00:00:00 --- log: started forth/04.07.15 00:01:19 --- quit: crc (Client Quit) 00:02:33 --- join: crc (crc@0-1pool176-13.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 00:03:25 http://cryptomancer.narod.ru/foto/way2work/ 00:46:51 --- join: arke (arke@melrose-251-251.flexabit.net) joined #forth 00:46:56 HI! 00:47:26 Hello arke 00:47:46 greetings 00:47:50 whats new? 00:48:03 I'm attempting to get RetroForth/Windows to work as a DLL :-) 00:49:53 Cool :) 00:50:00 * arke should try RetroForth 00:50:17 Version 7 is getting better & better 00:51:17 The DLL version can initialize itself, and put up an interpreter console if desired. Now I'm trying to make it possible to pass strings to interpret/compile 00:51:39 whats it written in? 00:51:45 pure asm+forth? 00:51:46 Assembly & Forth 00:51:48 Yup 00:51:56 Coolness. 00:52:07 how much is it like my favorite forth? :) 00:52:28 What's your favorite Forth? 00:53:25 It's a minimal Forth, easy to expand & learn ;-) 00:53:40 Exactly 100 words, constants, and variables 00:54:31 http://retroforth.org/board/index.php?board=4;action=display;threadid=74;start=0#lastPost for a list of them 00:54:56 (They might change slightly before RF7 is officially released) 00:57:01 Yay! I can pass counted strings to it and have them interpreted :-) 00:59:02 The windows version has a lot more code involved in it :-( 00:59:12 Linux = 628 lines 00:59:28 FreeBSD = 650 lines 00:59:47 BeOS = 619 lines 01:00:00 Windows = 6,718 lines 01:00:02 :-) 01:00:20 I should try to pare that down a bit... 01:01:36 o.O 01:03:05 The minimal codebase for Windows is 2,026 lines; which is still too big IMO 01:03:19 yeah.. 01:03:29 why is it so much? 01:03:39 But that's mostly macros & equates that the assembler needs to build the import/export tables 01:04:28 oh.... 01:04:30 eww. 01:04:44 If I had the patience, I'd wade through those files and eliminate all of the equates, etc that I don't use. 01:05:06 Maybe I'll get around to it before I finish RF7 :-) 01:05:40 * crc ponders whether to release RetroForth/Windows as a single application or a DLL and wrapper... 01:05:54 the latter might be better 01:05:58 albeit slower 01:06:01 but then again# 01:06:11 6k lines ... it doesn't really matter anymore after that :) 01:06:19 Not that much slower 01:06:44 At <6k for the DLL, it loads pretty quickly 01:08:01 The BeOS port is the slowest I've tested (my performace tests don't work under Windows...) 01:08:16 Of course the BeOS port also uses LIBC, so that's to be expected... 01:08:27 eww ;) 01:08:47 http://bash.org/?361585 01:09:33 :-) 01:10:32 --- quit: arke ("Lost terminal") 01:12:10 --- join: arke (arke@melrose-251-251.flexabit.net) joined #forth 01:12:27 sorry about that 01:12:59 np. 01:13:23 la dee doo 01:16:35 moo 01:16:38 talk to me 01:16:58 About what? 01:17:22 #forth @ #forth#users @ FOR DUP + ME TALK NEXT DROP 01:17:29 anything .) 01:17:53 * crc is (attempting to) writing documentation for RetroForth 01:18:00 I HATE WRITING DOCUMENTATION 01:18:37 so do i ;) 01:18:52 later# 01:19:01 later# ? 01:19:46 --- quit: arke ("'Do you have mayonaise?' 'Yes, would you like some?' 'No, I just wanted to get an overview of the situation here. Gimme some ) 01:45:31 Serg: nice pics 01:49:19 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 02:28:44 --- quit: crc (Read error: 110 (Connection timed out)) 02:53:06 fridge: sub-30$ digicam, 640x480, matchbox-size 02:54:20 have you read ALT's (ACDSEE does not put comments on page ;( 02:54:23 ? 03:07:46 Serg: I had missed those 03:08:28 this is my waypoints on the way to work 03:08:34 incomplete ;) 03:08:56 i should write what whis is Moscow, Russia ;) 03:09:45 do you live in the apartments in 001.jpg? 03:10:19 almost - my windows are behind me 03:10:51 i'll shoot my place today ;)) 03:13:07 * Serg looks at photos i took yesterday, right after buying it 03:45:45 --- quit: Research (Remote closed the connection) 03:46:33 the subway looks more like a dining hall hehe 03:46:56 --- join: Research (XINU@12-222-128-22.client.insightBB.com) joined #forth 04:03:19 fridge: ? 04:03:35 sometimes it's like can of worms 05:03:01 --- join: tucknip (pickroll@dialup-4.152.183.179.Dial1.Atlanta1.Level3.net) joined #forth 05:08:26 Serg: with all the lighting 05:08:55 sydneys rail system is bleragh bland and ugly 05:10:05 bleragh bland ?? 05:11:27 * Serg gonna buy one more cam, saw off it's lens and place the CCD behind a good old 35mm telethoto lens ;)) 05:16:11 now it's f=4.5mm 05:17:54 and "equivalent" is maybe 28mm 05:17:58 hmm 05:18:04 * mur wants mamiya 05:18:16 it's considered to be rolls royce of cameras 05:18:17 ig uess 05:18:32 i could loan one from school though 05:18:53 if i put the CCD behind "classic" 50mm, i'll have over 280 mm "equivalent" - but aiming gonna be difficult 05:19:52 mur: mamia rocks ;)) i want DLR one ;)) 05:21:00 i want to one with movable hmm axis 05:21:01 older one 05:21:14 what i want more - cheap scanning view camera, 120 Mpixel for the price of flatbed scanner 05:21:46 factory-made, not home conversion 05:21:52 hmm 05:22:09 the large negative mamiyas are, when negative is scanned 180 MPix 05:25:19 hehe... 05:25:43 A4 back and sheet of Unibrom photo-paper ;)) and F/64 05:25:59 and reversing b/w process 05:26:09 scanning cameras are so slow 05:26:18 not very usable, becasue they take long time 05:26:44 and you cant do as many effects, like delayed light and take quick pictures 05:27:14 would i be given a task to design commercial 30 Mpix under 200 bucks.. 05:27:50 i'll take a focal shutter and put sensor on it's curtain ;))) 05:28:30 it'll be very fast scanner, compatible w/ all 35mm lenses - no crop factor 05:29:14 can you create me 30 mpix camera that costs less than 200 usd? 05:29:15 :) 05:29:26 M39 Leica screw mount and adaptors for all kinds of lenses ,) 05:29:32 not "on the knee" 05:29:38 hmm? 05:29:45 i'm not familiar with that expression 05:29:51 do you mean from the beginning? 05:29:55 from scratch 05:30:19 no, i mean "w/o proper equipment, at home, in garage" 05:33:36 --- part: tucknip left #forth 05:35:06 btw, if sensor gonna be moved by cocked spring, scan cam can be made quite cheap and mobile 05:36:28 toy Mamia for 50$ ;)) 05:36:39 like my matchbox ;) 05:55:10 Privet, Serg :) 05:58:43 hi 05:58:53 i got tforth but did not try yet 05:59:04 did u seen my pics from matchbox camera ? 06:02:03 No 06:04:51 So, does that camera fit in a matchbox? ;) 06:07:46 Picture is still loading, but it seems pretty OK. 06:13:01 have no matchbox here to test ;)) 06:15:25 50x38x16mm 06:16:26 * Serg measured tiny camera by huge 5m metal measure tape 06:16:56 Heh, cool 06:17:20 That's certainly small enough to attach to a balloon ;) 06:17:48 but not controllable ;( 06:17:59 and inferior quality 06:20:47 i attempted self-shot but missed hard ;)) 06:21:42 :D 06:22:18 no, later one almost hit ;)) 06:23:43 it's tended to exposure errors - don't shoot against the light 06:27:04 Hehe 06:29:53 but even the worst photo can be saved by Photoshop ;) 06:38:37 * Serg makes "artist painting" out of grainy photo 06:41:56 wide angle portraits look funny ;)) 06:42:47 Hehe 06:44:17 and even more - under heavy "unsharp mask" filter 06:54:02 at home, i'll shoot thru monocular 06:54:21 i once shot thru telescope w/ friend's cellphone cam 06:56:58 Whoa 06:57:01 Did it work? :) 06:57:09 yes 06:57:35 we saw a circle w/ far away waterpump tower (it was at countryside) 06:57:45 hand-held, no mounting 06:58:44 :) 07:01:43 hm, it refuses to shoot at too dark or w/ lens obscured 07:13:51 beat it 07:26:07 --- quit: Serg () 07:47:14 --- join: warpzero (~warpzero@mi197.dn189.umontana.edu) joined #forth 07:59:50 --- join: wossname (wossname@HSE-MTL-ppp77700.qc.sympatico.ca) joined #forth 08:08:36 --- join: juhammed (~o@dsl-olugw3p33.dial.inet.fi) joined #forth 08:12:25 --- quit: Research (Read error: 110 (Connection timed out)) 08:47:08 --- quit: wossname ("asdf") 09:16:52 --- join: Serg (~z@193.201.231.126) joined #forth 09:17:28 re 09:17:47 Hi :) 09:17:50 i matched my cam w/ matchbox - exactly same size ;)) 09:18:17 outer size ;) does not fit in by 1-2 mm ;)) 09:19:09 Cool :) 09:19:30 i shot my home, Pravda and the place of big fire i mentioned once... but the damn thing either hung or battery discharged 09:24:00 ooops... 09:27:34 * Serg charges my AAA cells ripped from old Siemens cellphone at trash work 09:29:57 :) 09:30:58 great rain, or even thunderstorm, is coming ;) today i seen the tree uprooted by the wind 09:31:28 Heh.. we had some heavy winds the other day, but today it's clear and sunny 09:31:40 * Serg is reading tforth.txt 09:33:51 Hehe, OK :) 09:34:05 * Robert just added recurse, tail and tail; 09:34:06 nice ;) 09:34:21 Tell me if you think anything is missing 09:35:20 A! A@ AC! AC@ - for access beyond it's segment, mainly to screen but othr peeks/pokes too 09:35:41 How would they look? 09:36:11 Look in mem.asm 09:36:19 * Serg looks gp-forth doc ;)) 09:36:22 I have far! farc! etc. 09:36:33 Also functions for block copying over segments. 09:36:36 hey I440ri wanna talk to you 09:36:59 ah, good 09:42:15 word not found gfx-sync ??? 09:42:43 yey 09:42:45 :p 09:42:48 Hmm.... 09:43:14 Serg: It SHOULD be, I'll tets 09:43:15 test 09:43:34 by F3, not found too ! 09:43:39 Over here it's found. 09:43:43 F3? 09:43:54 Norton's Viewer 09:44:06 Hrm. 09:44:17 Which version did you download? 09:44:34 I uploaded the new gfx code yesterday, right after we talked about it 09:44:40 i hit the link today maybe about 10:00 by Moscow 09:45:06 i see it in gfx.asm 09:45:12 but not in tf.com ! 09:45:25 Maybe I forgot to reassemble. 09:45:31 Do you have NASM? 09:45:37 sto-o-op ! 09:46:33 What? :) 09:46:44 glitch ! 09:46:54 What do you mean? 09:46:55 it was in tforth.com but not in tf.com 09:47:12 i hit tforth.com again - and it appeared in tf.com ! 09:48:01 shroedinbug ;( 09:49:07 * Serg sees no pairing control in loops.f - bad, this may lead to hardly trackable errs 09:49:27 but... let's it stay so ! 09:50:03 my GP-FORTH does catch mismatches 09:52:04 GP: >: test 10 0 do [ s. ] loop ; 09:52:05 4665 18804 4669 - Ok 09:52:37 tforth: : test 10 for [ s. ] next ; 21c7 09:53:26 Hehe 09:53:37 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 09:53:37 That's a nice idea, but... slow and large ;) 09:53:42 heh, w/o commandline i'm like w/o files at all 09:55:02 What do you mean? 09:55:07 echo include test.f | TF.COM - not works too ;(( 09:55:35 i mean, i want to write my code in files and invoke compiler in batch mode, not dialog 09:55:35 Well, you can fix that. 09:55:51 Do something like... 09:56:41 : new-key 0 fkey drop ; ' new-key is key 09:57:19 So read from stdin instead of the BIOS keyboard routine. 09:58:42 Adds an extra newline in interactive mode (double buffering - DOS + tforth), but maybe you can live with that. 09:58:54 echo include test.f | TFC.COM - black screen, Ctrl-Alt-Del aka 'kill' 09:59:28 Hehe, I noticed ;) 09:59:35 it's "fload test.f" btw 09:59:40 But still doesn't work very well. 10:01:36 * Serg eats, pause 10:09:10 where is cmdline buff ? how can i interpret it ? 10:11:59 I think it's at offset 81 in your segment. 10:12:32 Do this: 10:12:46 tf.com hello 10:12:50 80 print 10:12:56 That should print " hello" 10:30:08 really ! 10:30:37 let me guess farther... 10:31:30 word not found: process 10:31:51 but it's in main.f ! 10:34:16 how the _hell_ is it not found, if it worked to PROCESS my input ????? 10:35:21 hey, Robert, you have shroedinglouse in TFORTH !!! 10:35:42 i seek in TF,com and see no PROCESS 10:35:55 i hit tforth.com second time - and it appears in TF.COM !!! 10:36:39 wha-a-a-a-at ??? 10:36:53 i see it by view, but "not found" !!! 10:39:53 ' process . da1 10:40:06 80 process 10:40:07 Word not found: process 10:40:22 * Serg calls for ghost busters 10:41:32 Hmm... 10:42:04 process is used differently. 10:42:14 ...I think. But I'm not sure, let's check. 10:42:43 Hrm, you're right, it should work 10:43:00 Even though you'd have to make those into separate words... 10:43:00 does '80 proces' work for you ? 10:43:23 It shouldn't work very well, process needs ONE word 10:44:40 so how may look cmdline executor ? 10:45:06 work or not work, it's not even found ! 10:45:13 You'll need to parse the string at 80, and give one word at the time to "process". 10:45:32 Type: 10:45:34 : x process ; 10:45:38 And you see that it's found. 10:45:44 but how can i give it to WORD ? 10:46:22 Write a command-line parser or something. I'll see if I can fix this 10:46:29 : x process ; 10:46:30 80 process 10:46:30 Word not found: process 10:46:30 ' process . da1 10:46:48 : and ' see PROCESS, but not direct interpretation 10:47:07 I found the error, btw 10:47:14 The error message in "process" is wrong. 10:47:21 --- join: I440r_ (~mark4@64.47.44.254) joined #forth 10:47:29 It doesn't print the argument, but the "word" buffer. 10:47:30 how so ? 10:47:36 what is the right error ? 10:47:48 hi I440r 10:49:16 hi serg 10:49:37 Now it works 10:49:52 Had to modify some things in compiler.asm 10:50:12 * Serg is shaking shroedinglouses out of Robert's TFORTH ;)) 10:50:16 Now this works: 10:50:22 : x ." Foo" cr ; 10:50:26 s" x" process 10:50:29 Foo 10:50:53 nice ! 10:51:34 OK, i gonna shoot off some heads in Far Cry and then sleep ;)) 10:51:38 Hehe 10:51:40 Night Serg :) 10:51:48 g'nite, bye ! 10:51:52 --- quit: Serg () 11:15:57 --- nick: lalalim_ -> lalalim 11:41:56 --- quit: slava (bear.freenode.net irc.freenode.net) 11:45:53 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 11:45:54 --- quit: I440r_ ("brb") 11:46:32 --- quit: slava ("Leaving") 12:03:57 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth 12:03:58 hi 12:04:07 Hi slava :) 12:04:10 yo 12:04:20 i'm working on catch/throw. 12:04:42 its similar to how forth does it, except you can throw anything (like a string, or a list etc) and the catcher can check the type of the thrown object and act accordingly. 12:05:18 Cool 12:05:34 I implemented those features in my Forth the other day, quick two-line hacks. 12:05:46 yeah, this is quite minimal too. 12:05:56 But at least it works and does its job. 12:06:02 there is a 'catch stack' that is pushed/popped when you enter a block protected by 'catch' 12:06:40 No fancy throwing around, designed to use an "errno" variable or something if you need that. 12:07:10 i like simplifying code. 12:08:44 i use catch for two things -- restoring the interpreter prompt after an error, and recovering the http server from something like a broken connection 13:39:53 --- log: started forth/04.07.15 13:39:53 --- join: clog (nef@bespin.org) joined #forth 13:39:53 --- topic: 'FORTH programming language. Info: http://forth.bespin.org/resources/introduction ANSI: http://www.taygeta.com/forth/dpans.htm FIG83: http://forth.sourceforge.net/standard/fst83/ Links: http://members.dsl-only.net/~loophog || Kestrel project weblog: http://www.falvotech.com/weblog || Note: this chan is publicly logged.' 13:39:53 --- topic: set by kc5tja on [Thu Jul 01 21:50:42 2004] 13:39:53 --- names: list (clog tathi crc chandler Topaz warpzero lalalim zardon slava juhammed qFox breno @ChanServ titanstar ianp cmeme ayrnieu Robert onetom skylan [Forth] fridge @I440r Klaw mur Fractal) 13:42:40 crc, try the recursive fib 'benchmark' 13:44:22 Code? 13:45:10 I use: 13:45:11 : .fib >r over + swap r> loop .fib ; 13:45:11 : fib 0 1 rot .fib drop ; 13:45:11 ." Fib (45) = " 45 fib . cr 13:45:20 To calculate fib 13:45:34 (RetroForth doesn't have conventional looping words) 13:45:39 recursive fib looks like this: 13:46:03 : fib dup 1 <= [ drop 1 ] [ pred dup fib swap pred fib + ] ifte ; 13:46:08 just change ifte to IF/ELSE/THEN 13:46:11 my guess: 13:46:19 : fib dup 1 <= if drop 1 else pred dup fib swap pred fib + then ; 13:47:22 What is pred? 13:47:39 I don't have ELSE either 13:47:53 crc: bah, yours is fine :-) 13:48:50 crc, pred is 1- 13:49:01 tathi, yes but the binary recursive one makes a semi-useful benchmark :) 13:49:57 My routines calculates 45 FIB in about 0.002 seconds :-) 13:50:25 Err, 0.012 seconds; sorry 13:51:05 slava: recursive or looping, who cares? 13:51:38 tathi, the doubly-recursive fib is much slower 13:51:45 its O(2^n) wheras crc's version is O(n) 13:51:51 so its good for benchmarking the interpreter itself 13:52:40 But if you're actually using it, you'd code a faster algo anyway! 13:52:42 yeah, but unless you do something really dumb, a forth interpreter should be faster than you need it to be anyway. 13:52:53 so as I said, who cares? 13:53:32 I try to only worry about speed when it turns out (in practice) to actually be a problem. 13:54:12 --- join: Tomasu (~moose@S010600045a4c73cc.ed.shawcable.net) joined #forth 13:55:01 --- quit: Topaz (Remote closed the connection) 14:07:33 --- join: tucknip (~Huck@170.141.109.33) joined #forth 14:13:07 * crc just realized that the version of FIB I had posted will not work anymore... This is the correct one: 14:13:07 : .1 >r over + swap r> repeat .1 ; 14:13:07 : fib 0 1 rot .1 swap drop ; 14:13:09 :-) 14:21:37 what is .1? 14:21:57 :) 14:22:18 --- join: kuvos (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 14:22:32 oh i c 14:23:04 Welkom terug, vos ;) 14:24:25 I could do something like this: 14:25:05 : fib 0 1 rot : .1 >r over + swap r> repeat .1 swap drop ; 14:27:01 RetroForth lets me nest definitions :-) 14:27:47 that's confusing 14:28:10 why? 14:28:48 so the first use of the word ends the definition? or what? 14:28:56 I don't find it any more confusing than I do the use of [ ] in Factor 14:28:58 Consider this: 14:29:06 : hello s" Hello, 14:29:14 Oops 14:29:55 : hello-world s" world!" : hello s" Hello, " type type ; 14:30:06 hello-word would display 'Hello, World!' 14:30:15 I can also use 'hello' separately: 14:30:22 " serg" hello 14:30:27 'Hello, serg' 14:31:07 This feature is taken from SEDForth 14:31:29 Normally each definiton would be on a separate line, with indents: 14:31:55 : hello-world s" World!" 14:31:56 : hello s" Hello, " type type ; 14:32:16 crc: ';' ends all "open" words, then? 14:32:21 Yup 14:33:19 interesting... 14:33:25 kind of strange, though. :-) 14:33:27 It's one word, with multiple entry points 14:33:38 You can get creative with multiple entry/exit points :-) 14:33:49 oh! duh. 14:33:55 I'm with you now. 14:34:06 Just not used to seeing multiple entry points in a punctuated forth. 14:35:21 * tathi wishes you could do multiple entry points on PPC... 14:35:54 Why can't you? 14:36:23 well, you can...but only in words that don't call any other words. 14:36:57 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 14:37:14 because the part of a "call" that saves the return address to a stack is inside the called function. 14:38:01 so generally each entry point needs: 14:38:21 mflr r0 #get return address from link register 14:38:34 stwu r0,-4(r1) #push it onto the stack. 14:39:42 does that make sense? 14:39:52 Yes, though I don't like it 14:40:06 * crc likes having 'call' push the return address to the stack 14:40:36 yah, well, it's a RISC thing. 14:40:44 I think MIPS does the same. 14:41:07 I gather the point is that you can avoid the overhead of writing the return address to memory when you don't need to. 14:41:18 But it is rather a nuisance. 14:41:23 I agree 14:44:11 RetroForth won't be ported to another CPU for a while anyway. When it is, it'll be a custom CPU done in FPGA :-) 14:45:32 I wish you could get FPGAs in a package that didn't have so damn many pins. 14:45:53 --- part: tucknip left #forth 14:51:17 me too 14:51:55 I don't know what they use them for normally, but it seems like way more pins than you need for a simple microprocessor. 14:52:48 Ah well. 14:52:59 I agree. If I was using them, I'd try to implement several CPU's on chip, running in parallel :-) 14:53:11 I was thinking about that :-) 14:53:20 At least they'd be put to good use then. 14:54:09 --- quit: Topaz (Connection timed out) 14:54:43 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 15:06:03 --- quit: Topaz (Remote closed the connection) 15:06:31 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 15:23:43 --- quit: Topaz (Read error: 110 (Connection timed out)) 15:24:12 --- join: Topaz (jonny@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 15:29:55 --- join: crc_ (crc@0-1pool88-140.nas48.philadelphia1.pa.us.da.qwest.net) joined #forth 15:31:45 --- quit: crc (Nick collision from services.) 15:32:02 --- nick: crc_ -> crc 15:55:37 --- quit: qFox ("meh ircii's netsplit detection appearantly sux ^^") 15:58:42 --- quit: Topaz ("Leaving") 16:13:44 --- quit: warpzero ("Tried to warn you about Chino and Daddy Gee, but I can't seem to get to you through the U.S. Mail.") 16:33:53 --- join: I440r_ (~mark4@64.47.44.254) joined #forth 16:35:15 Hi I440r_ 16:35:19 hi tathi 16:35:23 how goes the port ? 16:35:29 Well... 16:35:30 :) 16:35:31 im currently in calif and cant get in here much 16:35:38 yeah, you said. 16:36:17 I actually managed to find a workaround/fix for the problem with putting " and ; in strings. 16:36:34 cool 16:37:03 But then I found that gas was putting a symbol in the ELF file for vlink, and relocations for all the references to it. 16:37:16 So basically all the header link fields were getting the same address. 16:38:10 lol 16:38:15 fsckign hate gas 16:38:16 yup. 16:38:43 took me way too long to figure out why I was getting a stack overflow while still trying to relocate the first header. 16:38:47 :) 16:39:10 I found a workaround to get it to generate separate relocations for each link, that wasn't _too_ hard. 16:39:13 But... 16:39:43 it turns out there's a bug in gas where doing an align screws up the next invocation of .set 16:39:50 (which is like nasm's %xdefine) 16:39:56 Or something. 16:40:27 So I'm currently working down to a minimal example that demonstrates the problem so I can send the binutils guys a bug report. 16:40:36 Grr... 16:41:01 lol 16:41:04 i fsckign HATE gas 16:41:12 no argument there. 16:41:16 its too much of a fscking pain in the ass 16:41:22 its not intended to be used as an assembler 16:41:27 its a fucking c compiler rear end 16:41:28 yeah, that's the problem. 16:41:35 yup 16:42:00 I'm seriously starting to consider just giving up and just hacking together a quick PPC assembler in C. 16:42:50 write the assembler in c. then use that to write an assembler in asm 16:42:52 :) 16:42:56 then use that to port isforth :P 16:42:58 hehe 16:43:06 I mean, even if the binutils guys (or me) come up with a patch, then anyone who wants to use isForth on PPC would have to upgrade to the latest binutils 16:43:13