00:00:00 --- log: started forth/04.07.12
00:06:18 --- quit: segher (Read error: 104 (Connection reset by peer))
00:06:29 --- join: LOOP-HOG (~jdamisch@sub22-119.member.dsl-only.net) joined #forth
00:09:24 <LOOP-HOG> <TABLE ALIGN=LEFT>
00:09:26 <LOOP-HOG> puke
00:09:50 --- quit: LOOP-HOG (Client Quit)
00:11:03 <kc5tja> OK, I'm off to bed now.
00:11:19 <foobar_> bye
00:12:00 --- join: segher (~segher@blueice2n1.de.ibm.com) joined #forth
00:12:04 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
00:13:22 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth
00:23:10 --- quit: Serg ()
02:15:11 --- join: arke (~arke@melrose-251-251.flexabit.net) joined #forth
02:15:20 <arke> HI
02:16:13 <zardon> hi
02:20:06 <arke> °-°
02:20:08 <arke> ^
02:41:58 --- quit: arke ("leaving")
03:13:33 <aum> what's the better stack signature for binding a gui event callback: 
03:13:41 <aum> (1) ( widget xt -- )
03:13:42 <aum> or
03:13:47 <aum> (2) ( xt widget -- )
03:13:48 <aum> ?
03:17:56 <foobar_> only onetom would know...  :)
03:20:10 <aum> ahh, got my first gtk+ test program working (in forth)
03:20:37 <foobar_> great! i never have.  which forth?
03:20:44 <aum> ficl
03:20:59 <aum> well, my extensively hacked fork of ficl
03:21:05 <foobar_> ok
03:22:36 <aum> getting the callbacks working was 'fun'
03:22:39 <aum> was the hard part
03:22:54 <aum> the rest is easy - just adding another couple of hundred words ;p
03:25:27 * aum goes off to play rtcw:et to rest his brain
03:30:44 --- part: foobar_ left #forth
03:40:26 <segher> ( widget xt -- ... ) sounds the most natural to me
04:34:39 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
04:53:41 --- join: foobar_ (fooboo@dialup-4.154.47.228.Dial1.Atlanta1.Level3.net) joined #forth
04:57:38 --- part: foobar_ left #forth
05:59:58 --- quit: tathi ("leaving")
08:05:40 --- join: tucknip (~Huck@170.141.109.33) joined #forth
08:33:20 <qFox> ok i think i'm going to need help
08:33:36 <qFox> i'm trying to build a forth on top of another forth
08:33:46 <qFox> now i'm quite done with the new inner interpreter
08:33:53 <qFox> but how the hell do i innitiate the loop?
08:34:06 <segher> just call the inner interpreter
08:34:08 <qFox> i mean, what do i do when i execute?
08:34:30 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
08:34:33 <segher> dunno.  depends on what your inner interpreter looks like exactly
08:35:11 <qFox> when it starts, the new forth is going to be a entirely native system. the only thing used by the system underneath are a handfull of primitives (machine forth if you will)
08:35:42 <qFox> and i dont really have a problem with executing words in the new forth, or creating the primitives
08:35:50 <qFox> but how do i execute the primitives from the new forth
08:35:56 <qFox> i just cant hack it 
08:36:04 <segher> hee, een nederlander
08:36:12 <qFox> affirmative
08:36:32 <qFox> guess you're not serg :p
08:36:38 <segher> just tick-em, and EXECUTE em
08:36:47 <qFox> yes and no.
08:36:49 <segher> no, i'm segher.  that's my name.
08:37:02 <qFox> my dictionary is also incompatible with the old forth
08:37:10 <Robert> Yes, because you're a HEATHEN.
08:37:14 <segher> that shouldn't maatter
08:37:16 <Robert> Who doesn't appreciate my work :P
08:37:22 <segher> heh
08:37:25 <qFox> affirmative
08:37:35 <segher> robert: so who are you?
08:37:43 <Robert> How do the threading words look, qFox?
08:37:59 <qFox> Robert> dea's
08:38:05 <qFox> that way i can do indirect threading
08:38:11 <Robert> segher: I'm Robert, your humble servant and code monkey. Been here for a few years now. http://robert.zizi.org/ is my page.
08:38:11 <qFox> and yet be direct
08:38:12 <segher> robert: oh, you were calling qfox a heathen, not me.  okie :-)
08:38:12 <qFox> :p
08:38:23 <Robert> segher: Heh, that's correct.
08:38:42 <Robert> qFox: Have you written "next"?
08:38:54 <qFox> Robert doesnt realize he ought to be happy at last SOMEBODY is using a creation of his... ;)
08:39:13 <Robert> There's a difference between using and abusing, qFox. :)
08:39:16 <qFox> hmmm, i've written a loop. 
08:39:22 <qFox> accept
08:39:28 <qFox> get word untill end of buffer
08:39:31 <qFox> find word in dictionary
08:39:37 <qFox> write word if compiling
08:39:40 <qFox> or .... execute
08:39:44 <qFox> but thats where i'm stuck
08:39:56 <qFox> at end of input, back to accept
08:39:56 <Robert> Have you read BYOF and Moving Forth?
08:40:02 <qFox> no.
08:40:12 <segher> and MAF
08:40:23 <qFox> i'm not sure why this is bugging me so much
08:40:32 <qFox> i had no problem building the forth in machine forth
08:40:36 <qFox> this shouldnt be any different :\\
08:40:38 <Robert> Maybe that's a good idea. They include pseudocode for indirect threading primitives.
08:40:48 <Robert> Like execute, next, etc. 
08:41:01 <Robert> "next" being the interpreter internal word, not for/next.
08:41:07 <qFox> i know
08:42:22 <Robert> Goed
08:42:34 <Robert> Hrm! 
08:42:49 * Robert discovers the bug might be in a place he never thought of.
08:45:48 <qFox> hm, i think i'm missing call and ret
08:47:22 <qFox> in fact
08:52:41 --- quit: SDO ("Vision[0.9.6-0203]: i've been blurred!")
08:52:47 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
08:59:51 --- quit: ChanServ (Shutting Down)
09:08:42 --- join: ChanServ (ChanServ@services.) joined #forth
09:08:42 --- mode: irc.freenode.net set +o ChanServ
09:09:58 --- quit: SDO (Connection timed out)
09:59:39 --- join: Serg (~z@193.201.231.126) joined #forth
10:12:23 --- nick: TomasuDlrrp -> Tomasu
10:12:25 * Tomasu is back (gone 11:08:26)
10:19:52 --- part: tucknip left #forth
10:20:10 * Tomasu is away: dentist
10:21:28 <Serg> dhamn, #xml is dead !
10:21:47 * Serg throws depth charge in #xml to waken it
10:22:25 <Robert> Hey segher 
10:22:26 <Robert> er
10:22:28 <Robert> Serg I mean
10:30:26 * Serg is brute-forcing obscure XSLT specs
10:35:32 <Serg> CRACKED !!!
10:35:40 <Robert> :)
10:36:29 <Serg> <xsl:template match="p|ul|ol|li">
10:36:30 <Serg> <xsl:element name="{name()}"><xsl:apply-templates/></xsl:element>
10:36:30 <Serg> </xsl:template>
10:36:51 <Serg> this piece of XSLT transforms some tags to self, w/o attributes
10:40:07 <Serg> did you ever crack things by debugger ?
10:40:08 <Serg> and disassembler ?
10:40:28 <Robert> Hex editor 
10:40:41 --- join: wossname (wossname@Toronto-HSE-ppp3699167.sympatico.ca) joined #forth
10:40:41 <Robert> And instruction reference ;)
10:40:46 <Serg> way too hardcore ;)
10:41:07 <segher> *you had an instruction reference*?!?  that's CHEATING
10:41:18 <Serg> i need to crack old Fraunhofer MP3 coder, as rotten as 1998
10:41:18 <Robert> segher: I know :(
10:41:37 <Serg> no trace of it at Astalavista
10:41:41 --- join: tucknip (~Huck@170.141.109.33) joined #forth
10:41:48 <Serg> or is where any free drop-in replace ?
10:42:32 <Serg> i set the break to ExitProcess, but this gave me _really_ nothing
10:43:44 <Serg> it is limited to 30 seconds of encoding, so i want tit to be limit-free or at least more than 10 minutes
10:44:18 <wossname> you need to use this ancient encoder ?~
10:44:27 <Serg> yes !
10:44:58 <Serg> coz i need to recode 20G of music and give back the HDD i borrowed from work
10:45:13 <wossname> new ones are no good for this ? :(
10:45:38 <Serg> _nothing_ of new soft can recode filetree saving tags
10:46:17 <Serg> i asked in alt.music.mp3, they pointed me at dbPowerAmp music converter
10:46:31 <Serg> it's built-in Lame codec barfs all the way
10:46:40 <segher> what os/machine do you want?
10:46:47 <Serg> it can also accept Blade and Fraunhofer
10:47:10 <Serg> PC, any OS - agree to setup _anything_ on spare HDD to complete task
10:48:09 <Serg> Blade not works w/ so low bitrate i need, and Fraunhofer - demo only, fail to find kraked on net
10:48:19 <wossname> i would search for the number 30 in the binary and then give up when thousands of matches were returned. i advise you this course of action
10:48:36 <Serg> screw !
10:49:12 <Serg> i tryed _all_ 1E ( hex 30 ) and gained _nothing_
10:49:16 <segher> serg: http://www.rjamorim.com/rrw/l3enc.html
10:50:04 <Serg> segher: noticed
10:50:16 <Serg> is L3 Lame or Fraunhofer ?
10:50:37 <segher> FhG IIS
10:50:37 <Serg> i need standalone EXE w/ certain cmdline syntax ;((
10:51:42 <segher> it's the thing.  or get mp3enc.
10:51:42 <tucknip> there is a standalone Lame exe
10:52:43 <Serg> hell on it, i think of writing Perl script under Linux, but what w/ tags ????
10:53:18 --- join: warpzero (~warpzero@mi182.dn184.umontana.edu) joined #forth
10:54:01 <segher> just use Vorbis, instead.  you'll be happy.
11:05:06 <tucknip> usage: lame [options] <infile> [outfile]
11:10:47 --- join: Sonarman (~matt@adsl-64-160-165-47.dsl.snfc21.pacbell.net) joined #forth
11:11:06 <Sonarman> keybnoard noty workingh righhty
11:11:18 <Sonarman> *spilled lemonade on ity yestyerday
11:12:26 <Robert> Poor Sonarman.
11:12:42 <Sonarman> tyhatyy's noty enough symaptyhy!
11:12:51 * Robert cuddles Sonarman a bit.
11:13:05 <Robert> Is that enough, or should I blow you while I'm at it?
11:13:17 <Sonarman> well, while you're aty ity...
11:13:35 <Sonarman> anyway, i'm ghetytyingh quitye familiar wityh tyhe bnackspace key
11:13:55 <Robert> :D
11:14:10 <Robert> Where's your backup keyboard?
11:14:24 <Sonarman> my sisyter's room
11:14:27 --- join: slava (~slava@CPE00096ba44261-CM000e5cdfda14.cpe.net.cable.rogers.com) joined #forth
11:14:29 <slava> forthers!
11:14:33 <Sonarman> *hi slava
11:14:37 <Robert> slavas!
11:14:57 <chandler> factorers!
11:15:04 <Robert> Yes. :)
11:15:07 <slava> chandler, there's only 1 :)
11:15:14 <Robert> So far.
11:15:19 <slava> and doublec but he's not here
11:15:28 <Robert> Your world domination plans haven't reached the second stage, yet.
11:15:30 <Sonarman> slava: i'm ghoingh tyo tyell doubnlec you don'ty consider him a factyorer :P
11:15:47 <slava> Sonarman, are you using the Windows spelling checker there? :-P
11:15:47 <Sonarman> *ity will bnreak his hearty
11:15:55 <Sonarman> *'I AM TYEH JEFF K
11:16:03 <Robert> Heh.
11:16:10 <Sonarman> *keybnoard probnlems, actyuallty.
11:16:15 <slava> LOL
11:17:05 <Sonarman> *understyandabnly, entyeringh passwords is a pain
11:17:07 <chandler> fbnord!
11:17:13 <Sonarman> *lol
11:17:34 <slava> chandler, i found a bug in the C syntax highlighting in lisppaste
11:17:40 <slava> chandler, it doesn't highlight default: keyword :-)
11:17:48 <chandler> default is a keyword?
11:17:54 <slava> in switch statements
11:17:57 <chandler> oh
11:18:09 <chandler> that sounds like something I'm too lazy to fix :-)
11:18:13 <slava> heh
11:23:14 --- join: Sonarman_ (~matt@adsl-67-113-234-87.dsl.snfc21.pacbell.net) joined #forth
11:23:20 --- quit: Sonarman (Nick collision from services.)
11:23:28 --- nick: Sonarman_ -> Sonarman
11:24:24 <Sonarman> *beep*
11:24:33 <Sonarman> replacement text input module installed.
11:24:44 <slava> haha
11:25:09 <slava> chandler, ping
11:31:20 <chandler> slava: pong
11:31:37 <slava> chandler, how hard would it be to add forth/factor syntax highlighting (it will almost identical for both) to lisppaste
11:31:52 <chandler> shouldn't be too hard at all
11:32:01 <chandler> writing a syntax coloring mode is pretty simple
11:32:07 <slava> cool
11:33:08 * Serg is back
11:34:23 <slava> hi Serg
11:36:34 <slava> any ideas how to rewrite this without using custom stack words: http://paste.lisp.org/display/1707
11:39:10 --- join: Research (~Research@12-222-128-22.client.insightBB.com) joined #forth
11:39:22 * Serg works on my "Build XML homepage easy" article
11:39:54 <Sonarman> hi Research
11:41:17 --- join: thin (thin@bespin.org) joined #forth
11:42:06 --- part: thin left #forth
11:42:29 <Serg> Robert: do u need my control words for tforth ? i may write as excercise
11:44:57 --- quit: Serg ()
11:56:35 <slava> anybody using windows?
11:56:44 <chandler> at work, unfortunately yes
11:56:52 <slava> do you have a compiler?
11:56:59 <chandler> "yes"
11:57:07 <slava> want to test something?
11:57:18 <chandler> hit me
11:57:39 <slava> factor.sf.net/CFactor.zip -- see if this compiles (you can use build.sh as a guide, but the build process is basically 'compile all .c files').
11:58:07 <chandler> w00t!
11:59:43 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
12:01:15 <chandler> clearly I don't have a C compiler actually installed
12:01:20 <chandler> Compiling...
12:01:21 <chandler> Error spawning cl.exe
12:01:23 <chandler> sorry 
12:02:02 <wossname> i'll give it a shot
12:03:21 <chandler> hm, I tried to build it on my sun box
12:03:25 <chandler> Loading factor.image
12:03:25 <chandler> IMAGE ABORT: Bad magic number 202182159
12:03:39 <slava> it won't work on sun
12:03:46 <slava> oh i meant C compiler sorry :)
12:04:02 <slava> i haven't done 64-bit or big-endian support yet.
12:04:21 <chandler> oh, must be big-endian stuff
12:04:22 <chandler> *thwap*
12:04:26 <chandler> that means it won't work on my ppc at home
12:04:28 <chandler> *double thwap*
12:05:18 <slava> if you run the image file through a byteswapper it will work
12:05:18 --- join: titanstar (runehol@wirth.ping.uio.no) joined #forth
12:05:30 <slava> there's dependencies on byte order other than in the image load/save
12:05:33 <Sonarman> i wrote a byteswapper yesterday
12:05:45 <Sonarman> what a funny coincidence
12:05:48 <Sonarman> he he ha ha ha
12:05:59 <chandler> slava: works well on my i686 though!
12:08:05 <chandler> heh, I don't know factor, but this doesn't look right:
12:08:08 <chandler>     0] : foo 1 2 + ;
12:08:08 <chandler> MEMORY ABORT: vector type check 1075097619
12:08:16 <slava> :; don't work right at the moment
12:08:19 <slava> i broke something last night
12:08:25 <slava> and notice the lack of error handling, it just exits :)
12:08:31 <chandler> neat!
12:08:35 <chandler> :-)
12:08:39 <wossname> runs fine
12:08:45 <wossname> what should i type into this program
12:08:45 <slava> try entering: garbage-collection
12:08:55 <slava> wossname, its similar to forth, 2 2 + .
12:09:00 <wossname> ok
12:09:01 <slava> .s works too
12:09:14 <chandler> lots of shouting about arrays
12:09:15 <slava> anyway if you want to play with factor its best to use the java interpreter for now
12:09:20 <slava> chandler, debug output :)
12:09:34 <chandler> heh. write a CL VM for it! :-)
12:17:39 --- quit: SDO (Success)
12:59:43 --- quit: segher ("Leaving")
13:11:18 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
13:24:29 --- quit: SDO ("Vision[0.9.6-0203]: i've been blurred!")
13:24:43 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
13:41:49 --- quit: wossname ("brbrbrbreeboot")
13:42:29 --- quit: SDO ("Vision[0.9.6-0203]: i've been blurred!")
13:42:36 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
13:45:30 --- join: wossname (wossname@HSE-MTL-ppp78177.qc.sympatico.ca) joined #forth
14:02:15 --- quit: SDO (Connection timed out)
14:07:23 * Tomasu is back (gone 03:47:14)
15:03:02 --- quit: Sonarman (Read error: 110 (Connection timed out))
15:03:21 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth
15:03:25 --- mode: ChanServ set +o kc5tja
15:05:28 <slava> hi kc5tja
15:08:15 <Robert> Hi
15:10:27 --- quit: wossname (":blah:")
15:11:28 <kc5tja> Greetings.
15:12:07 <kc5tja> Still working on my patent search mini-project.  Up to 1450 patents searched so far, out of 1710.
15:14:39 <Robert> Heh
15:15:03 <kc5tja> I think the TTA is going to be the clear winner for the microprocessor architecture for my box.
15:15:28 <kc5tja> But, watch, patent #1710 in the list will be for a TTA, with my luck... :/
15:16:37 <kc5tja> I also was reading up on the history of the Macintosh computer too.
15:17:03 --- join: Sonarman (~matt@adsl-64-160-166-207.dsl.snfc21.pacbell.net) joined #forth
15:17:12 <Robert> Interesting?
15:17:14 <kc5tja> It's interesting to see how that computer evolved, and how its growth eventually brought the death of the Lisa.  But, yet, in so many ways, the Lisa still lives on in the MacOS libraries...*cough*...er..ToolBoxes.
15:17:46 <kc5tja> It is intriguing to think that the original Macintosh was intended to use a 6809 microprocessor (yep, an 8-bit CPU with 64K address space!).
15:19:19 <kc5tja> And as far as graphics, it was going to have 256x256 pixel monochrome graphics (versus 512x384).
15:19:27 <slava> haha
15:22:46 <Robert> kc5tja: What would you recommend for graphics/text display, when it comes to low-budget homebrew computers? I don't really have a TV to play with.... How about LCDs? Computer monitors?
15:23:57 --- part: tucknip left #forth
15:24:37 <kc5tja> I personally would recommend VGA.
15:25:22 <Robert> Is it somewhat easy to interface?
15:25:22 <kc5tja> VGA monitors are so ubigitous that they're dirt cheap.  And if it's a multisync (which most are today), you won't likely blow one up if you end up driving it at the wrong frequency.  At best you'll get a distorted display, or no display at all.
15:25:30 <kc5tja> Robert: It's easier to interface than a TV is.
15:25:36 <Robert> Cool!
15:25:51 <Robert> VGA monitors, that's a "usual" computer monitor, right? *clueless and ashamed of it*
15:25:56 <kc5tja> There are six signals to a VGA interface: Red, Green, Blue (all analog), HSYNC, VSYNC (digital), and ground.
15:26:10 <Sonarman> could you seriously blow up an old VGA monitor by sending it the wrong frequency?
15:27:17 <kc5tja> IBM released the VGA video card specification in 1986 or so, to compete with the Amiga's then brain-smattering graphics capability (of course, the VGA still paled in comparison).  But the VGA had some good technical advantages: 60Hz refresh rates with a 640x480 display, non-interlaced (the Amiga was still using NTSC and PAL video frame rates, which caused a lot of flicker).
15:27:46 <kc5tja> The disadvantage to using VGA is that you need a much higher video bandwidth.  It's approximately double that of NTSC or PAL.  But other than that, it's SOOO much easier to work with.
15:27:53 <kc5tja> Sonarman: Yes.
15:28:09 <kc5tja> Just as you can trivially damage a TV by feeding it video signals that were not properly in spec.
15:28:16 <Sonarman> i mean, i know you can break one, but would it really burst into flames?
15:28:33 <kc5tja> "blow up" is a colloquialism.
15:28:36 <Sonarman> ok
15:28:48 <Robert> How modern do monitors have to be to have multisync?
15:28:53 <kc5tja> No flames, no explosions, but you will likely hear fizzing, popping, and probably seeing quite a bit of smoke.
15:28:56 <Robert> I really don't want to fry any of them....
15:29:04 <kc5tja> If prolonged, it could potentially catch fire, but it's not likely.
15:29:08 <slava> when setting up linux for the first time back in 97, i enterd in the wrong sync values for my monitor and it made a really loud squealling noise
15:29:17 <slava> my i switched it off and tried again with the right values and it worked
15:29:17 <Robert> Heh.
15:29:26 <kc5tja> Robert: If a monitor displays anything beyond 640x480 resolution, it'll be multisync.
15:29:40 <Robert> Hrm, there went plan A...
15:30:00 <Robert> How tolerant are the modern monitors then?
15:30:21 <Robert> I mean, will they stand even way-off sync rates?
15:30:54 <kc5tja> Robert: Some may still make that squealing noise that Slava spoke about.  That's the horizontal flyback transformer that is working hard to keep resonance with the video signal, and failing.  But if a singal falls too far out of spec, the monitor will just turn off the display, showing "LOST SIGNAL" or whatever on the screen.
15:31:36 <Robert> Ah, cool. 
15:31:44 <kc5tja> Let's put it this way: I can't be 1000% sure of any of this.  I'm only basing this on what I've seen thus far.
15:31:53 <Robert> Sure, I won't sue you. :)
15:31:56 <kc5tja> I've never had a monitor permanently damaged by feeding it incorrect sync rates.
15:32:01 <Robert> Just wanted to know they're reasonably tolerant.
15:32:17 <kc5tja> Compared to ordinary TVs, they're battle tanks.
15:33:24 <Robert> You don't happen to have a link to some doc about the VGA standard?
15:33:26 <kc5tja> Transport triggered architecture it is.
15:33:31 <Robert> Google gives me junk.
15:33:42 <kc5tja> Out of 1710 filed patents, absolutely none have anything to do with the concept.
15:33:49 <Robert> Nice. :)
15:34:06 <kc5tja> Robert: Give me a second.
15:34:14 <kc5tja> Actually, what are you searching in Google with?
15:34:22 <Robert> "VGA specifications"
15:34:30 <Robert> Any other ideas?
15:35:34 <kc5tja> VGA timing diagram
15:36:25 <Sonarman>                 with a 640x480 display, non-interlaced (the Amiga was still 
15:36:31 <Sonarman> heh. sorry
15:37:28 <kc5tja> OK, I found a text file I can send you.  /msg me your e-mail address,
15:37:55 <Robert> Thanks :)
15:38:13 <Robert> Didn't find much there either - most stuff is about VGA -> PAL/NTSC converters...
15:39:52 <kc5tja> Of course, the alternative option is to review the timing parameters using an oscilloscope, monitoring the various control signals generated by the computer.  :D
15:40:29 <Robert> *shrug*
15:41:10 <Robert> Hm, another problem: VGA contacts look pretty odd. Can't find any on Mouser or Futurlec. Any ideas where I can get those?
15:41:35 <kc5tja> DB-15 pin connectors?
15:42:03 <kc5tja> I don't trust Futurelec -- you sent me there once before, and their selection is so limited that I don't consider them a serious distributor.
15:42:27 <Robert> They got some stuff though.
15:42:42 <Robert> I'll search for DB-15 on Mouser
15:43:34 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth
15:45:16 <Robert> Hrm.
15:45:31 <Robert> Seems like most of those have 2 rows (8+5) not 5+5+5
15:45:35 <Robert> er, 8+7
15:46:48 <kc5tja> I would have to spend some time finding them.
15:46:51 <kc5tja> I know they carry them.
15:46:59 <kc5tja> People make projects with them all the time.
15:49:08 <Robert> Yeah, I thought they'd be more common
15:49:12 <Robert> But I'll search
15:49:26 <Robert> Sure they're called DB-15?
15:49:58 <kc5tja> http://www.mouser.com/index.cfm?handler=displayproduct&lstdispproductid=253912&e_categoryid=614&e_pcodeid=57105
15:50:09 <kc5tja> HD15 Recepticle
15:50:15 <kc5tja> receptacle rather.  Sorry
15:51:14 <Robert> $11.2 each? Hrm...
15:51:24 * Robert starts thinking about taping needles together. ;)
15:51:37 <kc5tja> That's for a receptacle, not a plug.
15:54:39 <kc5tja> But, yes, that's how much I will have to pay (on average) for one of those plugs.
15:55:01 <kc5tja> Now you see why I'm having so much difficulty keeping the Kestrel within a specific budget?
15:55:06 <Robert> Yes, heh.
15:55:18 <kc5tja> (which may yet STILL not be possible, despite all the high integration I'm working towards!)
15:55:33 <Robert> How come they're so incredibly expensive? I mean, you get DSUB25s for like 50 cents...
15:55:51 <kc5tja> I can't answer that.
15:56:49 <Robert> A conspiracy...
15:57:03 <kc5tja> It could well be, I don't know.
15:57:19 <kc5tja> Maybe what I should do is use an RJ-45 connector for the video signals.
15:58:10 <kc5tja> And rely on enterprising individuals to wire up their own VGA-to-RJ45 adapter cables (or sell them myself or something).
15:58:50 <kc5tja> I mean, I already feel bad as it is not selling RAM with the computer stock.  :(
16:00:05 <Robert> You don't? Why?
16:00:09 <kc5tja> $$$
16:00:24 <Robert> Ah... You were aiming at like a MB of SRAM, right?
16:00:36 <kc5tja> Originally, yes.
16:01:34 <kc5tja> But with the new processor architecture, I see no reason not to just use the same SDRAM sticks that ordinary PCs use.
16:02:08 --- join: doublec (~doublec@coretech.co.nz) joined #forth
16:02:21 <kc5tja> This is especially the case with the TTA architecture, since a single Forth operation is now 32-bits to 64-bits long, instead of 6-bits.
16:02:47 <Robert> Heh, then you get TONS of memory, that's pretty nice.
16:03:51 * kc5tja nods
16:03:58 <kc5tja> But then again, you'll NEED it.  :)
16:05:09 <Robert> Hehe, I guess that depends on ehat you do. x86 instructions are already pretty long (in average), and you can still fit a lot in 64k.
16:05:59 <kc5tja> Robert: Most x86 opcodes average two to three bytes.  The TTA is like a traditional RISC processor, where each instruction is always a fixed size (32-bits in my case).
16:07:31 <Robert> That's just a ~2x factor.
16:08:01 <kc5tja> Yeah.
16:08:03 <kc5tja> It's not that bad.
16:09:33 <kc5tja> But consider, my current code size for the ROM image is about 930 words of memory.  512 are reserved for font data, so that leaves 418.  Assuming 3 packed instructions per instruction word on average (I don't know the real statistics, so...), moving to a TTA architecture would give a finished binary size of 5016 words consumed.
16:09:43 <kc5tja> Err...oops.
16:10:05 <kc5tja> That's 1254 words consumed (again, ignoring font data).
16:10:24 <kc5tja> That's a 3x increase in code size.
16:10:55 <Robert> Yeah. :/
16:10:59 <kc5tja> Of course, this isn't going to be that universal; once a program is of a certain size, it'll grow less sharply (due to the heavy amount of code factoring; subroutine calls, for example, are *already* one full word in size anyway).
16:11:33 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
16:11:38 <kc5tja> re SDO
16:11:56 <kc5tja> SDO: Good (partial) news -- no patent issues should be encountered with the use of a transport triggered processor architecture.
16:12:09 <SDO> howdie
16:12:34 <SDO> great
16:12:38 <SDO> this is good good news.
16:12:44 <kc5tja> I just scanned 1710 patents and only two out of that set are pertinent to VLIW architectures; and none of them relate to TTAs at all.
16:13:06 <kc5tja> It will, however, require a fundamental change in how I build the computer . . . again.  :D
16:13:17 <kc5tja> And then I started reading on the history of the Apple Macintosh.  I don't feel so bad.
16:13:32 <kc5tja> Mac had to go through five major hardware revisions before it finally settled on its launched design.
16:13:57 <zardon> Kestrel will be a TTA?
16:14:54 <kc5tja> zardon: Since I discovered a patent violation, it looks like that will be the case.  Unless Chuck's IP holding company can pull through with a pardon in my favor, I will be switching to a stack-optimized TTA.
16:15:10 <zardon> thats cool
16:15:25 <zardon> TTA is interesting
16:15:29 --- quit: qFox ("meh ircii's netsplit detection appearantly sux ^^")
16:15:32 <kc5tja> zardon: My plans are for a two-bus, 32-bit wide TTA (e.g., two moves per instruction).
16:16:32 <kc5tja> However, I also am looking at a 3-bus system too.
16:16:48 <zardon> why not just go ternary :)
16:17:09 <kc5tja> zardon: Because then I only have 32 registers (instead of 128) that I can work with.
16:18:08 <zardon> i was refering to a balanced trinary system, versus binary
16:18:14 <kc5tja> Besides, it shouldn't really matter -- each generation of the CPU is likely to not be compatible with its predecessor anyway (since TTAs whole point is to expose the architecture as much as possible), so you'll be dependent on the compiler for most things anyway.
16:18:14 <zardon> to really make it interisting
16:18:33 <kc5tja> Because interesting as it may be, it's patently not useful.
16:18:39 --- quit: warpzero (Remote closed the connection)
16:19:39 <kc5tja> And it'd require twice the chip real-estate to boot, since all chips made today are binary.
16:19:44 <kc5tja> But whatever.  i won't continue with this.
16:21:16 <kc5tja> Anyway, the interesting thing about the TTA architecture is that, upon initial investigation, it looks like it's actually easier and faster to build the TTA to interpret direct threaded representation of software than it is to interpret regular program code.
16:22:59 <Robert> Neat!
16:23:14 <kc5tja> I actually did not expect this to happen.  But if true, then this chip will be far better suited to processing higher level languages than almost any other chip on the planet, except the PowerPC.
16:26:05 <kc5tja> The "NEXT" prolog could be written like this: (w := next_t ; pc := next_t ), which in a single cycle, causes a jump to the next Forth primitive, while concurrently loading the W register (which makes CREATE/DOES> words very easy to support).
16:26:32 <kc5tja> next_t would, when read from, fetch the next virtual instruction (pointed to by the VIP register), and post-increment VIP after the fetch.
16:29:49 <kc5tja> In contrast, a cycle would be required for a subroutine call, and another cycle would be required for a subroutine return.
16:34:22 --- quit: SDO (Connection timed out)
16:35:21 --- join: foobar_ (fooboo@dialup-4.152.156.96.Dial1.Atlanta1.Level3.net) joined #forth
16:42:18 --- quit: Tomasu (Read error: 60 (Operation timed out))
16:46:37 <zardon> why do you say that PowerPC is so well suited to HLLs?
16:54:08 --- join: Tomasu (~moose@S010600045a4c73cc.ed.shawcable.net) joined #forth
16:56:39 <kc5tja> Because it can do VERY fast subroutine calls.
16:56:55 <kc5tja> Two cycle overhead best-case (one for call, one for return).
16:57:32 * kc5tja points to HerkForth -- the other day Herkamire compared his Forth to GCC; and it beat it by one second in a recursive fibonacci benchmark.
16:58:15 <kc5tja> So if you combine lots of execution units plus very low overhead subroutine calls, you get a system that is extremely agile.
16:58:22 <zardon> yeah
17:00:28 <zardon> but non-leaf subroutines are another matter
17:03:27 <kc5tja> OK, I think I have settled on a way to not use threaded interpretation, and be very fast.
17:03:27 <kc5tja> But we'll see.  I need to test my theories by simulation.
17:03:27 <kc5tja> I'll modify my current MachineForth compiler for the FTS1001 so that it puts out code for a hypothetical TTA optimized for MachineForth.
17:03:27 <kc5tja> But first, I need to get some food, work on some financial issues, and I'll be back in a bit.
17:03:27 <kc5tja> yeah, that's why I'd *rather* stick with a system where each TTA instruction maps to a single Forth primitive.
17:05:14 <kc5tja> Anyway, I'll be back in a bit.
17:06:35 --- quit: ChanServ (bear.freenode.net irc.freenode.net)
17:07:32 --- join: jc (~jcw@65.3.39.49) joined #forth
17:07:36 --- part: jc left #forth
17:08:54 --- join: ChanServ (ChanServ@services.) joined #forth
17:08:54 --- mode: irc.freenode.net set +o ChanServ
17:10:49 --- quit: Research (Remote closed the connection)
17:22:54 <aum> shit o fucking hell - GTK+ is **BLOATWARE** !!!!
17:24:49 <aum> distributing a windows version of a forth app using GTK+ is 'fun' - 160k for the forth app, 6+MB for all the DLLs
17:26:17 <foobar_> ugh
17:32:13 <kc5tja> Heheh :D
17:35:40 --- join: Sonarman_ (~matt@adsl-66-124-254-111.dsl.snfc21.pacbell.net) joined #forth
17:39:49 --- quit: Sonarman (Read error: 110 (Connection timed out))
17:40:10 --- nick: Sonarman_ -> Sonarman
17:57:14 --- join: LOOP-HOG (~jdamisch@sub22-119.member.dsl-only.net) joined #forth
17:59:45 <LOOP-HOG> hi
18:00:45 <Robert> Hi LOOP-HOG 
18:01:23 <LOOP-HOG> anything new?
18:02:47 <Robert> Not from me, but kc5tja discovered his new CPU design *won't* cause people to sue him.
18:03:06 <LOOP-HOG> thats nice
18:11:30 <LOOP-HOG> i'm selling a few games on ebay
18:12:09 <Robert> What kind of games?
18:12:21 <LOOP-HOG> shooters
18:12:26 <Robert> For PC?
18:12:31 <LOOP-HOG> arcade shooter for sega saturn
18:12:59 <LOOP-HOG> i'm trying to get up enough money for a digital camera and a domain name and parking space
18:17:30 <LOOP-HOG> i'm hoping for like $250, but i'd be lucky
18:18:14 <LOOP-HOG> both titles have sold before on ebay for $180 and $70 respectively
18:18:24 <LOOP-HOG> keep my fingers crossed ;^)
18:19:40 <kc5tja> Robert: Correction -- not LIKELY to cause people to sue me.
18:19:50 <LOOP-HOG> whats not in it?
18:19:54 <kc5tja> Just because I haven't found a patent with my feeble attempts, I'm not a patent lawyer.
18:20:01 <LOOP-HOG> if somebody wants it bad enough they will sue
18:20:28 <LOOP-HOG> people sue over dumb crap all of the time
18:20:30 <kc5tja> LOOP-HOG: They can't sue, because there is prior art.
18:20:41 <kc5tja> Well, they CAN sue, but they won't win.
18:20:52 <kc5tja> But that's the ticket, if they do sue, it'll be to just eliminate me from the market.
18:21:05 <kc5tja> But realistically speaking, I don't see that happening.  I'm too small a player.
18:21:19 <LOOP-HOG> ok
18:21:29 <kc5tja> And if I ever do get to be a really big player, I'll have the resources to file for a patent myself.
18:22:28 <kc5tja> I will probably also be making a papier mache case for my computer too.
18:22:30 <kc5tja> brb -- phone
18:22:47 <LOOP-HOG> k
18:29:17 <kc5tja> back
18:29:31 <LOOP-HOG> i'm up to nerdly doings myself
18:29:44 <kc5tja> One thing that fascinated me with respect to the transport triggered architecture is that the circuit may be so simple that I could possibly get away with an asynchronous design, right from the get-go.
18:30:03 <LOOP-HOG> i'm going to have a booth at the local science muesum this saturday
18:30:19 <LOOP-HOG> kc, facinating
18:30:45 <Robert> LOOP-HOG: $180 for a GAME?
18:30:54 <LOOP-HOG> Radiant Silvergun
18:30:58 <kc5tja> If I do implement this as an asynchronous design, then there stands to reason that system performance may be rather substantial -- somewhere at least in the 40 to 80 MIPS region.
18:30:58 <LOOP-HOG> yes
18:31:12 <LOOP-HOG> kc, i'll take it
18:31:33 <LOOP-HOG> the best shooter ever made
18:31:41 <kc5tja> The only thing is, if I go that fast that soon, I'll have nothing to upgrade to in later models.  :)
18:31:45 <LOOP-HOG> either that or Tempest 2000, either or
18:31:57 <kc5tja> Dude, Tempest is the ultimate shooting game.
18:31:59 <LOOP-HOG> kc, thats not a problem around here
18:32:03 <kc5tja> Haven't played Tempest 2000 though.
18:32:10 <LOOP-HOG> Atari Jaguar
18:32:27 <LOOP-HOG> i'
18:32:44 <kc5tja> Well, I figure if people need more performance, then it's time to look not towards higher speed instruction execution, but to increased levels of parallelism, either from instructions OR from multiple microprocessors.
18:33:02 <LOOP-HOG> i'm trying to get enough money so i can get a digital camera to take pics at the muesum, and then i was thinking about getting a pow-wow account for a space links site
18:34:30 <LOOP-HOG> kc, the upgrade can just include a coprocessor comunication bus or something, you can stack cpus
18:34:38 <LOOP-HOG> Transputer
18:41:45 <kc5tja> LOOP-HOG: No, it's not so simple.
18:41:53 <kc5tja> The more CPUs, the more bandwidth to main memory you need.
18:42:14 <LOOP-HOG> i though that each cpu had its own chunck of memory
18:43:14 --- quit: tathi ("leaving")
18:43:25 <kc5tja> Not sure about the Transputer, but what you're describing with separate memory can just be had by buying more than one Kestrel and linking them together.
18:46:00 <kc5tja> But, we shall see.  It's best to start off with a reasonable machine first, and then worry about future details later on.
18:46:11 <kc5tja> there will be at least one upgrade, though, and that is a machine with expansion slots.
18:46:30 <LOOP-HOG> i'll probably take one of those too
18:46:30 * fridge got a new job
18:46:32 <kc5tja> (although I personally probably won't use it, I'll at least develop one and market it for those who would like one)
18:46:36 <fridge> YAY
18:46:37 <LOOP-HOG> fridge, yha
18:46:37 <kc5tja> Congrats fridge!
18:46:53 <LOOP-HOG> unless its a horriable job, then well, just stick with it you need the $$
18:46:56 <LOOP-HOG> ;^)
18:46:59 <fridge> no more silly upside down hours for me
18:47:30 <kc5tja> fridge: History shows that unexpandable machines (or, more accurately, machines without slots) generally actually sell more than those that do.  iMac was a huge success, the Amiga 500 outsold the A1000 and A2000 combined, etc.
18:47:46 <fridge> ahh, to have a semblance of a sleeping pattern
18:47:48 <kc5tja> and let's not forget the Commodore 64.  22 million machines, still the most successful computer in all of computer history.
18:47:49 <LOOP-HOG> can't you have an expandable unexpandable
18:47:53 <LOOP-HOG> SegaCD
18:48:21 <kc5tja> fridge: I have no sleep schedule anymore.
18:49:12 <LOOP-HOG> people, just sleep when you can :^)
18:57:38 --- part: foobar_ left #forth
19:05:17 <LOOP-HOG> i'm probably going to goto powwow in about an hour and get that site
19:05:27 <LOOP-HOG> unless anybody can think of a better deal
19:11:12 <kc5tja> I have no idea what your goals are, so...
19:11:18 <kc5tja> I don't even know what powwow is.
19:11:40 <LOOP-HOG> $7.77 / month web server
19:11:57 <kc5tja> Ahh
19:11:57 <LOOP-HOG> $95 for 1 year with a domain name thrown in
19:12:08 <LOOP-HOG> this would be for a space links site that i would
19:12:24 <LOOP-HOG> advertise at the science muesum in relation to my X-Prize display
19:18:28 <kc5tja> BTW, what do you mean by "You can't have an expandable unexpandable?"
19:19:20 <Sonarman> he was asking if you could. he said "can't you"
19:19:40 <kc5tja> Oh, Lysdexia.
19:19:58 <kc5tja> Well, you can, but only to a limited extent.
19:20:56 <Sonarman> are yous till planning on having a "user port"?
19:42:46 <Robert> Ah... catch/throw working. Now I can finally get some rest ;)
19:43:45 <kc5tja> Sonarman: Yes, absolutely!
19:44:13 <kc5tja> A parallel port with at least 16 user I/O pins, plus a serial expansion bus in the same spirit as the Commodore IEC serial bus, only a lot faster.
19:48:47 <Sonarman> cool
19:49:46 <Robert> A/D / D/A?
19:51:49 <kc5tja> I don't think I'll have any A/D converters.
19:52:00 <kc5tja> But it should have at least two DACs, for the audio output.
19:54:20 * Robert wants to play a bit with the A/D on the PIC12F675 chips. Maybe it's easier to do data transfer over radio that way than with the receiver being responsible to feed the chip a square wave.
20:00:18 <kc5tja> I'm planning on building a circuit that lets the Kestrel act as a software-controlled, successive approximation ADC, by re-using one of its DACs.
20:00:33 <kc5tja> Of course, you can't play audio while this is happening.  :D
20:01:12 --- quit: doublec ("Leaving")
20:01:47 <kc5tja> Anyone know of an affordable, 16-bit DAC that can be used, that hopefully won't suck up too much board space?
20:02:01 <Robert> Or maybe the user could just attach a small A/D to the user port... Those PICs aren't really expensive.
20:02:08 * Robert has a few samples.
20:02:10 * kc5tja nods
20:02:14 <kc5tja> PICs are pretty cheap.
20:02:35 <kc5tja> Actually, what is the resolution on those PICs?
20:03:14 <Robert> 10 bits.
20:03:18 <kc5tja> hmmm
20:04:06 <kc5tja> Do they also have a DAC?  If so, what resolution is the DAC?
20:04:19 <Robert> Sample time is ~18µs.
20:04:31 <Robert> No DAC. :(
20:04:46 <kc5tja> How many ADC channels?
20:04:57 <Robert> 4
20:05:02 <kc5tja> !!!!!!!
20:05:08 <kc5tja> I didn't expect it to have that many.
20:05:10 <kc5tja> Impressiv.e
20:06:05 <Robert> Not bad for stuff they're giving away for free...
20:06:07 <LOOP-HOG> i'll probably need to sell my vr-goggles too
20:06:13 <Robert> :/
20:06:25 <kc5tja> Robert: How many pins on the package?  And is it a DIP?
20:06:30 <Robert> 8-pin DIP.
20:06:41 <kc5tja> Egads.
20:06:51 <kc5tja> I might consider adding a number of ADCs then.  :D
20:07:06 <Robert> Oh, yay. :)
20:07:22 <kc5tja> Robert: However, that leaves only two pins left for other I/O to those chips!
20:08:26 <Robert> Using an external oscillator?
20:08:38 <kc5tja> Robert: I might.
20:08:49 <kc5tja> That means only two of the four ADCs would be usable.
20:08:56 <Robert> I think the internal ones are +/- 1%, good enough accuracy?
20:09:23 <kc5tja> If they are that good, why would anyone want to use an external oscillator?
20:09:49 <Robert> Well, if you'd like to use a 3.6864 xtal for serial port use...
20:10:10 <Robert> Or simply need accurate timing, like in the frequency counter...
20:10:56 <kc5tja> Or, I could just use a bigger PIC chip.
20:11:01 <Robert> Yep.
20:11:31 <LOOP-HOG> i'm not going to use powweb, because in the customer directory every customer website is index.php and none of them come up
20:11:53 <kc5tja> heh
20:12:36 <LOOP-HOG> how about easyspace.com
20:14:01 <LOOP-HOG> $45.00 / yr   no php, no cgi, no aisp, free domain name...just what i wanted
20:14:34 <Robert> Transfer limits?
20:22:22 <LOOP-HOG>  lunaroutfitters.com
20:23:01 <kc5tja> Heh, sounds like a camping store, that specializes in lunar-grade camping gear.  :D
20:23:53 <LOOP-HOG> you can get like fusion reactors and control systems for life support 
20:23:58 * kc5tja proceeds to work on finishing up the cursor scrolling stuff before releasing 0.1.9.
20:23:58 <LOOP-HOG> and rocket engines
20:25:43 <LOOP-HOG> all they say is "free bandwidth"
20:40:42 <kc5tja> Don't expect a lot of support, and certainly don't expect it to have 99.9999% up-time either.  :D
20:42:35 <LOOP-HOG> no
20:43:04 <LOOP-HOG> well i hope that domain name i chose wasn't too cheesey.
20:47:04 <LOOP-HOG> if nothing comes of it, thats less than half of what i thought i was going to pay
20:47:48 <LOOP-HOG> $3.75/mo
20:52:07 <Sonarman> kc5tja: just out of curiosity, did you write the "SRAM for hobbyists" sction of the wikipedia SRAM article?
20:56:12 <Sonarman> kc5tja: what happened to your patent-skirting idea of having a pre-decoder in the processor that would translate packed forth instructions into TTA instructions?
20:56:30 <kc5tja> SRAM for Hobbyists??
20:57:48 <Sonarman> nevermind, i just asked because it looked like your style of writing :)
20:58:39 <kc5tja> As far as the pre-translator is concerned, that may still be an option, BUT, I want to start with the basics first.
20:59:07 <Sonarman> Ok
20:59:17 <kc5tja> GOD DAMN THIS THING.  All I did was add code to bump the cursor, and now it freakin crashes.
20:59:21 <kc5tja> I hate crap like this.
21:00:03 --- join: ball (~ball@dialup-4.158.0.23.Dial1.Chicago1.Level3.net) joined #forth
21:00:25 <ball> LOOP-HOG: did you look at null.html yet?
21:07:05 <LOOP-HOG> i fixed and uploaded the fixes for all pages to both servers
21:07:16 --- quit: kc5tja (bear.freenode.net irc.freenode.net)
21:07:16 --- quit: titanstar (bear.freenode.net irc.freenode.net)
21:07:16 --- quit: ianp (bear.freenode.net irc.freenode.net)
21:07:16 --- quit: ball (bear.freenode.net irc.freenode.net)
21:07:16 --- quit: lalalim (bear.freenode.net irc.freenode.net)
21:07:17 --- quit: cmeme (bear.freenode.net irc.freenode.net)
21:07:29 --- join: lalalim (~lalalim@217.94.174.32) joined #forth
21:07:29 <LOOP-HOG> <TABLE ALIGN=LEFT>   evil
21:07:39 <zardon> hi lalalim
21:07:41 --- join: titanstar (runehol@wirth.ping.uio.no) joined #forth
21:07:41 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth
21:07:41 --- join: ianp (ian@inpuj.net) joined #forth
21:07:41 --- join: ball (~ball@dialup-4.158.0.23.Dial1.Chicago1.Level3.net) joined #forth
21:08:21 <ball> LOOP-HOG: did you look at the one I did though?
21:08:43 <LOOP-HOG> yes i did
21:09:06 <ball> Did it help?
21:09:13 --- join: cmeme (~cmeme@216.184.11.30.unused.swcp.com) joined #forth
21:09:15 <LOOP-HOG> yes thanks
21:09:34 <ball> Great
21:09:36 <ball> He cmeme
21:09:51 <ball> LOOP-HOG: be sure to check out http://www.w3c.org/
21:10:17 <LOOP-HOG> i think i bookmarked it, thanks
21:10:27 <LOOP-HOG> definitely a link worth having
21:10:52 --- join: SDO (~SDO@67-23-111-213.clspco.adelphia.net) joined #forth
21:10:54 <ball> They define the official standards, and publish them on that site.
21:10:56 <ball> Hi SDO
21:11:26 <SDO> hello
21:14:52 <kc5tja> re
21:16:57 --- mode: ChanServ set +o kc5tja
21:21:01 * Tomasu is away: night
21:21:10 --- nick: Tomasu -> TomasuDlrrp
21:22:53 <kc5tja> Damn, I think I'm overflowing my return stack.
21:24:08 --- quit: aum ()
21:25:39 <LOOP-HOG> why is that, some low level word now crashing your system every time you execute
21:25:51 <LOOP-HOG> funney when you clobber your own words that way
21:27:28 <kc5tja> Yeah.
21:27:41 <kc5tja> I'm not at all sure why it's happening.
21:28:05 * ball blames BPL
21:29:37 <LOOP-HOG> when you decompile the word it goes into a long decompile
21:29:39 <LOOP-HOG> and its bunk
21:29:54 <LOOP-HOG> maybe you have an array or string that is overflowing
21:29:59 <LOOP-HOG> do you have any auxilary stacks?
21:30:06 <LOOP-HOG> what else could overflow?
21:30:39 --- quit: SDO (Read error: 110 (Connection timed out))
21:31:58 <ball> The Thames
21:32:16 <kc5tja> LOOP-HOG: No arrays.  No strings.  No other stacks.
21:32:30 <kc5tja> And each word is kept very, very short.
21:32:43 <kc5tja> But because I factor so much, it's easy to overflow the return stack.
21:32:57 <kc5tja> There is one easy way to find out though.  (Change the emulator!)
21:32:59 <LOOP-HOG> how deep is your return stack?
21:33:14 <kc5tja> 16
21:35:36 <kc5tja> OK, maybe I DIDN'T overflow the return stack.
21:35:40 <kc5tja> This is MOST bizarre.
21:38:09 <kc5tja> Fixed it.
21:38:30 <LOOP-HOG> guud
21:38:37 <kc5tja> I had '; instead of ;', which caused the definition of the word to fall into the next word, which just called the previous word.  Thus, an endless loop.
21:39:14 <LOOP-HOG> they look much alike
21:39:39 <kc5tja> I type fast too.
21:39:40 <kc5tja>  :D
21:39:49 <LOOP-HOG> sometimes i come up with names like that too
21:39:58 <LOOP-HOG> ">
21:40:03 <LOOP-HOG> output to file
21:40:11 <LOOP-HOG> ">' output to file with space, or something like that
21:45:31 <kc5tja> OK, got CR, EMIT, and TYPE all scrolling properly.
21:46:12 <kc5tja> Time to FINALLY bundle it up, and release it.
21:46:21 <kc5tja> Oh, wait, I have to include the changes from Sonarman first...
21:46:26 <ball> kc5tja: is that on your own platform?
21:46:32 <kc5tja> ball: My own platform??
21:47:24 <ball> kc5tja: the one we discussed the other day
21:47:30 <kc5tja> This is on the Kestrel.
21:48:24 <ball> Aye, that's the one.
21:48:57 <kc5tja> However, the software ought to be fairly easy to port.  (Despite the hard-coded constants for now.  I'll get to those in a later refactoring.)
21:49:49 <ball> heh, yeah, weed those out ;-)
21:49:53 <kc5tja> With about two or three months of effort, I suppose this can be brought up on an Intel- or PowerPC-based computer relatively easily.
21:53:26 <ball> Nice.
21:58:35 <kc5tja> Sonarman: Mind if I use your real name in the list of credits for your contribution?  /msg me if you don't mind.
22:14:45 --- join: O3BEPH (~z@212.34.52.140) joined #forth
22:15:17 <ball> Hell O3BEPH
22:15:24 --- nick: O3BEPH -> Serg
22:15:28 <Serg> hi Ball
22:15:51 <Serg> chatting on several nets w/ MIRC is difficult ;))
22:16:05 <LOOP-HOG> alot of folks around here have names like droids, never mind that
22:16:18 <LOOP-HOG> it's all good here
22:16:19 <Serg> no-no
22:16:25 <LOOP-HOG> :^)
22:16:38 <Serg> that nick was for RU and is meaningf00l ;))
22:17:27 * ball is also KB9YLW ;-)
22:17:41 <LOOP-HOG> radio, duh
22:17:49 <Serg> it means "berserk potion", slang name for quad damage powerup in shooter games
22:17:50 <LOOP-HOG> got pj in the earss
22:18:32 <LOOP-HOG> and small furry cyborgs
22:18:41 <kc5tja> LOOP-HOG: I don't think pajamas are supposed to be worn in the ears.  :D
22:18:54 <LOOP-HOG> lol
22:18:56 <LOOP-HOG> nm
22:19:30 <ball> Some sleepers benefit from ear plugs ;-)
22:23:04 <Serg> bananas in ears ?
22:24:38 <ball> Not in my ears
22:27:01 --- join: lalalim_ (~lalalim@p508AA4C2.dip.t-dialin.net) joined #forth
22:29:08 <ball> Hello lalalim
22:40:51 --- quit: lalalim (Read error: 110 (Connection timed out))
22:56:37 --- quit: LOOP-HOG ()
22:58:35 * ball thinks about setting fire to his study
23:00:27 --- quit: slava ("Leaving")
23:29:57 --- join: Sonarman_ (~matt@adsl-64-171-254-146.dsl.snfc21.pacbell.net) joined #forth
23:30:24 <ball> wb Sonarman_
23:30:57 <ball> I'm off to bed anyway
23:31:06 --- part: ball left #forth
23:32:07 --- quit: Sonarman (Nick collision from services.)
23:32:12 --- nick: Sonarman_ -> Sonarman
23:32:39 <kc5tja> Sonarman: Just so you know, Herkamire is working on a Kestrel emulator, that is written in C.
23:32:58 <kc5tja> Anyway, I'm going to be off to bed here in a bit.  Just need to finish the upload of the new site.
23:34:29 <Sonarman> cool
23:37:46 <kc5tja> New site uploaded.
23:37:53 <kc5tja> 4th screen shot is linked in.
23:37:59 <kc5tja> 0.1.9 is now available for global download.
23:38:00 <kc5tja> :)
23:38:49 <kc5tja> I haven't linked in Herkamire's code because, while it compiles on my box, it doesn't run (it chokes).  So I have to debug it a bit.  Besides, I'd rather it be a separate package anyway.  Especially with me doing development work on the TTA architecture.
23:39:10 <kc5tja> And now, since I have to work tomorrow, I'm off to bed.
23:39:17 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK")
23:40:39 --- quit: Sonarman ("leaving")
23:59:59 --- log: ended forth/04.07.12