#pentagram@irc.freenode.net logs for 11 Dec 2002 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:47:46] <-- DarkeWork has left IRC (Remote closed the connection)
[00:47:53] --> DarkeWork has joined #pentagram
[00:52:49] --> Kirben has joined #pentagram
[00:52:49] --- ChanServ gives channel operator status to Kirben
[01:52:50] <-- DarkeWork has left IRC (Remote closed the connection)
[01:53:13] --> DarkeWork has joined #pentagram
[02:54:02] <-- DarkeWork has left IRC (Remote closed the connection)
[02:54:06] --> DarkeWork has joined #pentagram
[02:59:16] --> sg1-cashman has joined #pentagram
[03:56:58] <-- sg1-cashman has left IRC (Read error: 104 (Connection reset by peer))
[05:47:39] <-- DarkeWork has left IRC ("DarkeWork has no reason")
[06:43:16] --> Darke has joined #pentagram
[06:43:16] --- ChanServ gives channel operator status to Darke
[08:01:43] --> wjp has joined #pentagram
[08:01:43] --- ChanServ gives channel operator status to wjp
[08:16:36] <wjp> <Colourless> i get the feeling that wjp is attempting to destroy the universe... he made a heap of directories and committed some source files :-)
[08:16:39] <wjp> foiled again :-)
[08:24:21] * Darke snickers.
[08:26:26] * Darke considers that now he's got just enough of his decompiler working that it 'does something', he should get just enough of a compiler working that understands it. *grin*
[08:26:59] <wjp> I'd like to have a semi-working compiler available :-)
[08:27:31] <wjp> I think I'll maybe try to implement a single 'console-output' intrinsic or something, and see if I can get some usecode scripts to, um, produce some console output :-)
[08:29:01] * Darke would suggest implementing the Bark() intrinsic to just pipe the string to con. *grin* That way you could fiddle with a few of the smaller functions that just dump a single string out, or spit out some text (books and such).
[08:29:48] <wjp> something like that, yes
[08:33:20] <wjp> of course this would require coding the string opcodes
[08:34:51] * wjp seems to remember reading some design notes on this in the logs or in docs
[08:35:51] <wjp> ah, a log snippet in the docs, even :-)
[08:37:34] * Darke snickers. There is a few cut&pastes of logs that he did to make some 'documentation' such that it is. *grin*
[08:38:24] <wjp> particularly the 'segment:offset' thing, with segment = SEG_STRING, SEG_OBJECT or something
[08:38:43] * Darke remembers he also needs to do some cleanup of Disasm.cpp, it's rather... messy.
[08:38:55] <wjp> you don't say :-)
[08:39:13] * wjp needs to read up on lists/stringlists
[08:39:14] <Darke> Which one was this?
[08:39:22] <wjp> in ucmachine.txt
[08:39:31] <wjp> "From IRC about pointers"
[08:40:15] <Darke> Also, if you're referencing the 'inline' opcode documentation, make sure to reference the stuff in Convert.h rather then Disasm.cpp, the former is more up to date.
[08:40:15] <Darke> Ahh.
[08:42:43] <Darke> I'm not actually sure that pointer stuff is what we're 'planning' anymore, that was added quite a while ago, and we've learned quite a bit more about How Things Work so the logic of it may not be as sound anymore. You'll probably have to quiz Colourless on this since it's his brainchild. *grin*
[08:43:10] <wjp> :-)
[08:43:16] <wjp> will do
[08:43:35] <Darke> Feel free to quiz me if there's anything in particular you might want to know about lists/strlists. *grin*
[08:44:38] <wjp> I wish I had an u8 usecode disassembly here :/
[08:45:20] <Darke> Lacking the usecode files?
[08:45:39] <wjp> I don't have any of U8's data files here :-(
[08:46:24] * Darke ahhs.
[08:46:32] <wjp> anyway, I seem to remember you could store a list/slist (reference?) in a 16 bit local var?
[08:46:48] <wjp> so that would probably mean we need some kind of (s)list 'heap'?
[08:48:25] <Darke> Yes. lists/slists are stored in a 'Yamm', only the offset from the start of the Yamm is stored in the variable.
[08:48:39] <wjp> yet another memory m?
[08:48:45] <Darke> As you might expect, the total size of the data is 16k.
[08:48:50] <Darke> Yet Another Memory Manager I guess. *grin*
[08:49:40] <Darke> My current guess is there's a single Yamm per thread. Though there 'might' be a global one.
[08:50:27] <Darke> I would suspect the strings, lists, slists and anything else 'large' is stored there.
[08:50:33] <wjp> do we need pointers to (s)lists?
[08:50:36] <Darke> As in, there's only one Yamm for all fo them.
[08:50:51] * Darke nods. You need them for passing them between functions.
[08:50:56] <Darke> And spawning.
[08:52:38] <wjp> looks like I'll have look at opcode 6C a bit for this
[08:53:49] <wjp> s/have/have to/
[08:53:57] <wjp> I seem to be skipping words an awful lot lately
[08:55:19] <Darke> 6C is certainly the more... 'complex' of them. *grin*
[09:01:28] <wjp> hm, judging by the comments there we may need to 'enhance' 6C a little bit
[09:01:37] <wjp> (since it misses list-element-size?)
[09:05:33] --- wjp is now known as wjp|work
[09:13:10] * Darke nodnods. It assumes any lists passed to it are two bytes long. Which works for word lists and string lists, but they apparently 'missed' the fact that it probably won't work for dword lists. Probably 'cause they never used them. *grin*
[09:47:50] <wjp|work> something to keep in mind if we ever decide to convert all usecode scripts to some 'custom' pentagram usecode format
[11:20:08] --> sg1-cashman has joined #pentagram
[11:54:07] <-- sg1-cashman has left IRC ()
[11:55:33] --> sg1-cashman has joined #pentagram
[11:59:07] <-- sg1-cashman has left IRC (Client Quit)
[12:01:57] --> sg1-cashm has joined #pentagram
[12:18:41] <sg1-cashm> Cashmans turn to go and ZZZzzzzzzzzzzzzzzzzzzzzzzzz
[12:19:35] <-- sg1-cashm has left #pentagram ()
[13:13:59] <-- wjp|work has left IRC ("gtg")
[14:18:46] <-- Darke has left IRC ("*pawwave*")
[14:38:17] --> Colourless has joined #Pentagram
[14:38:17] --- ChanServ gives channel operator status to Colourless
[14:59:44] <Colourless> about the usecode pointers, we still need to do something similar, but we need to know exactly what sort of data it can get pointers too
[15:01:02] <Colourless> if i'm not mistaken it can get pointers to lists, pointers to strings and pointers to the stack.
[15:01:24] <Colourless> now, i believe that Darke know more about the implmentation of lists and strings than i d
[15:01:25] <Colourless> o
[16:35:39] <-- Colourless has left IRC ("casts invisibility")
[16:46:47] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[16:48:20] --> wjp has joined #pentagram
[16:48:20] --- ChanServ gives channel operator status to wjp
[17:04:57] --> Dominus has joined #pentagram
[18:11:27] * wjp is happily continuing work on the usecode machine...
[18:11:39] <wjp> I hope I'll get something slightly working soon...
[18:12:00] <wjp> all this work without any kind of visible output is getting a bit annoying :-)
[18:12:32] * Dominus really wants to have something to look at soon!
[18:12:58] <wjp> pentagram does produces a binary, currently
[18:13:02] <wjp> but it doesn't do much :-)
[18:13:09] <Dominus> he he
[18:13:33] <wjp> some running usecode should be doable by the end of tonight
[18:13:49] <Dominus> great
[18:14:27] <wjp> if I don't get distracted by things like homework ;-)
[18:14:34] <Dominus> he
[22:11:15] <wjp> well, that was enough for tonight
[22:11:23] <Dominus> he he
[22:11:29] <wjp> I did actually run some usecode :-)
[22:11:41] <wjp> (only some basic pushing and popping, but still ::-) )
[22:11:45] <wjp> s/::/:/
[22:12:12] <wjp> I've been browsing disassembly for the rest of the evening... *yawn* :-)
[22:12:29] <Dominus> good god...
[22:12:55] <wjp> well, not _all_ of the evening, but still quite some time
[22:14:40] <wjp> weird, geocrawler's email address has just been removed from pentagram-cvs because of excessive bounces
[22:15:04] <Dominus> huh?
[22:15:35] <wjp> but I guess the new SF ML archives work differently
[22:15:41] <wjp> anyway, time for me to go
[22:15:50] <Dominus> k
[22:15:58] <Dominus> sleep well
[22:16:08] <wjp> OTOH... *thinks of something*
[22:16:17] <Dominus> he he
[22:20:54] <wjp> right, really going now :-)
[22:20:55] <wjp> g'night
[22:21:02] <-- wjp has left IRC ("Zzzz...")
[22:21:13] --> Darke has joined #pentagram
[22:21:13] --- ChanServ gives channel operator status to Darke
[22:21:25] * Darke wavies to Dominus. Hi!
[22:21:48] <Dominus> he he
[22:21:53] <Dominus> you can't be darke
[22:22:05] <Dominus> you must be wjp as he just signed off
[22:22:31] <Dominus> (otoh the ip address looks aussie)
[22:27:28] * Darke snickers. He obviously has 'good' timing then. *grin*
[22:27:34] --> sg1-cashm has joined #pentagram
[22:27:39] --- Darke is now known as DarkeWork
[23:00:08] --> sg1-cashman has joined #pentagram
[23:00:08] <-- sg1-cashm has left IRC (Read error: 54 (Connection reset by peer))
[23:20:45] <-- sg1-cashman has left IRC ()
[23:27:24] --> sg1-cashm has joined #pentagram
[23:30:21] --> Kirben has joined #pentagram
[23:30:22] --- ChanServ gives channel operator status to Kirben
[23:30:23] <-- sg1-cashm has left IRC (Read error: 104 (Connection reset by peer))