00:00:00 --- log: started forth/06.03.31
00:00:13 <snowrichard> JasonWoof: sent you an email through your site
00:02:11 --- quit: snowrichard (Client Quit)
00:06:15 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
00:10:37 <snowrichard> hello
00:31:42 --- quit: snowrichard ("Leaving")
00:36:20 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
00:36:38 --- quit: snowrichard (Read error: 104 (Connection reset by peer))
00:36:42 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
00:36:47 --- quit: snowrichard (Read error: 104 (Connection reset by peer))
00:48:27 --- quit: arke (Remote closed the connection)
00:54:13 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
01:04:43 --- quit: snowrichard ("Leaving")
01:51:58 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
01:53:56 <snowrichard> hi
01:55:55 --- quit: snowrichard (Client Quit)
04:47:49 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth
04:57:25 --- join: rehges (n=segher@blueice2n1.de.ibm.com) joined #forth
05:32:02 --- join: Cheery (n=Henri@a81-197-45-47.elisa-laajakaista.fi) joined #forth
05:38:05 --- join: PoppaVic (n=pete@0-1pool47-154.nas30.chicago4.il.us.da.qwest.net) joined #forth
06:26:58 <PoppaVic> hmm
06:53:06 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #forth
06:56:45 <JasonWoof> tathi: what sort of changes to fovium were you contemplating?
07:03:34 <PoppaVic> JasonWoof: I think he was trying to clarify better bytecoding and funcs.
07:08:47 <PoppaVic> prolly adding "dictionary" concepts, too
07:18:15 --- join: arke (i=c_walton@lnx101.hrz.tu-darmstadt.de) joined #forth
07:40:14 <tathi> JasonWoof: oh, I was just messing around
07:41:15 <tathi> I was going to document it, and then I just got experimenting instead
07:42:10 <PoppaVic> experimenting is fun
07:43:02 <tathi> although...if we changed to something other than absolute addressing for branches, then opcode 0 could just be 'branch'.
07:43:47 <tathi> PoppaVic: the bytecoding part of fovium is just fine :)
07:43:56 <PoppaVic> Yes, I saw that
07:44:14 <PoppaVic> I was just wondering if you'd lookup/dictionary and window
07:44:41 <tathi> if I understand what you mean by "window", then absolutely not.
07:45:30 <tathi> but I'd almost certainly add the dictionary stuff as opcodes for a scripting-language thingy.
07:45:46 <PoppaVic> I think you have to
07:46:13 <PoppaVic> It's like writing both a forthish "assembler" and the code that does the "asm-code"
07:47:28 <tathi> you don't have to
07:47:37 <tathi> not if you're just writing a VM
07:47:58 <tathi> you just need a way to load the code to get you started with
07:49:07 <PoppaVic> I know, understood.
07:49:18 <tathi> oh, ok then
07:49:38 <PoppaVic> The VM is like writing directly in the MC. THe dictionary-side sorta' adds "ASM"
07:50:27 <PoppaVic> btw... I've been trying to better and better codify ideas, and just learned (verified) something I suspected...
07:51:10 <PoppaVic> it looks like all the new 64-bit folks use 64-bit ptrs, but 32-bit integers.
07:52:53 <PoppaVic> I've been trying to work up sensible sizes and func/macro/accounting where it matters.
07:55:35 <tathi> huh
07:55:45 <tathi> I thought win64 was the only platform that used LP64
07:56:11 <PoppaVic> oh, lord no - I apparentl;y stirred up a nest of hornets in ##C
07:56:41 <tathi> very odd.
07:56:57 <PoppaVic> I get told int64 is a waste, so they use int32, and ptr64 - but asking about PADDING really pissed off some folks.
07:57:23 <tathi> wait, but are they saying that the C compiler's 'int' is 32 bits?
07:57:28 <PoppaVic> basically, no one is supposed to care
07:57:29 <tathi> or that they explicitly use a 32-bit type?
07:57:55 <PoppaVic> tathi: on the 64-bit systems, apparently an int is 32, a size_t is 64 and a void* is 64
07:58:25 <PoppaVic> which makes me believe we've got one helluva mess about ranges and sizes.
07:59:09 <oxygene> PoppaVic: not on all 64bit systems
07:59:24 <PoppaVic> stdint.h should have been introduced YEARS ago, and folks need to realize that the int/uint/short/long is just a macro
07:59:47 <PoppaVic> oxygene: well, shit - now what? I just had three or four guys try to rape me for even asking!
08:00:34 <oxygene> PoppaVic: use u?int{8,16,32,64}_t, forget about the native types (int etc)
08:00:44 <tathi> that's very odd -- all the sites/mail-lists that I've seen claim that all C compilers except win64's use the LLP64 model (64-bit ints).
08:01:03 <PoppaVic> btw, tathi - I'd love if we could debate typedef vs macro in terms of immediate, compile, etc
08:01:29 <PoppaVic> oxygene: yes, I understand that - then I asked about PADDING and got hollered at.
08:02:11 <tathi> well, if you care about padding, by definition you're doing something platform specific.
08:02:11 <PoppaVic> tathi: apparently, most guys see "long long" as 64-bit, and some even see long as 64, and plain int is 32
08:02:33 <PoppaVic> tathi: hey, I was just curious. FFI with structs will do that.
08:02:38 <tathi> I guess things have changed then.  I don't keep up with gcc at all :)
08:03:31 <PoppaVic> all I can say about it is: you guys are nuts - you don't even know yer fucking ranges, let alone padding, offsets, etc
08:03:32 <tathi> lately they seem to keep releasing newer, bigger, buggier versions, with more error messages
08:03:44 <PoppaVic> Yeah, I've seen that
08:03:53 <PoppaVic> ..gforth's latest had a similar issue
08:13:31 <PoppaVic> tathi: the latest gforth despises doubles, let alone floats
08:37:07 --- quit: PoppaVic ("bbl")
08:44:47 --- nick: Raystm2 -> nanstm
08:59:02 <JasonWoof> tathi: I think it'd be cool to change the branch addressing thing
08:59:50 <JasonWoof> I like the idea of using xor
09:00:44 <tathi> yeah.  I also realized that the small slot is at the end, which is kind of annoying.
09:01:01 <tathi> but I don't see a good way (in software) to put it at the beginning
09:01:44 <JasonWoof> why is that annoying?
09:01:59 <JasonWoof> because it does no good that the branch instruction fits?
09:02:06 <tathi> yeah
09:02:24 <tathi> oh, and we should probably change the opcode of the call instruction for the same reason :)
09:03:11 <JasonWoof> oh, weird, I forgot it was 2
09:03:39 <tathi> I must have been thinking the small slot was first.
09:03:59 <tathi> because I remember thinking that call would be really common, and putting it there on purpose :)
09:04:15 <JasonWoof> no, we picked the opcodes before we realized that we needed to be word-aligned for returns
09:04:55 <tathi> yeah, ok, but I still put it there thinking it would be good for it to fit in the small slot
09:05:09 <JasonWoof> well, I think it's cool that (if we switch to xor) the last two bits will sometimes be enough for the branch address
09:05:15 <JasonWoof> if foo then
09:05:16 <tathi> swapping it with return probably makes the most sense
09:05:20 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
09:05:21 <tathi> :)
09:05:46 <snowrichard> hi tathi
09:05:54 <tathi> hi
09:06:24 <JasonWoof> return sounds good
09:06:48 <snowrichard> Hi Jason
09:06:56 <JasonWoof> hi snowrichard  :)
09:07:30 --- join: PoppaVic (n=pete@0-2pool236-233.nas22.chicago4.il.us.da.qwest.net) joined #forth
09:07:31 <snowrichard> I got fovium to compile with Suse10, but I don't have an image to try it with
09:08:09 <PoppaVic> linux?
09:08:14 <snowrichard> yes
09:08:18 <JasonWoof> snowrichard: I e-mailed you about this
09:08:21 <PoppaVic> fovium is pretty plain
09:08:33 <snowrichard> oh ok I just got up haven't checked yet
09:08:39 <PoppaVic> actually, it's even lower than "plain"
09:08:56 <tathi> that was the general idea :)
09:09:02 <PoppaVic> right
09:09:09 <JasonWoof> the bug was fixed in svn, but I didn't make another tarball until today
09:09:15 <PoppaVic> tathi: except, you sorta' mixed "levels" a few cunthairs
09:09:25 <tathi> PoppaVic: how's that?
09:09:33 <PoppaVic> well, system()
09:09:54 <tathi> oh.  yeah, that was kind of necessary.
09:09:56 <PoppaVic> there is internal, there is core, and there is file-system.
09:10:04 <JasonWoof> PoppaVic: how do you expect to interract with the OS?
09:10:11 <JasonWoof> ppc has a system call instruction
09:10:12 <PoppaVic> this is sorta' the organized-insanity of posix
09:10:20 <PoppaVic> JasonWoof: that's nice
09:10:34 <PoppaVic> meanwhile, the OS is not the kernel, or the filesystem.
09:11:02 <PoppaVic> JasonWoof: from forthers, I'd expect better organization/layering, is all.
09:11:09 * JasonWoof doesn't understand anything PoppaVic says
09:11:23 <PoppaVic> JasonWoof: dat's nice - wuv you
09:11:23 <JasonWoof> PoppaVic: what's the alternative?
09:11:28 <PoppaVic> layering
09:11:33 <PoppaVic> levels
09:11:36 <tathi> that's not an answer.  be specific.
09:11:48 <JasonWoof> what's wuv?
09:11:54 <PoppaVic> fovium is SUPPOSED to be - essentially a superCPU.
09:11:58 <PoppaVic> ..and emulator.
09:11:59 <tathi> er...no.
09:12:13 <JasonWoof> no
09:12:25 <JasonWoof> fovium is supposed to be a consistent runtime environment
09:12:56 <tathi> for the primary purpose of allowing Jason's colored forth system to run most anywhere
09:13:12 <JasonWoof> emulating a CPU without any access to the outside world is totally useless
09:13:18 <JasonWoof> I need to interract with the user
09:13:31 <JasonWoof> specifically I want keyboard input and graphical output
09:13:46 <PoppaVic> I got news for you: that's how chips work
09:14:19 <JasonWoof> you're how chips work
09:14:29 <PoppaVic> so, if adding a layer literally means more-supports, it's not a "bad thing" - particularly for thinking levels/devices
09:14:43 <JasonWoof> 71! wow, it's nice out
09:14:48 <JasonWoof> I'm outa here
09:14:53 <PoppaVic> Yeah, it's NICE out there
09:15:09 <JasonWoof> PoppaVic: what the crap are you talking about?
09:15:15 <JasonWoof> say something specific
09:15:26 <PoppaVic> nm.. I'll go back to my corner
09:15:36 <JasonWoof> "layer" is a very vague and general term that could mean just about anything in this context
09:15:45 <PoppaVic> yep
09:16:02 <JasonWoof> so why do you say it as though it convyed some meaning?
09:16:13 <JasonWoof> or are you just flapping your mouth?
09:16:17 <PoppaVic> "context"
09:16:21 <snowrichard> well I've got to go to my Mom's and pick up my allowance and rent check, so catch you later
09:16:27 <PoppaVic> yep, just flapping.
09:16:43 --- quit: snowrichard ("Leaving")
09:17:21 <tathi> PoppaVic: if it makes you feel better, think of system() as in() and out() rolled into one, with some fancy peripherals on the other end ;)
09:18:40 <PoppaVic> ahhh
09:18:49 <PoppaVic> that might help, when I view it
09:20:21 <tathi> it isn't really, but it's easier to handle from both ends (the VM and the bytecode that runs on it) than regular i/o or mmapped i/o.
09:20:54 <tathi> for our purposes, it seems like the most sensible design decision, even if it does look a bit out of place on such a bare-bones VM.
09:22:04 <PoppaVic> right... I can sorta' see it, tathi 
10:15:01 <slava> hmm, system() should not be a bytecode
10:15:11 <slava> use an FFI to call whatever POSIX stuff you need
10:20:04 <PoppaVic> yeppers.
10:45:29 <PoppaVic> slava: interesting == "what FFI at level-zero" versus "level-N"
10:45:40 <PoppaVic> ..and, "define a level"
10:47:18 <PoppaVic> This, (to me), always suggested a Z80-like "windowing" of what folks called "opcodes"
10:47:48 <PoppaVic> ..and, remember: the Z80 was predicated on "the 8080"
10:55:03 --- join: pizza_ (i=pizza@poipu/supporter/pizza-milkshake) joined #forth
10:55:05 <slava> factor has a bunch of primitives for writing data to the code heap, and for changing a word's xt to a code heap location
10:55:19 <slava> then the ffi can be built 'from scratch' by the compiler
10:57:45 <PoppaVic> yes, as long as it is a codified-methodology
10:58:10 <PoppaVic> did you get into stack/entry typing yet?
11:08:11 --- quit: PoppaVic ("Pulls the pin...")
11:09:17 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
11:09:43 <snowrichard> hi
11:23:49 --- quit: tathi ("leaving")
11:44:50 <snowrichard> JasonWoof thanks for the email. I got the new tarball
11:49:10 --- join: Ray_work (n=Raystm2@adsl-68-90-195-65.dsl.rcsntx.swbell.net) joined #forth
11:49:16 --- join: Ray-work (i=Raystm2@adsl-68-90-195-65.dsl.rcsntx.swbell.net) joined #forth
11:49:19 <snowrichard> Hi ray
11:49:52 <snowrichard> is it Ray_work or Ray-work :) 
11:49:56 <Ray-work> Hi snowrichard!
11:50:05 <Ray-work> Taking a little lunch break here 
11:50:14 <Ray-work> it's both.
11:50:27 <Ray-work> what are you doing today?
11:50:39 <snowrichard> downloaded the new fovium to look at
11:50:56 <snowrichard> and its payday fo rme
11:50:58 <snowrichard> for me
11:51:06 <Ray-work> Oh cool, JasonWoof's vertual machine forth.
11:51:19 <Ray-work> ya payday here as well.
11:51:37 <Ray-work> I could use some more money. :)
11:51:51 <Ray-work> I inherited some back taxes.
11:52:03 <snowrichard> I can't complain -- the va gives me 2480 a month, non-taxable
11:52:22 <Ray-work> cool. but did you have to get hurt to get that?
11:52:27 <snowrichard> well yes
11:52:36 <Ray-work> Purple Heart?
11:52:48 <snowrichard> no, it wasn't war time or in a combat zone
11:52:54 <Ray-work> ah
11:53:15 <snowrichard> just after the Iranian Hostage crisis
11:53:26 <Ray-work> do you mind me asking what the disablility is? Obviously you have your brain power.
11:53:41 <snowrichard> it was a head injury, they say I'm schizophrenic
11:53:52 <Ray-work> oh heck that was what early 80's?
11:54:04 <snowrichard> yeah 81
11:54:25 <snowrichard> 6 months in Navy hospital
11:54:26 <Ray-work> I graduated in 81 and was 17 in boot camp. Thought I was gonna join you there.
11:54:44 <Ray-work> I was in USN as well.
11:54:45 <Ray-work> What was your rating?
11:54:50 <snowrichard> DP2
11:55:08 <Ray-work> what's that? Data Processor?
11:55:12 <snowrichard> yeah
11:55:15 <Ray-work> cool.
11:55:17 <snowrichard> trained monkey
11:55:20 <Ray-work> hehe
11:55:33 <Ray-work> How does a DP2 get a head injury, Hard Code?
11:55:39 <snowrichard> computer operator.
11:55:59 <snowrichard> the ship was doing propulsion test, flank reverse and hard right rudder
11:56:05 <snowrichard> hit a locker with my head
11:56:15 <Ray-work> whoa no kidding.
11:56:23 <Ray-work> sorry man.
11:57:14 <Ray-work> I was an Underway Replenishment Specialist, I held 4 ratings.
11:57:19 <snowrichard> it was very repetitive work.  Making flight recorder tapes, loading them back into the system, making backups.
11:57:32 <Ray-work> I see.
11:57:38 <Ray-work> Carrier then?
11:57:46 <snowrichard> yeah CVN-69
11:58:06 <snowrichard> I got some good photos of sunrise during UnRep one day
11:58:08 <Ray-work> 69 which is that? 
11:58:12 <snowrichard> Eisenhower
11:58:19 <Ray-work> the IKE okay :)
11:58:31 <Ray-work> cool about the photos.
11:58:45 <Ray-work> I was on AE24 PYRO.
11:58:53 <Ray-work> west coast.
11:59:08 <snowrichard> I have no idea which ones serviced us.
11:59:31 <Ray-work> I know, they steel into the night. :)
11:59:33 <snowrichard> the ship was nuke, but they had to load JP-5 periodically , and the groceries of course
11:59:41 <Ray-work> right.
12:00:00 <Ray-work> that nuke stuff used to be 'real' fun.
12:00:23 <Ray-work> men would freak out when they knew it was on board.
12:00:33 <Ray-work> Never really bothered me tho.
12:00:48 <snowrichard> worked one time in the storeroom during a replenishment.  dropped a jar of mayonaise. :) 
12:01:05 <Ray-work> yikes now that is a real bio hazzard.
12:01:15 <Ray-work> :)
12:01:33 <Ray-work> prob'ly where you schiz came from :)
12:02:31 <snowrichard> remember them pushing  a pallet of soft drinks over the side because it had roaches in it
12:02:54 <Ray-work> oh man have i got a story for ya...
12:03:26 <Ray-work> The Enterprise was gonna pull into San Fran for fleet week during low tide.
12:03:49 <Ray-work> in the Bearing straights we took everything they could send us, to lighten the load.
12:05:58 --- quit: Ray_work (Read error: 110 (Connection timed out))
12:06:54 <snowrichard> and?
12:17:09 --- quit: Ray-work (Read error: 104 (Connection reset by peer))
12:26:42 <Quartus> Classified, apparently.
12:34:54 <JasonWoof> slava: I have no interest in a ffi
12:35:10 <JasonWoof> for fovium
12:35:27 <Quartus> wait, where's PoppaVic to talk about layers and context?
12:35:56 <JasonWoof> I just want a dirt simple runtime environment that is exactly the same on all platforms
12:36:03 <Quartus> And vocs and semantic?\
12:38:37 <Quartus> I especially liked
12:38:43 <Quartus> <PoppaVic> "context"
12:39:03 <Quartus> Clean. Simple. Leaves nothing behind but the great scent of Brut.
13:07:02 --- quit: Cheery ("Leaving")
13:30:21 --- quit: rehges ()
14:11:33 <snowrichard> hi
14:22:20 --- quit: snowrichard ("Leaving")
14:37:56 <slava> JasonWoof: what if somebody wants to write one as an optional library?
14:38:00 <slava> JasonWoof: how would it 'plug in' to the vm?
14:42:00 --- mode: ChanServ set +o crc
15:06:41 <JasonWoof> drat. there's a bug in fovium in the other-endian detect code
15:07:31 <JasonWoof> slava: though a system call
15:09:16 <slava> will you be able to implement fovium in itself?
15:45:34 --- join: Amanita_Virosa (n=jenni@adsl-69-154-178-250.dsl.hstntx.swbell.net) joined #forth
15:45:50 <slava> hi
15:45:59 <crc> good evening
15:46:18 <Amanita_Virosa> heya
15:46:22 <Amanita_Virosa> the weirdest forth i've ever written...
15:46:35 <Amanita_Virosa> is a three-stack forth that stores blended datatypes on the two datastacks, and is used to drive a general-purpose assembler.
15:47:00 <Quartus> Blended like a fine scotch?
15:47:09 <slava> mmmm
15:47:49 <Amanita_Virosa> hehe i don't know about that, blended more like...  "numbers and symbolic labels"
15:48:49 <Quartus> Not nearly as good as the scotch, I fear. :)
15:49:32 <Amanita_Virosa> probably not
15:50:05 <Quartus> What's weird about it, aside from the extra stack?
15:54:04 <Amanita_Virosa> well, the fact that the stack can contain symbolic labels is kinda weird
15:54:12 <Amanita_Virosa> and the fact that it's a purely interpreted language
15:54:21 <Amanita_Virosa> or the fact that it's actually just the backend of an assembler... that's pretty weird too
16:03:28 <Quartus> Were I to tackle an extra stack, I think I'd approach it by adding the stack to an existing Forth. Were there advantages to building a new Forth for the purpose?
16:04:12 <Amanita_Virosa> it went quite the opposite way around
16:04:37 <Amanita_Virosa> the forth was built, but it needed a third stack, because the return stack stores string pointers, which are incompatible with the blended types on the data stack
16:06:17 <Quartus> Hmm.
16:07:57 <Amanita_Virosa> and the strange requirements on the language limited what existing forths (ie, none) could be used for it)
16:08:26 <Quartus> Must be strange.
16:09:22 <arke> hi all
16:09:38 <Amanita_Virosa> well to give you an idea, it's called from an instruction match module... so it calls a forth function for each line of assembler
16:09:41 <Amanita_Virosa> it's a strange setup
16:09:42 <Amanita_Virosa> but it works well
16:11:10 <arke> that is quite strange ;)
16:11:20 <Quartus> The Forth assemblers I've done work the same way, with the exception that the dictionary more-or-less *is* the instruction match module, each mnemonic determines the instruction context and proceeds accordingly.
16:11:34 <Amanita_Virosa> sure, i know how forth assemblers work
16:11:36 <Amanita_Virosa> and i've written one
16:11:39 <Amanita_Virosa> but this isn't a forth assembler
16:11:46 <Quartus> Sounds similar in nature, though.
16:11:47 <Amanita_Virosa> it takes regular format assembly
16:11:58 <Amanita_Virosa> and it's designed to allow the internal instruction table to be reprogrammed
16:12:07 <Amanita_Virosa> yes, it's not too different
16:12:22 <Amanita_Virosa> tho it also generates full relocatable code, multisection object files, the works
16:12:35 <Quartus> Multipass, I assume. 
16:13:28 <Amanita_Virosa> not exactly
16:13:32 <Amanita_Virosa> the assembler is one-pass
16:13:49 <Amanita_Virosa> tho the resulting object file might require interlinking
16:14:26 <Quartus> Forward branches to as-yet-unresolved labels?
16:15:30 <Amanita_Virosa> that just generates unresolved reloc entries
16:15:40 <Amanita_Virosa> you can link those in one of your linking passes
16:16:09 <Quartus> Perhaps no requirement to optimize forward short branches for the target architecture, then.
16:16:41 <Amanita_Virosa> that can be done explicitly by the programmer if desired
16:16:43 <Amanita_Virosa> but generally no
16:16:59 <Quartus> Not a tragedy. :)
16:17:07 <Amanita_Virosa> its meant to be generic and flexible moreso than optimal
16:17:29 <Quartus> Ok.
16:37:58 <JasonWoof> slava: sure, you could implement fovium in fovium
16:42:19 <JasonWoof> should be pretty easy
16:42:23 <JasonWoof> there's onl 64 instructions
17:02:10 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
17:14:14 --- quit: Amanita_Virosa ("Wewps")
17:21:30 <snowrichard> hi
17:22:44 --- quit: snowrichard ("Leaving")
17:29:06 --- join: snowrichard (n=richard@adsl-69-155-177-154.dsl.lgvwtx.swbell.net) joined #forth
17:30:58 --- quit: uiuiuiu (Remote closed the connection)
17:31:02 --- join: uiuiuiu (i=ian@dslb-084-056-241-002.pools.arcor-ip.net) joined #forth
17:35:26 <snowrichard> hi
17:39:32 --- part: snowrichard left #forth
17:51:02 <JasonWoof> : fv-gen-table 0 64 for dup op, ,; 1+ next drop ;
17:51:36 <crc> what is ,;
17:51:42 <JasonWoof> postpone ;
17:52:10 <crc> ok
17:52:35 <JasonWoof> that wouldn't work for all opcodes
18:01:47 --- join: LOOP-HOG (n=chatzill@sub22-119.member.dsl-only.net) joined #forth
18:01:52 <crc> hi LOOP-HOG
18:02:07 <LOOP-HOG> hi
18:02:16 <LOOP-HOG> what's going on?
18:02:58 <crc> I'm working on documentation
18:03:13 <LOOP-HOG> on what?
18:08:06 <crc> the handbook for RetroForth 9.1
18:10:18 <LOOP-HOG> OH
18:10:20 <LOOP-HOG> neat
18:22:50 --- nick: nanstm -> Raystm2
19:42:00 --- join: segher_ (n=segher@dslb-084-056-142-124.pools.arcor-ip.net) joined #forth
19:48:15 --- join: scope__ (n=junk@njd.paradise.net.nz) joined #forth
19:48:18 --- nick: scope__ -> scope
19:54:13 --- quit: segher (Read error: 110 (Connection timed out))
20:55:30 --- quit: LOOP-HOG ("ChatZilla 0.9.61 [Mozilla rv:1.7.5/20041217]")
22:33:26 --- quit: segher_ (niven.freenode.net irc.freenode.net)
22:33:30 --- quit: uiuiuiu (niven.freenode.net irc.freenode.net)
22:33:31 --- quit: virl (niven.freenode.net irc.freenode.net)
22:33:33 --- quit: crc (niven.freenode.net irc.freenode.net)
22:33:33 --- quit: Jim7J1AJH (niven.freenode.net irc.freenode.net)
22:33:33 --- quit: saon (niven.freenode.net irc.freenode.net)
22:33:33 --- quit: madwork (niven.freenode.net irc.freenode.net)
22:34:16 --- join: crc (i=crc@pdpc/supporter/active/crc) joined #forth
22:34:16 --- join: segher_ (n=segher@dslb-084-056-142-124.pools.arcor-ip.net) joined #forth
22:34:16 --- join: uiuiuiu (i=ian@dslb-084-056-241-002.pools.arcor-ip.net) joined #forth
22:34:16 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #forth
22:34:16 --- join: Jim7J1AJH (n=jim@221x115x224x2.ap221.ftth.ucom.ne.jp) joined #forth
22:34:16 --- join: saon (i=1000@unaffiliated/saon) joined #forth
22:34:16 --- join: madwork (n=foo@derby.metrics.com) joined #forth
22:34:27 --- mode: irc.freenode.net set +o crc
23:02:30 --- join: Robert (n=robert@unaffiliated/robert) joined #forth
23:04:37 <Robert> Was reading in comp.lang.forth about someone wanting to use Forth instead of XML, and someone else pointing out the security implications...  I'm not interested in the "XML" part, just wondering what kind of solutions you know of for a secure Forth (as in one having to execute untrusted code) system.
23:06:48 <oxygene> Robert: create a sandboxed inner interpreter that disallows access to some dictionaries. then move all critical words to such a protected dict
23:09:46 <virl> I would/will use a system which doesn't support such dictionaries.
23:32:31 <Robert> How about if there is a need (for example, for a script) to do general-purpose computations? Then it would most likely need access to @, ! or equivalent words. Have you played with bounds checking, or special words to access variables with?
23:39:15 <virl> hmm? well, I don't have played with it. I only guess at the moment.
23:45:11 <virl> but what is the problem with bound checking? or for @, !? ok, when you have the raw access with it to memory then it's a pita..
23:46:07 <Robert> I suppose other ways to handle variables and arrays could be better.
23:46:19 <slava> @ and ! are too low-level
23:46:41 <virl> well, no they aren't too low level..
23:46:51 <Robert> For this application they are.
23:46:58 <virl> well, no...
23:47:09 <virl> it depends on their implementation...
23:49:38 <virl> for example, they could also check each time when they are used if it's allowed to access a memory cell.
23:50:28 <virl> and then they don't do this operation
23:51:12 <virl> or what should be better suited? a word named secure-! or what? which does the same what I said?
23:51:24 <virl> well, that's crazy or stupid.
23:51:37 <slava> a word nth ( i array -- elt ) which bounds-checks an array index and returns the element
23:52:33 <virl> disgusting...
23:53:12 <slava> why?
23:55:02 <virl> iirc, it was asked what would be 'secure' and not data integrity, I think they both don't have anything todo with each other.
23:55:29 <virl> well, you probably don't like that.
23:59:31 <slava> can xell run programs yet?
23:59:59 --- log: ended forth/06.03.31