation 19:09:27 DocPlatypus: you may try isforth if you use linux :) http://www.isforth.com/ 19:09:39 Zarutian: sounds cool ! 19:10:10 I am deciphering a paper written by one Ueli Mourer called "Secure Multi-Party Computation Made Simple" and trying to figgure out how to implement it 19:10:19 xyh: heard of it. and it's actually "use GNU/Linux" or "use a GNU variant"... I really think "use Linux" makes as much sense as "drive a DieHard car battery to work" 19:10:31 it's drive a car to work, who cares who made the battery 19:11:02 I do because I live in an climate where there is lot of ambient temperture variation. 19:11:18 Polar car batteries for the win! 19:11:32 yeah, still, point is I'm going to be looked at weird if someone asks what I drive and I reply "DieHard" 19:11:45 or even, who made the starter motor, etc 19:12:11 yebb Linux is only the kernel GNU provides the rest afaik 19:13:50 --- quit: xyh (Ping timeout: 240 seconds) 19:24:15 --- quit: TodPunk (Read error: Connection reset by peer) 19:26:12 for xyh when he comes back: indeed it is. Spefically for when wants to run code on few cloud providers with them not being able to look at what is actually happening inside the forth vm (unless they collude) 19:30:06 he pinged out 19:30:36 so you are working on an SMPC in Forth? 19:32:47 other way around. I want to run Forth ontop of SMPC. 19:33:34 oh 19:33:36 as the SMPC will be bandwidth intensive between the nodes taking part. 19:34:46 and Forth is small enough to fit in 128 KibiBytes memory space while being rather powerfull yet easy to implement simple primitives for. 19:36:17 how much do you know of k-of-k random sum secret sharing of values in glaois fields? 19:37:38 I haven't done a whole lot of crypto 19:38:59 I have some idea about Galois fields, K-of-K doesn't ring a ball 19:39:01 bell* 19:39:42 Galois* (sorry I can never remember where ever the a comes before the l or not) 19:40:33 you are familiar with n-of-m secret sharing schemes? k-of-k basically means that all shares are required to reconstute the secret value 19:40:36 I have the occasional fit of CSWS too so it's all good 19:40:58 (Can't Spell Worth S#$%) 19:41:23 N-of-M I think I've heard of before 19:41:47 --- join: TodPunk (~Tod@50-198-177-186-static.hfc.comcastbusiness.net) joined #forth 19:42:14 probably from shamir secret sharing scheme that makes polynominals for each share. 19:42:22 yeah 19:42:27 I think I had compiled and played with it a bit 19:43:51 well in Ueli's scheme one gets multiplication and addition gates/operators which means for galois fields with cardinality two which is basically binary one gets AND and XOR gates on boolean wires 19:45:01 ok 19:45:32 now as far as I understand his paper multiplication involves communication between nodes. Thankfully no round trip time but only latency time. 19:46:10 as far as actually reading the white papers and trying to write code based on them... that's where I get in over my head 19:46:47 I can understand all the write-ups on the infamous Therac-25 problems... and I used to be able to figure out the more technical stuff 19:47:00 but at my age... I find myself swimming in babble 19:47:30 well squeasing what the hell he is meaning from the sigma equations he gives is a bit challange for me. 19:47:52 I do understand sigma notation but it's damned confusing at first 19:48:16 and there's an equivalent pi notation for multiplication, which I didn't know about until I needed to write about it 19:48:25 err write *using* it 19:49:17 so, the netlatency issue with AND gates means that I need to keep my logic and virtual devices as 'shallow' as possible 19:50:08 ok 19:50:41 sounds like you are going to have your hands full 19:50:55 I thought that the LUTs that I was using to implement addition (two 2x6 bit and one 2x5bit) would require at least two delays 19:51:33 I managed to half it. 19:53:36 so if maximum latency between nodes is 50 millisec then one addition by the virtual ALU would take 150 millisec instead of 300 millisec 19:54:20 oh, this thing will be slow as hell 19:57:08 completely depending on the latency between the nodes. 20:06:07 can't be any crazier than some of the code I wrote 20:06:38 math functions for numbers larger than two cells (triples, quads, even hex (6 cell) and octo (8 cell) after I really got bored) 20:06:55 some of those words I wrote are rather ugly 20:07:17 when you realize you need a 4-rot you know you're in trouble 20:07:34 : 4-rot 11 roll 11 roll 11 roll 11 roll 11 roll 11 roll 11 roll 11 roll ; 20:21:45 --- join: xyh (~xieyuheng@121.33.190.134) joined #forth 20:22:13 for xyh when he comes back: indeed it is. Spefically for when wants to run code on few cloud providers with them not being able to look at what is actually happening inside the forth vm (unless they collude) 20:22:34 but now I must be off to bed. Will probably hang around tomorrow. 20:22:58 --- quit: Zarutian (Quit: Zarutian) 20:29:03 hi, friends :) how do you think of the following to documentations :: http://cicada-language.github.io/documentations/2014/09/29/about-semantic-and-syntax.html and :: http://cicada-language.github.io/documentations/2014/09/30/about-interpreter-and-compiler.html 20:29:15 I need advice and criticism, I am going for lunch, and I will be back in 2 hours :) 20:29:20 --- part: xyh left #forth 20:54:17 In the actual field of computer science the terms "syntax" and "semantics" have pretty clearly-defined meanings. 20:54:43 Hell, in the realm of *LINGUISTICS* (where CS stole the terms from) the terms are pretty clearly defined. 21:06:10 For compiler vs. interpreter, yeah, the terms are getting so muddy that I question whether they have much intrinsic value beyond a VERY rough system of categorizing things. 21:49:43 --- join: xyh (~xieyuheng@2001:250:3002:5550:6ea1:cc0f:bcb2:b187) joined #forth 21:50:19 ttmrichter: in the realm of *LINGUISTICS*, the terms are defined differently by different authors. Chomsky different with Saussure, and both of their definitions also different with older school of comparative linguistics. and in computer science the situations are the same. 21:50:36 ttmrichter: about compiler and interpreter, IMPO the classification of the implementation techs are clearly different. 21:51:46 xyh: This is true of any discipline in history. 21:51:55 Meanings drift. 21:52:09 In the maths even the term "formula" has many different meanings. 21:52:15 Or "number" for that matter. 21:52:46 If you want a fixed, immutable, 100%-agreed upon definition for ANYTHING, ANYWHERE you are going to be badly disappointed in life. 21:53:22 yes, I know this :) this is why I try to define it by myself 21:55:30 just to make the meaning of these terms to be more clear in my context 23:45:48 A good model to use is lexis/syntax/semantics/pragmatics. 23:46:51 Lexis is a definition of what "words" (for want of a better term -- smallest unit of meaning) your language permits. This would include both the predefined units of meaning and the patterns of allowable extensions (like the rules for making identifiers, say). 23:47:20 Syntax is a definition of how "words" can be put together to make legal (not necessarily meaningful!) expressions in your language. 23:48:04 Semantics is what those expressions actually mean and, as is usual in the academic world, come in various flavours. The two big ones in use are denotational and operational, however. I won't get into those right now; entire books can be written on those. 23:48:20 (Indeed I have a 500-or-so page book that just does an OVERVIEW of this! :-o) 23:49:24 Semantics deals, typically, with the portion of meaning that is less context sensitive. The more easily predictable stuff. For the other weirder stuff, especially in natural language, you have pragmatics. That's where things get really hairy. Thankfully most sane computer languages rarely have a need for pragmatics. 23:50:40 Strictly defining the difference between semantics and pragmatics is difficult because really they're facets of the same thing. 23:50:51 But an illustration makes the point: "How's your son?" "He's grown another foot." 23:51:58 Most conventional semantic analysis would choke on that pair of sentences. Pragmatics enters when you've got indirect, invisible referents that provide unseen context for meaning, in this case for disambiguating between "foot" the appendage and "foot" the measurement. 23:52:20 If you've got this kind of problem in a programming language you generally have a bad programming language. (Like C++, in fact.) 23:59:59 --- log: ended forth/14.09.28