| --- Log opened Fri Dec 18 00:00:25 2009 |
| 00:04 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has quit [Connection reset by peer] |
| 00:05 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has joined #code |
| 00:37 | | Attilla [The.Attilla@FBC920.58502B.1ADD6E.CC89EF] has joined #code |
| 00:37 | | mode/#code [+o Attilla] by Reiver |
| 01:44 | | Attilla [The.Attilla@FBC920.58502B.1ADD6E.CC89EF] has quit [[NS] Quit: ] |
| 01:52 | | Derakon[AFK] is now known as Derakon |
| 02:15 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has quit [Connection reset by peer] |
| 02:16 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has joined #code |
| 03:02 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: *hums* Can't stay now!] |
| 04:51 | | McMartin [mcmartin@Nightstar-ec4e50af.pltn13.sbcglobal.net] has joined #code |
| 04:52 | | mode/#code [+o McMartin] by Reiver |
| 04:57 | | GeekSoldier_ [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has joined #code |
| 04:57 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has quit [Connection reset by peer] |
| 06:26 | | Derakon is now known as Derakon[AFK] |
| 06:30 | < gnolam> | OH JESUS CHRIST OPENOFFICE |
| 06:30 | < gnolam> | _Not funny_. |
| 06:31 | < gnolam> | It crashed and pretended to have corrupted the document I've been writing continously since... um... last night. |
| 06:40 | < Alek> | ouch. |
| 06:40 | | * Alek patpats gnolam. |
| 06:42 | < Alek> | this is why I want a utility that automatically not only backs up whatever document I'm working on every 10 mins or so (like Office does normally if you have it selected, and I think OO does too), but also makes multiple copies in multiple locations, in separate files, so you have multiple backups going back at least an hour, just in case it's a slow corruption. |
| 07:06 | <@ToxicFrog> | while true; do git commit -a -m "Automatic backup made at $(date)"; git push; sleep 600; done |
| 07:15 | <@ToxicFrog> | There you go, automatic backup every ten minutes from whenever you started this, automatically mirrored to origin. |
| 07:17 | <@Vornicus> | Only works if you save often enough. |
| 07:17 | <@ToxicFrog> | If you aren't in the habit of saving after every sentence, turn on autosave in your editor. |
| 07:17 | <@Vornicus> | The primary difficulty is you get going and forget to hit save. |
| 07:18 | <@ToxicFrog> | if neither of those are options, an external backup utility is not going to help you at all. |
| 07:18 | <@Vornicus> | Most editors I know put autosaves in some other directory entirely. |
| 07:18 | <@ToxicFrog> | ...most editors I know don't. |
| 07:19 | <@ToxicFrog> | But if they do, ok; version that as well. |
| 07:24 | <@Kazriko> | Yay, the IXP425 network driver demon has been slain... |
| 07:27 | <@ToxicFrog> | Anyways, the point I was trying to make is that we already have software that stores previous versions of what you were working on in multiple places; it's called a version control system. |
| 07:27 | <@ToxicFrog> | The only part that's missing is the "automatically", and that's trivial even on windows. |
| 08:04 | | You're now known as TheWatcher |
| 08:21 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 08:58 | | Rhamphoryncus [rhamph@Nightstar-a62bd960.abhsia.telus.net] has quit [Client exited] |
| 09:07 | | You're now known as TheWatcher[afk] |
| 10:04 | | Attilla [The.Attilla@FBC920.58502B.1ADD6E.CC89EF] has joined #code |
| 10:04 | | mode/#code [+o Attilla] by Reiver |
| 10:46 | | AnnoDomini [annodomini@Nightstar-3cdc6c0c.adsl.tpnet.pl] has joined #code |
| 10:47 | | mode/#code [+o AnnoDomini] by Reiver |
| 13:17 | | Tarinaky [Tarinaky@Nightstar-e2f9fdaf.adsl.virginmedia.net] has quit [Client closed the connection] |
| 13:18 | | Tarinaky [Tarinaky@Nightstar-e2f9fdaf.adsl.virginmedia.net] has joined #code |
| 13:32 | | You're now known as TheWatcher |
| 15:33 | | AbuDhabi [annodomini@Nightstar-5de733ad.adsl.tpnet.pl] has joined #code |
| 15:35 | | AnnoDomini [annodomini@Nightstar-3cdc6c0c.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 16:07 | | * Vornicus randomly wanders the Civilization 4 scripts. "CvDanQuayle.py" |
| 16:39 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 17:51 | | celticminstrel [celticminstre@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 17:54 | | Attilla [The.Attilla@FBC920.58502B.1ADD6E.CC89EF] has quit [Ping timeout: 121 seconds] |
| 18:02 | | Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has joined #code |
| 18:03 | < Derakon[work]> | Looking for theories, folks. We have a rotatable display. When rotated 90? (so that it's 1280x1920 instead of 1920x1280), a single component of the UI flickers. When the display is not rotated, there is no flickering. |
| 18:03 | < Derakon[work]> | What could account for this? |
| 18:05 | <@Vornicus> | The UI component - what kind of component is it? |
| 18:05 | < Derakon[work]> | A wx OpenGL canvas. |
| 18:05 | < Derakon[work]> | The specific bit that is flickering is some text drawn to the canvas. |
| 18:05 | <@Vornicus> | hm.... |
| 18:05 | < Derakon[work]> | The histogram that is also drawn to the canvas does not flicker. |
| 18:07 | < Derakon[work]> | Hmm...I just took a look at the code drawing said component, and naturally it looks fishy. But I don't see how it could work fine in one orientation and flicker in the other. |
| 18:08 | < Derakon[work]> | Waugh, the Ubuntu pastebin requires a login now. |
| 18:08 | | * Derakon[work] fails Starforge's humanity test the first time around, manages http://pastebin.starforge.co.uk/80 |
| 18:09 | < Derakon[work]> | The text being drawn starts at line 73. Note the flush and swapbuffers immediately preceding. |
| 18:15 | | * Vornicus ... dunno. |
| 18:16 | | * Derakon[work] vanishes for a bit as he unplugs his laptop from the wall. |
| 18:20 | | Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has quit [Ping timeout: 121 seconds] |
| 18:25 | | Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has joined #code |
| 18:25 | < Derakon[work]> | Did I miss anything? |
| 18:26 | <@jerith> | Yes. |
| 18:26 | < Derakon[work]> | What did I miss? |
| 18:26 | <@jerith> | But we won't tell you what you missed. |
| 18:26 | <@jerith> | :-P |
| 18:26 | <@jerith> | 20:20 -!- Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has quit [Ping timeout: 121 seconds] |
| 18:26 | <@jerith> | 20:25 -!- Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has joined #code |
| 18:26 | < celticminstrel> | You missed nothing. |
| 18:27 | | * Derakon[work] drops a jello anvil on Jerith. |
| 18:29 | | Finale [c0cb88fe@Nightstar-14e5d099.mibbit.com] has joined #code |
| 18:31 | <@Vornicus> | jello anvils. |
| 18:31 | < celticminstrel> | Heehee. |
| 18:31 | <@Vornicus> | So is that an anvil made of jello, or an anvil designed for forging jello. |
| 18:31 | < Derakon[work]> | The former, but the latter would be pretty amusing too. |
| 18:32 | < Finale> | ahaha |
| 18:33 | | Attilla [The.Attilla@FBC920.174237.0E5F2F.82D567] has joined #code |
| 18:33 | | mode/#code [+o Attilla] by Reiver |
| 19:12 | | * Derakon[work] spends twenty minutes trying to figure out why his newly-built version of the program is behaving just like the previous version, before realizing he forgot to update the .bat file to use said new version instead of the old version. |
| 19:16 | <@TheWatcher> | Ooop |
| 19:39 | | Rhamphoryncus [rhamph@Nightstar-a62bd960.abhsia.telus.net] has joined #code |
| 20:18 | | GeekSoldier_ [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has quit [[NS] Quit: Praise "BOB"!] |
| 20:40 | < Rhamphoryncus> | Anybody got a 16 core linux box I could borrow? >.> |
| 20:53 | < Rhamphoryncus> | hrm I'm seeing march for AMD's 12 core release |
| 20:55 | < Finale> | nifty |
| 21:02 | < Rhamphoryncus> | but I want it nowwwwww! |
| 21:05 | < Rhamphoryncus> | So I can benchmark standard python vs python-safethread where it might actually be worth using |
| 21:17 | < Derakon[work]> | What is python-safethread? |
| 21:17 | < Derakon[work]> | Does it get rid of the GIL? Please? >.> |
| 21:18 | | GeekSoldier [Rob@Nightstar-e86e3e0d.ip.cablemo.net] has joined #code |
| 21:19 | < Rhamphoryncus> | Derakon[work]: yes |
| 21:19 | < Derakon[work]> | Woot. |
| 21:19 | < Rhamphoryncus> | But making an alternate GC system (rather than simple refcounts) has been the focus of my pain |
| 21:20 | < Derakon[work]> | Is this Python 3 only? |
| 21:20 | < Rhamphoryncus> | it's a major fork of earlier python 3 |
| 21:20 | | * Derakon[work] nods. |
| 21:20 | < Derakon[work]> | Well, I wish you luck. :) |
| 21:20 | < Rhamphoryncus> | heh |
| 21:21 | < Rhamphoryncus> | Even an 8 core box could help a lot... |
| 21:22 | < Derakon[work]> | Get a Power Mac~ |
| 21:22 | < Derakon[work]> | Actually, Vorn has a fairly new Power Mac; dunno if it's 4 or 8 cores though. |
| 21:22 | < Derakon[work]> | You might ask him. |
| 21:22 | < Rhamphoryncus> | meh, I just don't want to be upgrading every 6 months |
| 21:22 | < Derakon[work]> | Heh. |
| 21:23 | < Rhamphoryncus> | 12 core seems like a good stepping point |
| 21:25 | < Rhamphoryncus> | maybe come march I'll get myself a whole new computer. Go for really high resolution |
| 21:27 | < Rhamphoryncus> | WQXGA :) |
| 21:27 | < Derakon[work]> | Heh. |
| 21:31 | < Rhamphoryncus> | Looks like that's the best generally available. There's medical greyscale that's higher, a CRT that was higher, and the discontinued IBM T220 family |
| 21:34 | < gnolam> | Mmm, radiography displays... |
| 21:34 | < Derakon[work]> | I got one of the giant 28" displays from NewEgg for $260, so I'm set for pixels for now. |
| 21:35 | | * Derakon[work] tries to figure out how to get Visual Studio's nmake to generate a debug build. |
| 21:38 | < Rhamphoryncus> | gnolam: greyscale might mess with my syntax highlighting ;) |
| 21:38 | < Rhamphoryncus> | Not to mention my porn! |
| 21:43 | < gnolam> | Solution: switch to mime porn. |
| 21:44 | < gnolam> | But there are wicked medical color screens as well. |
| 21:47 | < Finale> | I want a screen wall with a dot pitch below naked-eye visibility. >_> |
| 21:48 | < Finale> | imho, the dot pitch and possibly atmospheric blur generation are key to true graphic realism. |
| 21:50 | < Rhamphoryncus> | Finale: yeah, I always drooled over the T220 |
| 21:50 | < Derakon[work]> | I'm not going to worry about dot pitch until we get proper resolution-independent displays. |
| 21:51 | < Derakon[work]> | As it stands, getting a better dot pitch invariably results in my needing to squint to see what I'm working with. |
| 21:51 | < Finale> | it IS nice. I think we passed them in the stores this summer... |
| 21:53 | < Finale> | proper resolution independence would require at least one of a HUGE pixel array, in the millions per side probably, a programmable shifting array, or a staggered array... |
| 21:54 | < Finale> | hm. T220 has a 20k:1 dynamic contrast. I've seen 40k and even I think 50k... |
| 21:54 | < Finale> | but 20k is still better than most. XD |
| 21:55 | < Finale> | .282 dot pitch... also not the best. <_< |
| 21:55 | < Finale> | 2ms. yeah, that's good. |
| 21:55 | | * Finale likes. |
| 21:55 | <@McMartin> | Derakon: Is using devenv directly not an option? |
| 21:57 | < Derakon[work]> | McM: I'm basically clueless about building on Windows. I'm happy enough that I managed to replicate the previous dev's build environment to the point that I can make a working executable; now I get to figure out what it's actually doing. |
| 21:58 | < Derakon[work]> | FWIW, here's the makefile being used: http://derakon.dyndns.org/~chriswei/temp2/makefile.txt |
| 21:58 | < Derakon[work]> | (Tweaked so that the Y: mounted drive is actually locally mounted, mind you) |
| 21:59 | | * Derakon[work] eyes Windows, which just prompted him to delete a file he'd told it to delete, just because that file is a program and if he deletes it he won't be able to run it OMG!) |
| 22:02 | < Finale> | it's like that, yeah. |
| 22:02 | < Rhamphoryncus> | And there's a bajillion other files it won't prompt about that'll also stop a program from working |
| 22:02 | < Finale> | for the lusers. |
| 22:02 | < Finale> | yeah. |
| 22:02 | < Derakon[work]> | I don't generally use the term "luser". |
| 22:02 | < Derakon[work]> | Specific people may earn it, but being uninformed about computers is not itself a crime. |
| 22:03 | < Finale> | of course not. |
| 22:04 | <@ToxicFrog> | Being uninformed about computers and then rummaging around in their innards, however, is analogous to being uninformed about internal combustion engines, and opening up your car to disconnect random hoses. |
| 22:04 | < Derakon[work]> | My point is that such a message is nominally not directed at the luser (i.e. the intentionally clueless, refuses-to-learn type), but rather at the uninformed in general. |
| 22:04 | <@ToxicFrog> | And in either case the mechanic is entirely justified in smacking you. |
| 22:04 | < Finale> | mh. point. |
| 22:04 | < Derakon[work]> | Of course, as Rhamph noted the message is basically useless anyway. |
| 22:05 | <@McMartin> | I'm not used to that message. I'm used to the reverse, which is fair for even careless mechanics. |
| 22:05 | < Rhamphoryncus> | The uneducated users shouldn't have direct access to programs. They should simply be running them |
| 22:05 | <@McMartin> | "That program you just deleted was actually a symbolic link. You aren't deleting *anything*. Capiche?" |
| 22:05 | | * McMartin ponders a resource hacking program that turns "[ OK ]" into "[ Capiche ]" |
| 22:06 | <@ToxicFrog> | Rhamphoryncus: yes, but this kind of breaks down on a system where you need to be root all the time to do anything. |
| 22:06 | < Rhamphoryncus> | .. yes, that's stupid too |
| 22:06 | <@McMartin> | TF: That seriously hasn't been true since Vista. |
| 22:06 | | * Derakon[work] is on XP for this~ |
| 22:06 | <@ToxicFrog> | Quite. |
| 22:07 | <@ToxicFrog> | Conversely, it was true on every windows system before Vista. |
| 22:07 | <@ToxicFrog> | Which a lot of people are still using, too. |
| 22:07 | <@McMartin> | It's possible to not do it on XP, it's just that nobody does >_> |
| 22:07 | < Namegduf> | It's not really practical to do it on XP. |
| 22:08 | <@McMartin> | As of SP1 on Vista, even UAC is fairly reasonable. |
| 22:08 | <@McMartin> | The big problem with it is that the workflows for partial privilege don't remotely match POSIX, which means a bunch of things become *significantly* more annoying. |
| 22:09 | <@McMartin> | I'm sure the reverse would be true as well if anyone ever didn't design for POSIX first. |
| 22:09 | <@ToxicFrog> | The impression I've gotten from Vista is that they've solved the "not being administrator means you can't do anything" part without solving the "but you still need root access for far more stuff than is reasonable" part. |
| 22:10 | <@McMartin> | Mmmm |
| 22:10 | < Derakon[work]> | Well, a certain amount of that isn't up to Windows to solve. |
| 22:10 | <@ToxicFrog> | Hence the complaints about UAC tripping on everything ever. |
| 22:10 | <@McMartin> | I'm not convinced that's true, actually, either. |
| 22:10 | < Derakon[work]> | It involves getting the developers to update their programs to not assume privilege they don't have. |
| 22:10 | <@McMartin> | If POSIX prompted for auth every time a setuid program was run, I think you'd have a similar situation. |
| 22:10 | <@ToxicFrog> | Hmm |
| 22:11 | | * Derakon[work] heads into the lab. |
| 22:11 | <@McMartin> | If you take an architecture of "setuid programs that then drop privileges" and blindly port the architecture to Windows, you will end up getting UAC freaking out a lot *and* have a much larger trusted code base, both of which are bad. |
| 22:12 | <@McMartin> | If, on the other hand, you have an always-peon client and a heavily restricted always-admin service, UAC will prompt you exactly once - when the service is installed for the first time. |
| 22:12 | <@McMartin> | And no matter what some people say, demanding confirmation on installs is a perfectly legitimate decision. |
| 22:16 | <@ToxicFrog> | That I agree with. |
| 22:17 | <@McMartin> | But yeah, doing platform-independent service designs for POSIX and Windows is generally going to ensure one or the other loses. |
| 22:17 | <@McMartin> | Because POSIX's model is that if a process starts privileged, it will eventually drop privileges and keep them dropped forever after, while the Windows model involves a permanent admin thread that spawns impersonation threads that let it assume the identity of its clients as needed. |
| 22:18 | < Namegduf> | Because running in the background 24/7 for occasional usage is an awesome and infinitely scalable approach. |
| 22:18 | < Namegduf> | Unless you can request the service start without privileges or something weird. |
| 22:19 | <@McMartin> | COM requests to svchost.exe make them start as needed, AIUI, but my understand mostly comes from using Process Explorer to hack at other products. |
| 22:19 | <@McMartin> | *my understanding |
| 22:20 | <@McMartin> | I'm not sure offhand if standard users can do sc start/sc stop on some services |
| 22:20 | <@McMartin> | If they can, then there basically is a 1-1, though you'll have a pointless simulation layer if anything important happens after privilege-drop on the admin side. |
| 22:21 | <@McMartin> | If you do have important stuff there, you're better off redesigning your protocol. |
| 22:21 | <@McMartin> | (Where this has been burning me: writing installers. >_<) |
| 22:22 | <@McMartin> | Or rather, customers who are misusing nonadmin accounts, but this basically means I still have to deal with it. |
| 22:22 | <@McMartin> | Shockingly, when you do the equivalent of su to install something, and you tell that program to install for "just you" instead of "all users", the "you" is the person you su-d to. |
| 22:23 | < Finale> | hehe |
| 22:24 | <@McMartin> | Note that this doesn't happen in sudo, but nonadmin accounts in Vista+ aren't allowed to do the sudo equivalent ("admin" there is closer to "wheel" than "root") |
| 22:28 | < Finale> | wheel? |
| 22:30 | < Rhamphoryncus> | wheel is a group used for root access |
| 22:31 | < Rhamphoryncus> | or.. something. I don't recall the specifics |
| 22:32 | <@McMartin> | It's often the group that gets to run sudo. |
| 22:32 | <@McMartin> | When there isn't an actual group "sudoers" |
| 22:33 | < Finale> | admin in xp, at least, can have bugs where various programs won't recognize that it has admin privileges. XD |
| 22:33 | <@McMartin> | Ah, yes |
| 22:33 | | * McMartin actually wrote a little 12-line program to test for adminness two different ways, Just In Case. |
| 22:40 | < Finale> | my antivirus suite, for example, requires admin privileges to update its libraries. >_< |
| 22:40 | <@McMartin> | ... wouldn't that be the equivalent of writing to /usr/share? |
| 22:42 | < Finale> | and half of them refuse to update because they can't see my privileges. XD |
| 22:42 | < Finale> | they used to update just fine, then suddenly stopped. >_< |
| 22:42 | < Finale> | this isn't the only problem they have, so just stay the heck away from CA. <_< |
| 22:42 | < Finale> | the only reason I'm using it is cause dad bought it. and it does have some better functions than AVG. |
| 22:50 | | Reivthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has quit [Ping timeout: 121 seconds] |
| 22:51 | | Orthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has joined #code |
| 22:52 | <@McMartin> | CA is still better than Norton or MacAffee or Symantec. >_> |
| 22:55 | < Finale> | eh, McAfee worked pretty well while we had it. |
| 22:55 | < Finale> | better than CA. |
| 22:55 | < Finale> | and isn't Norton=Symantec now? |
| 22:59 | <@McMartin> | Not sure, tbh |
| 23:00 | | Orthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has quit [Client closed the connection] |
| 23:05 | < Finale> | oh well, I don't really care. >_> |
| 23:06 | | * McMartin uses AVG, himself, but ehn |
| 23:08 | | Orthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has joined #code |
| 23:09 | < Rhamphoryncus> | Huh. Guess I can't argue with a 13% boost |
| 23:11 | <@McMartin> | Hm? |
| 23:22 | < Rhamphoryncus> | Switching safethread from using a hashtable to a plain array |
| 23:22 | < Rhamphoryncus> | For storing asynchronous refcounts |
| 23:22 | | Orthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has quit [Client closed the connection] |
| 23:25 | < Rhamphoryncus> | Meaning each thread has its own array of refcounts for shared objects, which uses a lot of memory |
| 23:29 | <@McMartin> | Aha |
| 23:30 | | Orthia [Orthianz@Nightstar-899214c3.xnet.co.nz] has joined #code |
| 23:31 | <@ToxicFrog> | safethread? |
| 23:33 | < Derakon[work]> | Python with better multithreading support. |
| 23:41 | | Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has quit [[NS] Quit: Leaving] |
| --- Log closed Sat Dec 19 00:00:26 2009 |