[00:02] Join: CL joined #corewars [01:54] MSG: Ping timeout: 252 seconds [02:07] Is there a Windoze program that will allow me to see (i.e. examine MARS addresses) how a warrior works? [02:08] that needs a single-stepper to be useful [02:10] corewin [02:10] www.geocities.com/corewin2/ [03:07] thank you [05:51] Join: fiveop joined #corewars [07:37] hmm, why do some warriors have source code available, and others do not? Is there some command to Sal that hides the source code? [08:44] duh, I can always see my source! [08:44] MSG: Read error: Operation timed out [08:47] Join: asw joined #corewars [10:44] Join: tyomitch joined #corewars [10:44] who supports pMARS? is there someone? [10:45] Nobody really. The original group doing it died somewhat. Anton Marsden(?) moved it over to an SF project, but then he left the scene as well. On the other hand, it's open source :-) [10:46] I mean, when I encounter something that looks like a bug, what should I do? [10:47] Ask here or on r.g.cw is probably your best bet. [10:47] btw, is the 1996 version the latest release of pMARS? [10:49] 0.9.2 was released from sf in 2000. [10:50] I downloaded mine from koth.org; it's 0.8.6 [10:50] so, my problem is, when I have this code: c equ 1 \n equ mov 2,3 \n add 0,c (\n is for newline) [10:51] according to the doc, it must assemble as ADD 0,1 \n MOV 2,3 [10:51] but instead it assembles as MOV 2,3 \n MOV #0, #0 [10:52] and the ADD gets completely lost. [10:56] Yep, I see that as well. Looks like a bug to me... [10:56] r.g.cw may be able to help you more, unless there's anyone else reading around here at the moment. [10:57] I've never really tried too much with multi-line EQUs (I had to look up the syntax to see what you were doing :-) ) [11:22] What I am doing is: I was trying to make an assembler alternative to pMARS, and I found that the macro language of pMARS is neither documented in full nor implemented correctly. [11:25] How can I write to rec.games.corewar? Is there a public NNTP server? (Sorry if that's a noobish question) [11:26] Google Groups? [11:28] Isn't it read-only? [11:33] Nope. [11:51] I've posted the message, hopefully it will get through to the newsgroup. [11:54] * pak21 reads news. [11:54] * pak21 notices a message entitled "How do pMARS macros work?" [11:54] Looks good :-) [13:36] old school #1 is a plain stone, quite obvious. [14:11] MSG: [14:42] Join: Mizcu joined #corewars [14:50] MSG: [16:30] Join: Fizmo joined #corewars [16:30] hi [16:36] Join: sf joined #corewars [16:36] Hello [16:40] hi sf [16:40] Are there any plans to add age limits to the beginner hill soon? [16:41] There are 7 warriors older than 100 and another 2 older than 50 [16:41] Joonas wanted to implement it until end of 2005 :-( [16:42] Fizmo!!! [16:42] You said the name no-one should say :( [16:42] ups [16:42] shame over me [16:43] seems not my best day today [16:43] Why? [16:43] I downloaded all you music :) [16:45] :) [16:45] and? [16:45] comments? [16:46] I haven't put it on my MP3 player yet [16:46] I'll do that later :) [16:47] ok [16:47] ;-) [16:47] Anyone with a warrior over ago 100 could voluntarily retire it ;) [16:50] I removed Dragonfly because it's been briefly on some pro hills [17:04] i dont think i need to kill blindfolded to make more space; it can be defeated by almost any stone [17:04] though inversed might/could remove those two top-warriors [17:06] Blur 2 ought to be taken off [17:09] yes [17:10] MSG: [17:30] Join: Core29 joined #corewars [17:32] Hi Core29 [17:36] hi =) [17:41] hi [17:45] gotty go now [17:45] maybe back later [17:45] * Fizmo waves [17:45] MSG: [17:48] * sf waves [17:49] Join: willvarfa joined #corewars [17:50] Hi Will [17:51] hello folks [18:05] Join: Fluffy joined #corewars [18:06] :-) [18:08] pak21: I think, that your idea to use a switch in PyCorewar is the best solution (i.e. behave like the standard specifies or how pMARS does it) [18:08] I've thought about it quite a lot of time. [18:09] Hi Fluffy [18:09] It seems, that all implementations follow pMARS' behaviour when evaluating A- and B-mode and not one (!) follows the standard [18:09] Hi sf [18:10] Since nobody has a standard-compatible MARS we should change/update the standard [18:10] but I doubt, that it is doable [18:10] * Fluffy is again talking to himself ;-) [18:10] has He Who Must Not Be Named commented on what is right? [18:10] yes, pMARS is wrong :-) [18:10] at least on "jmz.b < 0, # 1" [18:11] that's, what he's said [18:11] exhaust/exhaust-ma is wrong, too [18:11] ... and exmars [18:11] (still haven't checked fmars) [18:12] This "bug/feature" seems to affect about every 50th fight or so for tiny [18:12] lol wow [18:12] I've run more than 2000 fights and got about 30 problems [18:13] The other problem is, that the way the standard explains everything is quite logical [18:14] but not pMARS' behaviour [18:14] Personally I would have no problem to change the standard ... [18:14] *with changing [18:15] It would create the least amount of work [18:16] only some few lines to change in the standard and almost nothing in existing MARS [18:17] MSG: Ping timeout: 252 seconds [18:17] Join: willvarfa joined #corewars [18:17] :) [18:17] for this particular opcode? [18:17] or in general? ;-) [18:17] As far as I could understand it, it only depends on the modes [18:17] and only, if B-mode is immediate [18:17] and A-mode is { } < > [18:18] hmm care to walk me through the problem? [18:18] no problem [18:18] * willvarfa searches for some sim.c file that is online so we can both step through it.. [18:18] take the one from exhaust [18:18] It is easy to understand the code [18:19] ttp://www.cs.helsinki.fi/u/jpihlaja/exhaust/exhaust.html [18:19] h + ... [18:19] hmm can't get to my redcoder cvs browser [18:20] is there any that I can browse in a web-browser? [18:20] the example code in the standard? [18:20] http://corewar.co.uk/icws94.txt [18:21] But I can explain it without code as well [18:22] ah the spec I've read [18:22] k [18:22] (I wrote a MARS or two in the past...) [18:22] but is there a sim.c (which'd be wrong) that I can see in a browser? [18:22] don't think so [18:23] one moment [18:23] Join: Fizmo joined #corewars [18:23] hi [18:23] be back [18:23] Hi [18:23] Hi Fizmo [18:24] willvarfa: http://corewars.jgutzeit.de/temp/sim.c [18:24] it is the sim.c from exhaust [18:25] willvarfa: ready? [18:25] go on [18:26] line 582 [18:26] ra_a, ra_b, rb_a, rb_b [18:26] theses values are the a-values and b-values of the effective addresses of the a- and b-field [18:26] i.e.: [18:27] 1. current insn is read into internal registers [18:27] 2. a-mode is evaluated [18:27] 3. a-value and b-value of effective a-address is loded into ra_a, ra_b [18:27] 4. b-mode is evaluated [18:28] 5. a-value and b-value of effectve b-address is loaded into rb_a, rb_b [18:28] that's the abbrev. theory [18:28] right [18:28] now for the fine points [18:28] line 667 [18:28] start of a-mode evaluation [18:28] better 678 [18:29] ups [18:29] I've forgot line 659/660 [18:29] the a-value and b-value of the executed insn is read into internal registers [18:29] (in_a, in_b) [18:30] yeap [18:30] line 676: value of in_a is stored into rb_a!!!! [18:30] that's only needed, if B-mode is immediate [18:30] we'll come back later to it [18:31] untill line 714 a-mode is evaluated [18:31] ! got to go [18:31] and core is changed (if necessary) [18:31] * willvarfa waves [18:31] MSG: Quit: Chatzilla 0.9.68.5.1 [Firefox 1.5.0.1/undefined] [18:31] hehe [18:32] * Fluffy sighs [18:43] * sf waves [18:43] * Fluffy waves to sf [18:44] * Fizmo has made an update [18:44] Fizmo: URL? [18:44] http://www.corewar.info/newsletter.htm [18:44] Core Explorer und Nano Warrior are now available on my site [18:45] actually as txt only [18:45] :) [18:45] MSG: Quit: I was using TinyIRC! Visit http://www.tinyirc.net/ for more information. [19:36] Join: Mizcu joined #corewars [19:38] Hi Mizcu :-) [19:39] hi F [20:25] Join: John joined #corewars [20:25] Hi [20:25] Hi John [20:27] Join: John_ joined #corewars [20:28] Hi John_ [20:28] ;-) [20:28] :-) [20:29] Updated nanohof online [20:29] Updated bibliography online [20:29] :-) [20:29] MSG: Ping timeout: 252 seconds [20:29] Nick Change: John_ changed nick to John [20:30] Join: John_ joined #corewars [20:31] Bvowk's down to 3 warriors on nano now! [20:31] And inversed has his first entry to the nano HoF :-) [20:34] MSG: Ping timeout: 252 seconds [20:34] Nick Change: John_ changed nick to John [20:35] hi John [20:35] Hi Christian [20:41] Join: John_ joined #corewars [20:41] John: What character would you use for preincrement and postdecrement [20:42] I wouldn't encourage the use of more addressing modes ;-) [20:42] John: Still there? [20:42] why? [20:43] I mean, do you have any special reason for that [20:43] But otherwise, I would use the same as Stefan Strack in Corewar Pro [20:43] which were? [20:43] I mean: what did he use? [20:43] I can't remember now [20:43] MSG: Ping timeout: 252 seconds [20:43] Nick Change: John_ changed nick to John [20:43] How about: << >> {{ }} ? [20:44] [] [20:44] yes, that's right, Mizcu, but how about the other possibilities [20:45] {} are bad! [20:45] we need 4 different characters for that [20:45] /\ [20:46] hmm ... mov.i / 0, \ 1 [20:46] more addressing modes would make writing warriors way easier [20:46] I suggested to use just an invert character [20:46] like § [20:46] so it's not good! [20:46] mov.i # 0, <§ 1 [20:46] < <§ > >§ { {§ ..... [20:46] yes, is confusing [20:47] mov.i # 0, << 1 [20:47] mov.i # 0, }} 1 [20:47] hmm [20:47] Join: John_ joined #corewars [20:47] John_: Please look at the logs [20:47] What do you think? [20:48] mov.i # 0, [ 1 [20:48] mov.i # 0, ] 1 [20:48] that more addressing modes are unneccessary if not damaging the fun in writing warriors! [20:48] I think strack used / \ [20:49] John_: but we would need two more chars to distinguish between A- and B-field [20:49] MSG: Ping timeout: 252 seconds [20:49] Nick Change: John_ changed nick to John [20:50] or you distinguish >{ for A- and B-Field and /\ for post- and pre [20:50] I don't see why it shouldn't be [ ] [20:50] You could also give the option of using an invert character [20:51] It doesn't matter if there are numerous representations in the source code [20:51] They'll all compile to the same thing [20:51] Like SEQ / CMP [20:51] Might be a good idea to it that way [20:51] now about opcodes: [20:51] spz - split if zero [20:51] spn - split if not zero [20:52] Implement all of http://corewar.co.uk/opcodes.txt [20:52] ijn [20:52] spz/spn would replace normal spl almost completely [20:52] new ins. should add, not replace [20:52] ijn - increment and jump, if not zero [20:53] Mizcu: Why would spz/spn replace spl? You still have to decide, what to "scan", which might not always be a good idea [20:54] Since papers are quite strong warriors, can you think of any instructions, which would increase the strength of scanners and stones? [20:54] conditional moves [20:55] Mizcu: Where do you store the condition? [20:55] mov if target is zero/non-zero [20:56] moz.i # 0, $ 1 [20:56] it would only move, if B-field ($ 1) is zero [20:56] sounds interesting [20:56] hehe, imps, that die, if they encounter an opponent [20:57] perfect troll, i say [20:57] how would you use moz/mon is an scanner/stone? [20:57] *is->in [20:58] *an->a [21:00] John: About your list of opcodes ... most propositions aren't that good [21:01] opcodes.txt updated ;-) [21:01] Join: sascha joined #corewars [21:01] John: Do you have any favorites? [21:01] Hi Sascha [21:01] Fluffy: not really [21:01] Hi2all [21:01] hi Sascha [21:01] many of them are a bad idea for one reason or another [21:01] what do you think of spz/spn? [21:02] Oh, i'm jumping into a "what do you think about a new standart" discussion ;-) [21:02] John: what would be against spz/spn/ijn? [21:02] spz spn could make interesting bombs [21:02] Sascha: Bingo :-P [21:03] what about a stack? [21:03] ijn would make a good decoy against djn streams :-) [21:03] I wouldn't try to drastically change cw by new opcodes [21:03] just to give some new possibillities of snippets [21:03] sascha: we already have p-space [21:03] so spz/spn/ijn would be fine for me [21:04] p-space won't give an individual stack for *each* process though :-P [21:04] now that might be handy [21:04] and what would you place in the stack [21:05] insn? values? [21:05] both? [21:05] isn't the most popular hill,the non-pspace one? Doesn't seem like new operands are popular. [21:05] instructions, then we could execute code on the stack ;-) [21:05] there could an insn-stack and a value-stack [21:05] (actually, I wrote an 8086 program once which created some code on the stack, then jumped to the stack) [21:05] there is more valour in making strong single warrior than choosing bunch of components and make a new P'3 -tabke for it [21:06] Hi CL [21:06] John: Then an insn-stack would be some kind of protected area [21:06] John: Nobody could kill you there [21:06] hello [21:07] we have no need for another data/instruction layer [21:07] If a few extra opcodes were created, we could implement Darwin in Redcode [21:07] heh [21:07] CL: You're right, the most popular hill is the 94nop ;-) [21:09] As I said, I wouldn't go to change cw too much by new opcodes [21:10] Nevertheless, we also have read- and write-limits which aren't discovered so far [21:10] New opcodes might make an interesting tournament round [21:10] John: I definitely will implement ijn/spn/spz in PyCorewar [21:10] (just to see, what you can use it for) [21:10] hmm, that would be a good playground to see what happens with new opcodes [21:10] I think Steve Gunnell has a pmars patch for IJN [21:11] Yes, i think we should think about a set of new opcodes and adressing modes and start a nice rf-round..after the round we will know more [21:11] Someone is sure to find a way to exploit them :-) [21:11] wjm [21:12] it would take publishing the winners code to entice others [21:12] Well, looks like RF26 will be running with PyMars :-P [21:12] wjm - win, if John Metcalf ;-) [21:12] and Fluffy organize ;-) [21:12] Hmmm... when was the last time I won something? :-( [21:12] Fizmo: Would like to :-) [21:12] I don't have time anymore [21:13] hey P.Kline is back one 94nop with a 2by4k-clone ;-) [21:13] Corewar seeks the Superstar [21:14] Jesus Christ Superstar [21:14] John I think you're under the last candidates ;-) [21:14] Never found a warrior of this "Jesus" in planars archives :-) [21:15] Time to write a warrior [21:15] should be aged at least 2006 [21:15] :-P [21:21] Bvowk's nano evolver doesn't seem to be sending much :-( [21:24] MSG: Quit: humhum [21:35] MSG: Read error: Connection reset by peer [21:43] * sascha waves [21:43] time to sleep [21:43] * John waves [21:43] me, too :-) [21:43] * Fluffy waves [21:43] Part: sascha left #corewars [21:43] MSG: Quit: spz.b # 0, < 1 [21:47] ok, time to go for me as well [21:47] * Fizmo waves [21:47] bye [21:47] MSG: [21:47] * John waves [21:48] http://corewar.co.uk/lphof.txt updated [21:59] Join: John_ joined #corewars [22:01] MSG: Ping timeout: 252 seconds [22:01] Nick Change: John_ changed nick to John [22:33] MSG: Ping timeout: 252 seconds