[01:33] MSG: Ping timeout: 240 seconds [01:33] Join: Caelian|w joined #corewars [02:08] Join: Caelian_ joined #corewars [02:11] MSG: Ping timeout: 240 seconds [06:09] Join: fiveop joined #corewars [08:19] Nick Change: Caelian_ changed nick to Caelian [09:19] ugh [09:20] why couldnt they just use ordinary power-button like everyone else? [09:21] anyways, back to work, just had to burn openbsd.. [10:05] Join: Core_old_ joined #corewars [12:35] MSG: Ping timeout: 240 seconds [13:10] now i just need a command-line torrentclient [13:28] though id rather dig my eye out than install windowing system or java [13:29] The original BT client is a simple command-line job [13:30] Can do ncurses interface if that's not too heavy for you :-) [13:34] hmm, should i run both computers with their own ip's or use NAT and a million port-redirects [13:35] (DMZ support in the box sucks) [13:48] and figure out the firewall on that thing, once i get an ssh in, i dont have to swap any wires between these 'puters [15:05] why a million port redirects? [15:05] you've got one box that you use for everythign right? [15:05] just redirect it all to it [15:21] Join: Core_old_ joined #corewars [15:23] MSG: Client Quit [15:32] Join: cc_toide joined #corewars [15:34] MSG: Ping timeout: 240 seconds [15:38] MSG: Quit: Ex-Chat [15:39] bvowk: DSL box only supports single ports in port-redirection, and its DMZ still sucks [15:39] oh. [15:39] I thought you were talking about your new server [15:39] * bvowk preachs the gospel of pf, and tells of its good works in the wilder parts of the intarweb! [15:42] Join: Caelian|w joined #corewars [15:45] so far only problem with BSD has been my own unability to use *nix:es, but biggest problem has been hardware in its own way [15:47] hell, BSD even supports its soundcard straight, its about cables [15:48] i need to get PSU-cable extender [15:49] and not to touch the cable between MB and the hub of powerbutton, audio connector and two usb's [16:00] Nick Change: cc_toide changed nick to cctoide [19:08] :) [19:11] Anybody alive? [19:11] meh [19:12] Did anything noteworthy happened? [19:12] nope [19:14] :( [19:17] Millions of RAM sticks got violently bludgeoned by information, were refreshed and then got amnesia. [19:17] Among other things. [19:18] * Fluffy shrugs [19:18] * bvowk uses a taser on fluffy [19:18] as long as my benchmarks run fine, I'm satisfied [19:19] think of the RAM [19:20] Any special reason to think of it? [19:21] it's useful and possesses an acronym propicious to double entendres? [19:26] that's no *special* reson [19:35] Join: Ant_222 joined #corewars [19:36] Hi all! Any C programmers here? [19:36] hi [19:36] maybe.. [19:36] it depends on your problem ;) [19:36] and the solution! [19:36] The problem is very simple. [19:37] then solution isn't [19:37] In file a.c I have a function defined. [19:37] there's your first mistake! ;) [19:37] (sorry continue) [19:37] yes, ignore bvowky! [19:37] In b.c I don't declare this function nor include any .h files [19:38] Why is this function visible from b.c? [19:38] "visible"? [19:39] C won't bother you for actual functions until you try linking.. [19:39] Hi Ant (sorry, still no splitter under works, been kicking puters around most of the day) [19:39] it links ok [19:39] Hi, Mizcu [19:40] you're compiling with both files on the command line I take it then. [19:40] Yes and what? [19:40] (so the C compiler can see all the symbols) [19:40] I still don't see a problem [19:41] Hmmm [19:41] if you compile them to serparate .o files, you'll likely see the error [19:41] and thus, the need for either extern'ing or #including the definitions. [19:41] Yes, and I'll be able to define same-named function in my files. [19:41] you can always do that :) [19:42] extern in for variables not function, right? [19:42] if you ask, C will give you a warning.. [19:42] but you can always brutally shoot yourself in the foot :) [19:42] extern is for symbols.. [19:42] be they variables or functions [19:42] bvowk: no, not always. With both files specified in commandline, I can't have functions with the same name in the files. [19:43] What is the difference between externing a function and just #including it? [19:44] externing is a little like taking your compiler aside and saying "Hey, I've got this thing you should know about, just take my word for it on the details ok?" and #include is where you actually define it for real. [19:44] you can use extern to call C libs from C++ too. [19:44] Ant_222: sth. like: cc a.c b.c -o a.out ? As soon as you link them together you can't have two equal indentifiers, because "a.out" can't decide which to use. [19:45] But shouldn't linming be done according to .h files? [19:45] linKing [19:46] no, why should it. the identifier would exist in both files nonetheless [19:46] when you're linking, you're stuffing all the symbols together in a big wad [19:46] under C++ they fix having multiple symbols named the same thing in different files with name mangling [19:47] C doesn't do that, so your symbols all come together in one big sticky wad of evil. [19:48] So, how do I compile them so that the scope is determined by .h files only, and if don't want to "export" a function I just don't put it in the .h file. Don't tell me it's impossible. [19:49] #define/#ifndef [19:49] so you have two functions that are named the same thing being combined into the same program? [19:50] (and they're *DIFFERENT*) [19:50] bvowk: two modules which have same-named functions, just like same-named methods in different objects in C++. [19:50] I translate: just use two different names ;) [19:50] Fluffy: define what? [19:51] ant: well, you can't have two identical symbols. you *CAN* overload a function name by having them require different argument.. [19:51] you #define something (or not) and if it exists you use one funktion and if not use use the other [19:51] Dependencies forming a [19:51] complete graph [19:51] fluffy: that won't work [19:51] bvowk: huh? [19:51] ...are not incouraged by structured programming [19:52] My symbol generator is in bad mood today... [19:52] fluffy: if you ifdef two different functions, it wont be happy.. you'll get one or the other, likely not both [19:52] unless you inlined the one.. but thats a request, not an order. [19:52] I thought that's exactly what Ant_222 wants [19:53] yeah, but I don't think he'll have control over the one he gets :) [19:53] I won't my units be like objects, having theor own methods (private if not #included), named as I want. [19:54] Ant_222: What exactly prevents you from using two different names? [19:54] so, why don't you use C++ instead. [19:54] your classes will wrap the namespaces to behave like you want. [19:55] and the compiler will name mangle things in ways that'll really drive you mad down the road :) [19:56] Fluffy: I just don't like this everything-visible-from-everywhere architecture, I want normal structured programming. [19:56] Ant: C predates that. [19:56] Then you shouldn't use C [19:56] Hence, C++ [19:56] or Python [19:56] or Ruby [19:57] or perl if you're into self harm ;) [19:57] not much difference there [19:57] In Pascal and Modula I can export part of functions, keeping the rest of them inside the module. [19:57] eeek! [19:57] You've said the P-word *and* the M-word! [19:57] Yep! [19:58] And the C-word: COBOL [19:58] nobody cares about cobol.. there's nobody sick enough to use it without being paid huge sacks of money in return.. [19:58] but people use the P-word for fun. [19:58] Ant_222: Try brainfuck :) [19:58] I know [19:58] No! [19:59] It's inversed's privilege to use languages like that [19:59] P****? [19:59] What? [19:59] yeah.. what? [19:59] What what? [19:59] I won't say it [20:00] I've been tortured long enough with those programming langues [20:00] *languages [20:00] * bvowk recommends ant learn python instead of C [20:00] unless you're planning on hacking unix.. [20:00] then suck it up ;) [20:00] * Fluffy recommends not to trust bvowk, but learn C or Python nonetheless ;) [20:01] make that: C *and* Python [20:01] yeah, I'll second that. [20:01] * bvowk goes for some food [20:01] So, I variables that are neither #included nor externed are not visible, and functions are... I don't get the reason behind it. [20:02] variables in C are usually local.. which is why you won't see them.. [20:02] they exist really only in the context of your function [20:02] I mean those outside any function. [20:02] those are visible.. they're global. [20:03] without extern? [20:03] depends a little on the implementation.. but usually what happens if you've got a variable thats assigned in the data section of your program on loading.. and anything that refers to that name at link time gets that spot.. [20:04] My compiler it doesn't see it until I extern it. [20:04] TCC [20:05] huh? [20:05] doesn't see it or doesn't warn you about it colliding? [20:05] Doesn't see: undeclared [20:05] but it *IS* declared [20:05] In another file, yes [20:05] oh god no, now i just remembered why i thought that my compression-idea was between genious and disaster [20:06] was about to ask how to approach certain problem, and then the memory hit me [20:06] Mizcu: when trying to remeber C, I implemented an algorithm that fills those tables. [20:06] Ant_222: yes, i have a working one too [20:06] Using 2-dimentional array? [20:06] yes [20:07] Bad idea. [20:07] The actual size of table is the input string's lenth time 2. Thus we have 2*n vs n^2. [20:07] time->times [20:08] thats not an issue to me atm [20:11] The main problem is the algorithm, not the implemetation? [20:11] i have no need for that table in my current thought of compression style (cant call it an algorithm really..) [20:13] Whenever I feel the urge to try a new "compression algorithm" I have a look at the current PAQ version and bury my idea very quickly. [20:14] umm, yeah, taking ten times the memory and 20 times the time for 10% better compression [20:14] yeah, but any other combinations are already out there [20:15] combination? [20:15] ant: looks like you're right about the variables.. there's a couple caveats there.. [20:15] http://irc.essex.ac.uk/www.iota-six.co.uk/c/d5_variable_scope.asp [20:16] Mizcu: 10% less compression, but 20 times less memory; 10% less compression, 20 times more memory, ... [20:17] but that doesnt stop one from experimenting [20:18] yes, but having a *working* algorithm might be good idea then [20:18] well, theres no new algorithms around there currently anyway, so its about implementing their variations [20:19] then try to speed up PAQ [20:20] Why is that? Inversed has invented (yes, fully from scratch) a novel image compression algorithm. [20:20] a) takes too high maths for me b) needs pruning [20:20] Ant_222: Are you sure that it is new? [20:21] Yes! Ask him. [20:21] Ant_222: http://www.cs.fit.edu/~mmahoney/compression/ [20:21] Multimedia compression and general datacompression are different [20:21] Yes, but... [20:21] bvowk: that page says: "[extern] This can also be used for function, which we'll see later on" What do they mean? [20:22] Bah, Fluffy is not fun, i hoped to get a discussion with opinion about the compressor, but not with somebody having a chip stuck on the ass on the subject [20:23] what compressor? I thought that you don't have one [20:23] its an idea i am unsure about [20:24] then why don't you tell the idea, that you are unsure about? [20:25] And yes, of course I'll steal and patent it, if it is worth the hassle [20:27] In short, data is read, and if its is unique, it is left alone and remembered in dictionary (LZW), but also fed to fixed predictor or dozen. These predictors all make predictions about following data based on data they have been fed earlier; if there are no good predictions, the output data is handled through LZW. But if one of the predictors hit a match, then instead is input the predictor's code. [20:28] Once decompressing, plain data and data after decompression is also fed to the predictors, and since they recieve the same data, they make the same predictions. [20:29] bad: A) the predictors are fixed B) it takes time to "learn" a data C) if predictors only predict one symbol at a time, then there is not much to improve in use of size [20:29] why don't you try LZP then [20:29] ? [20:30] i couldnt understand Campos's article about it [20:30] there is something fundamental i miss when readin it [20:30] then read Charles Bloom's original papers about it [20:31] D) if multiple symbols are predicted, there need to be some offset about how many symbols are output [20:31] Bloom: http://www.compression.ru/download/articles/lz/bloom_lzp_pdf.rar [20:32] www.cbloom.com [20:32] already downloaded from the mainpage, and was about to whine about me not supporting postscript [20:32] but that pdf will be useful [20:35] I use PDFCreator (OSS from sourceforge) to make PDFs and distill PS into PRF [20:40] Join: Fizmo joined #corewars [20:40] hi hi [20:40] _hello [20:42] hi Ant [20:42] you're new here? [20:42] Yep [20:43] It was Mizcu's puzzle that led me here. [20:44] ohh, I see [20:45] See something interesting? [20:45] hehehe [20:45] :-P [20:45] Something fun is before your eyes? [20:46] just the stats of the download rate for my songs at mp3.de [20:46] ;-P [20:47] Growing exponentially? [20:48] hehehe, would be nice if so [20:48] but it is still linear with not very high slope [20:50] The 1st-order derivative is low, positive and constant. [20:54] yep [20:55] THhe 2nd-order is zero. [20:56] ugh, the explanation is retard simple but yet i fail to understand perfectly [20:56] I am now steadily comming to the conclusion that normal structured programming is impossible in ANSI C. [20:57] i think ill need to go asleep [20:57] Good night, Mizcu. [20:57] ... 1. No nested function/procedures with scopes limited to the parent function/procedure. [20:58] 2. Very weak typization and implicit typecasting [20:59] 3. No "export" only part of functions (unit's interface) and leave the other available only from withing the file (am I right here?). [20:59] other -> others; No -> No way to [21:10] C is from the 1970's Ant.. [21:10] The latest ANSI C standard is younger, AFAIK. [21:11] you know, back when men were men, and computers were measured by tonnage.. [21:11] yes, but that doesn't make it any more modern! [21:11] well, I guess it does.. [21:11] but most of what you're looking for is in C++ [21:12] C was basicly made to be a step up from assembly when implementing unix [21:12] Pascal was developed in 1970 as well, though it much more _structured. [21:12] C++ � I don't want to do OOP. Just structured programming. [21:12] pascal wasn't made by engineers, it was made by computer scientists. [21:12] * bvowk ducks [21:16] what version of tcc are you using? [21:18] 0�9�23 [21:18] 0.9.23 [21:18] hrm. [21:18] that doesn't actually mean anything to me ;) [21:19] I'll admit to not having used TCC since my dos days.. [21:19] Why did you ask? [21:19] unless you count borland builder.. [21:19] which is technically just TCC with a fancier gui. [21:19] Oh... What programs did you write in that blessed time (of DOS) [21:20] TCC � is not TurboC it is Tiny C Compiler, for win32 and linux. [21:20] ahh. [21:20] I didn't write anything interesting in the time of dos.. but thats alright, because it was dos, I don't think anyone did ;) [21:22] Hey! [21:22] The were so many excellent games [21:23] yeah, there were a few. [21:26] Played Alone in the Dark I with good sound and in a dark room? [21:28] I was more into syndicate [21:28] and doom [21:28] MSG: Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.4/2007051502] [21:29] I liked syndicate too, but X-COM 1 drew more of my attention [21:32] I knew someone who was addicted to xcom in the worst possible way.. [21:32] but I never really got into it [21:32] I played stars! on windows 3 too.. [21:32] that's still a fun game [21:32] even tho I don't have any humans to play with these days :( [21:37] ant222: bvowk.c, line 3: undeclared identifier stars! [21:40] never played stars!? [21:40] tragic. [21:42] Yeah.. What's it about? [21:43] http://en.wikipedia.org/wiki/Stars%21 [21:45] Oh, PBEM. Never could understand why a game would have a PBEM mode and not have simple hot-seat... [21:46] Seems interesting [21:52] it did have hot seat.. [21:52] basicly, it worked over anything you could exchange the files over.. [21:52] Is it downloadable for free? [21:56] dl-able from Underdogs... [21:57] ..No. [21:58] no, I don't think it is.. [21:58] you could likely find a copy somewhere.. [21:58] I think they still sell it too.. [21:58] I bought (another) copy in like 2004 [22:02] BTW: Do you know that there's an OSS game intended for recreating the original XCOM's battlescape with all kinds of multiplayer [22:02] ? [22:18] http://wiki.gible.net/index.php?title=Downloads [22:22] DL-ing... [22:35] Ok. version 2.6 up and running! [22:39] Maybe I'll contact you when ready to play. OK? [22:40] sure [22:40] it'll take a while.. the game is kinda deep [22:44] That good. [22:46] I guess playing with more than two humans is even more interesting... [22:48] Looks interesting [22:49] Ok, gotta go to bed: Bye! [22:50] MSG: Quit: sleep()