| --- Log opened Tue Oct 10 00:00:59 2006 |
| 00:01 | | Reiver is now known as Reivette |
| 00:02 | | Reivette is now known as Reiver |
| 00:09 | | ToxicFrog [~ToxicFrog@Admin.Nightstar.Net] has joined #code |
| 00:09 | | mode/#code [+o ToxicFrog] by ChanServ |
| 00:36 | | Vornicus [~vorn@ServicesOp.Nightstar.Net] has quit [Quit: Leaving] |
| 00:39 | | Vornicus [~vorn@67.50.40.ns-3674] has joined #code |
| 00:39 | | mode/#code [+o Vornicus] by ChanServ |
| 00:44 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has joined #code |
| 01:25 | | ThaquiWork is now known as Thaqui |
| 01:41 | | Reiver is now known as ReivClass |
| 02:44 | | himi-heading-home is now known as himi |
| 02:45 | | himi is now known as NSGuest-329 |
| 02:45 | | NSGuest-329 is now known as himi |
| 02:59 | | ReivClass is now known as Reiver |
| 03:28 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has quit [Quit: 8%] |
| 04:18 | | * Reiver accosts jerith idly. |
| 04:52 | | McMartin [~mcmartin@Nightstar-8843.dsl.pltn13.pacbell.net] has joined #code |
| 04:52 | | mode/#code [+o McMartin] by ChanServ |
| 04:52 | <@McMartin> | "The Shadows Shadowed Rating is a Shadows Shadowed Outcome that varies." |
| 04:52 | | * McMartin should maybe reconsider his variable naming scheme. |
| 04:54 | <@Vornicus> | i agree |
| 04:59 | <@McMartin> | Mmm. I need a better name for my final scene, though. |
| 04:59 | <@McMartin> | "Endgame" is kind of... anticlimactic. |
| 05:13 | | * McMartin works out a nifty trick. |
| 05:14 | <@Vornicus> | show uuussss |
| 05:15 | <@Vornicus> | <pstickne> Does Ruby have anything like weak hashes? |
| 05:15 | <@Vornicus> | <Vornicus> What makes weak hashes weak? |
| 05:15 | <@Vornicus> | <cschneid_foo> Vornicus, sitting in front of computers all day |
| 05:16 | <@McMartin> | A weak hash will let you garbage-collect its keys or values. |
| 05:16 | <@McMartin> | And remove the relevant entry at garbage-collection time. |
| 05:16 | <@Vornicus> | Yeah |
| 05:17 | <@Vornicus> | I got that later |
| 05:17 | | * McMartin wonders if this would be worth actually turning into an I7 extension, but he kind of doubts it. |
| 05:17 | <@McMartin> | To say win message: (- deadflag = 2; -) |
| 05:17 | <@Vornicus> | I thought maybe he meant "I can look up a key that doesn't exist and still get a sane value back" |
| 05:18 | <@McMartin> | Now you can say, for instance: end the game saying "You have pwn3d the world, go you[win message]". |
| 05:18 | <@McMartin> | And you'll get the option for the AMUSING menu, just as you would for "end the game in victory". |
| 05:18 | <+himi> | A hash of weak references . . . |
| 05:18 | <@Vornicus> | I just thought it was funny. |
| 05:19 | | * Vornicus doesn't quite understand the usage of weak references - seems to him that trying to talk about an already-GC'd thing would, uh. suck. |
| 05:20 | <@McMartin> | Vorn: It's generally for instrumentation. |
| 05:21 | <@McMartin> | So, if I want to have a separate index to "everything the program really wants to deal with" |
| 05:21 | <@McMartin> | Then when the program I'm watching lets stuff leave its scope -- that is to say, the only links left to it are the ones in my watching program, I want to clean up and then trigger collection. |
| 05:22 | | * McMartin isn't doing this in his prototype just now, but he really should be. |
| 05:22 | <@McMartin> | Failure to use weak references in places where you really should be is how you get memory leaks in garbage-collected languages like Java or presumably Ruby. |
| 05:23 | <+himi> | Also something like elements tied to entries in a database backend - you want a cache, but you don't care too much if they get collected after a while and you have to regenerate them |
| 05:24 | <@Vornicus> | Ruby tends to have the same problem as early Java did, in my experience - it leaks and leaks until it runs out and /then/ goes looking. |
| 05:24 | <+himi> | Hey, GCs are hard |
| 05:24 | <@McMartin> | Java and *I think* OCaml are the only GCed languages that do pre-emptive GCing. |
| 05:25 | <@McMartin> | Hell, I think Perl is still using pure reference counting, which isn't even correct |
| 05:25 | <+himi> | Perl is still using Perl 5, which is now more than a decade old |
| 05:25 | <@Vornicus> | GCs are hard, but also mostly a solved problem. Refcounts or whatever that got replaced by first, then if you're /still/ having problems, mark/sweep. |
| 05:26 | <+himi> | And OCaml does incremental GC regularly, so yeah, it's preemptive |
| 05:26 | | * himi was being somewhat sarcastic when he said GCs are hard |
| 05:29 | <@McMartin> | Vorn has just described the system that does have the problem he notes in Ruby. |
| 05:29 | <@Vornicus> | really? |
| 05:30 | <@McMartin> | Yes. |
| 05:30 | <@McMartin> | Mark and sweep involves walking the entirety of core. |
| 05:30 | <+himi> | Yay! |
| 05:30 | <@McMartin> | You have to use generational or "the train algorithm", with which I am kind of unfamiliar, to get workable incremental collection. |
| 05:30 | <+himi> | At least it's not stop and copy . . . |
| 05:31 | <@Vornicus> | stop and copy? |
| 05:31 | <@McMartin> | Generational has some of the same problems stop and copy has, but it suffers a little less. |
| 05:31 | <@McMartin> | Stop and copy: Divide core in half. Instead of mark-and-sweep, instead copy over all live objects to the other half. |
| 05:31 | <@Vornicus> | ...wait, that's "hang on, I'm out of memory, let me copy what's live and dump the rest"? |
| 05:31 | <@McMartin> | Yeah, and you only get half the core to work in, since you need that other half as workspace. |
| 05:32 | <@McMartin> | This has the added advantage of removing all memory fragmentation, I note. |
| 05:32 | <+himi> | (incidentally, that issue I was seeing with CSIRO's Websphere/Vignette setup was probably due to the lack of a generational GC in the JVM being used) |
| 05:32 | <@McMartin> | A really lame version of stop and copy is actually totally reasonable for a lot of subshell-based systems like emacs. |
| 05:32 | <+himi> | Yeah, the fragmentation cleanup is nice, but the memory copy is heavy |
| 05:32 | <@McMartin> | Unless you don't have a lot of objects that need saving. |
| 05:33 | <@McMartin> | Serializing state to disk and then reloading it all under the hood every 15 minutes or so would solve a lot of Word's bloat problems~ |
| 05:33 | | * himi believes GNU emacs uses plain stop and copy, though it may be more sophisticated in v21 |
| 05:33 | <@McMartin> | stop and copy should be sufficient for Emacs, really. |
| 05:34 | <@McMartin> | How many object roots does it really have, after all? |
| 05:34 | <+himi> | Never looked at elisp beyond the minimum needed for fiddling with my .emacs, so I can't comment usefully |
| 05:35 | | * Vornicus hunts around for the train algorithm |
| 05:36 | <@McMartin> | It's better at spatial locality than generational, but worse at time locality. |
| 05:36 | <+himi> | But a sophisticated GC will use a combination of algorithms to tune things for the expected uses, so someone trying to gank, say, the OCaml GC would probably have a nasty time of it |
| 05:37 | <+himi> | Which is probably why so many languages start out doing refcounting |
| 05:37 | <@Vornicus> | nerb |
| 05:37 | <@Vornicus> | SLEP |
| 05:38 | | * Vornicus nerbSLEP |
| 05:38 | <@Mahal> | nini Vorni |
| 05:39 | | Vornicus is now known as Vornicus-Latens |
| 05:57 | | Thaqui is now known as ThaquiWork |
| 06:07 | <@McMartin> | "Hi, I'm writing my first IF, and I want to randomly change the background color every turn." "No. No you really don't want to do this." |
| 06:10 | | himi [~fow035@152.83.10.ns-12024] has quit [Connection reset by peer] |
| 06:22 | <@Reiver> | ... |
| 06:22 | <@Reiver> | But why? |
| 06:22 | | Chalcedon is now known as ChalcyOut |
| 06:22 | <@McMartin> | The answer appears to be that he's trying to use it to create an art gallery piece. |
| 06:26 | <@Reiver> | O.o |
| 06:26 | <@Reiver> | What? |
| 06:27 | | * McMartin dunno. |
| 06:37 | | ChalcyOut [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout] |
| 06:37 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 06:37 | | mode/#code [+o Chalcedon] by ChanServ |
| 06:58 | | EvilNiNiLord is now known as EvilDarkLord |
| 07:59 | | You're now known as TheWatcher |
| 08:30 | | Mahal is now known as MahalZzz |
| 08:30 | | MahalZzz [~Mahal@Nightstar-11770.worldnet.co.nz] has quit [Quit: It's hard to be mad at someone who misses you while you're asleep. ] |
| 08:58 | | McMartin [~mcmartin@Nightstar-8843.dsl.pltn13.pacbell.net] has quit [Operation timed out] |
| 09:00 | | McMartin [~mcmartin@Nightstar-8843.dsl.pltn13.pacbell.net] has joined #code |
| 09:00 | | mode/#code [+o McMartin] by ChanServ |
| 09:13 | | Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 09:13 | | mode/#code [+o Chalcy] by ChanServ |
| 09:14 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Killed (NickServ (GHOST command used by Chalcy))] |
| 09:14 | | Chalcy is now known as Chalcedon |
| 09:29 | | You're now known as TheWatcher[wr0k] |
| 09:37 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: ] |
| 09:38 | | Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 10:08 | | Chalcedon is now known as ChalcyZzz |
| 11:46 | | ThaquiWork is now known as Thaqui |
| 11:47 | | Thaqui is now known as NSGuest-334 |
| 11:47 | | NSGuest-334 is now known as ThaquiWork |
| 11:48 | | ThaquiWork is now known as Thaqui |
| 12:52 | < EvilDarkLord> | Silly question. How do I detect linebreaks in PHP ( \n ) using regexps? |
| 12:54 | < EvilDarkLord> | Neither '\n' nor '/\n/' seem to do it. |
| 13:23 | | ChalcyZzz [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout] |
| 13:38 | | Thaqui is now known as ThaquiWork |
| 13:39 | | Reiver is now known as ReivZzz |
| 13:39 | | ThaquiWork is now known as ThaquiSleep |
| 13:51 | | ReivZzz is now known as ReivSLEP |
| 14:05 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has quit [Ping Timeout] |
| 14:17 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has joined #code |
| 15:23 | <@ToxicFrog> | EDL: as a general rule, detecting linebreaks in regexes is made of pain and suffering except in the forms $ (end of line) and ^ (start of line). |
| 15:24 | <@ToxicFrog> | That is, you can detect \nfoo, and foo\n, but not foo\nbar except by using two seperate regexes. |
| 15:26 | <@ToxicFrog> | Depending on PHP's regex API, you may be able to change this by messing with REG_NEWLINE, REG_NOTBOL and REG_NOTEOL, but only maybe. |
| 15:37 | <@TheWatcher[wr0k]> | TF: ... I'm not sure about PHP, but somehting like $body =~ /foo(\w+)\n(.*?) bar/ will work fine in perl |
| 16:34 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has quit [Ping Timeout] |
| 16:35 | | You're now known as TheWatcher |
| 16:44 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has joined #code |
| 17:19 | | You're now known as TheWatcher[afk] |
| 17:22 | | Vornicus [~vorn@Nightstar-18307.slkc.qwest.net] has joined #code |
| 17:22 | | mode/#code [+o Vornicus] by ChanServ |
| 17:58 | <@Vornicus> | My coat contained a furnace where there used to be a guy |
| 18:50 | | ChalcyZzz [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code |
| 18:52 | | ChalcyZzz is now known as Chalcedon |
| 18:54 | | You're now known as TheWatcher |
| 19:05 | <@Vornicus> | gnar. I can't find a way to get this file without manually entering a password. |
| 19:05 | <@TheWatcher> | Hrm? |
| 19:06 | <@Vornicus> | I have a file on a machine with ssh/sftp. I need a password for either. |
| 19:06 | <@TheWatcher> | Can't use public cert auth for the ssh? |
| 19:07 | <@Vornicus> | ...having never heard of public cert auth, no. |
| 19:07 | <@Vornicus> | What do I need to do to set it up? |
| 19:08 | <@TheWatcher> | on the client: |
| 19:08 | <@Vornicus> | I have another machine that I have total control over, but it's not publicly routable. |
| 19:08 | <@Vornicus> | which is where I have to get the file /to/ |
| 19:10 | <@TheWatcher> | Okay, the idea is that on the client you set up a public/private keypair, then copy the public key to the server's allowed_keys file, then the client can connect to the server without the need to enter a password |
| 19:10 | <@Vornicus> | aha |
| 19:11 | | * TheWatcher digs |
| 19:11 | <@TheWatcher> | http://sial.org/howto/openssh/publickey-auth/ |
| 19:11 | <@TheWatcher> | I think that'll do it for you |
| 19:12 | <@Vornicus> | shiny |
| 19:12 | <@Vornicus> | okay I might be able to do that |
| 19:24 | <@Vornicus> | will doing this for ssh allow me to use scp or sftp the same way? |
| 19:27 | <@TheWatcher> | Yes |
| 19:28 | <@TheWatcher> | (I use the same technique for automatic remote backups on some of my servers - uses scp to do transfers) |
| 19:30 | <@Vornicus> | awesome |
| 20:02 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has quit [Ping Timeout] |
| 20:13 | | EvilDarkLord [althalas@Nightstar-17046.a80-186-184-83.elisa-laajakaista.fi] has joined #code |
| 20:20 | | aoanla [~sam@Nightstar-26565.range81-158.btcentralplus.com] has joined #code |
| 20:28 | <@ToxicFrog> | Vornicus: public-private key authentication is pretty tasty, yes. |
| 20:29 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has joined #code |
| 20:41 | | * aoanla ponders. |
| 20:41 | | aoanla [~sam@Nightstar-26565.range81-158.btcentralplus.com] has quit [Quit: Leaving] |
| 21:02 | | aoanla [~sam@Nightstar-26565.range81-158.btcentralplus.com] has joined #code |
| 21:03 | | * aoanla notes that asymmetric key authentication only really works with at least one trusted authority. |
| 21:04 | <@ToxicFrog> | I trust myself; my systems trust me; problem solved. |
| 21:04 | < aoanla> | I meant in the Wider World of People. |
| 21:04 | < aoanla> | Or, actually, in e-Science >.> |
| 21:05 | <@ToxicFrog> | Heh. |
| 21:05 | | * ToxicFrog just uses it so he can easily log into Orias from the other systems on his LAN. |
| 21:06 | < aoanla> | Well, that's also good, yes. |
| 21:07 | < aoanla> | And it /is/ lovely that SSH will use asymmetric crypto for that, since it's much better than passwords. |
| 21:11 | | Mahal [~Mahal@Nightstar-11770.worldnet.co.nz] has joined #Code |
| 21:11 | | mode/#code [+o Mahal] by ChanServ |
| 22:17 | | MahalPC [~Mahal@Nightstar-11770.worldnet.co.nz] has quit [Quit: Leaving.] |
| 22:30 | <@Vornicus> | HI MY NAME IS VORN AND I'M RETARDED |
| 22:30 | <@McMartin> | HI VORN |
| 22:30 | < Janus> | Hi Mr. Vorn! |
| 22:31 | <@TheWatcher> | Hello Vorn |
| 22:31 | <@TheWatcher> | (um, no you're not) |
| 22:31 | | Mahal [~Mahal@Nightstar-11770.worldnet.co.nz] has quit [Quit: It's hard to be mad at someone who misses you while you're asleep. ] |
| 22:31 | < EvilDarkLord> | Y halo thar monsieur vorn. |
| 22:31 | <@Vornicus> | Meet Chalain. He's my boss. |
| 22:31 | | * aoanla notes that none of us are travelling backwards in time. |
| 22:31 | <@Vornicus> | He's even more retarded. |
| 22:31 | | Mahal [~Mahal@Nightstar-11770.worldnet.co.nz] has joined #Code |
| 22:31 | | mode/#code [+o Mahal] by ChanServ |
| 22:31 | < aoanla> | (Sorry, a joke there for the EM Theorists in the channel) |
| 22:32 | <@TheWatcher> | Vorn: why is this, then? |
| 22:32 | | * McMartin would smite, but having just linked the CSE Band... |
| 22:32 | <@Chalain> | Yeah, well at least *I* don't walk away leaving my keyboard unguarded. :-) |
| 22:33 | <@TheWatcher> | ... ah-ha |
| 22:33 | <@McMartin> | http://www.cs.washington.edu/orgs/student-affairs/cseband/ idly. |
| 22:33 | | * McMartin suggests "Theory Girl" for this channel |
| 22:33 | <@Chalain> | Actually, I took the time to change the current channel. Had I just started typing, it would have gone to #ruby-lang on freenode. >:-) |
| 22:33 | | * aoanla is still eyeing that code you posted, McM. |
| 22:33 | <@TheWatcher> | It's an interesting snippet |
| 22:33 | <@Chalain> | HOWEVER, a) That would be mean, b) a lot more people would see it, and most importantly, c) they wouldn't appreciate the joke. |
| 22:33 | < aoanla> | (It looks like it does something horrible to a hash-table, but...) |
| 22:34 | <@McMartin> | That's about as much as I could work out. |
| 22:34 | <@McMartin> | It was the "muhahahahahaha" comment that was being called out, tohugh |
| 22:34 | <@McMartin> | though |
| 22:35 | | * TheWatcher is still trying to sort it in his head, knows it's a hash copy but the presenc eof the @{} cast is messing with his head a bit |
| 22:36 | <@McMartin> | Is it a hash flattener? |
| 22:37 | <@TheWatcher> | That'd be my guess, yes. |
| 22:37 | < aoanla> | Oh, good, it was mine, too. |
| 22:38 | < aoanla> | Well, I thought it was either that, or something so hideous that I'd never understand it. |
| 22:40 | <@TheWatcher> | Although if it wasn't 10:40 and I was pondering bed, I'd be tempted to write somethign to confirm that, because the comment was justified.. |
| 22:52 | | You're now known as TheWatcher[T-2] |
| 22:59 | | You're now known as TheWatcher[zZzZ] |
| 23:00 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has quit [Quit: ~] |
| 23:04 | | aoanla is now known as caps[t-2] |
| 23:08 | | caps[t-2] [~sam@Nightstar-26565.range81-158.btcentralplus.com] has quit [Quit: Leaving] |
| 23:14 | | ThaquiSleep is now known as Thaqui |
| 23:34 | | ReivSLEP is now known as Reiver |
| 23:45 | | mode/#code [+ooo Chalcedon Thaqui ToxicFrog|Stuff] by Reiver |
| 23:45 | | mode/#code [+vv Ev3 EvilDarkLord] by Reiver |
| 23:47 | | Chalcedon is now known as ChalcyUni |
| 23:58 | | Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has joined #code |
| --- Log closed Wed Oct 11 00:00:17 2006 |