00:00:00 --- log: started forth/10.08.27 00:18:59 Back. 00:19:54 --- quit: lambert_ (Quit: Quitte) 00:21:28 --- join: mathrick (~mathrick@emp.nat.sdu.dk) joined #forth 00:29:15 ok 00:31:19 so i'm thinking of a kernel implemented on linux x86, with a dstack of 4 registers only 00:31:56 call-ret next 00:36:37 What do you mean under "dstack of 4 registers"? 00:37:13 using only eax, edx, ecx, ebx 00:37:51 I know what registers exist on i386, what do you mean under "dstack of 4 registers"? 00:37:53 hw esp and ebp would be used like non-forthy coding 00:38:19 there's no dstack as in other forths 00:38:59 (I assume that "dstack" means just the stack or data stack.) 00:39:04 the latter 00:39:16 ...Which alone is inconvenient term. 00:39:20 ok 00:39:25 They are the same thing in Forth. 00:39:51 but its forth-use prevents normal stack usage for call-ret 00:40:32 If I understand you right, you're just going to use other non-conventional registers for stack operations. 00:40:39 Nothing new. 00:40:59 how do you mean non-conventional? 00:41:43 esp and ebp are meant to be used for stack(s). 00:41:48 yes 00:42:01 What you want is to use other registers. 00:42:27 yes, reducing pushes and pops 00:42:35 It's like 68K or ARM with less registers with reserved use. 00:42:53 but both of those have far more regs 00:43:49 2x isn't "far more", esp. on 68K where you have them in separate files. 00:44:18 but double what the x86 has is alot 00:44:31 or third 00:44:49 Not for stack based VM. 00:44:53 yea 00:44:59 anyway 00:45:18 In fact, it is irrelevant, you have enough shadow registers nowadays. 00:45:55 i've got alot of the standard code-words defined and i'm trying to figer out the inner, outter and CFA 00:46:14 my .text segment assembles to less than 1kb so far 00:46:22 CFA is obsolete. 00:46:26 why? 00:47:31 Because it is specific to particular implementation techniques 00:47:31 which may be not so effective. 00:47:54 ok; who developed that so i can search for info? 00:48:39 There's not so much information, you can go through almost all Forth sites in a day. 00:48:50 ok moving on 00:49:27 in mine, cfa is the asm-label of the .text code bc code and data cant mix in the same segment 00:49:32 CS != DS 00:50:37 Segment selectors are irrelevant if you're talking about linux. 00:50:54 so that, CF and PF are in .text for asm-code, or in .data for forth-coded words; lf and nf always in .data 00:51:00 yea, i know 00:51:16 but they are still technically used to separate code from data 00:55:12 It will be much easier, if you stop using obsolete terms and go with modern ones. 00:55:21 such as? 00:55:38 Such as "execution token" 00:55:50 i dont use that term :) and havent 00:56:06 i'm not ANSified :) 00:56:16 This doesn't matter here. 00:56:22 k 00:56:28 For instance. 00:56:54 You say that PF is in .text or in .data, 00:57:00 yea 00:57:23 and the confusion is because PF is obsolete term which is 00:57:24 derived for models where code and data are interleaved. 00:57:43 k, and what is current usage? 00:57:50 No PF now. 00:58:08 There's execution token, and it relates to code. 00:58:14 so how are the word fields distinguished? 00:58:19 Other things are data. 00:58:36 There're no fields in a "word". 00:59:00 is this part of the ANS model to make forth more portable across processors? 00:59:01 There's dictionary entry, and it stores execution token. 00:59:36 Not only more portable, it is to make more effective implementation strategies possible. 01:00:14 after the four implementation models already known? 01:00:20 dtc/itc etc 01:00:51 You can't sanely separate code and data with your CF/PF. 01:01:00 why not 01:01:08 You yourself tell that: 01:01:20 "CF and PF are in .text for asm-code, or in .data for forth-coded words" 01:02:18 This problem comes from the fact that "PF" denotes code and data at the same time for different classes of words. 01:02:30 what problem? 01:03:30 That you have to deal with disjunction, note your own word "or" above. 01:04:07 ok, i dont see an entanglement problem between ml and forth words 01:04:42 ie, ml is aligned to abcdexxx boundary (3bit) 01:04:57 What is "ml" above? 01:04:59 that indicates that the target is .code 01:05:01 machine language 01:05:08 the binary code 01:07:09 on x86, code and data cannot exist within the same cacheline without incurring a time penalty during execution 01:07:17 stalling 01:07:29 Yes, that's why CF/PF terms are obsolete. 01:07:46 ANS model is conceptually simpler. 01:08:01 but i thought ans didnt specify internals 01:08:04 That's why anti-ANS folks can't grok. 01:08:07 that it is higher level 01:08:18 Yes, that's why it can present simpler model. 01:08:19 i only lack time spent reading their docs 01:10:24 Instead you're going to waste more time implementing conceptually heavier model, nice. 01:11:51 asau` do you have any constructive information available to impart? 01:18:25 are you going to answer? 01:21:32 hmm, your silence is deafening. thanks for the words. 01:21:51 anyone else have information to impart? 01:24:13 I'm at work. 01:24:25 If you think that parallel programming is easy, you're wrong. 01:24:59 i'm not discussing parallelism 01:25:08 that's not relevent to what i'm doing 01:25:21 nor do i think it's easy. 01:25:40 I'm at work still, I'm not paid for chatting, sorry. 01:25:46 i get it 01:28:00 --- quit: gnomon (Ping timeout: 260 seconds) 01:28:19 --- join: gnomon (~gnomon@CPE0022158a8221-CM000f9f776f96.cpe.net.cable.rogers.com) joined #forth 01:32:51 Constructive answer is this: 01:33:15 . 01:33:20 instead of implementing yet another Forth with inferior quality, 01:33:37 you would rather join some open source project and help 01:33:37 i'm not really interested in negative deconstructive statements. 01:33:37 improve existing implementation. 01:33:57 Believe me, there's enough to do. 01:35:27 For instance, writing web interface isn't trivial. 01:35:49 Writing it so as to follow traditional Forth propaganda is even less so. 01:37:24 I don't know how to do that in Forth, which approach to follow, 01:37:40 I don't have time to spend on it, and this is very practical problem. 02:52:41 --- quit: cataska (Quit: leaving) 03:03:22 --- part: tripFantastic left #forth 03:44:49 ASau`: that's very interesting problem 05:55:41 saper: remind me to commit ficl changes today later in the evening or tomorrow. 05:58:59 ASau`: I'll have a look at it again, since there's more 06:24:32 --- join: fantazo (~fantazo@178-191-165-230.adsl.highway.telekom.at) joined #forth 06:24:38 --- join: fantazo_ (~fantazo@178-191-165-230.adsl.highway.telekom.at) joined #forth 06:33:38 --- quit: fantazo_ (Quit: Verlassend) 07:04:32 --- quit: fantazo (Remote host closed the connection) 08:59:31 --- quit: martinhex (Remote host closed the connection) 08:59:46 --- join: Monevo (~nal@adsl-207-204-133-223.prtc.net) joined #forth 08:59:53 --- join: martinhex (~mjc@93-97-29-243.zone5.bethere.co.uk) joined #forth 10:16:40 --- join: ygrek (debian-tor@gateway/tor-sasl/ygrek) joined #forth 11:14:40 --- quit: Monevo (Quit: Monevo) 11:18:18 --- quit: mathrick (Read error: Connection reset by peer) 11:55:58 --- join: Monevo (~nal@adsl-207-204-133-223.prtc.net) joined #forth 11:57:16 --- join: mathrick (~mathrick@users177.kollegienet.dk) joined #forth 13:18:41 --- quit: DocPlatypus (Read error: Connection reset by peer) 13:49:19 --- join: DocPlatypus (~skquinn@dsl253-084-021.hou1.dsl.speakeasy.net) joined #forth 13:51:21 --- quit: DocPlatypus (Read error: Connection reset by peer) 14:07:55 --- join: DocPlatypus (~skquinn@dsl253-084-021.hou1.dsl.speakeasy.net) joined #forth 14:33:50 --- quit: DocPlatypus (Read error: Operation timed out) 14:46:22 --- quit: ygrek (Remote host closed the connection) 14:47:50 --- join: DocPlatypus (~skquinn@dsl253-084-021.hou1.dsl.speakeasy.net) joined #forth 18:42:03 --- quit: gogonkt (Ping timeout: 258 seconds) 18:43:32 --- join: gogonkt (~info@183.27.208.128) joined #forth 19:11:53 --- join: WarWeasle (~brad@c-98-253-6-86.hsd1.in.comcast.net) joined #forth 19:42:37 --- quit: WarWeasle (Ping timeout: 245 seconds) 21:21:30 --- quit: DocPlatypus (Ping timeout: 258 seconds) 21:35:52 --- quit: crc (Ping timeout: 265 seconds) 21:36:37 --- join: crc (~charlesch@184.77.185.20) joined #forth 22:16:51 --- join: DocPlatypus (~skquinn@dsl253-084-021.hou1.dsl.speakeasy.net) joined #forth 22:35:28 --- quit: DocPlatypus (Ping timeout: 276 seconds) 23:12:22 --- join: ygrek (debian-tor@gateway/tor-sasl/ygrek) joined #forth 23:59:59 --- log: ended forth/10.08.27