00:00:00 --- log: started forth/04.08.30
00:04:44 --- join: Serg[work] (~z@212.34.52.140) joined #forth
00:16:50 <qFox> damn cows.
00:16:50 <qFox> 09:16:21 IdleBot@#idlerpg@freenode: qFox was attacked by a rabid cow. This terrible calamity has slowed them 0 days, 04:10:37 from level 43.
00:25:35 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth
00:25:42 --- mode: ChanServ set +o kc5tja
00:25:51 <kc5tja> Back briefly.
00:26:08 <qFox> gone for a few days. back wednesday or thursday. later
00:26:17 <kc5tja> Hehe :)
00:26:22 <kc5tja> see you qFox :)
00:26:24 <qFox> intro camp uni
00:26:25 <qFox> :D
00:26:27 <qFox> bye
00:26:34 --- quit: qFox ("gone for a few days")
00:32:44 --- join: crc (crc@27-pool1.ras11.nynyc-t.alerondial.net) joined #forth
00:32:46 <kc5tja> re
00:32:47 <kc5tja> crc
00:32:51 <kc5tja> oops
00:32:54 <crc> Hi kc5tja
00:32:56 <kc5tja> you get the idea.  :)
00:33:00 <crc> :-)
00:33:06 <kc5tja> ENTER, Space, it's all the same.  :D
00:33:14 <crc> right
01:12:52 <kc5tja> off to bed.
01:12:59 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
01:30:07 --- join: proteusguy (~proteusgu@66.168.185.144) joined #forth
01:42:15 --- join: mur_ (~mur@smtp.uiah.fi) joined #forth
01:53:02 --- quit: mur (Read error: 110 (Connection timed out))
02:53:10 --- quit: mur_ (Remote closed the connection)
02:53:42 --- join: mur (~mur@mgw2.uiah.fi) joined #forth
04:20:47 --- quit: crc (Client Quit)
04:23:06 --- join: mur_ (~mur@mgw2.uiah.fi) joined #forth
04:34:56 --- quit: mur (Read error: 110 (Connection timed out))
04:49:21 * Tomasu is away: night
04:49:29 --- nick: Tomasu -> TomasuDlrrp
05:22:28 --- join: mur (~mur@smtp.uiah.fi) joined #forth
05:34:18 --- quit: mur_ (Read error: 110 (Connection timed out))
06:19:44 --- join: mur_ (~mur@smtp.uiah.fi) joined #forth
06:30:56 --- quit: mur (Read error: 110 (Connection timed out))
06:31:43 --- join: I440r_ (~mark4@216-110-82-203.gen.twtelecom.net) joined #forth
06:47:52 --- join: zruty (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth
07:03:15 <I440r_> slava
07:03:21 <I440r_> i know whats wrong with ./isforth
07:21:14 --- quit: I440r_ (Read error: 110 (Connection timed out))
07:21:23 --- join: I440r_ (~mark4@216-110-82-203.gen.twtelecom.net) joined #forth
07:24:09 --- join: mur (~mur@smtp.uiah.fi) joined #forth
07:34:59 --- quit: mur_ (Read error: 110 (Connection timed out))
07:53:58 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth
07:53:58 --- mode: ChanServ set +o Herkamire
07:58:30 --- quit: Serg[work] ("going home...")
08:27:41 --- quit: zruty ("Lost terminal")
09:14:05 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth
09:14:12 --- mode: ChanServ set +o kc5tja
09:40:29 --- join: mur_ (~mur@uiah.fi) joined #forth
09:41:23 --- quit: proteusguy (Read error: 104 (Connection reset by peer))
09:42:03 --- join: proteusguy (~proteusgu@66.168.185.144) joined #forth
09:46:52 <Herkamire> I'm begining to think that getting herkforth to bootstrap from source is more trouble than it's worth
09:48:08 --- quit: mur (Read error: 110 (Connection timed out))
10:12:03 <kc5tja> What do you mean?
10:15:21 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
10:22:54 <tathi> I440r_: I told ya you should add a .bss section... :P
10:36:08 * kc5tja thinks he found his problem.
10:36:12 * kc5tja tests...
10:36:32 <slava> hi
10:37:40 <kc5tja> FOUND IT!!!
10:37:49 <kc5tja> The -IF word was generating the wrong opcode.
10:37:50 <slava> :) :) :)
10:37:55 <kc5tja> $7A instead of $79.
10:38:27 <tathi> which is what?
10:40:14 <kc5tja> Apparently, the $7A opcode tests *PARITY* and not sign.
10:40:34 <kc5tja> All is well again.
10:41:01 <tathi> oh maan
10:41:07 <tathi> well, at least you found it
10:41:23 <kc5tja> Yep
10:41:48 * tathi goes to try mowing the lawn again.
10:41:52 <slava> I440r, did you fix it?
10:42:36 <kc5tja> I knew my word parsing code wasn't b0rked.  :)
10:42:45 <kc5tja> So now it can properly parse words.
10:42:58 <I440r_> kc5 then why does...
10:43:00 <I440r_>  mov eax, 1
10:43:03 <I440r_>  sub eax, 2
10:43:03 <kc5tja> bash-2.05b$ ./a.out
10:43:03 <kc5tja> FTS/Forth System 2 for Linux
10:43:03 <kc5tja> Copyright (c) 1989-2004 Falvo Technical Solutions
10:43:03 <kc5tja> Type `license' to see license.
10:43:09 <I440r_> NOT generate any sign flag ?
10:43:11 <kc5tja> ok> This is a test.This
10:43:11 <kc5tja> is
10:43:11 <kc5tja> a
10:43:11 <kc5tja> test.
10:43:21 <kc5tja> It does.
10:43:50 <slava> kc5tja, :)
10:44:27 <kc5tja> Parsing Forth words: 7 lines of code.
10:45:04 <kc5tja> I was pretty sure that I was going insane yesterday evening.  Good grief.
10:45:12 <I440r_> lol
10:45:24 <I440r_> thats how isforth devel is all the time with fbsd
10:45:40 <I440r_> first a syscall doesnt behave as its supposed to. then another. then another
10:45:50 <slava> heh
10:45:53 <I440r_> and this last one... the syscall was 100% correct!
10:45:55 <slava> libc4life
10:46:07 <kc5tja> I mean, I've coded the x86 platform since the 8088 -- I knew that the SUB instruction properly set the flags.
10:46:09 <I440r_> freekin bsd needing a bss in order to do a break
10:46:10 <I440r_> duh!
10:46:58 <kc5tja> I440r_: That's good to know -- I'll be sure to remember that when I get to that point in the Linux version of FTS/Forth.
10:47:30 <I440r_> linux was easy - linux will do a sys brk even if you have no bss - linux is a little smarter - fbsd is a bit more anally retentive about it
10:47:41 <slava> I440r, its a "real unix" :)
10:47:53 <I440r_> lol
10:48:16 <I440r_> yea its a "lets see how many times we can fsck up a coders sanity" unix
10:48:41 <I440r_> lol
10:49:02 <Herkamire> kc5tja: :) that explains why you were getting different results from -1 than -2
10:49:13 <kc5tja> Yes.
10:50:11 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
10:50:27 <kc5tja> Damn -- listening to Rochmonov right now, and this music is just plain awesome.  There are no other words that can describe it.
10:50:32 --- join: futureX (identity@d197.wildapache.net) joined #forth
10:50:38 <kc5tja> Rochmonanov rather.
10:51:41 <futureX> Anyone know why forth doesn't fare so well on this page: http://www.bagley.org/~doug/shootout/craps.shtml ?
10:53:20 <I440r_> bigforth
10:53:21 <I440r_> nuff sed
10:53:39 <slava> i thought bigforth was fast
10:54:02 <I440r_> they should do that test using marcel hendrix's iforth
10:54:03 <futureX> bigforth...hmm
10:54:27 <I440r_> that would prolly put the score closer to gcc
10:54:32 <I440r_> maybe even better than
10:54:34 <futureX> But forth can't get faster than C, would that be accurate?
10:54:44 <kc5tja> It actually fares rather well, considering how gforth actually compiles its code.
10:54:46 <I440r_> it depends on the forth
10:55:28 <slava> futureX, why not?
10:55:43 <slava> esp. if your C is gcc :)
10:55:48 <kc5tja> futureX: It depends on the Forth, of course, but it is also true that most CPUs today are optimized for C performance, not Forth performance.  Therefore, it is a good approximation at the big-picture level.
10:55:56 <I440r_> futurex how long does it take you to compile yoru linux kernel ?
10:56:01 <futureX> I don't know. You guys are the forth programmers. I'm just asking questions :)
10:56:11 <I440r_> if linux were coded in forth my compiler would compile it in 3 seconds and it would be a hell of alot smaller
10:56:17 <I440r_> :)
10:56:26 <I440r_> futureX, is that your page?
10:56:30 <futureX> no
10:56:33 <I440r_> ok
10:56:33 <kc5tja> However, Forth's claim to speed is purely coincidental.  It was never really intended to run as fast as it actually does.  It's purely a side effect.
10:56:59 <kc5tja> Forth's beauty is how much more productive a programmer becomes (once you're comfortable in your Forth environment that is) versus, say, a C programmer.
10:57:04 <futureX> I just saw that page and I was wondering, because I had the impression that forth is more low-level than C.
10:57:06 <kc5tja> Much of this has to do with the succinctity of Forth source.
10:57:54 <I440r_> it is and it isnt
10:57:54 <kc5tja> futureX: It's about the same level as C.  But its execution model is markedly different.
10:58:00 <futureX> Here's bigforth: http://www.bagley.org/~doug/shootout/lang/bigforth/
10:58:02 <I440r_> forth can be lower level than c and higher level than ada
10:58:23 <slava> ada is not high level its just crap
10:58:24 <futureX> A lot of red numbers.
10:58:35 <kc5tja> slava: Ada has a LOT of NICE features in it.
10:58:43 <I440r_> lol
10:58:54 <I440r_> kc5 thats the problem with ada
10:58:56 <I440r_> featureitis
10:59:33 <slava> lisp pulls of featureitis with style
10:59:34 <kc5tja> I440r_: In a sense, yes.  But they were put there for a reason.  Ada strongly favors formal methods for software verification, and it CAN statically determine the correctness of a very large number of programs you throw at it.
11:00:01 <kc5tja> slava: More accurately, Lisp implements a framework with which all features employ and interoperate with.
11:00:21 <slava> kc5tja, well i meant CL in particular
11:00:45 <slava> a CL implementation is in the order of 100,000 lines of code at least, just to implement all the fetaures in the spec.
11:00:48 <slava> but its a very nice language.
11:01:07 <kc5tja> slava: That's why I like Scheme.
11:01:15 <kc5tja> It's nice and small, well, compared to Lisp.
11:01:20 <kc5tja> But Lisp is the more practical language.
11:01:51 <slava> yes.
11:02:00 <slava> personally i don't like scheme lumping functions and variables in one namespace.
11:06:26 <Herkamire> lines of code is so arbitrary
11:06:49 <Herkamire> the gforth fib code is all spread out with like 1 to three words per line, and another languages implementation is 3 long long lines
11:07:18 <slava> Herkamire, yes but if something is written in 100,000 of dense lisp code, there's no way to squash it to, say, 5000 lines
11:09:34 <kc5tja> Suddenly, I don't feel so bad about ANSI Forth anymore.  >:)
11:10:58 <Herkamire> just seems silly
11:10:59 --- quit: tgunr (Remote closed the connection)
11:11:09 <Herkamire> they could have done bytes of code (excluding comments and indentation
11:11:38 <Herkamire> what's ackerman's function
11:11:39 <Herkamire> ?
11:14:39 <I440r_> its generates huge numberes very quickly
11:14:58 <I440r_> i can see NO practical use for it what so ever
11:14:59 <slava> as quickly as 1 1000 << ?
11:15:04 <I440r_> tho - im told there are some
11:15:15 <slava>   ( 0 ) 1 1000 shift .
11:15:16 <slava> 10715086071862673209484250490600018105614048117055336074437503883703510511249361224931983788156958581275946729175531468251871452856923140435984577574698574803934567774824230985421074605062371141877954182153046474983581941267398767559165543946077062914571196477686542167660429831652624386837205668069376
11:15:22 <I440r_> lol
11:16:00 * slava calculates 1 100000 shift and fills up screen with random digits
11:16:09 <I440r_> lol
11:16:12 <slava> whoa its like 10 pages
11:16:15 <kc5tja> http://www.cs.washington.edu/homes/klee/misc/slashdot.html#reasons
11:19:21 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
11:20:55 <kc5tja> Man, I couldn't agree more with that article.
11:21:04 * kc5tja will post a link to that site on my page later today.
11:34:33 <Herkamire> anybody know how deep ackermann recurses for ackerman(3,8) ?
11:35:30 <Robert> No idea, maybe wikipedia will tell you?
11:37:07 --- quit: mur_ (Remote closed the connection)
11:37:43 --- join: mur (~mur@smtp.uiah.fi) joined #forth
11:37:55 --- quit: tgunr (Remote closed the connection)
11:38:30 <Robert> I like this reason the best: #3. Gullible editorial staff continues to post links to bogus pseudoscience articles by crackpots.
11:39:17 <kc5tja> I like #3 and #7.
11:39:26 <kc5tja> But they're ALL damn accurat.e
11:39:28 <kc5tja> accurate.
11:39:31 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
11:43:22 <Robert> I'm not a slashdotter, so I wouldn't know.
11:44:40 <Robert> There's a magazine over here specializing in crackpot theories. Sad part of the story is that it's the most popular "popular science" magazine in the country.
11:44:45 <Robert> So I guess it's a general trend :)
11:53:24 <Herkamire> ahh, ok, so the result of the ackerman thing is how deep it has to recurse
12:13:49 <Herkamire> recursion is fun
12:16:08 --- quit: tgunr (Remote closed the connection)
12:16:22 <Robert> Yes :)
12:17:48 <Herkamire> just wrote this:   : string++ b-@ 1+ dup b!a $3a <; $30 b!a string++ ;
12:18:02 <Herkamire> takes a pointer just past the end of a string in the 
12:18:06 <Herkamire> Address register
12:18:12 <slava> hi
12:18:24 <Herkamire> the string has a number on the end, and this word adds one to it
12:18:29 <Herkamire> hi slava
12:19:10 <Herkamire> I use it to generate a new filename for herkforth on bootstrapping
12:27:03 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
12:27:21 <kc5tja> Herkamire: I found out how Linux passes commandline arguments to processes for x86 versions of Linux.
12:27:49 <kc5tja> It passes the number of parameters and a character array on the stack.
12:28:10 <kc5tja> It's up to the process to pack this array into argv[].
12:28:21 <kc5tja> Also anything beyond the last argv[] is the environment variables.
12:29:18 <tathi> kc5tja: same on PPC, if I understand you right.
12:29:37 --- join: madwork_ (~madgarden@derby.metrics.com) joined #forth
12:29:40 <kc5tja> tathi: Herkamire says that Linux passes CLI arguments in CPU registers.
12:29:46 <tathi> nope.
12:30:31 <tathi> on entry to your program, you have argc on top of the stack, followed by pointers to the command-line arguments.
12:30:41 <tathi> followed by 0, then pointers to environment strings.
12:30:47 <kc5tja> Nope.
12:30:50 <kc5tja> Not what I'm talking about.
12:30:53 <kc5tja> That's how C does it.
12:31:09 <kc5tja> I'm talking about how the *kernel* interfaces to the process (assume C hasn't been invented)
12:31:13 <tathi> that's how the linux kernel does it on PPC
12:31:19 <tathi> kc5tja: I never ever use C anymore.
12:31:23 <tathi> never did, much.
12:31:35 <kc5tja> On x86, there is precisely one argc, and precisely one pointer to the parameters.
12:31:39 <kc5tja> The parameters are kept in a byte array.
12:31:40 <tathi> the strings themselves are actually allocated on the stack to.
12:31:44 <tathi> too
12:32:10 <tathi> so the strings are just there, one after the other?
12:32:15 <kc5tja> Yep
12:32:20 <kc5tja> Separated by $00 bytes.
12:32:57 <tathi> gotcha.
12:33:04 <kc5tja> You have to build the argv[] array manually in x86-Linux apparently.
12:33:11 <tathi> CLI args, followed by environment?
12:33:16 <kc5tja> Yes
12:33:44 <tathi> interesting.
12:34:17 <tathi> PPC Linux puts the strings on the stack like that, but then puts pointers to everything on the stack as well.
12:34:34 <tathi> funny that x86 doesn't.
12:40:58 --- quit: tgunr (Remote closed the connection)
12:41:53 <tathi> http://qualdan.com/misc/ppc-linux-stack
12:43:21 <slava> the freebsd java port is pathetic:
12:43:22 <slava> > Math.pow(-1.1,3)
12:43:22 <slava> 1.3310000000000004
12:45:41 <I440r_> #1. Two words: Jon Katz.    <-- is this guy in any way at all related to the pkzip guy ?
12:45:49 <I440r_> phill ?
12:46:37 <slava> i mean the linux jdk gives the correct result how could they possibly fuck up pow() while porting
12:46:39 <slava> i mean FUCK
12:46:44 <slava> retards
12:47:00 <I440r_> lol
12:47:05 <I440r_> itsd fbsd
12:47:07 <I440r_> its evil
12:47:10 <I440r_> hehe
12:48:14 --- join: mur_ (~mur@kyberias.uiah.fi) joined #forth
12:49:24 <kc5tja> I440r_: No.  PKZip is Philip Katz.
12:50:36 <I440r_> yea
12:50:38 <I440r_> same last name
12:51:02 <I440r_> what did phil katz die of ?
12:52:09 <I440r_> On April 14, 2000, Phil Katz was found dead in a Milwaukee motel room. An empty bottle of peppermint schnapps was still clutched in  his hand, and five more empties were scattered about the room. It was a miserable, lonely end for a man who had created a product now in use by millions of people;
12:52:29 <slava> are you serious?
12:52:32 <I440r_> yes
12:52:33 <slava> peppermint schnapps overdose?
12:52:38 <I440r_> heh
12:58:15 <Herkamire> that's too bad x86 doesn't give you pinters for the argv array
12:59:22 --- quit: mur (Read error: 110 (Connection timed out))
12:59:53 <kc5tja> Herkamire: Not for me; I'll just reference the byte array directly in Forth.  :)
13:00:02 <kc5tja> I consider NULLs to be spaces in ParseWord anyway.
13:02:03 <kc5tja> slava: In case you haven't noticed, many extremely bright people were found dead as a direct result of corporate strong-arming.  Gary Kildall, author of both CP/M and of GEM, commited suicide because of Microsoft.
13:03:19 <I440r_> HE was a fool though. when IBM went to go see him about a new OS he was too busy playing golf
13:03:25 <I440r_> he dropped the ball
13:03:29 <I440r_> unfortunatly :(
13:03:34 <kc5tja> HE HAD NO IDEA IBM WAS TO SHOW THAT DAY.
13:03:39 <kc5tja> NONE.
13:03:46 <kc5tja> IBM came completely unannounced.
13:05:30 <kc5tja> anyway, I'm off to get some food.  Be back in a bit.
13:14:06 * TomasuDlrrp is away: for real this time
13:23:33 --- quit: mur_ (Remote closed the connection)
13:24:08 --- join: mur (~mur@smtp.uiah.fi) joined #forth
13:26:16 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
13:26:51 <Herkamire> heh, I'd say he would be a fool to never play golf on the off chance that IBM might show up unannounced
13:27:13 <Herkamire> I assume that large corperations usually call ahead...
13:32:17 --- quit: tgunr (Read error: 104 (Connection reset by peer))
13:34:10 --- join: tgunr (~davec@vsat-148-63-4-107.c001.g4.mrt.starband.net) joined #forth
14:00:30 --- quit: tgunr ("Leaving")
14:25:24 <kc5tja> back
14:25:56 <kc5tja> Yeah, me too.
14:25:57 <kc5tja> :)
14:26:00 <kc5tja> I know I certainly do.
14:28:11 --- join: FlamingRain (Ecoder@c-24-129-95-254.se.client2.attbi.com) joined #forth
14:45:59 --- quit: FlamingRain ("Leaving")
14:46:19 <Herkamire> heh, I really need to make an undo feature in herkforth
14:46:31 <Herkamire> sometimes weird things happen if you hit the wrong key
14:49:20 --- quit: mur (Remote closed the connection)
14:49:52 --- join: mur (~mur@mgw2.uiah.fi) joined #forth
15:04:41 <kc5tja> Herkamire: Yeah, I've more than once accidentally damaged a block of source that I wish I could just type "UNDO" and have it work.  :)
15:04:57 <slava> programming without undo is painful
15:05:09 <kc5tja> slava: Puts hair on your chest.  :)
15:05:11 <slava> i have 1000 levels of undo/redo, as well as a kill ring for bringing back deleted text
15:05:54 <slava> kc5tja, :)
15:06:09 <kc5tja> All the more reason to have a humane editing environment.
15:06:36 * kc5tja is trying to hack the ALD debugger so it prints PC-relative addresses with their proper target addresses
15:14:57 <Herkamire> slava: ahh, but you didn't write your editor :)
15:15:07 <slava> Herkamire, but i did!
15:15:30 <Herkamire> slava: you did? I thought you were just working on a plugin for jedit
15:15:44 <slava> no, i wrote most of it
15:15:53 <slava> check jedit.org if you don't believe me :)
15:16:14 <Herkamire> so, have you been working on jedit for a long time before you started on factor?
15:16:19 <slava> yea
15:16:24 <slava> it was the first 'real' app i wrote
15:16:30 <Herkamire> cool
15:16:34 <slava> before jedit, i made a calculator applet
15:16:41 <slava> its not even online anymore
15:17:46 <Herkamire> I have an undelete
15:17:54 <Herkamire> but no general purpose undo
15:17:54 <slava> is that like undo?
15:17:58 <slava> oh
15:18:00 <Herkamire> only for deletions
15:18:19 <Herkamire> e.g. if you enter a word, then hit undelete, it will paste in the last thing you deleted
15:18:33 <slava> undo is pretty easy, really. especially with OOP -- have an abstract Edit class with undo/redo methods, and Insert and Remove subclasses. make instances of this, etc.
15:18:41 <slava> and in forth its similar too
15:18:50 <slava> you could create/does> some words i gues
15:18:55 <slava> does herkforth have create/does>?
15:18:59 <Herkamire> no
15:22:19 <kc5tja> Let's be a little bit more abstract here.  CREATE/DOES> is pretty limiting.
15:22:25 <kc5tja> What you want is a "Command Pattern."
15:24:07 <Herkamire> I think my undo will be implemented as a circular stack containing pairs of words.
15:24:21 --- join: warpzero (~warpzero@mi031.dn189.umontana.edu) joined #forth
15:24:25 <Herkamire> each pair will contain a word of data, and an execution token
15:24:44 <Herkamire> undo will just pop these onto the stack and execute it
15:24:51 <slava> cool.
15:24:53 <slava> one other minor thing jedit does with undo is 'compound undo'. eg, replace-all is wrapped in beginCompoundEdit()/endCompoundEdit() calls, so it can be undone as one step.
15:25:00 <slava> this is just a nested undo list
15:25:40 <Herkamire> my design does not handle all kinds of editos.
15:25:42 <Herkamire> idets
15:25:45 <Herkamire> edits
15:25:49 <Herkamire> sheesh
15:25:52 <slava> :)
15:26:17 <Herkamire> eg if there is more data than fits in 32 bits, (such as a rename)
15:26:44 <Herkamire> ahh rain :)
15:26:48 <Herkamire> maybe it will finally cool off a bit
15:27:25 <Herkamire> perhaps I could have variable length data on the undo stack
15:27:45 <Herkamire> there'd just have to be a count or something in there so I could tell I have the whole thing still.
15:28:23 <Herkamire> whatever. I'll tackle that later
15:32:03 <Herkamire> hmm... I can't have an undo that works for executing words
15:32:35 <Herkamire> eg, if they type drop and hit enter
15:33:03 <slava> when i make a stepping debugger for factor, it will have rewindable execution for code that doesn't side-effect heap objects
15:33:15 <slava> so you'll be able to 'undo' a drop (in the debugger, not the general case obviously).
15:33:26 <slava> but not undo storing into an array. that's too much
15:33:34 --- quit: proteusguy ("Leaving")
15:34:03 <Herkamire> ok
15:34:59 <Herkamire> I don't need it to do everything. I just want a simple way to undo things I did by hitting the wrong key.
15:35:12 <Herkamire> this would be especially useful and comforting for people who don't know herkforth as well as I do.
15:35:18 <slava> yup.
15:35:27 <Herkamire> eg, if you don't know what happened, you probably won't be able to figure out how to put it back
15:37:36 <Herkamire> right now I'm working on code to change the color of existing words.
15:37:57 <Herkamire> I've got the simple green/yellow (execute/compile) token change
15:39:20 <Herkamire> I think I'm now going to add support for changing the type of a word (eg constant <--> variable) by changing the color of the defining word.
15:40:06 <slava> it annoys me that the java crowd gets off so much on IDEs that generate code with templates.
15:40:34 <Herkamire> but I don't think I will go as far as making it so you can change a word from a normal word to eg a lit (which would involve changing all source tokens that reference that dictionary entry from eg compile-call color to lit color
15:40:49 <slava> brb dinner
15:41:03 <Herkamire> most anything having anything to do with java annoys me :)
15:46:42 <Herkamire> I can only think of one java program that I've been at all happy with
15:46:42 --- quit: madwork_ (Read error: 104 (Connection reset by peer))
15:48:18 <Herkamire> on the occation when they actually work at all, they are always slow, and almost allways have horrible horrible user interfaces
15:50:27 <tathi> have you actually tried java progams other than freenet?
15:54:14 <Herkamire> yeah
15:54:52 <Herkamire> magnus is the best juggling simulator I could find that works on my computer
15:56:31 --- 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.")
15:57:30 <Herkamire> I've tried a bunch of java stuff over the years
16:02:35 --- join: mur_ (~mur@uiah.fi) joined #forth
16:08:14 <Herkamire> ok, enough herkforth for now
16:08:37 <slava> time for... herkC!
16:08:47 <Herkamire> riiight
16:08:55 <slava> herkPL/I
16:09:11 <Herkamire> time for work
16:12:34 <Herkamire> heh, making calls goes sooo fast when nobody answers
16:12:54 --- quit: mur (Read error: 110 (Connection timed out))
16:22:42 <I440r_> lol pl/1 lol
16:22:50 <I440r_> that cant even get true/false right!
16:23:02 <I440r_> 0 = false, 1 = true, 2 = false, 3 = true
16:23:03 <I440r_> lol
16:23:08 <slava> hehehehe
16:23:09 <slava> lol
16:23:11 <slava> damn
16:23:30 <tathi> that's a neat system.
16:23:42 <slava> : not 1 + ; ?
16:23:51 <I440r_> er : not 1+ ;
16:23:52 <I440r_> lol
16:23:53 <slava> : or + ;
16:24:04 <I440r_> no because false + flase is still false :)
16:24:10 <slava> : and * ;
16:24:23 <tathi> I440r_: false or false is still false too
16:27:34 --- quit: tathi ("leaving")
16:28:46 <I440r_> :)
16:30:43 --- join: TheBlueWizard (TheBlueWiz@pc33dn1d.ppp.fcc.net) joined #forth
16:30:43 --- mode: ChanServ set +o TheBlueWizard
16:40:03 <slava> TheBlueWizard 
16:43:08 <TheBlueWizard> slava hiya
16:46:43 <kc5tja> re TheBlueWizard 
16:47:04 <TheBlueWizard> kc5tja hiya
16:48:47 <kc5tja> PL/1 is actually a neat language.  
16:48:56 <kc5tja> I really wish people would just stop blindly dismissing languages.  *sighs*
16:49:12 <I440r_> i wasnt dissin it actually
16:49:35 <I440r_> but the 0/1/2/3 thing is kinda humorous :)
16:49:43 <kc5tja> Maybe not (sure came across that way though), but there are too many who do.
16:50:19 <kc5tja> You know, if you logically extend it in the reverse direction, it fits with the -1=true, 0=false system.  :)
16:50:30 <I440r_> :)
16:51:17 <TheBlueWizard> hmm?
16:51:17 <Herkamire> -1 is odd whether you look at the bits as signed or unsigned
16:52:50 <kc5tja> I440r_: Hey, I made some changes to my ALD disassembler -- PC-relative addressing modes now display the target address in parenthesis.  However, the address is still "wrong" because the data buffer the disassembler library uses isn't loaded at the same address as the executable being debugged apparently.
16:53:09 * kc5tja is going to try and get some information from ALD's maintainer to see if he knows of a way around this.
16:53:29 <I440r_> is he "maintaining" it still ?>
16:53:38 <I440r_> if he would just finish the damned thing i would be much happier
16:53:51 <kc5tja> Last site update was 2004, so yes.
16:53:53 <I440r_> he was upset when i told him i used gdb to debug isforths kernel
16:53:57 <kc5tja> (not too long ago either)
16:54:06 <I440r_> thats the web page - not the util
16:54:18 <kc5tja> I have version 0.1.6
16:54:33 <I440r_> thats the latest
16:54:49 <I440r_> ald.sf.net says so anyway
16:55:01 <kc5tja> Well, I'm sure that if people actually contributed, hint hint, it might get done a bit faster.  :)
16:55:01 <I440r_> that was the version i looked at when i started isforth
16:55:32 <I440r_> i cant read his sources. i have almost no ability to read c sources
16:55:36 <I440r_> its like write only
16:55:52 <kc5tja> Most recent version: 0.1.6 (23 July 2004) 
16:56:09 <kc5tja> His C sources are not write only.
16:56:13 <kc5tja> You're just making excuses.
16:56:37 <I440r_> no - that might have a new date on it but its the same version i was looking at 3 years ago
16:56:54 <I440r_> and i cannot read c sources - or it takes a HUGE effort on my part
16:57:17 <slava> i can't read assembly ;)
16:57:17 <I440r_> any language that takes that ammount of effort to read isnt worth reading. i can read any assembler no matter how badly coded
16:57:20 <I440r_> same as forth
16:57:25 <slava> bullshit
16:57:36 <slava> i can show you some assembly that you have no idea what it does, i bet
16:57:48 <I440r_> ive never seen any assembler code i couldnt read and understand with just a little effort
16:58:06 <I440r_> ive seen single c functions i could not understand after hours of study
16:58:21 <I440r_> i can write c. if i write it i understand it 10 years later
16:58:45 <I440r_> oh. part of the reason i never used ald is because he never finished the curses mode
16:59:40 <I440r_> err this has a GTK user interface
16:59:50 <I440r_> man i KNOW i was looking at 0.1.6 way back but it didnt have gtk
17:03:11 --- join: slava_ (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth
17:03:13 <kc5tja> That's funny, because C is just a re-organized PDP/7 assembly language.
17:03:15 <kc5tja> Literally.
17:03:35 <I440r_> c is just ALL WRONG!!!!
17:03:41 <slava_> I440r, no its not
17:03:45 <I440r_> my mind shuts it out
17:03:49 <kc5tja> I can read C without problems.  And ALD's source is not in any way write-only.  There is a methodical convention used throughout the source -- something I don't often find in C sources.
17:03:49 --- quit: slava (Nick collision from services.)
17:03:53 --- nick: slava_ -> slava
17:04:00 <I440r_> it takes a HUGE effort for me to understand c sources
17:04:11 <slava> I440r, "for me" are the key words there
17:04:22 <I440r_> slava lol
17:05:04 <I440r_>     Print(ws,
17:05:04 <I440r_>           P_ERROR,
17:05:04 <I440r_>           MSG_PTERR,
17:05:04 <I440r_>           strerror(errno));
17:05:04 <I440r_>  
17:05:06 <I440r_> like
17:05:12 <I440r_> what the fuck is with that formatting
17:05:12 <I440r_> its
17:05:13 <I440r_> like
17:05:14 <I440r_> if
17:05:24 <I440r_> i 
17:05:24 <I440r_> talk
17:05:24 <I440r_> to
17:05:24 <I440r_> you
17:05:41 <slava> what does that have to do with C?
17:05:41 <I440r_> like
17:05:41 <I440r_> this
17:05:48 <I440r_> everything i see C coders do my mind just screams "god i hate this" at me
17:05:48 <kc5tja> I440r_: That's how your Forth code looks though!!
17:06:08 <slava> I440r, C is the only choice for many applications
17:06:11 <slava> not everybody wants to rewrite assembler for each cpu
17:06:23 <I440r_> you cant disassociate c from the way people write it
17:06:24 <I440r_> kc5 no
17:09:38 <I440r_> i also dont like the way im seeing people align ( ) stuff
17:09:43 <I440r_>  if ( .....
17:09:50 <I440r_>      )
17:09:57 <I440r_> erm undent that by 1 heh
17:10:03 <slava> I440r, hush now
17:10:12 <I440r_> it just fragments up the display and scatters everything all over
17:10:18 <I440r_> making a huge visual cluster fuck heh
17:10:40 <I440r_> if (x < 10) || (y > 5))
17:10:48 <I440r_> that looks much better. MUCH easier to read
17:10:58 <kc5tja> Then I never, ever, ever want to hear you complain about how it's "improper" to not divide the screen into two rough columns, one for comments, one for source.  Because, well, it just creates one big visual cluster fuck.
17:11:28 <slava> i managed to reuse this file written many years ago and now i understand it almost fully, if it was assembly i'd have no chance:
17:11:28 <I440r_> does it fuck
17:11:29 <slava> http://cvs.sourceforge.net/viewcvs.py/factor/Factor/Factor/native/s48_bignum.c?view=markup
17:11:31 <I440r_> code over here              comments over here
17:11:40 <I440r_> you can scan either one withotu the other one interfearing
17:12:16 <I440r_> slava - i would never ever ever be able to read your sources. not in a million years
17:12:23 <slava> that file is not my sources
17:12:25 <kc5tja> I440r_: Sorry, sir, but that is horridly interfering.  The code already tells me what it does.  I am not paying my electric bill to waste half my monitor's screen real-estate.
17:12:49 <slava> that's my point, it was written by others, in C, and i understand how it works
17:12:51 <I440r_>     ((BIGNUM_ZERO_P (x))
17:12:51 <I440r_>      ? (BIGNUM_ZERO_P (y))
17:12:51 <I440r_>      : ((! (BIGNUM_ZERO_P (y)))
17:12:51 <I440r_> && ((BIGNUM_NEGATIVE_P (x))
17:12:51 <I440r_>     ? (BIGNUM_NEGATIVE_P (y))
17:12:51 <I440r_>     : (! (BIGNUM_NEGATIVE_P (y))))
17:12:53 <I440r_> && (bignum_equal_p_unsigned (x, y))));
17:12:56 <kc5tja> Unless you have something really nifty to tell me, I don't want to see comments.  Period.
17:12:59 <I440r_> err thats obfuscated
17:13:01 <slava> no its not
17:13:13 <slava> use an editor that matches brackets and its quite clear what's going on
17:13:20 <I440r_> too many ? : statements in one line lol
17:13:23 <kc5tja> I can read that even WITHOUT such an editor.
17:13:24 <kc5tja> That's easy.
17:13:30 <I440r_> i would have to study taht for atleast a couple of hours to understand it
17:13:37 <slava> that's your problem not C's
17:13:59 <kc5tja> if X is zero, then return whether or not Y is zero, otherwise, return whether or not Y *isn't* zero.  Likewise for the negative senses.
17:13:59 <I440r_> rewrite it in assembler and i doubt i would have any problems understanding it
17:14:42 <slava> its 4000 lines of assembly, after gcc -S
17:14:54 <slava> that would be quite hard to understand
17:14:58 <I440r_> slava you have one function here with at least 8 different exit points lol!
17:15:07 <slava> I440r, and?
17:15:14 <I440r_> slava asm output from c compiler is NOT assembler
17:15:17 <I440r_> its c lol
17:15:23 <TheBlueWizard> I for one find the "? :" code I440r_ just posted fairly clear (nice lineup and indentation)
17:15:46 <slava> my only problem with that code is it was written in a slightly older style of C with char* instead of void* etc
17:15:47 <I440r_> slava do you follow the GNU coding standards 100% ?
17:15:55 <slava> I440r, no, i write with 8 spaces per tab
17:15:56 <I440r_> your using bnu bracing here too
17:16:05 <slava> I440r, i didn't want to mess with that file's formatting
17:16:12 <I440r_> not asking you to
17:16:14 <slava> i want to be able to import new versions from scheme48 with minimal changes
17:16:30 <slava> if you look at the other files in http://cvs.sourceforge.net/viewcvs.py/factor/Factor/Factor/native/ you'll see i use linux kernel coding style
17:16:57 <I440r_> linux kernel doesnt indent the braces from the statement and then indent the body from the braces
17:17:10 <slava> and neither do I
17:17:11 <I440r_> kernel doesnt indent the braces at all. it indents the body 
17:17:13 <slava> only the bignums code does that
17:17:22 <I440r_> thats not yours ?
17:17:23 <slava> because i don't want to redo the formatting each time the scheme48 people update it
17:17:25 <slava> no
17:17:29 <I440r_> oh
17:17:30 <slava> s48_bignum.c is scheme48 code
17:17:33 <slava> i said that like 5 times
17:17:34 <kc5tja> GNU's bracing policy is this:
17:17:35 <kc5tja> void
17:17:37 <kc5tja> foo(..)
17:17:38 <kc5tja> {
17:17:39 <kc5tja>   {
17:17:42 <kc5tja>   }
17:17:44 <kc5tja> }
17:18:39 <I440r_> thats good
17:18:44 <I440r_> do you do the...
17:18:47 <slava> i prefer \t to spaces
17:18:50 <I440r_> if ( )  {
17:18:51 <I440r_> }
17:18:52 <I440r_>  ?
17:18:53 <I440r_> ick
17:19:03 <slava> i do
17:19:04 <slava> if(foo)
17:19:05 <slava> {
17:19:09 <slava>         aaa
17:19:09 <slava> }
17:19:20 <slava> and i avoid nesting more than 3 levels or so
17:19:20 <I440r_> linus uses the k&r bracing
17:19:31 <kc5tja> No, I do this:
17:19:35 <kc5tja> if( ... )
17:19:36 <kc5tja> {
17:19:37 <kc5tja>   ...
17:19:37 <kc5tja> }
17:19:46 <I440r_> "Becahse as all right thinking people know K&R are right and k&r are right" he says
17:19:48 <I440r_> ick
17:19:56 <I440r_> thats how i brace if statements
17:20:02 <I440r_> i also brace ALL if statements
17:20:08 <I440r_> if (x=1)
17:20:11 <I440r_>   for (...)
17:20:16 <I440r_>    switch (foo)
17:20:19 <I440r_> fucking stupid
17:20:31 <I440r_> lol
17:20:45 <I440r_> i notice you dont brace your slingle line If's slava
17:20:46 <kc5tja> If the context is trivial, I won't brace, but I normally do brace everything.
17:20:47 <slava> i don't brace one line if's
17:20:50 <I440r_> i bet kc doesnt either
17:20:54 <kc5tja> I also explicitly brace if nested conditionals are used too.
17:20:55 <slava> I440r, yes
17:21:01 <slava> I440r, i avoid what they do in the s48 code though:
17:21:02 <slava> if(foo)
17:21:05 <slava>   while(...)
17:21:07 <I440r_> slava single line if's can contain single line statements
17:21:12 <I440r_> taht contain single line statements
17:21:15 <I440r_> ad infinitum
17:21:19 <slava> no, i don't do that either
17:21:24 <slava> if(foo)
17:21:27 <slava>         bar();
17:21:33 <slava> but for nested if's i brace
17:21:36 <kc5tja> There is no need to brace a single-line if statement.
17:21:38 <I440r_> do you do  
17:21:42 <I440r_> while (1);
17:21:42 <I440r_> or
17:21:46 <I440r_>  while (1)
17:21:47 <I440r_>   ;
17:21:57 <kc5tja> None of the above.
17:22:00 <slava> why would i have ; after while?
17:22:02 <kc5tja> #define forever while(1)
17:22:03 <kc5tja> ...
17:22:05 <kc5tja> forever
17:22:05 <slava> for(;;)
17:22:05 <kc5tja> {
17:22:07 <kc5tja>   ...
17:22:09 <kc5tja> }
17:22:18 <I440r_> kc5 slava is that as efficient as while 1 ?
17:22:20 <I440r_> same ?
17:22:23 <kc5tja> Same.
17:22:24 <slava> who cares?
17:22:38 <slava> i use a profiler to tell me where the bottlenecks are
17:22:47 <I440r_> "ineficiency is ok" is part of my problem with c
17:22:56 * TheBlueWizard hides in a little quiet corner
17:22:56 <slava> that's not what i said
17:23:02 <slava> i optimize code that is important to performance
17:23:07 <slava> if you have a statement that's executed once only
17:23:14 <slava> it doesn't matter if it takes 1 micro or milli second
17:23:47 * kc5tja agrees completely with Slava.
17:23:50 <I440r_> ill build a profiler into isforth eventually
17:23:56 <I440r_> kc5 yes. i didnt disagree
17:24:03 <I440r_> but the one thing i wont do is
17:24:07 <I440r_> for (;;) 
17:24:10 <kc5tja> FTS/Forth's code production is HORRIBLE as far as even naive native code Forths are concerned.  But it's still "fast enough" for me.
17:24:13 <slava> what the hell is wrong with for(;;)?
17:24:15 <I440r_> in one place and then a while (1) in another
17:24:18 <I440r_> slava nothing
17:24:19 <slava> oh sure
17:24:20 <slava> be consistent
17:24:39 <slava> i never do while(true) because back in 1998 java was actually faster at for(;;) than while(true) since it would check if true was true at each iteration lol
17:24:41 <slava> and the habit stuck
17:24:47 <slava> even in C and modern java
17:24:47 <kc5tja> I440r_: That's why you spend an hour or two pouring over the code you're working with to get a feel for the coding conventions used.  If there are none, make one.
17:24:48 <I440r_> slava ive used BOTH before - i was just asking if the for statement was as/more/less efficent than the  while
17:25:30 <I440r_> kc5 you can spend a lifetime reformatting other peoples sources
17:25:32 <I440r_> lol
17:25:37 <kc5tja> I440r_: Then do it.
17:25:42 <slava> I440r, or use an automated tool
17:25:44 <I440r_> no - i chose to NOT do
17:25:52 <kc5tja> Then that's not our problem.
17:25:54 <I440r_> i have no interest in touching any c
17:26:00 <slava> then don't bitch about it
17:26:01 <I440r_> its not MY problem either
17:26:02 <kc5tja> Then don't.
17:26:04 <slava> you're just appearing ignorant right now
17:26:09 <slava> i don't bash x86 asm beacuse i don't know it
17:26:17 <I440r_> err. i didnt say i dont know c
17:26:25 <I440r_> i said that everything i learn about it makes me hate it more and more
17:26:33 <slava> then don't learn about it and don't bitch about it
17:26:34 <I440r_> specially the way people format it
17:26:39 <slava> i don't spend time learning COBOL
17:26:58 <I440r_> slava this is both a free country and an open channel :P
17:27:15 <slava> I440r, its true, but bitching about what amounts to personal preference is kind of futile
17:27:26 <slava> if you have legitimate concerns about C, voice them, but if its formatting... well
17:27:28 <I440r_> i was asked. i answered
17:27:34 <kc5tja> But it IS your problem.  You're slamming ALD because it's "not maintained," but you won't contribute the effort to submit patches yourself to make it better for your needs, because it's written in write-only C.  Whatever.  I'm 90% done with the code hack, and I've e-mailed the author asking for help in resolving one last technical issue on what needs to be done to fix it (which I would have done if it were written in Forth too).
17:27:40 <I440r_> slava. formatting is only PART of it
17:27:56 <I440r_> kct ok. YOU submit patches to that
17:28:00 <I440r_> and every other util you use ?
17:28:05 <I440r_> or...
17:28:11 <I440r_> concentrate on your own projects ?
17:28:14 <kc5tja> I440r_: Yes.  It's what open source is all about.  :)
17:28:34 <kc5tja> I need a feature X to work on my project Y, so by spending time on X, I make development of Y faster in the long term.
17:29:19 <kc5tja> (The logic is sound: if I need feature X, and am willing to hack the code to enable it, then I will obviously need to use it more than once.  Hence, the time is well spent.)
17:29:59 * kc5tja would like to point out that the RONJA optical networking system is entirely open source hardware, made entirely with open source EDA tools.
17:30:18 <kc5tja> I know for a fact that only a few years ago, gEDA couldn't even render an AND gate without segfaulting.
17:30:26 <kc5tja> So yeah, sometimes submitting patches is a good thing.
17:30:28 <kc5tja> :)
17:33:24 <Herkamire> whenever you run accross the situation where a pice of free software that you like doesn't do what you want it to, you should always consider fixing/adding-to it
17:33:52 <I440r_> herk i wouldnt have any time left lol
17:34:18 <Herkamire> I said consider
17:34:21 <I440r_> which is part of the reson for isforth
17:34:30 <Herkamire> obviously choose the ones that are worth your time
17:34:32 <I440r_> isforth is so i can develop for linux without going anywhere near any c sources
17:34:47 <kc5tja> This was a significant-enough feature that I thought it was important to have.  It was the first thing I noticed when debugging the result of the FTS/Forth compiler.
17:34:57 <kc5tja> call +0xffffffe8
17:35:04 <Herkamire> kc5tja: heh
17:35:07 <kc5tja> Yikes!  I have no idea what address that is.
17:35:20 <kc5tja> It's nice to see what the relative offset is, especially in jump-table situations.
17:35:41 <kc5tja> But realistically speaking, having a real, target address is more helpful.  So, I now have it displaying a real target address like this:
17:35:45 <I440r_> the IDE that comes iwth my little 8051 board has a disassembly window
17:35:55 <I440r_> TOTALLY not symbolic disassembly
17:35:57 <kc5tja> call +0xffffffe8 (0x40078abc)
17:36:05 <I440r_> i asked the developers if they would ever make it symbolic
17:36:09 <I440r_> their response...
17:36:15 <I440r_> "whats symbolic disassembly?"
17:36:17 <slava> hahaha
17:36:23 <slava> I440r, come on its the 8051
17:36:25 <slava> what do you expect
17:36:35 <I440r_> when i explained it to them they sayd nah.. i dont think were gona add that
17:36:40 <kc5tja> Well, symbolic disassembly for me isn't at all important, especially since I simply don't have any symbol tables to speak of in my ELFs.
17:37:01 <I440r_> assemble it with a -g :)
17:37:08 <I440r_> of corse... that only works for kernel.com with me
17:37:16 <I440r_> after fsaving its TOTALLY stripped heh
17:37:17 <kc5tja> I440r_: The cross compiler is written in GForth.  It goes from raw source to raw binary.
17:37:20 <kc5tja> Instantly, too.
17:38:39 <I440r_> hrm. if i put a -j REJECT of an ip in my firewall and then ping -f it ?
17:38:40 <I440r_> lol
17:38:45 <I440r_> do i "reject" all the replies ?
17:38:46 <I440r_> lol
17:40:31 <kc5tja> Without more information, I can't answer that question.
17:58:21 --- quit: slava (Remote closed the connection)
18:02:39 <I440r> it was a joke (the -j reject i mean)
18:20:13 --- quit: mur_ (Remote closed the connection)
18:20:45 --- join: mur (~mur@smtp.uiah.fi) joined #forth
18:37:12 --- quit: I440r ("Leaving")
18:41:17 --- quit: [Forth] (Read error: 104 (Connection reset by peer))
18:50:40 --- join: I440r (~mark4@216-110-82-59.gen.twtelecom.net) joined #forth
18:58:01 --- join: madgarden_ (~madgarden@Toronto-HSE-ppp3707020.sympatico.ca) joined #forth
18:58:54 --- quit: madgarden (Nick collision from services.)
18:59:00 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth
18:59:03 --- nick: madgarden_ -> madgarden
19:40:24 --- join: Robert__ (~snofs@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth
19:41:54 --- quit: Robert (Read error: 104 (Connection reset by peer))
19:53:07 --- quit: slava ("Leaving")
19:56:46 --- join: mur_ (~mur@uiah.fi) joined #forth
20:00:37 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth
20:00:46 <slava> yo
20:02:46 <I440r> wb
20:03:01 <TheBlueWizard> re
20:03:09 <Herkamire> hi
20:04:37 <slava> xc
20:04:38 <slava> up
20:04:40 <slava> aq
20:05:02 <I440r> eh ?
20:05:19 <jDoctor> wtf
20:05:51 <jDoctor> eof
20:05:51 <slava> FORGET jDoctor 
20:05:55 <slava> : jDoctor i440r ;
20:06:11 <jDoctor> : slava "the weird one" print ;
20:06:16 <slava> that's not forth
20:06:23 <slava> : jDoctor ." should learn it" ;
20:06:28 <jDoctor> slava, thats not forth
20:06:31 <jDoctor> haha
20:06:32 <slava> that is
20:06:50 <jDoctor> that isnt. this isnt either. its english
20:06:56 * jDoctor sticks his toungue out
20:07:36 <jDoctor> : slava ."the weird one" slava ;
20:07:57 <slava> that's not forth either, unless you DEFER slava first :)
20:08:16 <slava> in forth a definition cannot refer to itself
20:08:19 <slava> unless you take extra steps
20:08:22 <Herkamire> must have space after ."
20:08:24 <jDoctor> it cant do that? man.
20:08:25 <slava> that too
20:08:35 <slava> jDoctor, : slava ." the weird one" RECURSE ;
20:08:47 <Herkamire> : slava ." blah " recurse ;
20:08:59 <I440r> recursion!
20:09:04 <I440r> do it itterativly
20:09:08 <slava> lol
20:09:11 <Herkamire> normal forths don't actually stick the word in the dictionary (at least the search path) untill the ;
20:09:16 <I440r> : slava begin ." blah" again ;
20:09:22 <I440r> that wont overflow the return stack
20:09:29 <jDoctor> : care? ." no, I don't   " recurse ;
20:09:32 <slava> I440r, some forths have tail call optimization
20:09:41 <slava> I440r, Herkamire's does
20:09:46 <I440r> Herkamire, : recurse last @ , ; immediate
20:09:47 <Herkamire> :)
20:10:10 <Herkamire> yeah, I don't think that feature is worth it.
20:10:18 <slava> recurse?
20:10:20 <slava> or tail call op?
20:10:29 <jDoctor> I think it is, if it was used from the foundation
20:10:33 <jDoctor> and not too late
20:10:33 <Herkamire> slava: I'm reffering to the thing about not putting a word in the search path until ;
20:10:35 <I440r> recurse doesnt give you anything you cant do without it
20:10:38 <slava> Herkamire, oh yeah
20:10:52 <slava> I440r, iteration doesnt give you anything you can't do without it
20:11:01 <I440r> herk forth does taht to stop you including an incomplete definition in the dictionary
20:11:05 <Herkamire> I like being able to recurse by using the word's name
20:11:12 <slava> yeah
20:11:19 <I440r> slava 99 times out of 100 the itterative method is faster
20:11:19 <Herkamire> and I like being able to use ; in the middle of a definition instead of having two different words for exit
20:11:29 <I440r> and recursion can be a complete bitch to read
20:11:39 <jDoctor> it is incorrect in society to do that though, you know.. like in a Dictionary, it shouldnt use the word to describe the word
20:11:46 * jDoctor gets hit by a book
20:12:03 <slava> I440r, colorforth only has recursion
20:12:08 <slava> I440r, and factor too
20:12:11 <I440r> : : create ;uses nest ] ;
20:12:18 <slava> wtf?
20:12:26 <Herkamire> jDoctor: hehe :)
20:12:28 <I440r> :)
20:12:43 <Herkamire> jDoctor: iteration in a dictionary is not useful. (but happens anyway)
20:12:53 <Herkamire> itteration in code can win fabulous prizes
20:13:01 <jDoctor> yah. I just noticed that
20:13:12 <jDoctor> it nullifies the entired definition in a dictionary
20:13:17 <jDoctor> yet it doesnt to a computer
20:13:19 <jDoctor> now why?
20:13:26 <slava> what the hell are you on about?
20:13:36 <jDoctor> nothing. 
20:13:44 <slava> :)
20:13:56 <Herkamire> I440r: yeah, it's true in a lot of cases that recursion is slower. (even in herkforth until I get it optomized) but it doesn't have to b e
20:14:04 <jDoctor> but I hate when in a dictionary, a word says look at the other word, and the other word says to look at the first!
20:14:42 <Herkamire> for now I just went with the quick "make it work". I've been pushing my "optomize later" rule a lot in an attempt to make herkforth useful
20:14:49 <jDoctor> oh, so I see my answer now. a Dictionary is pure information. in programming, its pure command. the dictionary does tell you to do anything
20:15:00 <kc5tja> FTS/Forth has recursion and only recursion.
20:15:02 <Herkamire> I'd rather dick around with stuff trying to make it fast, but that wouldn't be terribly useful
20:15:37 <Herkamire> in herkforth I have one counted loop
20:16:39 <Herkamire> if implemented well, these imo should compile to the same code:   : foo begin ... again ;   and   : bar ... recurse ;
20:16:57 <kc5tja> They do in my cross compiler.
20:17:07 <Herkamire> kc5tja: nice :)
20:17:33 <kc5tja> Herkamire: I've been depending on recursion for looping for . . . about a year now.  Where have you been?  ;-P
20:17:38 * slava ponders working some more on the docs.
20:18:03 * jDoctor goes back to making his bot work on irc
20:19:23 <Herkamire> kc5tja: yeah, I know. you're cool :)
20:20:32 --- quit: mur (Read error: 110 (Connection timed out))
20:21:45 <kc5tja> Hehehe -- the cool part is debatable.
20:21:52 <kc5tja> Most people think I'm just plain a geek.
20:21:54 <kc5tja> But I digress.
20:23:31 <Herkamire> not in here
20:23:39 <Herkamire> in here we think you're cool
20:23:56 <jDoctor> geeks are cool, man
20:24:28 <Herkamire> and there's girls that like geeks too :)
20:24:40 <I440r> lol
20:24:42 <Herkamire> perhaps not enough such girls... but they do exist
20:24:44 <I440r> i want two
20:25:17 <jDoctor> I didnt see those on getYourChickOnline.com ..
20:25:18 <TheBlueWizard> two? to better the odds of getting married?
20:25:27 <Herkamire> I440r: heh, I'm not sure they come in packages of two
20:25:27 --- join: zruty (~chuck@64-121-15-14.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth
20:25:46 <I440r> lol
20:25:51 <I440r> damn
20:26:20 <jDoctor> angelina jolie played one
20:26:37 <jDoctor> I was like wow... she said " 28.8 k modem"
20:27:16 <zruty> .o( leave it to a forther to say she said QUOTE SPACE ...
20:27:53 <jDoctor> wow. you really caught onto the subject fast
20:28:16 <TheBlueWizard> gotta go...all bye
20:28:22 --- part: TheBlueWizard left #forth
20:28:23 * zruty ponders making .o( a forth word, pepper my source with thought bubbles
20:28:41 <jDoctor> oh THATs what that was
20:28:55 <jDoctor> I thought it was a guy with a big nose and a missing eye
20:29:05 <slava> zruty, did you get anywhere with your factor experiemtns?
20:29:06 <zruty> heh
20:29:16 <jDoctor> certainly make sense that he didnt have a smile on his face
20:29:21 <zruty> slava: keep getting interrupted every time i sit down to start using it
20:29:32 <jDoctor> by what.. irc?
20:29:40 <zruty> work and gf
20:29:49 <jDoctor> gf ! teach her factor
20:29:58 <zruty> ye gads
20:30:09 <jDoctor> I taught mine ti-83 BASIC
20:30:10 <zruty> she's a PL/SQL developer
20:30:18 <jDoctor> oracle?
20:30:18 <zruty> you think she's gonna learn forth?
20:30:20 <zruty> yep
20:30:47 <jDoctor> I didnt even know that was a developer type
20:31:00 <zruty> from my perspective, it isn't
20:31:25 <zruty> page after page of bureaucratic crap.  the real skill is in the business analysis stuff, figure out where all the domain specific rules in the db are, etc
20:31:27 <jDoctor> but , hey, she might just be geeky enough to catch onto forth. you just gotta put it on her hallmark cards and stuff
20:31:40 <zruty> awful stuff for me, i'm a perl hacker by trade
20:31:51 <zruty> high level language guy.  like forth because it's orthogonal mostly
20:31:58 <jDoctor> perl is like.. I dont even know the words to say
20:32:15 <zruty> swiss army chainsaw with built in dvd player and cheesemaker
20:32:25 <zruty> that sum it up?
20:32:25 <jDoctor> yah. something like that
20:32:40 <jDoctor> and I once only knew how to use perl to filter several text files
20:32:43 <jDoctor> now I use unix
20:32:44 <zruty> it's better than awk
20:33:18 <zruty> i tend to use perl for shell scripts that have gotten too big
20:33:24 <jDoctor> well, I dont have big jobs of filtering anymore, but ya. I keep perl in mind for it
20:33:31 <zruty> my shell scripts always run faster tho because i go nuts with pipes
20:33:44 <jDoctor> me too, dude
20:33:46 <zruty> big pipelines on a 8 cpu box is a win
20:33:59 <jDoctor> whoa.. you really have this?
20:34:04 <zruty> not at home
20:34:08 <jDoctor> oh. lol
20:34:15 <zruty> at work.  that's a downgrade, used to work with 24 cpu boxen
20:34:26 <jDoctor> I have two 600mhz's in the closet. its a computer just for burning cd's.
20:34:39 <jDoctor> what do yall use at your work? sun?
20:34:41 <zruty> my home box is a 600 mhz vaio laptop
20:35:09 <zruty> nope, pc's.  never use the 8 cpu box anymore tho, it's purely a db server
20:35:23 <zruty> use a dual HT xeon now
20:35:26 <zruty> so 4 cpu's sorta
20:36:05 <jDoctor> cool. cool. for home, I conclude to apple winning 
20:36:16 * jDoctor is on one now
20:36:19 <madgarden> I like FOR/NEXT loops.
20:36:20 <I440r> zruty, damn im jelous
20:36:32 <I440r> i want to get some of those dell servers with dual xeon's in em
20:36:34 <slava> madgarden, i just made mine twice as fast
20:36:35 <zruty> would love a mac, but the price point just isnt there for me
20:36:40 <I440r> thery are only 3000 usd +
20:36:48 <madgarden> slava, your FOR/NEXT loop?
20:36:57 <jDoctor> what? no
20:37:00 <slava> madgarden, its called times, but yeah
20:37:02 --- join: Sonarman (~matt@ppp-66-124-255-188.dsl.snfc21.pacbell.net) joined #forth
20:37:03 <jDoctor> mine was 1,200
20:37:06 <zruty> FOR/NEXT?  foo, RECURSE
20:37:07 <madgarden> Cool... what changed?
20:37:21 <zruty> factor does tail recursion elimination, right?
20:37:23 <slava> madgarden, its implemented recursively, and not as a while loop which is implemented recursively
20:37:26 <slava> zruty, yes
20:37:37 <zruty> oh heck it has continuations, i forget :p
20:37:39 <slava> zruty, loops are actually higher order functions that recurse and call their arguments
20:37:59 <madgarden> Well, I'm fairly happy with the speed of my for/next loop/
20:38:14 <zruty> are there creating and parsing words ala CREATE..DOES> and the like?
20:38:44 <slava> there's parsing words, and creating words but they're different than forth
20:38:58 <zruty> cool.  will have to read more of that manual i guess :)
20:38:59 <slava> its closer to lisp
20:39:05 <zruty> macros?
20:39:18 <slava> no, just how the code is constructed from cons cells
20:39:56 <slava> if you have a number on the stack, [ + ] cons now gives you a quotation on the stack that adds that number to the top of stack when executed
20:40:13 <slava> 3 [ + ] cons is [ 3 + ] etc
20:40:30 <zruty> wow so it's like currying
20:40:56 <slava> [ and ] input a list literal. they are parsing words, similar to forth
20:42:05 <zruty> noticed that, [ ] for lists, { } for vectors
20:42:11 <zruty> any literal syntax for hashtables?
20:42:23 <slava> no
20:42:37 <slava> you can have a literal association list, and turn it into a hashtable though
20:42:44 <jDoctor> they are like.. binded and edited, right?
20:45:21 --- quit: ChanServ (ACK! SIGSEGV!)
20:46:21 <kc5tja> OK, my changes to ald are done, and submitted.  They'll appear in ALD 0.1.7.  :)
20:46:35 --- quit: jDoctor ("leaving")
20:46:42 --- join: ChanServ (ChanServ@services.) joined #forth
20:46:42 --- mode: irc.freenode.net set +o ChanServ
20:48:14 <slava> why is ChanServ actually in this channel
20:48:25 <slava> i never see it in other channels
20:51:08 <I440r> its an option :)
20:51:22 <slava> whats the benefit?
20:51:30 <I440r> err.... errr....
20:51:36 <I440r> ill get back to you on that one
20:51:39 <I440r> heh
20:52:35 <slava> -ChanServ- When GUARD is enabled, ChanServ will join and sit in your channel.
20:52:35 <slava> -ChanServ- One advantage to this is to keep your channel occupied when
20:52:35 <slava> -ChanServ- the last user leaves.
20:52:41 <slava> err
20:53:57 <I440r> oh yea
20:54:00 <I440r> topic is retained
20:54:04 <slava> oh
20:57:40 <zruty> hm, how do i make the factor docs?
20:58:00 <zruty> been so long since i've made anything from a .tex
20:58:14 <zruty> should just be latex devel-guide.tex right?
20:58:24 <zruty> then dvipdf devel-guide.dvi
20:59:44 <slava> pdflatex devel-guide.tex
21:01:22 <I440r> sc
21:01:48 <Herkamire> chanserv also guards ops
21:01:59 <I440r> correct
21:02:04 <zruty> bleh.  debian screws up latex permissions in font cache
21:02:16 * zruty makes it world writable.  live dangerously
21:02:23 <Herkamire> when a chan is empty, then the first one there gets ops, and can do whatever they like with it (eg password protect entry)
21:02:33 <zruty> too bad i can't make fbgs work
21:02:35 <kc5tja> I440r: Hey, if you want, I can send you the copy of ALD 0.1.7pre that I have, with my mods in it to properly display eIP-relative addresses.
21:02:47 <I440r> kc5 sure lol how u get that ?
21:02:49 <zruty> fbi and company dont like my neomagic fb
21:03:25 <kc5tja> I440r: Like I said, I e-mailed the author and submitted a patch.  It didn't quite fit first try, some things had to be changed to make it work with the 0.1.7 version.  But I have it fixed, and it actually works better than before on 0.1.6.
21:03:42 <kc5tja> Anyway, he sent me the pre-release of 0.1.7 to integrate my changes in.
21:03:52 <I440r> cool
21:03:56 <I440r> what ws your change ?
21:04:09 <I440r> how can you do that
21:04:25 <I440r> i can study c sources for WEEKS and not know enough about them to make any inteligent chages
21:04:35 --- quit: mur_ (Remote closed the connection)
21:04:42 <slava> here he goes again :)
21:05:15 --- join: mur (~mur@kyberias.uiah.fi) joined #forth
21:05:17 --- quit: Herkamire ("bed bed bed")
21:05:17 <I440r> but im totally stumped that anyone can look at c sources and make head or tail of them
21:05:47 <kc5tja> I440r: Well, like I said, I helped maintain about 700MB of C and assembly code for Hifn, so . . . it's very doable, however ugly the code may be.  :)
21:05:55 <zruty> huh the dvi renders less chunky than the pdf oddly
21:05:57 <kc5tja> When coding in C, you just gotta think like a C programmer.
21:06:06 <I440r> thats part of the reason i hate c. its the only language i know taht has that affect on me
21:06:14 <kc5tja> I440r: Anyway, /msg me your e-mail address so I can send it to you.
21:06:22 <I440r> im incapable of doing that lol
21:07:19 <zruty> most C++ using heavy templates has that effect on me
21:07:28 <kc5tja> I can't stand templates.
21:07:39 <zruty> touch anything and the linker will spit at you with venom and hate and errors
21:07:50 <kc5tja> Not that I don't know how they're used or anything, but I just think that people use them reflexively instead of thinking of cleaner, more efficient alternatives.
21:07:59 <zruty> oh templates are fine, C++'s implementation of them just sucks
21:08:40 <zruty> probably 95% of template use is better served with a language that has real algebraic types
21:09:14 <zruty> and metaprogramming with a real macro system
21:09:20 <kc5tja> On its way.
21:09:37 <kc5tja> Scala's "type parameters" blow "templates" clean out of the water.
21:09:37 <zruty> or something like openc++ which gives you real metaprogramming, not a macro hack
21:09:43 <kc5tja> Templates are, as far as I'm concerned, NOT fine.
21:10:26 <zruty> well, ML is hitting the big time with stuff like mldonkey
21:10:37 <kc5tja> The problem is that types need to be first-class objects when making a language that has true parametric polymorphism.
21:10:41 <zruty> tho mldonkey doesn't exactly help ocaml's reputation concerning documentation :p
21:11:10 <zruty> kc5tja: well first class or a separate language, there's no between
21:11:16 <zruty> ML uses a separate language
21:11:27 <zruty> coq uses first class types.  'course coq isnt turing complete
21:11:28 <kc5tja> See, I have major issues with "separate languages."
21:11:44 <zruty> it's actually something of a separate language in C too
21:11:53 <I440r> got it
21:12:04 <kc5tja> Multiple languages in the same source file leads to nasty, nasty syntax and often results in ugly kludges.
21:12:25 <slava> coq is just a terribly rude name
21:12:28 <zruty> not with ML
21:12:33 <slava> "hey man i'm programming some coq"
21:12:35 <slava> "cock?!"
21:12:37 <slava> "no, coq"
21:12:40 <zruty> you know what's a type declaration and you know what isnt
21:12:42 <slava> "cock?!"
21:12:49 <zruty> pronounced "coke" i believe
21:12:50 <slava> "see oh que!"
21:12:55 <I440r> kct what about asm and forth in same source file
21:13:10 <zruty> saying it's a separate langauge isn't like it's an embedded language
21:13:12 <I440r> err kc5 even
21:13:17 <zruty> it means it doesn't touch runtime semantics
21:13:18 <kc5tja> Since the assembler is valid Forth, you're coding Forth all the way.
21:13:26 <kc5tja> Especially if it's an RPN Forth.  :)
21:13:30 <I440r> lol
21:13:31 <zruty> ML declarations look like ML, they're just compile-time only
21:13:46 <kc5tja> zruty: Can you give a brief example?
21:13:56 <zruty> lemme switch over
21:14:23 <zruty> give you the canonical haskell quicksort.  cheating a bit, since haskell is the prettiest of the ML family
21:15:26 <zruty> fooey, it's not annotated
21:15:29 <kc5tja> I personally disagree; I like ML's syntax the best.  But, whatever you're comfortable with.
21:16:11 <I440r> how do you get the GTK user interface up ? :)
21:16:17 <I440r> or is that like the curses one
21:16:19 <I440r> vaporware
21:16:22 <zruty> chmod -R a+w /var/cache/fonts
21:16:25 <zruty> damn
21:16:39 <zruty> can't copy and paste between console and X
21:16:44 <zruty> i really hate *nix sometimes
21:17:16 <kc5tja> I440r: ?
21:17:27 <zruty> length :: [a] -> Int
21:17:29 <I440r> ald
21:17:41 <I440r> the curses version of ald was always disabled cuz he hadnt finished it
21:17:42 <zruty> that's a type declaration.  it's actually a separate language in its own right
21:17:44 <I440r> it never worked
21:17:53 <zruty> data constructors are another
21:17:53 <kc5tja> I440r: I never knew it had a GTK interface.
21:17:54 <I440r> i see a ui directiry in here
21:18:00 <zruty> data Maybe = Just a | Nothing
21:18:07 <zruty> er Maybe a that is
21:18:19 <kc5tja> zruty: OK, I do not consider that a separate language at all.  I consider that an intrinsic part of Haskell.
21:19:06 <zruty> it is intrinsic, but it's a separate language, so it gets verified first, then everything else is verified against it
21:19:11 <I440r> thers only two command line switches
21:19:14 <I440r> -h and -v
21:19:23 <I440r> so the gtk interface is still vaporware
21:19:35 <zruty> so you get parametric polymorphism without first class types
21:19:35 <I440r> and the curses interface looks to have been totally removed
21:19:48 <I440r> the lack of the curses interface is disappointing
21:20:03 <I440r> i would like to see registers in a windows on the screen
21:20:06 <I440r> disassembly
21:20:13 <I440r> and a highlighted like for IP
21:20:25 <I440r> and a display of the top few items of the stack
21:20:31 <zruty> speaking of interfaces, does factor have a C interface for stuff like SDL, curses, etc?
21:20:43 <zruty> i'm interested in SDL
21:20:51 <I440r> ill go single step some isforth code and take a look
21:20:52 <kc5tja> I440r: None of this is even documented on the ALD website.  Not sure where you've seen this before.
21:21:10 <I440r> kc5 3 years ago. the curses code was enabled with a -c
21:21:28 <I440r> but all it did was say  "the curses interface has been disabled because blah blah"
21:22:02 <I440r> a continuously scrolling display after each single step is a horrible interface for a debugger
21:22:16 <I440r> anywa
21:22:21 <I440r> thats not a disqualification
21:22:24 <I440r> tho
21:22:32 * I440r goes to single step some code
21:22:36 <kc5tja> I440r: It's the only kind of debugger I've ever used, right back to the Commodore PET days.  :)
21:22:39 <I440r> actually the man page is next to useless
21:22:51 <I440r> kc5 got a dos box handy ?
21:22:55 <I440r> get d86
21:23:10 <I440r> its the debugger i measure all debuggers against. and it has some HUGE flaws
21:23:20 <kc5tja> No, I don't.  I have used MSVC's debugging environment, which was nice, but damn, it took up too much screen space.
21:23:36 <I440r> msvc debugger hehe
21:23:43 <kc5tja> It's actually very good.
21:23:47 <I440r> no d86 is a dos based asm debugger
21:24:03 <kc5tja> MSVC does 32-bit assembly too.
21:24:06 <I440r> kc5 msvc++ also has a damned good code optimizer
21:24:14 <I440r> and i dont care what the gcc people a
21:24:16 <zruty> and it's a zippy compiler
21:24:25 <kc5tja> MSVC is a good product.
21:24:28 <I440r> gcc's optimizer wont even come close to vc++
21:24:42 <zruty> MS should make more stuff like msvc
21:24:51 <kc5tja> However, I've never experienced an MSVC program going significantly faster than a GCC program, or vice versa.  As far as optimizations are concerned, I'd say they're both at the same level more or less.
21:24:54 <zruty> 'course compilers were ms's thing before they were into OS's
21:25:02 <zruty> you could say msvc is like MS's original product
21:25:10 <zruty> that and BASIC
21:25:15 <I440r> ok heres my BIGGEST critisism of gdb
21:25:22 <I440r> you cant debug till you start to run
21:25:33 <I440r> and you CANNOT put a breakpoint on the entry point
21:25:54 <kc5tja> I440r: More accurately, you CAN, but it'll break GDB.  :)
21:26:00 <kc5tja> (or worse, be completely ignored)
21:26:03 <I440r> now for a HUGE flaw in ald
21:26:22 <I440r> you dont see a disassembly of an instruction until AFTER you step it
21:26:29 <kc5tja> Not true.
21:26:34 <I440r> thats an automatic disqualification right there
21:26:42 <I440r> i jsut did ald isforth
21:26:50 <I440r> and it didnt disassemble the first instruction
21:26:51 <I440r> i said
21:26:52 <I440r> step
21:27:01 <I440r> and it executed and disassembled the first instruction only
21:27:08 <I440r> i dont get to see it till ive executed it
21:27:15 <I440r> or tell me what im doing wrong
21:27:16 <kc5tja> No, not true.
21:27:21 <I440r> because the man page doesnt tell me shit
21:27:24 <I440r> nor does -h
21:27:27 <kc5tja> What you're seeing is the instruction *about* to be executed.
21:27:39 <kc5tja> Yes, it doesn't show the very first instruction.
21:27:43 <I440r> the one about to be executed AFTER i told it to singlestep once
21:27:56 <kc5tja> But after you step it, it shows the *next* instruction to be executed, which is correct behavior.
21:28:03 <I440r> so the first "step" is "disassemble first instruction"
21:28:12 <I440r> and the second step is "step first instruction"
21:28:15 <I440r> ?
21:28:34 <I440r> err no
21:28:38 <I440r> i can see tahts not the case
21:28:43 <I440r> from blank screen you say
21:28:54 <I440r> step and it steps the first instruction but doesnt show it
21:28:59 <I440r> im now on the second instruction
21:29:05 <I440r> disassembled 
21:29:06 <kc5tja> a.out: ELF Intel 80386 (32 bit), LSB - little endian, Executable, Version 1 (Current)
21:29:09 <kc5tja> ald> disas   
21:29:12 <kc5tja> 0800044F  89E6                 mov esi, esp                
21:29:15 <kc5tja> 08000451  81EC00040000         sub esp, 0x400              
21:29:17 <kc5tja> 08000457  E87AFEFFFF           call near +0xfffffe7a (0x80002d6)  
21:29:20 <kc5tja> ...
21:29:21 <kc5tja> ald> s
21:29:21 <kc5tja> eax = 0x00000000 ebx = 0x00000000 ecx = 0x00000000 edx = 0x00000000 
21:29:21 <kc5tja> esp = 0xBFFFF9F0 ebp = 0x00000000 esi = 0xBFFFF9F0 edi = 0x00000000 
21:29:21 <kc5tja> ds  = 0x0000002B es  = 0x0000002B fs  = 0x00000000 gs  = 0x00000000 
21:29:23 <kc5tja> ss  = 0x0000002B cs  = 0x00000023 eip = 0x08000451 eflags = 0x00000346 
21:29:26 <kc5tja> Flags: PF ZF TF IF 
21:29:28 <kc5tja> 08000451  81EC00040000         sub esp, 0x400              
21:29:31 <kc5tja> ald> 
21:29:33 <kc5tja> As you can see, the first instruction is mov esi,esp.
21:29:54 <I440r> kc5 you NEVER get to see the first instruction - my first instruction is a call init_mem - ive stepped it and im seeing the disassembly of the first instruction of that subroutine
21:29:56 <kc5tja> Yet, after hitting "s" and ENTER, the instruction that is displayed is the second instruction -- the instruction that has yet to execute, which is sub esp,0x400
21:30:00 <I440r> you do not get to see the first
21:30:09 <I440r> right
21:30:10 <kc5tja> So disassemble it.
21:30:15 <I440r> this is forgiveable
21:30:19 <I440r> but will need to be fixed
21:30:26 * kc5tja shrugs.
21:30:26 <I440r> when i say ald blah
21:30:36 <I440r> ald should auto disassemble the first instruction
21:30:40 <I440r> i think
21:30:49 * I440r continues
21:31:01 <kc5tja> Well, I'm not prepared to make that change.  It works well enough for me.
21:31:02 <kc5tja> :)
21:31:16 <I440r> Flags: PF ZF TF IF ID 
21:31:21 <I440r> err which flags are set
21:31:29 <I440r> not ALL of thse surely
21:31:56 <I440r> it should display the names of all the flags in lower case and uppercase the ones that are set - or it should highlight th eones that are set
21:31:58 <I440r> or something
21:32:31 <I440r> and there could be an option in here to NOT display selectros
21:32:36 <kc5tja> If you see PF ZF TF IF ID, etc, then yes, those flags are currently set.
21:32:47 <I440r> its very very very unlikely that anyone is ever going to be using ald in ring 0
21:32:57 <I440r> their presence is just noise
21:33:07 <kc5tja> I disagree.
21:33:33 <kc5tja> What if you're like me, who enjoys working on operating system software, and is using ALD over a serial port at the kernel level (as with tethered-GDB)?
21:33:50 <I440r> i said very very very unlikely
21:33:53 <I440r> and "an option"
21:34:08 <I440r> i doubt ald CAN be used to debug in kernel space
21:34:19 <I440r> so those registers are superfluous
21:34:28 <I440r> you cant change them except from ring 0
21:34:31 <kc5tja> Neither can gdb, technically, but yet, gdb has support for tethered mode.
21:34:47 <kc5tja> (and has been, and is, being used for that purpose.)
21:34:54 <I440r> also, ald is still disqualified until it does symbolic disassembly
21:34:56 <I440r> which it doesnt do
21:35:02 <kc5tja> Whatever.
21:35:22 <I440r> you try single stepping thru isforth dood without any symbols
21:35:31 <kc5tja> I somehow doubt your generated ELFs have debugging symbol information in them anyway.
21:35:44 <kc5tja> Well, that's not my problem.
21:36:02 <kc5tja> My goal is to get FTS/Forth to a point where I can do automated unit testing, so that I no longer *need* ald to begin with.
21:36:24 <I440r> gdb might be a pile of crap but its acutally easier to use for debugging my kernel simply because it displays symbol names
21:36:42 <kc5tja> gdb is the most difficult to use debugger to use.
21:36:52 <I440r> ive seen worse hehe
21:36:52 <kc5tja> (for assembly)
21:36:56 <kc5tja> No, I haven't.
21:36:56 <I440r> not many... but some
21:36:59 <I440r> i have
21:37:01 <kc5tja> And I've used many.
21:37:07 <I440r> but yes. gdb is a pile of crap
21:37:19 <I440r> ever use an nec 75x in circuit emulator ?
21:37:26 <I440r> you DONT get to see an instruciton till you step it
21:37:27 <kc5tja> No, and I have no need to.
21:37:38 <I440r> you break on the instruction FOLLOWING the breakpoint
21:37:40 <I440r> if at all
21:37:50 <kc5tja> Umm...well...an in-circuit emulator normally works that way anyway
21:37:56 <I440r> you cant put a breakpoint on the first instruciton of a routine (nor can you in gdb)
21:38:26 <I440r> put a breakpoint ont eh first line of a c function and it wont break there
21:38:29 <I440r> it will stop on the second
21:38:33 <I440r> or not at all
21:38:45 <kc5tja> I've set a breakpoint on the first line of a C program PLENTY of times.
21:38:58 <kc5tja> First line of main(), works just fine.
21:39:20 <kc5tja> Note: obviously, you can't set a breakpoint on a variable declaration, since it technically isn't executable code.
21:39:22 <I440r> maybe they have fix3ed that then
21:39:29 <I440r> thats not code :)
21:39:33 <I440r> not unless its
21:39:36 <I440r>  char x = 1;
21:39:38 <kc5tja> I440r: I've done it since Linux 1.3.8. :)
21:39:43 <kc5tja> Yes, you can break on those.
21:40:07 <I440r> kc5 3 years ago you could NOT put a breakpoint on the entry point
21:40:15 <kc5tja> Sure could.
21:40:20 <I440r> nor on the first instruction (valid instruction) of a function call
21:40:45 <kc5tja> Because main() is, contrary to popular belief, not the entry-point of a program.  Never has been.  It is called by libcrt.(s)o.
21:40:46 <I440r> i was emailing the developers of gdb and they were like "who cares"
21:40:57 <I440r> err i know that 
21:41:12 <I440r> im talking about the ENTRY point of the program
21:41:28 <I440r> gdb has this silly notion of "is executing"
21:41:36 <I440r> i.e. you have to "run" before you can stop
21:41:49 <I440r> and you cannot RUN and STOP on the same instructon
21:41:51 <kc5tja> That's because the debugged program runs in another process.
21:42:16 <kc5tja> Of course you can't.  That's contradictory.
21:42:33 <I440r> well why doesnt gdb automatically START the process being debugged yet NOT execute any of it
21:43:01 <I440r> well it means you cannot debug your entry point without doing trickery like making your first instruciton a NOP
21:43:02 <I440r> lol
21:43:07 <kc5tja> Because you may need to do additional configurations.  You can specify command line arguments via the "run" command, you can (IIRC) change environment variables, etc.
21:43:20 <I440r> which is bad because your bug might be related to location of a specific instruciton
21:43:24 <I440r> and the nop might move oit
21:43:39 <I440r> you can change env vars inside
21:43:43 <I440r> while executing
21:44:10 <I440r> ald has one thing going for it tho
21:44:14 <I440r> its a hell of alot faster 
21:44:15 <I440r> lol
21:44:24 <I440r> but it doesnt show the freekin stack
21:44:43 --- quit: zruty ("Lost terminal")
21:45:13 <I440r> it would also be nicer if it DID display all flags even ones taht arent set - that way your eyes would become used to the location of specific flags
21:45:19 <I440r> OFF flags can be in lower case
21:45:23 <I440r> or ON flags can be in bold
21:46:30 <I440r> no symbols makes things very difficult tho
21:46:32 <I440r> you cant say
21:46:46 <I440r> break mung_data_fudgificator
21:46:52 <I440r> and then run
21:47:12 <I440r> and single stepping to the part you want to debug can take years lol
21:47:29 <I440r> it would take an hour to single step through my header relocation 
21:47:39 <I440r> unless you just hit s and held cr down 
21:47:43 <I440r> but then you might overshoot
21:48:47 <I440r> but i guess im talking to myself
21:51:17 <kc5tja> Nahh
21:51:31 <kc5tja> My coding style tends not to rely on the presence of a debugger.
21:51:55 <kc5tja> Like I said, I use unit testing, and in some cases, even formal verification methods.
21:52:06 <kc5tja> I think today was the first time I've used a debugger in over 2.5 years.
21:52:16 <I440r> unit testing just tells you "something is wrong here"
21:52:24 <I440r> debugging can tell you "what is wrong"
21:52:35 <kc5tja> Not true.
21:52:40 <kc5tja> Not even close to being true.
21:52:47 <I440r> sometimes a debugger is a very usefull tool
21:52:51 <kc5tja> Sometimes.
21:53:01 <kc5tja> But proper coding practices can minimize its use.
21:53:23 <I440r> anyone who RELIES on a debugger or ai ICE should go find another career field tho lol
21:53:29 <I440r> yes
21:53:33 <kc5tja> Unit tests tell you what is wrong, and by definition, its rough location too.  Specific location is often determinable by visual context.
21:54:13 <kc5tja> Using partial ordering of unit tests (so that all lower-level tests run before higher level tests), you can pin-point bugs by process of elimination, and often, down to individual lines of source.
21:54:37 <kc5tja> And the nice thing, they're repeatable, so as you refactor your code, you can keep running those tests, to see if you have inadvertently broken anything.
21:54:40 <kc5tja> It's wonderful.
21:55:22 <kc5tja> Anyway, I think I've had a wonderfully productive day of coding.
21:55:44 <kc5tja> So since it's nearing the end of the day, I'm going to commit my changes to FTS/Forth's cross compiler, and go play a game.
21:56:06 <kc5tja> I promised a site update today, but I'll do that tomorrow.  I need to hack on my resume anyway, so lumping all my editing stuffs together seems to make some sense.
22:07:52 <I440r> :)
22:08:02 <I440r> i need to work on my power down mode :)
22:10:45 <I440r> kc5 why didnt you simply get rid of the silly branch disassembly showing offsets and just replace it with disassembly showing target address
22:11:10 <I440r> i dont care that this jnc is brancing 15 bytes back! - thats useless info
22:11:20 <I440r> your ( ) info should simply replace the offset
22:11:26 <I440r> THATS usefull info
22:11:38 <I440r> that was anotehr thing i didnt like about ald 3 years ago
22:12:33 <I440r> grr i dont even care about the OBJECT code of the instruction being displayed
22:12:46 <I440r> or its address - unless its a target
22:13:08 <I440r> if i need to know the instructions addres i look at IP
22:17:07 --- nick: Robert__ -> Robert
22:18:50 --- quit: Sonarman ("leaving")
22:22:02 --- join: Serg_penguin (~z@212.34.52.140) joined #forth
22:22:19 <I440r> hi serg lol
22:22:28 <I440r> you always come by right when im going to sleep hehe
22:24:04 <Robert> Hehe
22:24:12 * Robert is going to school in a while
22:24:17 <Robert> Mmm... 1 hour of travelling.
22:24:29 <I440r> OOOH i just invented a new programming language like brainfuck and whitespace hehe
22:24:36 <I440r> im gona work out the details tho first@
22:24:44 <Robert> :D
22:24:47 <Robert> What's it like?
22:24:51 <I440r> but itll be naturally obfuscated {:)
22:24:51 <Robert> I mean, the general concept
22:24:53 <I440r> not tellin!
22:24:57 <Robert> Bah
22:25:00 <I440r> heh
22:25:03 <I440r> gimme a day or 2
22:25:04 <Robert> Not until your patent is approved? :P
22:25:07 <I440r> and ill write it up
22:25:10 <I440r> lol
22:25:24 <I440r> damn.ya got me
22:25:29 <I440r> im gona make BILLIONS off this!
22:25:30 <I440r> hehe
22:25:46 <Robert> ;)
22:26:24 <I440r> neway
22:26:26 <I440r> nite nite
22:26:38 <I440r> way past my bed time
22:26:56 <Robert> Hehe
22:26:58 <Robert> Night :)
22:27:34 <Serg_penguin> hi
22:27:49 <Robert> Hi serg
22:27:59 * Serg_penguin just loaded film into my new toy ;))
22:28:20 <Serg_penguin> half-frame full-mechanic automatic camera 
22:30:35 <Serg_penguin> and writing an article on XML engine for static homepages
22:32:46 <Robert> Fancy ;)
22:33:31 * Robert is coding a few things for an article he found yesterday: http://en.wikipedia.org/wiki/Surreal_number
22:43:47 --- join: mur_ (~mur@uiah.fi) joined #forth
22:53:58 --- quit: mur (Read error: 110 (Connection timed out))
22:56:51 <slava> kc5tja, ping
22:57:06 <slava> kc5tja, on x86, are 32-bit addresses encoded in instructions always 32-bit aligned
22:57:15 <slava> kc5tja, in jumps even
22:57:18 <slava> not all instructions
22:57:59 <slava> kc5tja, ie, if i have a memory zone containing compiled code, can i do a simple iteration through each 4-byte word and do conservative gc
23:24:11 --- quit: mur_ (Remote closed the connection)
23:24:43 --- join: mur (~mur@smtp.uiah.fi) joined #forth
23:56:06 --- join: proteusguy (~proteusgu@66.168.185.144) joined #forth
23:59:44 <kc5tja> slava: No.
23:59:51 <kc5tja> Instructions can appear anywhere in memory.
23:59:59 --- log: ended forth/04.08.30