[00:05] *** ashim left [00:05] *** ashim joined [00:53] *** lupo33 left [01:34] *** sa1 left [01:34] *** sa1 joined [01:48] *** sa1 left [01:49] *** sa1 joined [03:05] *** navgeet joined [04:06] *** sa1 left [04:18] *** ksinkar joined [04:20] *** ksinkar left [04:26] *** navgeet left [05:12] *** ps_jinx left [05:17] *** ps_jinx joined [05:53] *** sa1 joined [05:54] *** aftab joined [06:13] *** aftab left [06:17] Gmail isn't working for me. [06:20] Works for me. Using 212. [06:21] Wasn't working for me either. Working on 247 now. [06:28] Works with 247. Thanks sa1 :). [06:29] HTTPS is down on 245 :) [06:58] 212 not working now. :( [07:44] *** sa1 left [07:47] *** aftab joined [08:20] *** lupo33 joined [08:24] *** sa1 joined [09:41] They have some idea, but nobody is really certain about these things. [09:52] *** aftab left [10:05] * artagnon shudders [10:05] artagnon: At the thought of what? [10:05] I just commented on a super-high-influence thread about Git Generation Numbers -- mainly a conversation between Linus, and the original authors. [10:05] :) [10:06] It took me some time to convince myself that what I was saying wasn't all garbage. [10:07] http://thread.gmane.org/gmane.comp.version-control.git/177146/focus=180783 [10:11] * lupo33 googles Git Generation Numbers [10:12] <$$$> You'll find nothing :) [10:12] Except this thread, and a previous thread like this. [10:13] http://news.ycombinator.com/item?id=2765844 [10:13] :). In the first and last paragraphs, you sound like me when I am talking to you or Sanjoy. The middle paragraphs, of course, I have no clue about. [10:13] * artagnon laughs [10:14] Yes, that was the point of showing you this :) [10:14] :) [10:15] It's perfectly normal behavior. This is how the community grows. [10:16] Note how I'm blatantly disagreeing with almost everyone except Linus :P [10:16] It's very important to disagree. [10:17] artagnon: I disagree. :P [10:17] * artagnon smiles [10:17] Also, note how most less-useful responses on the thread are completely ignored. [10:18] See the people Linus responds to. [10:25] *** ps_jinx left [10:27] * lupo33 was wondering if a we could write code to make a simple s/foo/bar-like command to replace foo by bar in the channel log (and if possible in the live channel feed too). [10:27] s/a// [10:28] lupo33: not possible in the live channel feed. [10:28] <$$$> sanjoyd: Thought so. [10:31] *** ps_jinx joined [10:34] sa1: Read Linus' first email to the Usenet group announcing Linux. [10:34] Sounds very similar :) [10:36] http://en.wikipedia.org/wiki/History_of_Linux#The_creation_of_Linux [10:36] Yeah, I have read it. But still, he announced a _kernel_, which makes the the rest of the email rather secondary.:P [10:40] The same of course could be the case with your email, though I don't understand it. [10:42] lol, I don't either -- and that isn't the point anyway. [10:42] yEAH [10:42] s/yEAH/Yeah [10:43] Sometimes, I wonder how people managet to hit CAPS LOCK -- I don't even have that key. [10:44] *** ps_jinx left [10:45] artagnon: I find it faster than Shift.:| My fingers have got used to it. [10:49] *** ps_jinx joined [10:55] * lupo33 is surprised to find that "/ME" works even on gchat. [10:55] * sa1 leaves lab. [10:55] *** sa1 left [11:13] Anyone for chai? I'm going to BCR to get a quick chai [11:14] * artagnon goes off [11:23] *** sa1 joined [11:53] /join #xmonad [11:54] <$$$> sorry for that :/ [12:17] *** sa1 left [12:20] *** sa1 joined [12:23] *** vickyc joined [12:23] *** sa1_ joined [12:32] *** ps_jinx left [12:37] *** ps_jinx joined [12:42] *** vickyc left [12:42] *** vickyc joined [12:44] *** sanket left [12:44] *** ps_jinx left [12:47] *** ps_jinx joined [12:58] *** ps_jinx left [13:04] *** ps_jinx joined [13:09] I just a very lol thing today. [13:09] *** ps_jinx left [13:09] Turns out that the work I did last year at BC helped my boss get a promotion :P [13:09] <$$$> *heard [13:14] *** ps_jinx joined [13:23] diffoperator: BC = Barcap? [13:23] Yes! I figured out what was wrong with my computer! :D [13:23] After going through many many GPF logs, I figured out that it was my RAM. [13:23] Ran memtest and figured out which stick :) [13:24] Oh, okay. You have the warranty now. [13:24] Ofcourse :) [13:25] At first, I thought it was my i7 core or hard drives. Even suspected the motherboard for some time. [13:26] I was desperate today -- can't have Chrome crashing all the time during my interview :) [13:27] *** vickyc left [13:28] artagnon: You are still interviewing? After that amazing Amazon offer? :D [13:28] The drive I have is almost scary. [13:29] *** vickyc joined [13:43] * yeban cheers for artagnon [13:43] * yeban dances :D/-< [13:43] * yeban dances :D|-< [13:43] * yeban dances :D-< [13:43] * yeban dances :D|-< [13:46] * diffoperator needs to get himself an SSD [13:49] :) [14:00] *** sanket joined [14:06] *** vinaykola joined [14:14] *** vickyc left [14:15] *** vickyc joined [14:27] artagnon: What time is your interview? [14:34] <$$$> yeban: 10:30 PM. [14:44] *** ps_jinx left [14:44] *** sa1_ left [14:45] *** sa1 left [14:46] *** ps_jinx joined [14:46] Dinner? [14:51] *** lupo33 left [14:51] *** vinaykola left [14:51] *** vickyc1 joined [14:52] *** vickyc left [14:52] *** sanket left [14:53] artagnon: Best of luck for the interview. :) [15:08] *** sanket joined [15:10] *** sanjoyd left [15:10] *** sanjoyd joined [15:19] sanjoyd: Are you writing GPython in C++? [15:19] diffoperator: yes. [15:19] eschewed C? [15:19] diffoperator: a lot has changed from what is there in the github account. I'll probably re-write history. [15:20] diffoperator: yes. I'm a C++ fan now. :D [15:20] are your latest commits in? [15:20] diffoperator: no. I'll push once the lexer is written. [15:20] diffoperator: or, actually I can push what I have now. [15:20] (you'll come back to java soon, the way you are going...) [15:20] yeah do that [15:20] diffoperator: you'll probably reconsider when you look at the code. :P [15:21] diffoperator: it does not work, barely compiles. [15:21] I'll push nevertheless. [15:23] diffoperator: pushed. [15:23] <$$$> ok [15:24] are you writing the lexer by hand? [15:24] diffoperator: yes. Actually I've cleaned up the one in CPython itself. [15:24] Now to tweak it until it runs. [15:25] The CPython one is a huge function, with quite a few gotos. [15:25] I've removed all that. [15:25] And I'm experimenting with the idea of turning the lexing + parsing the other way than usual. [15:25] i.e. the lexer calls the parser when it has a token, and not the other way around. [15:25] Right now, I don't know if this will work well. [15:26] huh? that is not the standard workflow [15:26] It shouldn't [15:26] You mean the lexing + parsing? [15:26] ashim: Thanks :) [15:26] 1 [15:26] No, not really. Things do work unexpectedly. For instance, v8 does not even have a lexer. [15:26] It parses JS directly. [15:27] Are there any proofs on the efficiency and correctness of this method? [15:27] diffoperator: which method? [15:27] I'm not following any method. [15:27] of parsing directly [15:27] no the standard way of doing things is to use compiler compilers [15:27] You'd have a hard time proving a C++ program. :D [15:28] diffoperator: they're too slow for a JIT. [15:28] wow, with templates and all [15:28] oh right my bad [15:28] thats standard compiler theory [15:28] Plus, compiler-compilers tend to work well with whitespace agnostic langauges. [15:28] IMO. [15:28] how do things work for JIT compilers? [15:28] I could not find a nice and elegant way to encode Python's indentation in lex. [15:29] is it possible to simply import V8? [15:29] i mean its architecture [15:30] or does the target language requires different way of doing things? [15:30] *** maxking joined [15:39] diffoperator: sorry, was on the phone. [15:39] diffoperator: give me a few minutes. [15:40] np [15:44] diffoperator: I do intend to base the code on V8's architecture. I've been hearing a bit on trace based JITs too; perhaps that will also be interesting to explore. [15:46] https://github.com/resistor/BrainFTracing [15:46] have you seen this? [15:47] I've not gone through the source. [15:47] I'm feeling too darn sleepy... [15:48] Since I left kgp I have become a wimp [15:48] In what sense? [15:56] in my capacity to give up on sleep [15:56] I have gotten too used to the "good life" [15:56] *** lupo33 joined [16:05] *** sa1 joined [16:12] *** theonlyone joined [16:12] Why so many channels? [16:13] theonlyone: There is only one. [16:13] Anyone here? [16:13] theonlyone: A lot. Try /names [16:13] So how many have joined KLUG this year? [16:20] theonlyone: There are around 45 people on the Google Group. [16:20] theonlyone: http://groups.google.com/group/kgplug [16:20] My bad. There are 79 members now. [16:21] The whole idea of "joining KLUG" is not very well defined. :) [16:21] Right. [16:22] It's not "joining" so much as "participating" -- hence the change in the side link. [16:24] *** artagnon left [16:24] *** artagnon joined [16:24] *** ChanServ sets mode: +o artagnon [16:24] *** ChanServ sets mode: -o artagnon [16:25] Why do i see npower here? [16:26] He passed out. [16:26] Yes he continues to exist ... [16:26] s/Yes/Yet/ [16:27] * crodjer does :)-=< [16:27] * crodjer does :)-|=< [16:27] * crodjer does :)-/=< [16:27] * crodjer does :)-|=< [16:27] * crodjer does :)-=< [16:28] artagnon: ping [16:28] He's got a girlfriend now. That disqualifies him. [16:28] theonlyone: Huh? [16:28] :| [16:29] np0w3r: has a girlfriend??!! [16:29] 1 [16:29] artagnon: no wonder he was looking different. [16:29] artagnon: In klug.github.com's there is a commit which says [rr: rebased]. [16:29] SHA: af50af4b [16:30] Now rebase can mean a lot of things, right? [16:30] His teeth look the same to me. [16:30] He was looking all shiny and stuff. [16:31] <$$$> (When he came to Kgp for the convo.) [16:31] lol :D [16:31] Wait, [16:31] John Doe to me show details Aug 24 8:26 AM me: http://alpha.dateiitians.com lol 8:28 AM John: wait, opening. but the name seems very promising me: lol Naveen: dude fuckin' awesome idea damn, i'm a li'l late :P me: No You are an ex-IITian. Still works out. 8:29 AM Naveen: no, i'm in a relnship :P me: :O Naveen: hmm, 1 day old? :D me: DateIITians offers you variou [16:39] yeban: artagnon is giving his interview, probably. [16:44] sanjoyd: Right. [17:01] *** lupo33_ joined [17:03] *** lupo33_ left [17:04] *** lupo33_ joined [17:04] Oops. Sorry. :/ [17:07] lupo33_: For what? [17:10] <$$$> I mis-typed the msg NickServ command, and thought it appeared on the feed(with my account password :P). Just checked the logs, it didn't. :) [17:12] *** maxking left [17:14] I see. [17:24] *** crodjer left [17:29] sanjoyd: Btw, what about your intern next summer? Thinking about it yet? :P [17:30] *** maxking joined [17:30] Will start applying soon. [17:30] In a week or two, i.e. [17:30] Hmm, nice. [17:35] *** theonlyone left [17:35] *** yeban left [17:35] *** lupo33 left [17:36] *** sa1 left [17:36] *** sanket left [17:36] *** maxking left [17:36] *** vickyc1 left [17:37] *** lupo33_ left [17:48] *** sanjoyd left [17:52] *** ps_jinx left [20:11] *** sa1 joined [20:15] * sa1 hopes artagnon's interview wasn't affected because of the power cut. [20:22] *** yeban joined [20:50] *** aftab joined [20:51] Just the last question of the first interview got screwed up. [20:51] Second interview re-scheduled. Aced everything else. [20:55] *** sa1 left [20:56] *** sa1 joined [20:59] * artagnon dozes off [21:02] artagnon: read the logs now. nice. :) [21:11] Just read the logs too. Was quite interesting. [21:11] * aftab calls it a day. [21:11] *** aftab left [02:14] Anyone up? [02:19] *** lupo33 joined [02:21] yeban: yes [02:21] sa1: No class? [02:23] Well yeah, there's one at 8:30, so I was about to go get ready. [02:24] * yeban has one at 9:30 [02:24] * yeban hates classes [02:26] yeban: The feeling is universal. :) [02:26] :) [02:28] * lupo33 missed the 7:30 double lecture. [02:42] * yeban cheers for lupo33 [02:53] * sa1 leaves. [03:27] *** vickyc joined [04:45] *** ksinkar joined [04:49] *** lupo33 left [05:01] *** crodjer joined [05:01] *** aftab joined [05:05] *** aftab left [05:23] *** sanjoyd joined [05:23] *** sanjoyd left [05:23] *** sanjoyd joined [05:58] *** ksinkar left [05:59] *** ksinkar joined [07:11] *** vickyc left [07:55] *** maxking joined [08:13] *** aftab joined [08:27] *** yeban left [08:48] * aftab is trying out wmii. [09:20] *** maxking left [10:38] *** aftab left [10:44] chai? [10:51] crodjer: Yeah, sure. I'll leave in 2 mins. [10:51] * crodjer leaves [10:52] Call for a meeting this Friday if you like -- I might not be in a position to attend it. [10:52] I'm not in a position to organize/ call it either -- I'm being weighted down by interviews. [10:53] Ask sanjoyd and yeban for the final say though; they're our community maintainers. [11:37] *** lupo33 joined [11:42] sanjoyd: Are you thinking of incorporating nanojit into gpython? [11:42] I was taking a look at the tracemonkey arch [11:42] they use nanojit for generating fast machine code and something called jstracer for hot code opt [11:52] *for monitoring, sorry [11:53] *** ksinkar_ joined [11:53] *** ksinkar left [12:06] *** vickyc joined [12:08] <$$$> diffoperator: I'll have a look at it. Does it look good, btw? [12:53] *** ksinkar_ left [12:55] *** ps_jinx joined [12:55] it looks usable [12:55] + it is maintained by adobe and mozilla [12:56] So essentially what the Firefox guys are planning to do is [12:56] javascript -> quickly generated but not that optimized code using nitro [12:57] nitro code -> slowly generated but fast machine code using nanojit [12:57] the last step applies for code that is marked as hot [12:58] MSR came up with something called SPUR. I will read that paper. [13:00] diffoperator: okay. I too will have a look. [13:03] *** ps_jinx left [13:15] *** ps_jinx joined [13:22] artagnon: ping! [13:33] sanjoyd: pong. [13:34] artagnon: got time to look at my resume? [13:35] artagnon: or later/ [13:35] ? [13:35] As in, should I ask you later? [13:35] Later, please. Tomorrow for sure. [13:35] * artagnon nods [13:35] Okay. [13:36] sanjoyd: I am currently reading nanojit code. When do you want to discuss this? [13:37] diffoperator: I still have not seen the code, maybe in two hours? [13:37] ok [13:38] I was actually thinking we should work on the lexer and parser; and focus on getting a CFG first. Once that is out of the way and we actually have something to play with, these things will make more sense. [13:38] It does make a lot of sense to get a library do the platform specific code gen, of course. [13:39] Ahh ... too many 'actually's. [13:39] yeah, definitely [13:39] can you prepare some sort of a rough roadmap? I have only used lex and yacc in the past [13:40] Okay. [13:40] I'll do that. [13:40] so if you intend to do things differently, i will need to read up on some theory [13:40] or code [13:40] Even I'm not very sure about many things right now. :) [13:54] np, we first try to get hello world to run [13:54] Exactly! :) [13:54] diffoperator: do you know if A Sarkar de-regges students? [14:21] *** ps_jinx left [14:26] *** ps_jinx joined [14:50] *** navgeet joined [15:03] Is Minix a high entry barrier project? [15:11] <$$$> lupo33: try to look at the source, and see how much sense it makes. [15:12] lupo33: you might want to contribute to some C project first, hacking on an OS can be pretty hardcore. [15:17] lupo33: It depends on how much time/energy you have to spare. [15:18] hacking on minix is like a high intensity, high reward activity [15:18] it will be painful but your maturity will develop rapidly [15:18] sanjoyd: turns out nanojit does not implement tracing, its just a really fast JIT with its own IR [15:19] diffoperator: oh, okay. [15:19] diffoperator: I was just talking about something that would, sort of, allow us to compose assembly in a platform non-specific way. [15:19] tracemonkey uses something called jstracer to analyse code that is "hot" and selectively JITs that [15:19] diffoperator: so I won't have to worry about what the hex code for mov is. [15:19] so we "could" use nanojit for that [15:19] diffoperator: oh, okay. [15:20] lexer/parser->cfg->nanojit? [15:20] diffoperator: but I still think we're thinking too far ahead. :P [15:20] diffoperator: yes, probably. [15:20] aye, lexer/parser first [15:20] diffoperator: *lexer* first. :D [15:20] diffoperator: I also want to implement some inline caching. [15:20] can we just take CPython and rip out its parsing/lexing infrastructure? [15:20] diffoperator: we can do that. [15:21] *** aftab joined [15:21] are we doing that? [15:21] diffoperator: if I don't get much done by this weekend, perhaps it would be intelligent to do so. [15:21] diffoperator, sanjoyd: OK. [15:21] diffoperator: it would be nice to have a faster and more streamlined parser in C++, I think. [15:21] ok, this weekend I will be travelling anyway [15:22] diffoperator: in fact, that is exactly what unladen swallow did -- start from the place where CPython has already generated the bytecode. [15:22] right, i remember [15:22] unladen swallow is written in C++ [15:22] ? [15:22] In fact, do you think it'll be better the other war around? [15:22] diffoperator: the LLVM parts are, I think. [15:22] what? [15:22] LLVM has a C api too, so I can't be sure. [15:22] s/war/way/ [15:23] i.e. first use the CPython frontend. [15:23] (By wrapping it around some classes etc.) [15:23] And start with the more fun parts. [15:23] If we see the need later, we can re-write the frontend in C++. [15:23] hmm we could [15:24] I think this makes more sense ... more importantly we won't be bored to death before we start seeing some results. :) [15:24] oh ok then [15:24] prepare a roadmap and let me know, i'll start experimenting with nanojit tomorrow [16:04] *** aftab left [16:17] *** maxking joined [16:18] *** maxking_ joined [16:19] *** maxking left [16:20] *** yeban joined [16:24] *** maxking joined [16:24] *** maxking_ left [16:25] I tried upgrading from debian stable to testing but size mi-match error is coming everytime [16:26] I tried changing the proxy and mirror as well but no use. [16:26] Any help? [16:33] maxking: use aria2c to download the files and then use apt-get to install them. [16:33] maxking: wait ... [16:34] *** navgeet left [16:34] http://pastebin.com/3UZbYmrB [16:34] maxking: ^ [16:35] Save this as fast-apt-get, chmod +x it, and then instead of sudo apt-get install foo, do fast-apt-get install foo. [16:35] sanjoyd: Thanx i'll try it then. [16:38] sanjoyd: I used to have the same problem when I was on Ubuntu. Changing the internet connection(as in using my USB modem) used to fix it. Is it a problem with our proxies? Do they cache corrupted files or something? [16:38] sa1: yes, yes. [16:39] sa1: rather, I think so, I think so. [16:40] pacman has been positively peaceful in this regard. :) Also allows me to use any downloader as the backend. [16:48] sanjoyd:after saving it as fast-apt-get then? [16:48] chmod +x fast-apt-get? [16:48] 22:09 Save this as fast-apt-get, chmod +x it, and then [16:48] instead of sudo apt-get install foo, do [16:48] fast-apt-get install foo. [16:48] :D [16:51] command not found [16:51] :O [16:51] If you have it in the current dir, you need to do ./fast-apt-get [16:51] Or you can put it somewhere on your path. [16:51] sanjoyd: http://viewer.opencalais.com/ [16:52] lupo33: check PM! [16:53] Illegal option- for ./fast-apt-get [16:53] maxking: works on my comp, try fixing it? [17:56] *** maxking left [19:08] Totally aced it. [19:08] Designed the backend for GMail in 10 mins :P [19:09] * artagnon goes out to get chai [19:40] *** vickyc left [19:43] *** vickyc joined [20:01] artagnon: congrats again. [20:12] sa1: Results aren't out yet, so I wouldn't be celebrating yet :P [20:12] Plus, there'll be 4 more interviews :) [20:13] ... but yeah, thanks :) [20:16] 4 more interviews? as in, 4 more Google rounds or other places? [20:18] artagnon: awesome! [20:19] sa1: Google rounds. [20:20] sanjoyd: 'twas a *cakewalk*. I can't wait to take on more interviews. [20:21] * artagnon dozes off [20:35] *** sa1 left [20:59] *** ps_jinx left [21:04] *** ps_jinx joined [21:09] *** sa1 joined [21:23] *** sa1 left [21:30] *** ps_jinx left [21:35] *** ps_jinx joined [22:23] *** maxking joined [22:40] Any one there? [22:51] *** maxking left