00:00:00 --- log: started forth/08.04.22
00:44:02 --- join: ecraven (n=nex@plc31-103.linzag.net) joined #forth
00:59:47 --- quit: proteusguy (Connection timed out)
01:00:40 --- join: proteusguy (n=proteusg@61.7.144.97) joined #forth
01:10:16 --- quit: arke (Read error: 110 (Connection timed out))
02:26:36 --- quit: nighty^ (Remote closed the connection)
03:42:27 --- quit: ecraven ("bbl")
05:05:31 --- quit: snoopy_1611 (Read error: 131 (Connection reset by peer))
05:10:22 --- quit: Snoopy42 (Read error: 145 (Connection timed out))
05:20:50 --- join: ecraven (n=nex@plc31-103.linzag.net) joined #forth
08:15:40 --- quit: Quartus` ("used jmIrc")
08:15:56 --- join: Quartus` (n=Quartus`@205.205.50.2) joined #forth
09:04:08 --- join: JasonWoof (n=jason@c-65-96-165-155.hsd1.ma.comcast.net) joined #forth
09:04:08 --- mode: ChanServ set +o JasonWoof
10:25:41 --- join: ygrek (i=user@gateway/tor/x-08a889b91ea26c16) joined #forth
10:41:35 --- join: forther (n=forther@207.47.34.100.static.nextweb.net) joined #forth
11:31:52 --- join: Snoopy42 (i=snoopy_1@dslb-088-068-212-174.pools.arcor-ip.net) joined #forth
11:33:38 --- join: Maki (n=Maki@adsl-224-84.eunet.yu) joined #forth
11:46:43 --- join: arke (n=arke@p57A755E5.dip.t-dialin.net) joined #forth
11:46:44 --- mode: ChanServ set +o arke
12:24:09 <Deformati> There is no way to do circular function calls in forth, is there?
12:24:27 <slava> what's a circular function call
12:24:54 <Deformati> : foo ... bar ;  :  bar ... foo ;
12:25:03 <slava> defer bar
12:25:07 <slava> : foo ... bar ;
12:25:10 <slava> : bar ... foo ;
12:25:14 <slava> err
12:25:18 <slava> : (bar) ... foo ;
12:25:21 <slava> ' bar is (bar)
12:25:31 <slava> or maybe its ' (bar) is bar
12:25:33 <slava> Quartus` knows
12:26:52 <gnomon> You can do it via vectored execution, or - if you're crazy - by backpatching your compiled code.
12:38:59 --- quit: madgarden (Read error: 110 (Connection timed out))
12:43:49 <Fractal> ya you prolly want vectored execution: http://www.forth.com/starting-forth/sf9/sf9.html (scroll down a page)
12:46:06 <Fractal> or what gnomon said.. something like [ here 0 , ]  and then later use ! to store the thread of the word to that location
12:46:43 <Fractal> (later being before you run that word obviously)
12:53:32 --- join: TreyB_ (n=trey@74.203.168.157) joined #forth
12:55:03 <Fractal> so would be like  : foo ... [ here 0 , ] ;  : bar [ here ] ... ;  swap !    (**untested**)
13:01:13 <Fractal> but ya use vectored exec unless the indirection is too inefficient
13:01:37 <slava> highly doubtful that it will be too inefficient with defer
13:02:02 <Fractal> ya prolly not.. keep it outta inner loopz etc
13:10:24 <Fractal> also if you need to do this many times, vectors are less space efficient
13:10:39 <slava> if your target has 4kb of ram, maybe
13:11:00 --- quit: TreyB (Read error: 110 (Connection timed out))
13:46:12 <JasonWoof> don't use ! to patch a call
13:46:27 <JasonWoof> isn't there a standard word like RESOLVE
13:46:42 <Quartus> yeah, you've got a really unreliable method going there; might work on some erector-set Forth, but it's hardly portable
13:47:02 <Fractal> ya will depend on thread model
13:47:03 <JasonWoof> slava's method is what you want
13:47:09 <JasonWoof> defer bar
13:47:14 <JasonWoof> : foo ... bar ;
13:47:26 <slava> Quartus`: if you're concerned abotu the overhead of a deferred call you've alredy started off on the wrong track i think
13:47:29 <JasonWoof> : real-bar ... foo ; ' real-bar is bar
13:47:50 <Quartus> overhead? right.  I don't disagree.
13:48:52 <Fractal> also will break if you use it in a loop  : foo begin [ here 0 , ] again ;
13:49:14 <JasonWoof> If you're trying to optomize to the point where a defered word would matter, you'd probably do better to just inline the whole thing
13:54:01 <Quartus> Fractal, it's a bad technique front-to-back.
13:55:55 <Fractal> quartus - for this application i'd prolly agree.. tho i'm hesitant to call the technique unconditionally bad.. threads give you lots of opportunities not avail in other langs
13:56:56 <Quartus> if you mean a specific kind of threaded implementation, even then there are better methods
13:57:48 <Quartus> Better than stuffing placeholder values into memory and tagging them up later.  Semantically meaningful ways, even.
13:58:20 <Fractal> well.. the technique should work with most types of threading I think.. you might have to use something other than ! tho
13:59:11 <Fractal> what type(s) of threading does quartus forth use?
13:59:21 <Fractal> and no offence but does anybody use palms anymore? :)
13:59:23 <Quartus> it's not portable; it's not readable; there are better alternatives.
13:59:33 <Quartus> Quartus Forth is a native-code generating system.
13:59:48 <Fractal> nice.. do you optimise?
13:59:55 <Quartus> To a point.
14:00:32 <Fractal> should try to dig up my IIIxe
14:03:27 --- quit: ygrek (Remote closed the connection)
14:04:55 <Fractal> prolly still has quartus installed :)
14:08:47 <Quartus> :)
14:09:01 * slava is fixing up the rss library
14:27:22 --- quit: Maki ("Leaving")
14:36:24 --- quit: forther ("Leaving")
15:04:41 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth
15:04:42 --- mode: ChanServ set +o tathi
15:16:04 --- quit: ecraven ("bbl")
15:48:44 <Fractal> so.. you guys don't like relying on indirect/direct threading.. any thoughts on trans-word programming like  : foo ... if ;  : bar then ... ;   (if conditional is false, branch to bar (not call it), with bar as an additional entry point)
15:48:57 <Fractal> ?
15:49:08 <slava> that's unreadable and uselesss
15:49:17 <Quartus> non-portable, confusing, and pointless.  A parlour trick without any value.
15:49:40 <Fractal> what are the portability restrictions?
15:50:05 <slava> its disallowed by the standard
15:50:11 <Quartus> It is indeed.
15:50:26 <Fractal> ANS?
15:51:10 <Quartus> that's the one.
15:53:50 <Fractal> what is rationale for disallowing it? so ; can use control stack?
15:54:08 <tathi> why would you want to do something like that?
15:54:09 <Quartus> The colon-sys and the control stack may both use the data stack.
15:54:18 <Quartus> Yes, why?  It's daft.
15:56:26 <Fractal> just curious.. i have used this and similar techniques before, but obviously not in portable code
15:56:48 <Quartus> Not sure who you're trying to impress with peculiar technique, but it's very poor practice.
15:56:52 <tathi> Used how?  For what?
15:57:28 <Fractal> for sharing parts of threads between words
15:57:36 <slava> that's why you define a new word
15:57:44 <tathi> define "thread"?
15:57:46 <Quartus> It's called 'factoring'.  
15:57:59 <slava> i hope this wasn't in real code. otherwise i pity your coworkers
15:57:59 <Quartus> He's using 'thread' to mean a sequence of compiled code in a threaded forth, tathi.  Confusing.
15:58:06 <tathi> ah, ok
15:58:18 <tathi> That was my first guess, but I wasn't sure.
15:58:18 <Fractal> sry what do you guys call that?
15:58:35 <slava> if you want to share a part of a word, define a new word with the shared part
15:58:45 <Quartus> Factoring.
16:07:52 <slava> that's what i do for a living :)
16:08:00 <Quartus> :)
16:23:19 <Fractal> anyways gotta run. thanks for your comments everyone! (PS. "unreadable" and "useless" are what python programmers say about lambda ;) peac30utz++
16:26:23 <slava> yeah, but you probably don't know python, lisp or forth
16:47:49 --- join: madgarden (n=madgarde@bas2-kitchener06-1096751243.dsl.bell.ca) joined #forth
16:56:04 --- quit: madgarden (Read error: 104 (Connection reset by peer))
17:43:33 --- join: madgarden (n=madgarde@bas2-kitchener06-1096751243.dsl.bell.ca) joined #forth
18:09:03 --- join: nighty^ (n=nighty@210.188.173.246) joined #forth
19:48:58 --- quit: JasonWoof ("bed")
21:56:00 --- quit: madwork (Read error: 110 (Connection timed out))
22:10:14 --- join: Aardappel (n=Aardappe@cpe-24-175-70-30.tx.res.rr.com) joined #forth
22:15:27 <Aardappel> so
22:15:55 <Aardappel> I would say the hardest thing for non-forth programmers when they try forth is mentally picturing the stack as you read through code
22:16:19 <slava> it takes some practice to read stack code but then its pretty easy once you get accustomed to it
22:16:20 <Aardappel> has there been any tool (editor? debugger?..) that helps with this somehow? visualizes it?
22:16:32 <Aardappel> sure yeah
22:16:38 <slava> i have a tool which checks stack effect declarations for validity
22:16:54 <Aardappel> though I must say with a lot of dups & swaps its easy to lose track
22:17:03 <Quartus> Then don't do so many.
22:17:07 <slava> dups and swaps are pretty simple
22:17:11 <Aardappel> haha
22:17:35 <Quartus> I'm serious.
22:17:36 <slava> when you start having too much rot, tuck and so on it can be a problem and i try to simplify code or use locals if it becomes an issue
22:21:28 <slava> Aardappel: also debuggers which step through code and show the stack at each point are very common
22:22:24 <Aardappel> ok
22:23:16 <Aardappel> any editors? I guess they would have to do some analysis to know the shape of the stack at a particular point
22:23:34 <slava> i'm planning on working on that at some point
22:26:12 <Aardappel> cool
22:26:31 <Aardappel> do you work on a particular forth implementation or what
22:30:03 <slava> both Quartus and I have our own implementations :)
22:30:23 <Quartus> slava has his own language that looks a little like Forth if you kind of squint
22:30:28 <slava> Quartus sells his for money and I use mine at work
22:30:52 <Quartus> I give away a version, too.
22:32:35 <Aardappel> cool
22:32:48 <Aardappel> I've written a bunch of compilers
22:33:13 <Aardappel> looking at making a stack based language with a more visual environment next
22:33:18 <Quartus> thank you
22:34:48 <Aardappel> thats funny... looking on your site... I think I may have actually tried your forth waaay back on my palm
22:35:52 <Quartus> I see.
22:36:02 <slava> how visual of an environment are you looking for?
22:37:15 <Aardappel> very visual
22:37:33 <Aardappel> it would still be forth-like in the sense that it is a sequence of stack ops
22:38:18 <Aardappel> but other than that, its a mouse driven editing environment, that uses lines to help view what the stack looks like at any point
22:40:49 <slava> here's a pretty old screenshot: http://factorcode.org/pics/new-ui-and-help.png
22:42:58 <Aardappel> that looks extensive :)
22:43:09 <Aardappel> I have heard of factor, but not looked into it yet
23:04:01 --- join: mdmorri1 (n=mdmorri1@user-1120gkv.dsl.mindspring.com) joined #forth
23:06:46 <mdmorri1> Their it's working now.
23:07:09 <mdmorri1> I just had to let it sit a few minutes.
23:08:55 <slava> mdmorri1: you were in ##forth formerly and now you're in #forth
23:09:05 <mdmorri1> Oh.
23:09:34 <mdmorri1> I got disconnected so I relogged in to the one I thought I was on.
23:10:09 <mdmorri1> I was wondering why the channel had so many more people on now.
23:10:28 <slava> so are you using os x 10.5.2?
23:11:13 <mdmorri1> Yes.
23:11:48 <slava> i think i've seen your bug before
23:12:14 <mdmorri1> Really? What exactly is the problem?
23:12:27 <mdmorri1> And is it going to do that every time I run it?
23:12:37 <slava> no, it seems rare
23:12:48 <mdmorri1> Cool.
23:14:34 <slava> i found a way to reproduce it on my imac
23:14:48 <slava> log in as two different users, start two copies of factor, switch users. one of the copies stops repainting
23:15:02 <mdmorri1> Wow.
23:15:36 <mdmorri1> I only have one account on my computer, so that isn't the only cause.
23:15:55 <slava> might be two different bugs, or two different ways of triggring the same bug
23:17:24 <mdmorri1> The cookbook says that this 
23:17:25 <mdmorri1> : b a 1 + ;
23:17:25 <mdmorri1> : a 2 ;
23:17:25 <mdmorri1> b .
23:17:25 <mdmorri1> Has slightly different behaviour than Forth, but the only difference I see is that it makes you click on continue when you redefine a word.
23:18:03 <slava> you missed the first line in the copy/paste?
23:18:12 <slava> the difference is that when you redefine a word, all callers will use the new definition
23:18:15 <mdmorri1> Oh wait. It printed 3.
23:18:38 <mdmorri1> I can'r redefine words?
23:18:48 <mdmorri1> Oh.
23:18:49 <slava> you can
23:18:57 <slava> in forth, you can't redfeine words, you can only 'shadow' them.
23:19:02 <slava> in factor you can redefine words
23:19:54 <mdmorri1> Like on the Jupiter ACE only it always does it?
23:20:14 <slava> i'm not familiar with the ACE, sorry
23:20:18 <slava> i'm much too young
23:20:35 <mdmorri1> I am too, but I downloaded an emulator.
23:21:19 <mdmorri1> It is ITC and has a command that replaces all references to word x with a new word.
23:21:53 <mdmorri1> Obviously this isn't ITC, but it's the same effect.
23:22:39 <mdmorri1> If only I could remember the name of the word that did that.
23:22:58 <mdmorri1> Is their a way to merely shadow words?
23:26:08 <mdmorri1> Factor only makes use of one native thread, and Factor threads are scheduled co-operatively. C library calls block the entire VM.
23:26:30 <mdmorri1> It might have been OS X blocking the Factor thread.
23:26:51 <mdmorri1> I don't know why though.
23:29:06 <slava> i'm an idiot
23:29:11 <slava> i figured out the problem.
23:29:23 <mdmorri1> Cool, what is it?
23:29:41 <slava> its really stupid. i wasn't implementing the drawRect: callback at all.
23:29:48 <slava> os x has backing store and rarely calls that function
23:29:56 <slava> if you have a lot of video ram and not many apps open, it never gets called
23:30:03 <slava> because i never saw this on my g5 with a high-end video card
23:30:22 <slava> but when switching users it looks like the video ram contents are flushed
23:30:26 <slava> and i bet this is the problem you were having too
23:31:02 <slava> i should have tested on an old g3 with software rendering or something, then i'd see this straight away
23:31:52 <mdmorri1> I did have a bunch of programs running when the problem happened. 
23:33:03 <mdmorri1> But I think I had closed most of the windows.
23:33:25 <mdmorri1> How much VRAM did your G5 have?
23:33:42 <mdmorri1> Because I have 256 Megs.
23:34:05 <slava> i think it had 256
23:34:17 <slava> i have 256 on my mac book and i've never seen it
23:34:40 <mdmorri1> I must of had more windows open than I thought.
23:34:42 <slava> i pushed a fix so if all unit tests pass there will be a new binary in about 2 hours
23:34:51 <mdmorri1> Cool.
23:34:58 <mdmorri1> So you wrote Factor?
23:35:13 <slava> yes
23:36:58 <mdmorri1> How do I use forget?
23:37:42 <slava> \ forget help
23:37:50 <mdmorri1> Thanks.
23:38:35 <mdmorri1> Word not found in vocabulary search path.
23:39:17 <mdmorri1> Nice Adventure reference.
23:39:25 <mdmorri1> You are in a twisty little maze of objects, all alike.
23:42:47 <mdmorri1> Oddly enough, I was writing an adventure description language.
23:42:58 <slava> heh
23:45:31 <mdmorri1> 2 Hours you said?
23:52:36 <mdmorri1> Why can't TextEdit save regular text files? I thought it was the new Simple Text.
23:55:00 <mdmorri1> Oh, you have to make it plain text first.
23:59:59 --- log: ended forth/08.04.22