00:00:00 --- log: started forth/06.02.24
00:09:48 <I440r> is it me or does loop-hog go off on a tangent in the channel and have a conversation with nobody? lol
00:33:33 --- join: skylan_ (n=sjh@dialup-216-211-5-79.tbaytel.net) joined #forth
00:34:11 --- quit: skylan (Remote closed the connection)
00:42:41 --- quit: moreCowbell ("Computer goes to sleep!")
01:46:26 --- quit: I440r (Read error: 110 (Connection timed out))
03:59:54 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth
04:59:52 --- join: rehges (n=segher@blueice1n1.de.ibm.com) joined #forth
05:27:41 --- join: PoppaVic (n=pete@0-1pool46-226.nas30.chicago4.il.us.da.qwest.net) joined #forth
07:07:26 --- join: Cheery (i=Henri@a81-197-45-47.elisa-laajakaista.fi) joined #forth
07:42:45 --- join: sproingie (n=chatzill@64-121-2-59.c3-0.sfrn-ubr8.sfrn.ca.cable.rcn.com) joined #forth
08:01:51 --- quit: PoppaVic ("Pulls the pin...")
08:03:36 --- join: PoppaVic (n=pete@0-1pool74-248.nas24.chicago4.il.us.da.qwest.net) joined #forth
08:16:17 <PoppaVic> tathi: btw, I think I resolved the 'see' coring out with a bus-error.
08:16:57 <PoppaVic> it's a stupid, low-level test, but it seems to work
08:20:05 <tathi> oh?
08:22:10 <PoppaVic> yeah, I was doing shit like "see oop" - and it'd die ugly. I just added a casting/compare before. It apparerently needs a sighandler, because his code is screwed.
08:22:52 <PoppaVic> I was running into word-ptrs that were around 0x46 
08:25:41 --- join: ravenEx (i=ravenEx@VPN-224-82.aichyna.com) joined #forth
08:31:55 --- quit: PoppaVic ("make a beer-run and call it a knight")
08:47:57 <Quartus> 'word-ptrs that were around 0x46'?
08:48:47 --- quit: sproingie (Remote closed the connection)
08:48:51 <tathi> I dunno
08:49:20 <tathi> he's right that it crashes if you do "see oop", but I didn't follow anything beyond that
08:49:31 <Quartus> In ##c today he actually corrected someone who said "your" to "you're".  I was expecting "yer", it was a big disappointment. :)
08:49:51 <tathi> lol
08:50:22 <Quartus> coz, you know, you've got yer commies with their autoshit coming at you from all directions.
09:03:30 <Quartus> Apropos of nothing, Julian Noble's double-double float stuff looks neat.
09:05:22 <tathi> Where's that?
09:11:36 <Quartus> comp.lang.forth, let me find the link
09:11:56 <Quartus> http://galileo.phys.virginia.edu/~jvn/
09:14:56 <tathi> ah, it was farther back on comp.lang.forth than I looked.  gotcha.
09:15:35 <tathi> Interesting
09:19:37 --- join: I440r (n=foo@ns.laserscope.com) joined #forth
09:20:06 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #forth
09:21:59 <tathi> I440r: have you ever used your ihex reader?
09:28:57 --- join: PoppaVic (n=pete@0-1pool74-44.nas24.chicago4.il.us.da.qwest.net) joined #forth
09:30:04 <PoppaVic> tathi: you alive?
09:32:07 --- quit: rehges ()
09:58:09 <tathi> PoppaVic: ping
09:59:56 <PoppaVic> sorry, patched ficl4 - as I said (well, not a patch), but I was thinking about their whole "instruction" foo.
10:01:28 <PoppaVic> Don't most of these - like ficl - use a miserable switch() for the "instructions" or a vtable?
10:03:42 <tathi> dunno.  I thought FICL was just indirect threaded
10:04:30 <PoppaVic> tathi: hmm... well, they use SOMETHING that has a huge number of cases under a switch... like a cheapskates vtable.
10:04:33 <I440r> tathi whats ihex reader ?
10:04:40 <I440r> ohhh 
10:04:41 <I440r> yes
10:05:07 <I440r> it is part of the 8051 disassembler heh - its NOT a very robust way of loading intel hex files but it works :)
10:05:15 <I440r> could probably be improved on
10:05:39 <PoppaVic> I always agree when ix86 is the endall
10:05:41 <tathi> I440r: it doesn't look like it should work at all
10:05:46 <I440r> it does
10:05:51 <I440r> what looks wrong about it ?
10:06:14 <tathi> : >digit upc $30 - dup 16 > 7 and - ;
10:06:30 <I440r> heh
10:06:34 <I440r> thats fine
10:06:58 <tathi> yeah, it'll just give you the wrong values.
10:07:16 <I440r> the $30 minus converts $30 to $39 into 0 - 9.  if the result is larger than 16 after that conversion we subtract a further 7
10:07:48 <I440r> dup 16 > returns a true or a false 7 and makes taht t/f into 7/0 respectivly
10:07:53 <I440r> which we then subtract
10:07:54 <tathi> oh, right, you'll never have values in the middle
10:07:59 <I440r> the code is fine 
10:08:03 <I440r> yup
10:08:07 <tathi> because it doesn't use those chars, ok.
10:08:11 <I440r> :)
10:08:30 <tathi> also : get-buffer 3 8192 allocate !> codebuff ;
10:08:37 <virl> hiho..
10:08:54 <I440r> erm
10:09:13 <I440r> is that 3 8193 or 38912 ?
10:09:32 <tathi> 3 8192
10:09:34 <I440r> i dont believe allocate chanved from two params to one. it was always just one
10:09:43 <tathi> AND it returns a flag
10:09:55 <tathi> so it would always store t/f to codebuff
10:09:56 <I440r> oh. it didnt used to lol
10:10:20 <I440r> put a drop in there or test the flag heh
10:10:37 <I440r> im not sure what that 3 is for - it might be a typo that got introduced
10:10:46 <tathi> yeah, that's not the problem, that's just as far as I got in reading the code before I got distracted.
10:10:49 <I440r> but teh code was working because i used by 8051 disassembler heh
10:11:03 <tathi> I tried it on a file and it didn't stop
10:11:10 <I440r> you forgot to take yor riddlin didnt you hehe
10:11:16 <I440r> you keep getting sidetracked :P
10:13:04 * PoppaVic sighs
10:13:53 <tathi> I see...FICL does have a switch, in ficlVmInnerLoop
10:14:16 <PoppaVic> yeah, several
10:14:21 <tathi> I don't know the source at all well
10:14:40 <PoppaVic> It's not insurmountable - by it makes life.... Interesting
10:14:51 <PoppaVic> by/but
10:22:31 <PoppaVic> tathi: you can see #Metabuilder and our download-url for the latest ficl4 that has your patches and my (immense) changes., btw
10:22:52 <PoppaVic> or, I can email you an attachment
10:23:47 <PoppaVic> ..Basically, all we now lack is "VOCS", I guess... I am NOT happy about the way they bend ideas/rules/std over the sofa.
10:24:04 <I440r> is the guy doing ficl folding your patches in ?
10:24:08 <tathi> I440r: ah, bother.  It's just endianness issues.
10:24:11 <PoppaVic> Mostly, I think they just provide some interesting ideas
10:24:23 <I440r> lol
10:24:35 <I440r> tathi hah i didnt think of that :)
10:24:43 <PoppaVic> tathi: well, of course: you need NBO before the sprintf/sscanf
10:25:24 <tathi> reading one char at a time is silly anyway, should just use fmmap or whatever that word is called.
10:25:26 <PoppaVic> generating the hex is nice. If it's multibyte, you need ordering.
10:25:36 <PoppaVic> mmap?
10:25:40 <I440r> tathi u shud know that svn is down and will be for a while. im no longer at that address so the IP is no longer mine
10:26:08 <tathi> heh
10:26:13 <PoppaVic> I can't fully agree... Just because the mmap advocates seem to use it loose and lamely
10:26:20 <tathi> I just assume svn is going to be down :)
10:26:21 <I440r> i moved into a new apartment but then had to drive from san antonio tx to san jose ca for a 4 month contract. i have svn  here with me so when i get internet it shud be back up on another temporary IP :)
10:26:47 <PoppaVic> tathi: np, I "speak C" ;-)
10:27:07 <tathi> I440r: my copy is now a darcs repo as well as an svn working copy :)
10:27:30 <I440r> heh
10:27:32 <I440r> whats darcs ?
10:27:42 <JasonWoof> rcs system
10:27:46 <I440r> ugh
10:27:50 <PoppaVic> blech
10:27:54 <I440r> what he said
10:27:57 <I440r> lol
10:28:06 <PoppaVic> cvs, rcs, it's all a mulch
10:28:07 <JasonWoof> hmm... s stands for system doesn't it :)
10:28:12 <I440r> they use PVCS here where im working - gona try get them to convert to svn :)
10:28:32 <tathi> I rather like darcs
10:28:40 <PoppaVic> well, it never seems to mean "source"
10:28:54 <JasonWoof> I find svn to be well worth it, even on projects where I'm the only one committing
10:28:56 <tathi> ghc runs almost everywhere these days
10:29:21 <PoppaVic> "ghc"? 
10:29:28 <JasonWoof> dillo google.com
10:29:36 <JasonWoof> oops
10:29:55 <tathi> Glasgow Haskell Compiler
10:30:05 <tathi> darcs is written in haskell.
10:30:07 <PoppaVic> ahh - die well ;-)
10:30:08 <JasonWoof> I've got rxvt and irssi on adjacent fkeys
10:30:27 <PoppaVic> anyway, ok - time to go make dinner. Stay well, folks
10:30:31 --- quit: PoppaVic ("Pulls the pin...")
10:33:09 <tathi> I440r: anyway, if you'd host a semi-reliably available svn server, I wouldn't have to use anything else ;)
10:34:27 <I440r> its not my fault~
10:34:38 <I440r> lol
10:34:50 <I440r> svn was reliable (ish) for a whole year!
10:34:50 <I440r> lol
10:35:08 <I440r> btw could you create reposatories on svn before ?
10:35:38 <tathi> you could perfectly well use dyndns or something instead of just the IP address.
10:36:02 <tathi> no, I don't think so.
10:36:03 <I440r> yea i know - but thats a pain in the ass lol
10:36:23 <I440r> yea - i was planning on adding the ability to do that but svn was a machine INSIDE my local network 
10:36:30 <I440r> i was forwarding to it from my firewall
10:37:05 <tathi> whatever.  I'm just using darcs. :P
10:37:37 <tathi> I tried bzr too.  It's more svn-like, but it's written in Python, so it's slow as mud.
10:38:56 <I440r> lol
10:44:09 <JasonWoof> "the first mark of a primitive society is that it thinks itself advanced"
10:48:43 <I440r> we are advanced. we are advancing to the rear
12:41:04 --- quit: tathi ("rebooting")
12:50:06 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth
13:19:55 --- quit: tathi ("leaving")
13:35:33 --- quit: uiuiuiu (Remote closed the connection)
13:35:38 --- join: uiuiuiu (i=ian@dslb-084-057-242-195.pools.arcor-ip.net) joined #forth
13:40:40 --- quit: Cheery ("Leaving")
14:17:47 --- quit: ravenEx (Read error: 131 (Connection reset by peer))
14:42:14 --- join: thinkinginbinary (n=tom@phoenix.thomastuttle.mooo.com) joined #forth
14:42:27 <thinkinginbinary> Quartus: Hey
14:43:03 <Quartus> Hey.
14:43:24 <thinkinginbinary> Quartus: Where've you been?  I haven't seen you around lately.
14:43:41 <Quartus> I think I was offline for a night or so.
14:43:54 <thinkinginbinary> Ah.
14:44:22 <Quartus> Maybe two.
14:47:39 <thinkinginbinary> I might *finally* get to buy some electronic stuff this weekend.
14:48:36 <Quartus> Soon the room will be filled with solder fumes!
14:48:40 <thinkinginbinary> lol
14:48:48 <thinkinginbinary> I'll open a window.
14:49:07 <thinkinginbinary> I can't wait, this is going to be really cool.
14:50:36 <Quartus> My only recent soldering experience was repairing a pewter figure.
14:51:17 <thinkinginbinary> I want one of these: 
14:51:31 <thinkinginbinary> http://www.thinkgeek.com/gadgets/tools/69d3/
14:51:32 <Quartus> Here's one:
14:51:33 <Quartus> ah
14:51:56 <Quartus> Oh yes. They sell those nearby; I've wondered if they work well.
14:52:28 <thinkinginbinary> Never tried one.
14:52:35 <thinkinginbinary> Thinkgeek refuses to stock them, but says they do :-b
14:52:50 <thinkinginbinary> Apparently you get 700+ joints on 4 AAA batteries.
14:53:35 <Quartus> I used to wire-wrap prototypes.
14:53:48 <thinkinginbinary> I'm going to pick up a breadboard.  /me is lazy.
14:55:21 <thinkinginbinary> Have you seen Intel's latest Pentium Extreme Edition?  It's dual core, with two 3.46 cores *each* with 2 MB L2 cache.  (65 nm process)  It's more than $1000, unfortunately.  (/me isn't getting one anytime soon :-()
14:55:32 <Quartus> I haven't.
14:55:38 <thinkinginbinary> They look cool.
14:55:51 <thinkinginbinary> http://www.newegg.com/Product/Product.asp?Item=N82E16819116246
14:55:53 <Quartus> Sounds like rather the opposite, big heat generator. :)
14:56:01 <thinkinginbinary> Oh yeah, 1066MHz FSB.
14:56:02 <thinkinginbinary> Oh they are.
14:56:55 <thinkinginbinary> Although apparently the 65nm process makes them cooler.  (Intel probably uses up all the headroom it gains by increasing core count and core and bus speeds.)
14:58:55 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth
15:23:02 --- quit: thinkinginbinary ("leaving")
15:29:09 --- join: danniken (i=CapStone@ppp-70-252-87-94.dsl.ltrkar.swbell.net) joined #forth
16:00:47 --- join: amca (n=plump@as-bri-3-82.ozonline.com.au) joined #forth
16:51:25 --- join: snoopy_1711 (i=snoopy_1@dslb-084-058-137-087.pools.arcor-ip.net) joined #forth
16:59:51 --- quit: Snoopy42 (Read error: 145 (Connection timed out))
17:00:16 --- nick: snoopy_1711 -> Snoopy42
17:07:32 --- quit: I440r ()
17:41:12 --- part: danniken left #forth
17:48:56 --- join: thinkinginbinary (n=tom@pool-68-163-202-241.bos.east.verizon.net) joined #forth
17:49:14 <thinkinginbinary> Quartus: Hey, I got a simulator for the PIC I'm going to be using (well, a similar one).  What are some good words to start with?
17:50:53 --- quit: amca ("d34d")
17:51:02 --- join: amca (n=plump@as-bri-3-82.ozonline.com.au) joined #forth
17:54:21 <Quartus> Hey. Words?
17:55:07 <Quartus> Well, think back to the Java Forth I helped you build, start with the dictionary, move to the interpreter, etc.
17:55:12 <thinkinginbinary> Quartus: OK.
17:55:40 <thinkinginbinary> Quartus: Well in this Forth the memory and stack management is going to be more difficult, so should I do that first?
17:55:44 <Quartus> Build the VM and associated primitives, build the dictionary and associated primitives, build the interpreter and associated primitives, then start adding new words as required.
17:55:52 <Quartus> The VM has to be built first, so yes.
17:56:08 <Quartus> That'll determine the memory model, the register allocations, the stacks.
17:56:18 <thinkinginbinary> Quartus: There is no VM this time (unless you mean threading).
17:56:29 <thinkinginbinary> Quartus: Or you just mean like stacks 'n' stuff?
17:56:38 <Quartus> There's always a VM.  Stacks, dictionary, etc.
17:57:45 <thinkinginbinary> Quartus: Oh.  Yeah.  I guess I just considered that the environment.
17:58:14 <Quartus> Six of one, half-dozen of the other.  It's the VM, even if the execution model is native code.
17:58:18 <thinkinginbinary> Yeah.
17:58:24 <Quartus> Which yours isn't, as I recall.
17:58:59 <thinkinginbinary> Well, it's indirect-threaded, so yeah, I guess the thread-following-thingy is a VM.
17:59:11 <Quartus> Part of it. The execution engine.
17:59:31 <Quartus> The 'inner interpreter', in antique lingo.
17:59:38 <thinkinginbinary> This is gonna *rock*.
18:00:18 --- quit: saon (Read error: 104 (Connection reset by peer))
18:00:24 <Quartus> Well, unlike many, you actually write and finish at least some of what you start, so there's a better chance of actual rocking. :)
18:00:37 <thinkinginbinary> Hahaha.  My brother thinks I don't.
18:12:01 <thinkinginbinary> So, NEXT is the traditional name for the thread interpreter?
18:12:39 <Quartus> I don't know that the thread interpreter has or needs a name, per se.
18:12:56 <thinkinginbinary> Quartus: Well, I gotta jump to it sometime.  So it needs a name.
18:13:16 <Quartus> Internally, sure.
18:13:34 <Quartus> Externally, it's EXECUTE.
18:15:46 --- join: saon (i=1000@unaffiliated/saon) joined #forth
18:16:42 <thinkinginbinary> What happens when native code needs to call threaded code?
18:23:01 <Quartus> Pass the xt to EXECUTE.
18:24:06 <thinkinginbinary> And then when that returns?  It will return to the previous threaded word, not the native code.  There are two separate return stacks.  Badness?
18:25:38 <Quartus> That's not quite how it works -- you'd do well to have a look at a reference implementation.  Perhaps eForth.
18:26:03 <Quartus> Didn't you already work this out with your Java Forth?  Calling native routines from the VM, and vice versa?
18:26:13 <thinkinginbinary> Yes, with a Java-only hack.
18:26:18 <thinkinginbinary> It used exceptions to return from native code.
18:26:49 <thinkinginbinary> Perhaps I'll just have to pay attention to how I call things in native code.
18:27:53 <Quartus> Always wise. :)
18:30:23 <Quartus> You could also code your system such that it's unnecessary to call threaded code from native code.
18:30:29 <thinkinginbinary> Yeah, that would be good.
18:31:26 --- quit: tathi ("foo!")
18:32:08 <Quartus> Or modify the inner loop to exit when a specific sentinel value shows on the VM return stack.
18:32:29 <Quartus> Or have a separate version of the inner loop just for that purpose.
18:33:12 <thinkinginbinary> Okay... As the interpreter is running, when it finds a threaded instruction, what does it do?  Push the current IP+2 on the RS, put the new address into IP, and loop back?
18:33:43 <Quartus> The inner interpreter, you mean? The execution engine?  It has to execute that instruction.
18:34:43 <Quartus> If it represents a call to another threaded word, it has to store an address on the VM return stack and nest into the thread.
18:38:41 <Quartus> One way to manage calling threaded code from native code is to set up what we used to call a 'trampoline' -- a modifiable thread with two instructions in it, the first being the thread you want to call, the second being a native routine that performs the neat trick of returning to your native code routine.
18:40:19 <Quartus> You modify the first instruction in the thread and call your inner interpreter with the address of the trampoline.
18:41:41 <Quartus> Still there?
18:45:22 --- quit: amca ("d34d")
18:49:28 <thinkinginbinary> Yeah.
18:50:21 <Quartus> Trampoline idea make sense to you?
18:50:26 <thinkinginbinary> Yeah.
18:50:39 <thinkinginbinary> Although the whole thing is still fuzzy in my mind.
18:50:45 <Quartus> Best to try it out.
18:52:59 <thinkinginbinary> Is "EXIT" the instruction that returns to the previous thread?
18:53:12 <Quartus> Yes.
18:53:28 <Quartus> Same as in your Java version.
18:59:35 <thinkinginbinary> Oh yeah, I remember... I used -1 as the sentinel in the Java version.
19:00:59 <Quartus> Stepping out for a few.
19:01:14 <thinkinginbinary> ok
19:32:00 <Quartus> Back.
19:32:23 --- join: arke (i=f2@bespin.org) joined #forth
20:03:08 <thinkinginbinary> Hey!
20:03:11 <thinkinginbinary> It pushed a number on the stack!
20:03:51 <thinkinginbinary> I wrote code that punches the thread for (literal) 257 EXIT into RAM, and then starts that thread.
20:04:00 <thinkinginbinary> (257 == 0x0102, so I can see if the byte order's right.)
20:11:26 <thinkinginbinary> And it carries right.
20:11:30 <thinkinginbinary> Alright, I gotta go to bed.
20:12:22 --- quit: thinkinginbinary ("leaving")
20:33:41 --- join: LOOP-HOG (n=chatzill@sub22-119.member.dsl-only.net) joined #forth
20:52:54 --- quit: LOOP-HOG ("ChatZilla 0.9.61 [Mozilla rv:1.7.5/20041217]")
21:18:24 --- nick: nanstm -> Raystm2
22:03:18 --- quit: EnragedTux ("Leaving")
22:58:48 --- join: rsyncx (n=bob@CPE000c41aac435-CM00111ae4f4cc.cpe.net.cable.rogers.com) joined #forth
23:26:21 --- join: Entopical (i=entop@d161-184-112-112.abhsia.telus.net) joined #forth
23:59:59 --- log: ended forth/06.02.24