00:00:00 --- log: started forth/04.06.05 00:01:17 re arke 00:01:29 whats up, mi amigo? 00:01:47 Reading up on Cicadas, for some strange reason. 00:07:26 cicadas leg technology has much insights that can be applied to the K! 00:21:10 hey guys. 00:21:13 heh 00:21:15 what do you think about L4? 00:21:19 is it worth using"?: 00:21:19 --- quit: Herkamire ("goodnight") 00:21:25 arke: I've heard some good things about it. 00:22:33 Mechanic, Photographer, Artist, Architect, Engineer, Builder, Set designer 00:22:34 but, nobody here has ever used it? :) 00:22:36 oops. 00:22:41 Wrong window. 00:22:49 but, it seems that L4 is supposedly pretty good. 00:22:52 and it's portable 00:22:53 = good 00:23:06 dunno if I wanna put F2 on it though. 00:23:08 arke: Read "Worse is Better." 00:23:56 hrm? 00:24:00 http://www.neocoder.net/nanos/ 00:24:04 how would that one be? 00:26:13 hrm, nanos sseems simple 00:30:41 There is a document on the net called "Worse is better," and it explains the most likely reason why something like C would win out over Lisp. The same arguments apply to other better-or-worse technologies. 00:30:45 I forget the URL. 00:30:55 * kc5tja is currently reading something though, and I don't want to lose my place to go look. 00:30:59 ok 00:31:23 I'll be available to look in a few minutes though, assuming this document is as short as I think it is... 00:34:23 found it. 00:37:48 interesting 00:57:30 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 02:43:52 --- join: Topaz (~top@exten-halls-131.soton.ac.uk) joined #forth 02:44:39 hmm, what exactly is an execution token (xt)? is it the address of the word definition, or the address of the dictionary entry, or the address of the code pointer, or the code pointer itself, or something? 03:14:12 ooh, i found a good explanation 03:14:19 http://www.zetetics.com/bj/papers/moving1.htm 03:57:10 --- join: crc (crc@0-1pool176-19.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 04:04:31 --- quit: warpzero (Read error: 60 (Operation timed out)) 04:16:19 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 04:37:00 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 04:38:44 Hi 04:38:48 Hamforth. 04:42:01 --- quit: SolarFire[ (Read error: 60 (Operation timed out)) 04:55:11 --- quit: crc ("Going to sleep now...") 05:02:51 --- join: schihei (~schihei@pD95F72F8.dip.t-dialin.net) joined #forth 05:03:19 Hello, knows somebody a program for Forth source which generates Documentations, like Doxygen or Javadoc? 05:25:42 Sorry, I don't. 05:32:24 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 05:33:42 hmmm, that's too bad :( Robodoc has written in its manual that it can support Forth, but all comment header types doesn't look really like forth. 05:36:33 Forth can look very different... 05:37:49 --- quit: Robert ("brb") 06:08:35 --- quit: schihei (Read error: 60 (Operation timed out)) 06:18:04 --- join: schihei (~schihei@pD9E5CCE2.dip.t-dialin.net) joined #forth 06:30:05 hwo do i simply convert n to r? 06:30:11 (a normal number to a float) 06:33:27 err...in forth? 06:33:31 yessir 06:33:53 although i just realized i dont even need floats right now, but i will need it later anyhow 06:34:49 Hmm, looks like in ANS, you have to convert it to a double-cell number, then to float. 06:35:08 so S>D D>F I guess. 06:35:13 huh, how. as far as conversion goes i could only see string>float 06:35:31 oh d means double? i thought double precision or something like a double float 06:35:50 well that makes it easier then :) 06:35:53 :) 06:36:32 --- join: ASau (~asau@158.250.48.204) joined #forth 06:37:03 Dobryjj vecher! 06:38:57 Doh! 06:39:09 ALLOT takes a byte count, NOT a cell count. 06:39:28 That would explain my problems. 06:39:30 Hi ASau 06:39:30 yep 06:40:10 and read-file returns 3 arguments, not 2, which explains my infinite looping 06:40:11 hehe 06:41:02 programming works so much better when you know what the words actually do :) 06:41:13 well, ansi says it should return 2 06:41:14 :\ 06:41:44 so i dont even know what the other argument represents, but i only need the "actual bytes read" one 06:42:12 Oh. What forth are you using? 06:42:16 win32forth 06:44:45 That's weird. 06:46:02 oh i meant file-position 06:46:56 i keep reading 3 bytes, and put the position back 2 bytes 06:47:11 so i need file-position and reposition-file 06:47:28 I see 06:48:15 huh, and reposition-file consumes 3 arguments?? 06:48:53 code SWAP --> [4] 116 3 1 116 06:48:53 : REPOSITION-FILE 06:48:53 --> [2] 116 0 06:49:49 according to the kernel, it should only take 2 :\ 06:49:59 (the sourcecode of win32forth for that word) 06:50:19 : reposition-file ( len-ud fileid -- ior ) 06:50:19 FILE_BEGIN SetFP >R 2DROP R> ; \ ior - 0 = success 06:51:24 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:51:36 ANS says position is a double-cell number. 06:51:52 REPOSITION-FILE ( ud fileid -- ior ) 06:52:00 oh! 06:52:46 high-low? ( high low fileid -- ior ) or the other way around? 06:54:02 hmm... 06:54:44 "On the stack, the cell containing the most significant part of a double-cell integer shall be above the cell containing the least significant part." 06:55:10 so ( low high fileid -- ior ) 06:55:31 k 07:37:50 --- join: MrReach (~spam@209.181.43.190) joined #forth 07:37:51 --- mode: ChanServ set +o MrReach 07:38:07 hihi 07:39:14 Good evening! 07:40:18 yes, 07:30 here 07:40:38 06:30 here. 07:40:56 --- join: murl (quinton@murl.registered) joined #forth 07:41:14 Good evening, murl! 07:41:29 you're in Hawaii?? 07:41:41 or do you mean 18:30? 07:42:19 PM, of course. 07:42:38 ok 07:43:25 I never have heard that "evening" means "a.m." 07:43:55 I may mistake here. 07:44:05 :) 07:44:48 heya, qFox 07:44:52 hi 07:45:14 I was confused ... time starting with "0" usually means it's in 24 times 07:45:16 time 07:45:46 otherwise you have to deliberately place the ante/post indicator 07:45:59 Well. 07:46:19 I'm not used to british/american standard. 07:47:22 what country are you in? how do they specify time? 07:47:41 * qFox mumbles ctcp time 07:47:49 Russia. 07:49:15 We write quite arbitrarily. 07:50:31 It's better to append time of day if it is not clear. 07:51:00 * MrReach nods 07:55:49 I have an interesting idea. 07:56:02 i have to go in 2 minutes 07:56:18 We live in almost UNIX world. 07:56:58 Maybe we should work out common UNIX ways to deal with simple Forth related tasks. 07:57:24 Blocks to lines, lines to blocks. 07:57:31 heh, there's so many tools provided with unix 07:57:36 --- nick: MrReach -> MrGone 07:57:45 Maybe with not full blocks. 07:58:01 We can use FF to separate them. 07:59:02 And maybe it would be good to create printing filter. 07:59:10 fold 07:59:59 fold can convert blokcs and files 08:01:39 Well. dd can do this too. 08:02:04 What about separating blocks with form feeds? 08:20:02 err 08:20:50 dd cant insret an eol at the end of every block line 08:20:58 or remove them from the flat file 08:23:51 Strange. 08:23:58 My dd did. 08:26:00 dd cbs=64 conv=(un)block (?) 08:26:53 err it cant insert a 0x0d every nth char of the output file 08:30:04 Hm. 08:30:26 "dd cbs=64 conv=unblock" does it pretty well. 08:30:40 DJGPP v.2 08:38:29 --- quit: schihei (Read error: 60 (Operation timed out)) 08:39:06 --- part: murl left #forth 08:46:38 --- quit: warpzero (Read error: 110 (Connection timed out)) 08:48:53 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth 08:48:58 --- mode: ChanServ set +o kc5tja 08:50:20 Dobryjj vecher, kc5tja! 08:51:04 Greetings. 08:54:52 its not possible to devide with two doubles? 08:55:00 double-numbers... 08:56:07 D/ doesn't work? 08:56:16 no 08:56:21 hmm 08:56:22 its not in the ansi either btw.. 08:56:30 ansi only has d*/ 08:56:37 which takes only one d, and two n 08:56:52 oh thats M*/ 08:57:06 Don't know then. 08:57:13 grr 08:57:14 I never actually use double-width numbers, so... 08:57:28 well i want to do a percentile for two filesizes 08:57:37 and filesizes come as double numbers 08:58:33 If you're willing to live with just 32-bit integers, you can use D>S to convert the double number to a single-width number. 08:58:50 yeah i know, but i'm not sure if that'll be accurate enough 08:59:26 but i guess 4gig is pretty big 09:00:01 Another solution is to convert it to floating point somehow. I'm not sure how though -- I also never use floating point. :) 09:01:00 hm wait, thats not a bad idea... d>f 09:03:24 qFox. Divide both doubles by the same number. 09:03:36 --- quit: Robert ("brb") 09:03:37 huh? 09:03:40 You don't need all digits, when you want percentage. 09:04:05 a/b = (a/c) / (b/c) 09:04:12 ic 09:04:33 When you have integer division, you get approximate answer. 09:04:36 That loses the low-end bits, though, which is where 90% of the information is stored, since I *really* doubt he has very many files bigger than 4GB. 09:05:30 kc5tja. Actually, the number isn't to be the same for all files. 09:05:37 if anyone dares to run this code on a 4gig+ file and doesnt mind waiting at all... :p 09:06:13 ASau: Doesn't matter. It still assumes the presence of a double-width division operator, which does not exist in ANSI Forth, apparently. 09:06:36 For each pair you're able to find c. 09:06:49 a/c is single! 09:06:50 He doesn't want to find c. 09:07:14 He just wants to divide a/b; such an operation should not involve an equation solver. 09:07:28 its ok. it works now :) 09:07:33 * arke is b4ck 09:07:44 And why even bother doing that, when you can just lop off the top 32 bits anyway (which are 99% of the time guaranteed to be zero ANYWAY)? 09:07:53 No need to divide anything. 09:09:10 Those number may have 35 bits. 09:09:30 Not for a single file, no. 09:09:34 When you get rid of 32 lower bits, you get too little. 09:09:51 32 lower bits? I was talking about the 32 upper bits. 09:10:00 f/ [ 100 s>d d>f ] fliteral f* f. ." %" 09:10:28 ok i'd prefer there to be no space before the %, but i can live wit hthat 09:10:34 qFox: You could have just said f/ 100e f* f. 09:10:48 oh 09:10:57 e? 09:11:06 kc5tja. HEX 09:11:15 ASau: Don't use hex. 09:11:18 Problem solved. 09:11:42 well i am in decimal mode anyhow, but e puts a number directly in float? 09:11:59 qFox: In this context, e means "with an exponent of". It is scientific notation. I'm sure you've seen numbers like 1.24e+5 or some such before? That's where it comes from. But since our exponent is zero, we just leave it ff. 09:12:01 * arke is writing an email to the NanOS guy 09:12:12 aha ok 09:12:53 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 09:13:25 hi Robert 09:13:40 Hi 09:13:41 Robert: http://www.neocoder.net/nanos/ 09:15:26 A swede.. 09:15:58 So, why did you show me this? 09:17:06 well... i created a almost correct compressor, using a simple linear algorithm. it obviously has variating compression rates, and it looses on winrar (although my few tests seem to show so). it varies for images (100 bytes loss/win for a 200k image). 09:18:03 Robert: well, I wanted to know waht you thought about it. 09:18:10 Robert: seems quite nice. 09:18:32 arke: Well, I don't know how well it works... 09:18:42 For now, I'm happy with Debian 09:18:48 the code checks a file per 3 bytes (position increasing 1 byte at a time), and checks if the three bytes are a linear (whatsthename?). so b2-b1 = b3-b2 09:19:08 Robert: well, I meant for F2.. 09:19:27 * arke is playing with AROS 09:20:34 the downside is that it requires at least 4 bytes per euhm... found match (one escape char, one begin char, one number thats the coefficient (?), and one number indicating how many bytes should be put in) 09:20:52 plus it creates some overhead for the escapechar 09:21:37 but the code does tackle single byte repeating strings (aaaaaaa), since their coefficient is 0 09:21:49 ohwell 09:21:55 oh god 09:21:57 Interesting. 09:21:58 this is so annoting 09:22:08 AROS queries the floppy drive every 2 seconds 09:22:14 making a very audible sound 09:22:16 WAY annoying 09:22:17 heh 09:22:33 arke: That's because AmigaOS can auto-detect when a disk is inserted. 09:22:46 The only way to get that same effect on the PC is to ping the drive and actually spin the motor. 09:23:05 On the Amiga, it involved only a click, but you can't move the heads in the PC's floppy without spinning up the motor. 09:24:10 Besides, it's safer for the drive to have a floppy in it at all times anyway. 09:24:30 Helps to keep the heads clean. 09:25:15 * arke is going to turn off teh floppy stuffs 09:27:24 bleh 09:27:31 had to add a dummy floppy 09:27:37 otherwise it crashes VMWare :P 09:27:57 What crashes VMWare? 09:28:06 Not having a floppy in it ought not to crash VMware. 09:28:13 booting up AROS without a floppy 09:28:16 I know lol 09:28:23 oh well 09:28:28 Then VMware is broken, I suspect. 09:28:29 I got it for free, so... 09:28:47 yeah, it's got its quirks, but it's one HELL of an effort that I'm very thankful for. 09:29:33 even though I stole it.. lol 09:29:41 I dunno -- I never liked VMware. 09:29:55 It works most of the time. 09:30:01 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 09:30:03 it boots my linux partition without a sound 09:30:04 It has NEVER worked for me. 09:30:18 the older VMWare used to give me all kinds of shit though 09:30:21 hi warpzero 09:30:29 Even the latest versions don't work for me. 09:30:39 They inevitably crash, hang, or otherwise choke on even the simplest things. 09:30:56 I find bochs to be vastly more reliable for my needs. 09:30:59 never crashed on me, btw. 09:31:19 bochs is more reliable becasuse it copmletely emulates, not protects. 09:31:24 bochs has nothing to protect 09:31:25 haha. i tried to swap two floats with 2SWAP, but didnt know the floats no longer existed on teh data stack (or didnt realize it), because the devision was going the wrong way. so the 2swap didnt matter anything and i didnt understand why :p 09:31:26 vmware does 09:31:26 :) 09:31:53 arke: No, it emulates physical resources. 09:32:09 It may provide a bridge to some devices, like the floppies, but everything else it emulates. 09:32:18 In fact, there are some CPU instructions that it has no choice but to emulate. 09:32:19 VMWare? 09:32:25 oh 09:32:27 yeah. 09:32:34 Because the x86 is borked in that a user-mode program can tell when it's not running in user mode (RETARDED!) 09:32:41 but, it lets most instructios run native 09:32:55 :P 09:32:55 It sure doesn't seem like it to me, but whatever. 09:33:19 for explorer.exe (xp pro) 09:33:19 Esc = 79 (O) 09:33:19 Saved bytes = 168044 09:33:19 Size inputfile: 1007616 09:33:19 Size outputfile: 763529 09:33:19 >> 75.7758 % 09:33:20 its kinda like plex86 in that respect. 09:33:25 All I know is that *EVERY* opportunity I've had to use VMware, I've always walked away disgusted. 09:33:43 whats teh cd command in amiga/aros? 09:33:50 arke: CD 09:34:03 arke: However, you don't need it. 09:34:15 Type the name of a directory, and it'll auto-detect it's a directory, and switch to it. 09:34:26 cool 09:34:33 is it case sensitivE? 09:34:34 Use / to go back one level. 09:34:43 depends on the filesystem. 09:34:59 Most filesystems for AmigaOS *preserve* case, but are not case sensitive when matching filenames. 09:35:07 But Unix-compatible FSes are case sensitive. 09:36:22 woah cool 09:36:34 you can have the desktop be a separate, moveable window 09:38:46 * arke is impressed by the whole system 09:39:10 *** Virtual machine kernel stack fault (hardware reset) *** 09:39:10 The virtual machine just suffered a stack fault in kernel mode. On a real computer, this would amount to a reset of the processor. It can be caused by an incorrect configuration of the virtual machine, a bug in the operating system, or a problem in the VMware Workstation software. Press OK to reboot virtual machine or Cancel to shut it down. 09:39:22 gotta hate that. 09:39:56 arke: I *told* you AmigaOS rocked. 09:40:25 Of all the OSes I've ever used, AmigaOS is the lightest weight, fastest, most user-friendly system I've *ever* used. Period. Hands down. No further discussion needed. 09:42:13 --- quit: warpzero (Read error: 60 (Operation timed out)) 09:44:31 cool. 09:45:20 That's what Dolphin is most inspired by. 09:45:29 Or, at least, was. 09:45:32 Wow. 09:45:34 * arke likes it 09:46:42 ooh 09:46:47 here's another thing I'm loving 09:47:17 hold the right button to get the menu for that particular app at the top (like MacOS, only that MacOS has it always focused) 09:47:45 Yep. 09:47:59 Note that you can also make MULTIPLE MENU SELECTIONS with the LEFT mouse button WHILE the right is held down too. 09:51:09 So AROS is self-hosting now? I remember when AROS was still hosted under Linux. 09:54:28 yep. 09:54:49 it runs native x86, m68k (or whatever the amiga CPU was), linux, and fbsd 09:56:42 Oh, so it still can run under Linux then. 09:56:43 kc5 i would agree with that statement 100% 09:56:59 the one about the amiga os being the best os ever created i mean 09:57:08 * I440r misses his amiga :( 09:57:35 * arke is loving AROS 09:57:48 * kc5tja will have to play with AROS again. 09:57:55 * kc5tja is currently (still) using amiwm for X11. :D 09:58:11 it's running under VMWare, with other VMs open too, a whole bunch of native stuff, no accelerated graphics, and it's BLAZING! 09:58:17 kc5tja: :) 09:58:29 One thing I really miss is having HUGE Workbench screens (like, on a 640x200 display, having a 1024x1024 Workbench screen). 09:58:48 you can have that in C 09:58:49 er 09:58:50 X 09:59:00 arke: Only if your wm supports it. 09:59:08 eh? 09:59:15 sure about that? 09:59:24 I'm fairly sure. I am not 100% sure. 09:59:31 Well, yeah, i suppose you can. 09:59:42 if you ctrl+alt+- now, you'll see the same effect 09:59:43 Because you can ctrl-alt-- and have it drop into low resolution mode. 09:59:46 scrolling with the mouse 09:59:48 yeah 09:59:50 exactly 10:00:14 But, I still miss the ease of configurability with the Amiga. That, combined with multiple screen support, is just sooooooooooooo niiiiiiiice. 10:00:24 multiple screen == depth-arrangable screens. 10:00:35 --- quit: tathi ("leaving") 10:00:53 depth arrangaeble? 10:00:56 The nice thing about amiwm is it emulates multiple, depth-arrangable, scrollable screens very fast. 10:01:18 arke: Notice that the Workbench screen titlebar has a depth-arrange gadget on it? 10:01:41 yeah. 10:01:49 Try dragging the screen's titlebar down. 10:01:50 :) 10:02:15 If you had another program running in a different screen, you'd reveal the screen immediately beneath the Workbench. 10:02:23 In other words, in AmigaOS, screens layer like windows do. 10:02:37 oh, you mean the wanderer being able to move like a window of its own? 10:02:52 oh....hrm. 10:02:55 I don't know what "the wanderer" is, so I can't answer that question. 10:03:07 the wanderer seems to be like the finder in macos 10:05:08 actually, if I may say that, AROS and my old Mac (1986, dunno what version, like 4?) are quite similar in interface 10:06:57 Ahh, so what they call Wanderer was called Workbench in AmigaOS. 10:14:35 --- join: arke_ (~Chris@wbar8.lax1-4-11-100-108.dsl-verizon.net) joined #forth 10:26:22 Dude, if AROS gets that TCP/IP implementation going, I just might consider tinkering with it instead of Linux for my desktop. That's assuming they have reasonable IRC, e-mail, and web browser support of course. Oh, and VIM. Gotta have VIM. 10:31:50 --- quit: arke (Read error: 110 (Connection timed out)) 10:35:00 of course, always gotta have VIM :) 10:35:03 --- nick: arke_ -> arke 10:35:23 are they working on eth right now? 10:35:56 TCP/IP isn't Ethernet. 10:36:03 They have Ethernet drivers as far as I can tell. 10:36:12 They don't have bsdsockets.library though. 10:36:21 There is a $575 bounty on it. 10:36:33 bounty? 10:36:57 Yes. People will pay a bounty if you port or write a new open source module for something within a specific period of time. 10:37:25 hmmm ic 10:37:51 All donations come from other members of the community. 10:38:14 cool. 10:38:23 can't wait until they have that then. 10:38:24 For example, if I were a business or whatever, and I wanted to see some accounting software, I might offer some cash as an incentive to get people to port something like OpenOffice to the OS of my choice. 10:38:32 Others who agree would add to the pool of cash, and the bounty goes up. 10:39:17 For example, there is a $100 bounty on getting AROS running under PowerPC Linux. 10:39:23 then a curses port and then port irssi 10:40:39 I'm not seeing any open bounties for that. 10:41:52 $100 is hardly worth my trouble for such a large project (or at least it sounds like large) 10:41:53 true. but if you have BSD sockets, all you need is curses and then you can port irssi, right? :) 10:42:04 --- join: warpzero (~warpzero@dsl.142.mt.onewest.net) joined #forth 10:55:02 unistd.h is there. 10:55:21 the headers are coming! the headers are coming! 10:55:23 hi kc5tja 10:55:30 hi kc5tja 10:55:32 hi warpzero 10:55:34 hi arke 10:57:29 qFox: Why? Just port the sockets API that exists in another OS, bundle it as an AmigaOS library, and release. 10:57:35 I can't imagine it being at all that hard. 10:57:50 re warpzero 11:00:06 kc5tja: he meant for teh PPC port 11:00:46 arke: Even then, I can't see it as being all that hard. 11:00:51 It *ALREADY* runs under x86 Linux. 11:01:17 ^_^^__^^___^^____^^___^^__^^_^ 11:04:16 what are we porting? 11:04:22 AROS 11:05:27 Well, *I* am not porting anything. 11:05:28 what is that? 11:05:34 But I'm just saying, it can't be hard to port. 11:05:36 (that hard, rather) 11:05:48 AROS is an open source clone of the AmigaOS 3.x-era operating system. 11:07:34 oh 11:07:40 i never used anything like that 11:07:47 im too young 11:08:02 i think the oldest os ive used was THEOS/OASIS 11:24:12 --- quit: Robert ("brb") 11:33:23 --- join: Robert (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:50:54 wb Robert 11:57:01 Time to work on re-organizing my website again... 11:57:06 Service calls are made through a call gate at selector 0008h(descriptor 1 in GDT). 11:57:10 what does that mean? 11:57:41 That means calls are made using a far call to address 0008h:00000000h. 11:58:15 To understand what precisely this means you will need to be aware of how Protected mode works in x86 processors. 11:58:25 well, i don't. :) 11:58:35 and especially how selectors work. 11:58:41 All I know is that segments don't map to physical memory, but to virtual memory 11:58:46 or selectors, yeah 11:59:01 Yes, but these selectors/segments (same thing) are referenced by *index* into a table. 11:59:42 13 bits comprises the index into either the global or local descriptor table, then there is a bit indicating which table to use, then the other two bits comprises the privilege level you claim to have. 12:00:16 into the GDT, correct? 12:00:22 In this case, yes. 12:00:25 or LDT, i guess. 12:00:27 No. 12:00:28 GDT. 12:00:36 Entry #1, global table, privilege level 0. 12:00:42 Because the lower three bits are cleared. 12:00:57 privilege 0 means full access? 12:01:55 Well, it's kind of hard to answer that question, but suffice it to say that the "requested privilege level" bits must match certain expectations (in tihs case, 00) for the call to succeed; otherwise, there will be a privilege violation. 12:02:41 0-3 are highest-lowest level though, right? or is it more of a "user" thing? 12:06:30 0 is the highest privilege, and 3 is the lowest, yes. 12:06:54 Though, in practice, most OSes only use rings 0 and 3, and completely ignore rings 1 and 2, because their behavioral semantics are not nearly as well specified. 12:08:16 kc5tja: waht' syour opinion on micro/nano kernels? 12:09:12 * arke wonders if he can convince pygmy to create 32-bit pmode code 12:09:29 I mean, not run it etc., just compile it. 12:09:41 ack 12:09:44 nevermind, pygmy is DTC. 12:09:51 hrm, actually, possible anywayh. 12:11:32 arke: Well, I got gforth (a DTC Forth implementation) to produce ELF executables for my FTS/Forth system. :) 12:11:38 It's up on my website, if you don't believe me. :-) 12:12:01 Although the future of my MachineForth systems appears to be BoxForth, of which my FTS1001 assembler will form the core of. 12:12:23 It ought to be very, very easy to adapt FTS1001's assembler to the x86 platform, which *will* obsolete my current FTS/Forth Cross Compiler. 12:12:51 Admittedly, I ripped most of the software infrastructure off from Chuck's MachineForth implementation, as documented by Jeff Fox. :) 12:13:26 arke: As far as micro/nanokernels, who knows? I'm in favor of them, but they have to be implemented just right. AmigaOS's exec.library is a microkernel, and is largely why AmigaOS kicked ass. 12:13:40 But when memory protection gets involved, some factors change. 12:13:59