00:00:00 --- log: started forth/01.11.08
03:18:16 --- quit: clog (Ping timeout)
03:18:16 --- log: stopped forth/01.11.08
03:19:57 --- log: started forth/01.11.08
03:19:57 --- join: clog (nef@bespin.org) joined #forth
03:19:57 --- names: list (clog plbofh @MrReach)
03:35:34 --- join: aum (aum@210-54-227-86.adsl.xtra.co.nz) joined #forth
03:58:01 --- quit: aum ()
04:14:48 --- quit: MrReach (Read error to MrReach[209.181.43.190]: Connection reset by peer)
04:16:20 --- join: MrGone (mrreach@209.181.43.190) joined #forth
04:39:12 --- join: edrx (edrx@200.240.18.94) joined #forth
04:39:28 --- quit: edrx (Read error to edrx[200.240.18.94]: EOF from client)
04:55:01 --- quit: MrGone (Read error to MrGone[209.181.43.190]: No route to host)
04:56:15 --- join: MrGone (mrreach@209.181.43.190) joined #forth
05:07:55 --- join: edrx (edrx@200.240.18.82) joined #forth
05:28:02 --- quit: clog (Ping timeout)
05:28:02 --- log: stopped forth/01.11.08
05:28:10 --- log: started forth/01.11.08
05:28:10 --- join: clog (nef@bespin.org) joined #forth
05:28:10 --- names: list (clog MrGone edrx plbofh)
05:50:52 --- nick: MrGone -> MrReach
05:51:06 --- mode: ChanServ set mode: +o MrReach
05:51:13 --- mode: MrReach set mode: +o edrx
05:51:18 <MrReach> blah
05:52:19 <MrReach> Hitachi has, apparently, built the super-duper microcontroller series with DSP and floating point capabilities, but neglected to write an assembler for them
06:07:58 --- quit: edrx ([x]chat)
06:57:33 --- nick: MrReach -> MrGone
10:55:06 --- join: Speuler (roberto@orange.icafe.spacenet.de) joined #forth
10:55:16 <Speuler> g'day
10:56:06 <Speuler> mrreach: i have put the beginnings of a forth interpreter, written in asm, together with the script to assemble it, using NASM, on http://blik.wox.org/software/forth
10:59:32 --- part: Speuler left #forth
12:12:11 --- join: Speuler (akhandel@preiselbeere.icafe.spacenet.de) joined #forth
12:13:10 --- nick: MrGone -> MrReach
12:13:12 <MrReach> hihi!
12:13:18 <Speuler> hi mrreach
12:13:28 <MrReach> reading those sources now
12:14:00 <Speuler> i don't have cvs on that server
12:14:16 <MrReach> I don't imagine you would
12:14:26 <MrReach> CVS is a complicated beast in any case
12:14:40 <MrReach> though not bad once running
12:15:01 <Speuler> the parts commented out i have pulled from previously written forth,. i think they can be recycled 
12:15:01 <MrReach> how come there are no notes about how its put together?
12:15:06 <MrReach> are they in German?
12:15:11 <Speuler> notes ?
12:15:43 <MrReach> in the past, when I wrote a system, it always started with a text doc explaining dictionary layout, threading, etc
12:15:50 <MrReach> I take it you don't do that?
12:15:57 <Speuler> i dod so once
12:15:59 <Speuler> did
12:16:10 <Speuler> but nowadays i start with the source, and document later
12:16:17 * MrReach laughs.
12:16:20 <Speuler> when i know how i'm going to put things together
12:16:32 <MrReach> easier to write than to document, the bane of forth programmers
12:16:33 <Speuler> to avoid discrepancies between docs and program
12:16:53 <Speuler> misleading docs are worse than no docs
12:17:16 <MrReach> that is quite true
12:17:32 <Speuler> i'd like to make the doc part of the source
12:17:40 <MrReach> so, which level of processor did you want the assembler to support?
12:17:49 <Speuler> so the tow things can be changed in one place
12:17:51 <Speuler> two
12:19:39 <MrReach> from what I see ... read1char will cause the line to accumulate until RETURN is pressed ... is this true?
12:21:30 <Speuler> read1char doesn't work yet
12:21:34 <Speuler> waits for a CR
12:21:40 <Speuler> shouldn't
12:21:51 <MrReach> yep, need to call the ioctl sometime during startup
12:21:59 <Speuler> idea is to buffer incoming string, and have key read from buffer
12:22:21 <Speuler> you know better than i do
12:22:53 <MrReach> really? you've never tripped over the line buffering of console input before?
12:23:06 <MrReach> I hate it ... screws up everything, even modem drivers
12:23:19 <Speuler> not with linux
12:23:48 <MrReach> ok, lemme see if I can find the kernel call for it
12:24:01 <Speuler> (not tripped because haven't had single character input running as far
12:24:02 <Speuler> )
12:24:11 <MrReach> which processor do you intend to write the assembler for?
12:24:32 <Speuler> i consider using something similar like i did with elastic:
12:24:41 <Speuler> a cpu-independant assembler :)
12:24:59 <Speuler> had virtual mnemonics which got translated to the target opcode
12:25:34 <Speuler> so, what is written in virtual assembly, should translate to anything for which the virtual mnemonics have translation rules
12:25:47 <Speuler> which is sometimes not so efficient like straight assembler,
12:25:53 <Speuler> bit very much more portable
12:26:22 <Speuler> one virtual mnemonic might be:   mov r1,r2
12:26:31 <MrReach> heh, since when were assemblers supposed to have portable source?
12:26:45 <Speuler> on intel, it might be   mov  ebx, eax
12:27:11 <Speuler> i don't know whether they are supposed to or not. if they are, i reserve me the right to ignore it
12:27:43 <Speuler> on 68000, it might be  move  A0, A1
12:28:06 <Speuler> on H8,    mov.w ER0, ER1
12:28:25 <MrReach> haha! are you listening to BBC World Service?
12:28:41 <MrReach> he just mentioned an "inscrutible sheep"
12:28:49 <Speuler> brb
12:29:00 <Speuler> need electricity again :)
12:29:16 <Speuler> service is just sorting glasses
12:29:28 <Speuler> don't want her to step down from the ladder :)
12:30:06 --- join: Speuler_ (roberto@banane.icafe.spacenet.de) joined #forth
12:30:09 <Speuler_> ahem
12:30:16 <Speuler_> other machine for now :)
12:30:53 <Speuler> here
12:31:00 --- quit: Speuler_ (Speuler_)
12:31:01 <MrReach> ok
12:31:16 <MrReach> you ought to install ICQ on one of those machines
12:31:38 <Speuler> there is icq installed on all of them
12:33:38 <Speuler> the logic of  0< 0= = etc has been carefully crafted for code-copying  
12:34:19 <Speuler> i hope i am proud on those for no bad reason
12:34:27 <MrReach> ok, am looking at syscalls now
12:34:42 <Speuler> that would be extremely helpful indeed
12:35:14 <Speuler> i was considering pulling out that code from another program
12:35:43 <Speuler> but that's a kind of work i don't like very much :)
12:36:18 <Speuler> the outer interpreter comes from an embedded implementation. there are still watchdog timer kicks in it
12:36:44 <MrReach> heh
12:36:58 <MrReach> why do you put the tasker at such a low level?
12:37:00 <Speuler> and the tasker might have to be adapted to take advantage of thread if that would be of any use
12:37:14 <Speuler> coming from pervious implementation too
12:37:22 <Speuler> where the interpreter was the os
12:38:54 <Speuler> same goes for quare root ... not really useful here but what the heck i thought. jsut stuff it in :)
12:39:13 <MrReach> yes, I saw that
12:39:19 <MrReach> rather raised my eyebrows
12:39:53 <Speuler> might throw it out again.
12:39:56 <MrReach> heh, the ioctls allow you get and set the keyboard LEDs
12:40:38 <MrReach> and make the little speaker beep with various tones
12:40:47 <Speuler> what docs are you reading now ?
12:41:08 <MrReach> man 4 console_ioctl
12:41:12 <Speuler> i was looking for better info on syscalls on my machine
12:41:29 <Speuler> (which is off-line now)
12:41:40 <MrReach> then you probably want Reckenwald's page
12:41:42 <Speuler> only got what has been installed already
12:41:58 <Speuler> i can take some doc home with me on a floppy
12:42:23 <MrReach> http://home.snafu.de/phpr/
12:43:38 <Speuler> do you, by any chance, know what scsi raid controllers are well supported with linux ?
12:44:04 <MrReach> heh, no
12:44:20 <MrReach> there's lots of docs about them, though
12:45:26 <Speuler> is there any hardware channel on this net ?
12:46:11 <MrReach> on openprojects? I should think so
12:46:19 <MrReach> try #linux
12:46:33 <MrReach> getting channel list now
12:47:13 <Speuler> brb
12:53:02 <Speuler> back now
12:53:11 <MrReach> ok
13:01:25 <Speuler> on linuxhelp they recommend adaptec raid controllers
13:01:42 <MrReach> that would make sense
13:02:04 <MrReach> GAAARGH! I *hate* setting line disciplines for consoles
13:02:16 <MrReach> I always forget how to do it, and spend hours refinding the info
13:17:53 <MrReach> blarg
13:18:00 * MrReach checks the eforth sources
13:37:43 <Speuler> ?key may be somewhat problematic too
13:37:58 <Speuler> as unix doesn't like apps to poll for keys
13:38:28 <MrReach> well, that can be circumvented via a 1-cell buffer
13:39:24 <MrReach> : key ( -- char ) key-buf @ dup -1 = if (key) then ;
13:39:29 <MrReach> I think that's right
13:39:49 <MrReach> nope ...
13:40:17 <Speuler> drop (key)
13:40:24 <MrReach> : key ( -- char ) key-buf @ dup -1 = if (key) else -1 key-buf ! 255 and then ;
13:40:40 <Speuler> @ 1+ ?dup if (key) ...
13:40:42 <MrReach> (key) physically gets the keystroke
13:41:05 <MrReach> ok
13:41:23 <MrReach> hrm ... that still doesn't fix the blocking if no keystroke is ready
13:41:29 <Speuler> could be "0" in key-buf in case of empty too
13:41:37 <MrReach> ioctl can be used to fix that, also
13:41:53 <MrReach> it is possible to receive a null byte in the input stream
13:42:05 <MrReach> so need a flag that is more than 8 bits
13:42:27 <Speuler> stick a fifo in between, and read # of fifo-elements with  ?fifo
13:42:45 <Speuler> asynchronously reload fifo
13:42:53 <MrReach> eforth appears to use non-blocking input
13:43:08 <MrReach> heh, then we're into signals ... *FUN*
13:43:30 <Speuler> no need for fifo ...
13:43:45 <Speuler> we type slower than programs reads keys ...
13:44:01 <MrReach> yes, the blocking when no key is ready is the only concern
13:44:06 <Speuler> problem is if there is no key, not if there are too many of them ...
13:44:25 * Speuler slaps himself with a steam roller
13:46:56 <Speuler> begin ?key until   is no good
13:47:20 <Speuler> but programmable
13:47:25 <MrReach> two things *MUST* be specified to the kernel:
13:47:42 <MrReach> 1. that buffering is turned OFF
13:48:01 <MrReach> 2. that a read where no input is available is non-blocking
13:48:04 <Speuler> begin ?key not while  (sleep-until-key) repeat 
13:48:05 <MrReach> oh, also ...
13:48:25 <MrReach> might want to tell kernel to use "RAW" input ... no translation
13:48:31 <Speuler> yes, good
13:49:23 <MrReach> I like menus that pop up by pressing ESC and F1 (even in text mode)
13:49:46 <MrReach> that's impossible with line-buffered input
13:49:59 <MrReach> ok, I can say empirically that "it can be done"
13:50:18 <Speuler> it has been done, so it is possible
13:50:26 <Speuler> zed does it
13:50:33 <Speuler> (editor)
13:51:31 <MrReach> iomode:A
13:51:31 <MrReach> 	push ebx
13:51:31 <MrReach> 	sys_fcntl [uref uPINP],F_SETFL,O_RDONLY|O_NONBLOCK
13:51:31 <MrReach> 	sys_fcntl [uref uPOUP],F_SETFL,O_WRONLY|O_SYNC
13:51:31 <MrReach> 	pop ebx
13:51:32 <MrReach>     ret
13:52:19 <MrReach> that's from Reckenwald's linux version of eForth
13:53:30 <MrReach> btw, signals are *WAY* cool
14:08:21 <Speuler> mrreach: you should still have the shell on blik.wox.org available to you
14:08:24 <MrReach> better, I think, to write the assembler as expected by other coders, then setting up macros for uniformity across platforms
14:08:35 <MrReach> I think I do, yes
14:09:06 <MrReach> this eforth is written in NASM, btw
14:09:16 <MrReach> you might want to look at the sources
14:09:20 <Speuler> by uploading the ssh public key, you autolog in and out from command line
14:09:42 <Speuler> maybe some kind of version control can be build around this ?
14:09:56 <MrReach> I use SecureCRT from my windows machine
14:10:20 <Speuler> hmm - wouldn't know about that one
14:10:24 <MrReach> eventually, we will want CVS, for sure
14:10:32 <Speuler> (only know putty on win boxes)
14:11:03 <MrReach> if you have a windows machine, you should try SecureCRT ... it's awesome software
14:11:10 <Speuler> would cvs on openprojects be an option ?
14:11:22 <MrReach> even has a "linux console" mode
14:11:28 <MrReach> fine with me
14:11:42 <Speuler> i have never set up a project on openprojects
14:12:09 <Speuler> been using TLIB a lot.
14:12:09 --- mode: MrReach set mode: +o Speuler
14:12:28 <Speuler> currently MKS which i don't like so such
14:12:30 <Speuler> much
14:12:31 <MrReach> yes, there are several CVS services
14:12:46 <MrReach> may I suggest that one of use install Zope on one of our machines?
14:13:04 <MrReach> integrated CVS w/ web access, mailing lists, and wikis
14:13:13 <Speuler> sounds way cool
14:13:23 <MrReach> heh, if you like Python
14:13:38 <Speuler> i think i like it
14:13:57 <MrReach> http://www.zope.org
14:14:00 <Speuler> (only had some short glimpses at)
14:14:23 <MrReach> python looks pretty good to me, but interpreted
14:14:37 <MrReach> haven't really looked at it in depth
14:14:46 <MrReach> OH! that reminds me! ...
14:15:19 <MrReach> have you considered using a token interpreter ... so that Forth binaries can be portable across platforms?
14:15:27 <Speuler> yes, i did
14:15:33 <Speuler> liked the idea
14:15:48 <MrReach> similar to java, but w/o the oerhead of typing
14:15:53 <Speuler> but wanted to get something conventional running quickly
14:16:00 * MrReach nods.
14:16:28 <Speuler> with option to remodel it once some base stuff works
14:17:03 <Speuler> returning to a known working state seems to me easier than getting some new model that far ...
14:17:20 <MrReach> ok, but you end up with cruft in the sources
14:17:52 <Speuler> as long the system is minimal, remodeling it shouldn't be such huge a task
14:18:26 <MrReach> minimal prims, system itself can be huge
14:19:06 <MrReach> oh, when I wrote my forth, it took (IIRC) 54 words to get to INCLUDED ... have you done it in less?
14:20:31 <Speuler> wouldn't know the word count on elastic when it compiled.
14:20:52 <Speuler> but the early version of it was my most compact forth as far
14:21:08 <Speuler> (that was before it became metastic)
14:21:10 <MrReach> that was the objective, in my opinion ... getting the system far enough along that I can extend itself
14:21:24 <Speuler> writren elastic, to remodel it to a metacompiler
14:21:30 <MrReach> in fact, you had to include SAVE-SYSTEM
14:21:37 <Speuler> then i dropped it ...
14:21:56 <Speuler> i didn't use included ...
14:22:13 <Speuler> but  type sources | elastic :)
14:22:26 <Speuler> loaded include that way too
14:22:33 <MrReach> haha!
14:22:54 <Speuler> elimiated the need for disk i/o at that stage
14:23:07 <MrReach> I think Ting did that, too, for a while ... led me to believe he was irritating
14:23:20 * Speuler can be irritating too
14:23:38 <MrReach> good solution, though, even if it's a bit kludgy
14:23:42 <Speuler> ask i440r :)
14:23:59 <MrReach> I don't have to ... he's easily irritated
14:24:24 <Speuler> i included an  i440r conditional in toyforth :)
14:24:42 <MrReach> heh
14:24:54 <MrReach> oh, do you intend toyforth to be ANS compliant?
14:25:20 <Speuler> lets say, leaning towards ans compat
14:25:32 <MrReach> ok
14:25:34 <Speuler> but not necessarily 100%
14:25:36 <Speuler> depends
14:25:54 <Speuler> not my intention to by counter-ans by all means
14:25:57 <MrReach> right, I presume you don't intend to implement all words, but those that you do will work as advertised
14:26:24 <Speuler> that sounds like a good base
14:27:00 <MrReach> do you intend to implement a full-scale testing framework?
14:27:18 <Speuler> i can live with the concept of naming words differently if they differ from ans
14:27:19 <MrReach> so that it borks if you break some bit?
14:27:41 <Speuler> would be helpful
14:28:21 <MrReach> I'm getting the impression that you're considering porting to a plethora of platforms ... am I mistaken?
14:29:10 <Speuler> i'd like to able to take it with me to other platforms, yes, but don't design especially for another platform right now
14:29:29 <MrReach> ok
14:29:39 <Speuler> wouldn't have started using asm otherwise
14:29:48 <MrReach> shouldn't be an issue so long as few prims
14:29:55 <MrReach> what would you have used?
14:30:09 <Speuler> (have done fairly little asm under linux, so it would be a good exercise too)
14:30:32 <Speuler> a target compiler written in AWK :)
14:30:42 <MrReach> YUCK!!!
14:30:53 <Speuler> hehe
14:31:04 <MrReach> that's as bad as m4
14:31:51 <MrReach> actually, you might be amused to know that I've been pondering how to implement a .SECTION directive in a forth metacompiler
14:32:59 <Speuler> what debugger would one use with a program, assembled with nasm ?
14:33:16 <Speuler> gdb ?
14:33:24 <MrReach> I would use embedded print statements
14:33:31 <Speuler> anyting more visual ?
14:33:53 <MrReach> isn't there a debugger IDE at the nasm site?
14:34:01 <Speuler> breakpoint, single stepping would be helpful sometimes
14:34:36 <MrReach> also, I think NASM output files are designed to work with EMACS and GDB
14:35:06 <MrReach> why not write a pentium assembler and disassembler ... and do it right?
14:35:14 * MrReach grins behind his hand.
14:35:48 <Speuler> cause the system would have to work in order to use its buils-in debugging facilities
14:36:36 <MrReach> do you even own a Windows machine?
14:36:37 <Speuler> that's when a debugger is less important
14:36:54 <Speuler> i have vmware on my notebook, with a w98 installation
14:37:03 <MrReach> ok, pretty close
14:37:21 <Speuler> mainly for downloading detail maps to the gps
14:37:38 <MrReach> please try to isolate all the i/o stuff into one section ...
14:37:47 <Speuler> and to show off with linux :)
14:37:59 <MrReach> I intend to port your code to Win32, and it should be trivial
14:38:55 <Speuler> shouldn't be too much of a hassle to keep the system specific part apart
14:39:30 <Speuler> could call all system specific code through one i/o function
14:39:58 <MrReach> many people don't realize that Win32 provides a console API ... it's started with one system call, and then the stdin/stdout/stderr works as expected, AND one can abandon the event-driven programming model (if desired)
14:40:05 <Speuler> procided, same cpu, same mode
14:40:46 <Speuler> would probably fairly minimal os interface anyway for the start
14:40:59 <MrReach> no need to bottleneck like that
14:41:21 <MrReach> better to have several words that are clearly expected to be system-dependant
14:41:58 <MrReach> even if I had to rewrite all 40 of the primitives, it still wouldn't be such a bad thing
14:42:18 <MrReach> but I'm only expecting to have to write 5-8 i/o prims
14:43:10 <MrReach> do you expect the kernel to link to glibc at runtime?
14:43:26 <Speuler> not planned at the moment
14:43:31 <MrReach> (I see that so far, you've only been using syscalls)
14:43:51 <Speuler> i tried glib exit
14:43:58 <Speuler> quit, that is
14:44:09 <Speuler> ahem .. bye
14:44:15 <MrReach> couldn't get it link??
14:44:22 * MrReach chuckles
14:44:43 <Speuler> i don't recall what the problem was :)
14:45:05 <Speuler> i think, core dump
14:45:21 <Speuler> well, that's a way to exit a program too
14:45:31 <MrReach> at the minimum, I'd recoomend linking enough to use LoadLib and FindProc
14:45:51 <MrReach> I think GNU provides a seperate, much smaller library to do that
14:46:06 <Speuler> would help immensely to interface to the higher-level system stuff
14:46:34 <Speuler> being able to call x functions is something i'd like to be able to do
14:46:47 <Speuler> but not given much thought on that yet
14:46:53 <Speuler> postponed 
14:46:57 <MrReach> ok, you're gonna need a LoadLib for sure, then
14:47:22 <Speuler> i've put it into the "optional" drawer
14:47:31 <Speuler> add-on
14:47:52 <Speuler> too early at this current stage
14:47:59 <MrReach> (unless you're gonna talk raw XProto, which I would recommend against)
14:48:32 <Speuler> i agree without restriction with you on this point
14:48:58 <MrReach> well, I've got to go renew my ID ... back in about two hours
14:49:04 <Speuler> don't want to reimplement the functionality of xlibs
14:49:10 <MrReach> (expires on my birday tomorrow)
14:49:27 <Speuler> provider's gift
14:49:36 --- nick: MrReach -> MrGone
14:49:37 <Speuler> i'll be gone in two hours
14:49:49 <Speuler> thanks for advice
14:49:55 <MrGone> ok, see you tomorrow, then
14:50:02 <MrGone> my pleasure
14:53:38 --- part: Speuler left #forth
15:30:48 --- join: qless (qless@clgr000977.hs.telusplanet.net) joined #forth
15:43:31 --- join: fare_ (Fare.LISPM@samaris.tunes.org) joined #forth
15:43:57 --- nick: fare_ -> FareWell
15:44:05 --- quit: FareWell (Connection reset by pear)
16:18:31 --- nick: MrGone -> MrReach
16:18:37 --- mode: MrReach set mode: +o qless
16:19:05 <qless> thx mrreach
16:19:41 <MrReach> np
16:50:55 --- join: XeF4 (xef4@apoc.undergrid.net) joined #forth
16:56:30 --- quit: XeF4 (changing servers)
17:14:17 --- quit: qless (Read error to qless[clgr000977.hs.telusplanet.net]: EOF from client)
18:39:09 --- join: I440r (mark4@1Cust240.tnt1.bloomington.in.da.uu.net) joined #forth
18:39:09 --- mode: ChanServ set mode: +o I440r
18:39:20 --- mode: I440r set mode: +o plbofh
18:42:42 --- mode: I440r set mode: +o clog
21:43:02 --- join: aum (aum@210-54-227-86.adsl.xtra.co.nz) joined #forth
21:43:34 --- quit: aum ()
23:51:26 <I440r> who aum
23:51:33 <I440r> ?
23:52:09 <I440r> well neway - time for zzz :)
23:52:12 --- quit: I440r ()
23:59:59 --- log: ended forth/01.11.08