00:01:43 -!- oerjan has joined.
00:02:50 <Rugxulo> yeah, I figured it was too late there
00:07:59 <oerjan> <Phantom_Hoover> OK, so can a von Neumann CA be represented as a non-planar graph?
00:08:22 <oerjan> um, the neighborhood graphs for _both_ von Neumann and Moore neighboorhoods are planar.
00:09:04 <oerjan> if you mean anything else by represented, please specify. also, please read the logs.
00:09:46 <oerjan> but definitely von Neumann
00:16:25 -!- Rugxulo has quit (Quit: Rugxulo).
00:22:05 -!- Oranjer has joined.
00:25:48 <uorygl> `translate Ellas comen unos tostitos.
00:34:31 <oerjan> uorygl: that lazy bum Gregor told me we should fix `translate ourselves
00:34:43 <oerjan> but i'll show him. i'll out-lazy him!
00:35:07 <oerjan> `translate I don't think that makes a difference
00:45:02 -!- Oranjer has left (?).
00:50:42 -!- coppro has quit (Remote host closed the connection).
01:04:55 -!- Oranjer has joined.
01:06:39 -!- pikhq_ has joined.
01:07:15 -!- coppro has joined.
01:08:41 -!- pikhq has quit (Ping timeout: 252 seconds).
01:15:06 -!- pikhq_ has changed nick to pikhq.
01:31:53 -!- oerjan has quit (Quit: leaving).
01:39:24 -!- Zuu has quit (Ping timeout: 276 seconds).
01:44:56 -!- Zuu has joined.
01:44:56 -!- Zuu has quit (Changing host).
01:44:57 -!- Zuu has joined.
01:47:54 -!- bsmntbombdood has joined.
02:37:34 -!- augur has quit (Ping timeout: 245 seconds).
02:43:25 -!- augur has joined.
02:53:24 -!- augur has quit (Ping timeout: 245 seconds).
02:57:06 -!- lament has joined.
03:08:40 -!- Oranjer has left (?).
03:13:40 -!- jcp has changed nick to javawizard2539.
03:13:44 -!- javawizard2539 has changed nick to javawizard.
03:13:46 -!- javawizard has changed nick to jcp.
03:19:42 -!- Sgeo_ has joined.
03:22:09 -!- Sgeo has quit (Ping timeout: 245 seconds).
03:25:25 -!- Oranjer has joined.
03:38:16 -!- augur has joined.
03:46:19 -!- augur has quit (Ping timeout: 245 seconds).
03:47:55 -!- augur has joined.
03:50:15 <augur> brussel sprouts are so good omg
03:51:40 <Ilari> Especially with butter, heavy cream or something like that? :->
03:53:24 <Gregor> That's "Brussel's sprouts"
03:53:38 <Gregor> Minus the apostrophe :P
03:54:00 <Gregor> It's hard to type Brussels without thinking it's a possessive.
03:55:19 <pikhq> Easy when you recall that Brussels is a place name.
03:55:20 <augur> brussels sprouts, broccolo, oil, salt, pepper, baked
03:56:22 <augur> to brustle -- to bro rustle, that is, to steal bros from a field as they graze on natty ice
03:57:25 <pikhq> What is this strange language you speak, augur?
03:57:43 <augur> broccolo, you know, one head of broccoli
03:57:47 <augur> as opposed to mean heads
03:57:50 <augur> which are broccoli!
03:58:55 <Gregor> (broccoli glans is the head of broccoli :P )
04:01:02 -!- sshc has quit (Read error: Connection reset by peer).
04:03:51 -!- augur has quit (Remote host closed the connection).
04:04:15 -!- augur has joined.
04:04:23 -!- oerjan has joined.
04:05:28 -!- augur has quit (Remote host closed the connection).
04:05:45 -!- augur has joined.
04:22:37 -!- pikhq_ has joined.
04:25:09 -!- pikhq has quit (Ping timeout: 276 seconds).
04:32:29 -!- pikhq has joined.
04:33:09 -!- pikhq_ has quit (Ping timeout: 264 seconds).
04:42:52 -!- pikhq_ has joined.
04:44:59 -!- pikhq has quit (Ping timeout: 276 seconds).
05:04:18 -!- Oranjer has left (?).
05:18:19 -!- sshc has joined.
05:19:28 <Gregor> xkcdsucks is gonna have a field day
05:19:58 <oerjan> i thought the comic itself was nice
05:20:09 <oerjan> but the alt text could have been better
05:24:08 <Sgeo_> The xkcdsucks for 744 isn't up yet
05:25:13 <Sgeo_> http://wikisuperosity.com/index.php?title=Special:RecentChanges Somehow, I don't think I'm going to bother to get involved with this wiki
05:32:04 -!- coppro has quit (Remote host closed the connection).
05:50:28 -!- oerjan has quit (Quit: leaving).
06:04:52 <AnMaster> <oerjan> very meta, xkcd <-- argh, parsing the alt text hurts
06:10:49 -!- coppro has joined.
06:56:27 -!- FireFly has joined.
07:04:14 -!- tombom has joined.
07:20:19 -!- kar8nga has joined.
07:51:13 -!- tombom has quit (Quit: Leaving).
07:59:40 -!- kar8nga has quit (Remote host closed the connection).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:06:03 -!- FireFly has quit (Quit: swatted to death).
08:27:34 -!- Sgeo_ has quit (Ping timeout: 245 seconds).
08:44:42 -!- augur has quit (Remote host closed the connection).
08:45:05 -!- augur has joined.
08:46:10 -!- MizardX has joined.
08:55:15 -!- Phantom_Hoover has joined.
08:59:37 -!- atrapado has joined.
09:00:56 <coppro> does anyone know of an X server that just forwards to another X server, but can have the target server disconnected and another one reattached later?
09:08:22 <AnMaster> coppro, eh, like screen for X?
09:08:37 <coppro> exactly like screen for X
09:08:53 <AnMaster> coppro, have a running xserver + X11vnc or such?
09:09:02 <coppro> that's too much work though
09:09:06 <AnMaster> or do you mean forward as in, reattach to desktop?
09:09:33 <coppro> yeah, it should run them in the local server
09:09:54 <AnMaster> I'm sure a lot of people would find it usefuk
09:10:00 <coppro> the toughest bit would be handling reattachment I think
09:10:22 <AnMaster> coppro, well, you need to hide deattch/reattach from the programs
09:10:33 <coppro> the program never notices
09:10:35 <AnMaster> since X can't handle that sort of stuff
09:10:45 <AnMaster> coppro, also opengl would be a pain
09:11:27 <coppro> stupid X persistent resources
09:11:27 <AnMaster> for example: nvidia drivers mmap stuffs iirc
09:12:26 <coppro> The local server would have to maintain anything like resources that might get buggered by a server restarting
09:12:48 <coppro> fortunately, an application already has to cope with being moved onto a new screen
09:13:05 <AnMaster> coppro, hm? that is the same X server isn't it?
09:13:17 <AnMaster> well, that way you don't have to change as much
09:13:39 <coppro> but so things like random "you lose OpenGL" are already accounted for (in theory)
09:14:03 <AnMaster> you don't lose opengl due to moving between screens do you?
09:14:03 -!- lament has quit (Ping timeout: 240 seconds).
09:14:22 <coppro> if they have different cards and one has no drivers, you might
09:14:27 <AnMaster> also, as I said, I'm pretty sure X11 uses shm for local clients. And for opengl it would be very slow to use the socket method
09:15:24 <coppro> then I'll need to spoof it into thinking it's nonlocal
09:15:35 <AnMaster> coppro, and that is quite a speed difference
09:15:36 <coppro> (if I do in fact do this)
09:15:58 <AnMaster> opengl over non-shm is not feasible I suspect
09:16:15 <coppro> AnMaster: This is mostly intended for network use anyways. OpenGL speed is not a concern
09:16:17 <AnMaster> coppro, of course you could use shm against the client but then socket to the server
09:16:34 <AnMaster> it depends on how much you are willing to do
09:16:41 <coppro> I'd have to research the spec more. It's quite possible OpenGL isn't really feasible to implement
09:17:08 <AnMaster> coppro, there are other stuff than opengl that could cause issues
09:17:22 <coppro> Anything that X persists would have to be intercepted by the proxy
09:17:38 <coppro> it would have to persist them itself, and establish new copies of them when it connected to a new server
09:17:55 <coppro> I don't think that matters?
09:18:08 <AnMaster> coppro, you would have to forward such info to the client I suspect
09:18:35 <coppro> any info that the proxy doesn't need to handle is forwarded directly
09:18:51 <AnMaster> have you ever tried a non-xinerama build of firefox on a dual head X server (using nvidia twinview, not multiple screens)?
09:19:00 <AnMaster> I had menus appear on the wrong screen
09:19:10 <AnMaster> like file menu showed up on the other monitor from the application window
09:19:46 <AnMaster> coppro, rebuilding with xinerama support fixed it
09:20:19 <coppro> the proxy would retrieve any screen/display/xinerama/whatever info upon connecting to a server, and inform the client about the changes. Anything beyond that is handled by the communication directly between the server and client
09:20:41 <AnMaster> coppro, there are some extensions that could cause issues still
09:20:54 <coppro> yes, well, those could be handled on a case-by-case
09:21:01 <AnMaster> what would you do if client used xrandr for example?
09:21:14 <coppro> forward directly to the server, do not pass go, do not collect $200
09:21:21 <AnMaster> coppro, "X RandR is used to configure which display ports are enabled (e.g. LCD, VGA and DVI), and to configure display modes and properties such as orientation, reflection and DPI."
09:22:03 <coppro> I don't see how it's an issue. The goal is to make the client act as if it's connected the target server, except that the target server might randomly change.
09:23:02 <coppro> and if it does, the proxy needs to handle any aspect of that change so that the client, other than possibly noticing a radical change in geometry, can proceed normally
09:23:02 <AnMaster> coppro, yes and that is the issue. You would have to notify it about xrandr changes and such to reset the client's state when you change the "host" server
09:23:09 <coppro> yes, I just explained that
09:23:25 <AnMaster> coppro, and what with all the X11 extensions that could be quite a messy task
09:23:26 <coppro> all geometry info is retrieved when connecting to a new server; clients are informed of the changes
09:23:35 <coppro> after that, geometry communication is server-client
09:24:11 <coppro> if a server is disconnected, nothing happens until a new one is connected, so it's like the application was just left idle normally
09:24:41 <coppro> the tricky thing is handling any persistence info, like window IDs
09:24:46 <coppro> or resources or something
09:25:06 <coppro> the proxy needs to create a virtual set of them that the client sees
09:25:31 <coppro> and reestablish them upon connecting
09:26:43 <AnMaster> coppro, what is the app grabs the mouse and hides the pointer
09:26:50 <AnMaster> will that be auto restored on reconnection?
09:27:13 <coppro> Not at first, for sure
09:27:20 <AnMaster> coppro, that could confuse the app
09:28:08 <coppro> AnMaster: I've seen it happen with badly-written applications before. The result is an immovable default cursor in the default position; the game's cursor works fine
09:28:43 <coppro> (this occurred when using Debian's virtual terminals to switch away from X and back to it)
09:29:37 <AnMaster> you know that is a linux feature, not specific to debain
09:29:37 <coppro> I don't know if other distros have it; I know Debian does
09:29:45 <coppro> wasn't sure on that one
09:29:53 <coppro> I could also imagine a situation where the default cursor is mobile, but the game is still tracking its own cursor. It would also be weird
09:30:02 <AnMaster> coppro, they would have had to patch to kernel for it ;P
09:30:15 <coppro> AnMaster: I wouldn't put them past it.
09:30:58 <coppro> anyways, I think I'd just see what events X normally passes if it loses screen control like that and replicate them upon a reconnection
09:32:04 <coppro> and then let badly-written applications eat cake.
09:32:25 <coppro> hmm... what should this be called... Also, Erlang.
09:32:46 <AnMaster> coppro, what? X11 protocol in erlang?
09:33:04 <coppro> Erlang's the obvious candidate for this!
09:34:06 <AnMaster> coppro, I would have said that C was
09:34:20 <coppro> for managing tons of connections simultaneously? no thanks
09:34:22 <AnMaster> since then you can use X11 header files for stuff
09:34:44 <AnMaster> coppro, you mean multiple apps per "bouncer"?
09:34:55 <coppro> AnMaster: Yes, it would be a full server
09:35:10 <AnMaster> there is dtach which is like screen but with most features except deattaching and reattaching removed
09:35:27 <coppro> yeah, sounds about right
09:35:37 <AnMaster> coppro, but I can't pronounce xdtach
09:36:33 <AnMaster> http://dtach.sourceforge.net/ <-- strangely enough this is not a project from the authors of dwm and slock (as far as I know). It would fit right into that category though
09:37:34 <coppro> ideally you could even run a DM through it, but that's another can of worms
09:37:42 <AnMaster> coppro, can't do XD-tach. XD(TM)(R) memory cards(TM)(R)
09:38:00 -!- augur_ has joined.
09:38:12 <AnMaster> coppro, oh yeah... how do you handle dbus?
09:38:17 -!- augur has quit (Read error: Connection reset by peer).
09:38:20 <AnMaster> coppro, a lot of apps these days use dbus
09:38:35 <AnMaster> coppro, a lot of those wouldn't work correctly then
09:38:51 <AnMaster> coppro, anything like gtk or qt apps would be a pain
09:39:11 <AnMaster> sure, simple motif apps or such, no problem I bet
09:39:54 -!- kar8nga has joined.
09:39:54 <coppro> You'd probably need a similar daemon that would do the same with dbus
09:40:17 <AnMaster> hm how does dbus interact with ssh X11 forwarding?
09:40:51 <AnMaster> coppro, so it might not be required
09:41:11 <coppro> or you could just run it locally and hope that keeping it disconnected from the target machine wouldn't explode things
09:41:24 <coppro> actually, no, you musn't actually need it
09:41:39 <coppro> or at least run it connected
09:41:56 <coppro> think of a non-ssh network X connection
09:42:02 <coppro> you could be connecting to /anything/
09:42:10 <coppro> including a Windows machine
09:42:40 <AnMaster> coppro, still what about interacting with window manager
09:43:05 <AnMaster> coppro, I suspect it might be problematic if you switch from a host with, say, metacity to one with kwin
09:43:15 <AnMaster> or even worse, from a "normal" one to a tiling WM
09:43:24 <AnMaster> in fact, tiling WMs open up a new can of worms
09:43:27 <coppro> AnMaster: Not much different than replacing a running WM
09:44:22 <AnMaster> coppro, that can cause buggy behaviour in my experience
09:44:36 <coppro> in fact, it would be slightly easier
09:44:45 <AnMaster> coppro, also, I can't imagine how replacing a non-tiling one with a tiling vm would work well
09:44:51 <coppro> since a reconnection would involve the proxy opening up a new set of windows
09:45:28 <AnMaster> coppro, in a tiling wm it might be problematic to allocate the correct window sizes
09:45:40 <coppro> that's the wm's problem
09:46:06 <AnMaster> pretty sure the X standard suggests it is the problem of the client to fit into the space given by the wm :P
09:46:36 <coppro> it is the client's problem to do that
09:46:50 <AnMaster> well you said "<coppro> that's the wm's problem"
09:47:03 <coppro> you said "allocate the correct window sizes" which is the wm's problem
09:47:10 <coppro> it is the client's problem to fit in them
09:47:17 <AnMaster> coppro, it might give you a smaller size back
09:47:48 <coppro> the proxy could help by forcibly resizing the window down
09:48:02 <coppro> after that, it's the client's call
09:48:12 <coppro> the window stuff will be the toughest
09:48:19 <coppro> and I don't use a tiling WM, which is also important
09:48:52 <coppro> plus, if this is as ridiculously useful as it seems like it would be, it will become and instant it, I will receive song and praise and volunteer developers
09:49:37 <AnMaster> coppro, you think there are enough people interested who know erlang
09:49:57 <AnMaster> coppro, also using erlang more or less makes opengl impossible
09:50:21 <AnMaster> you need very high performance for it probably.
09:50:36 <AnMaster> iirc you need to use shm to get reasonable speed
09:50:55 <AnMaster> sure you think about network but most people would probably like to use it locally
09:51:28 <coppro> HIPE if necessary. Shm is a serious cluster if you detach from a local server and move to a remote one.
09:52:23 <AnMaster> coppro, also the thing could use shm against the clients, and non-shm "upstream"
09:52:43 <coppro> yeah, I realize both of those things.
09:52:59 <AnMaster> plus I'm not sure how stable it is. As in, clients being supposed to use the stable APIs of libx11 (or more recently, libxcb iirc)
09:53:17 <AnMaster> how well defined is the actual protocol
09:54:00 <coppro> X is pretty well-defined
09:54:08 <coppro> I've read the spec before
09:54:44 <coppro> and it's got to be stable or else servers which implement the protocol directly would break
09:55:21 <yiyus> i don't see why tiling wms would be a problem, if a client request a fixed size is usually handled as "floating", else it will only see a resize request
09:57:20 <yiyus> at least that is how dwm and wmii work, and afaik most of tiling wms are based on them
09:58:24 <coppro> The X core protocol hasn't changed in 5 years, and then probably not by much
10:08:33 <AnMaster> will still be a very complex task
10:14:59 -!- coppro has quit (Read error: Connection reset by peer).
10:15:10 -!- Halph has joined.
10:15:22 -!- Halph has changed nick to coppro.
10:36:24 -!- coppro has quit (Read error: Connection reset by peer).
10:50:02 * Leonidas stumbled over this code http://okmij.org/ftp/Prolog/QueensProblem.prl
10:50:24 <Leonidas> while not that much esoteric, would anyone explain to me how permute works?
10:50:40 * Leonidas is trying to translate that into a more esoteric language
10:54:51 -!- kar8nga has quit (Remote host closed the connection).
11:12:02 -!- ais523 has joined.
11:36:53 <AnMaster> so dbus crashed, then gnome-settings-daemon
11:37:02 <AnMaster> now everything looks strange even after restarting those
11:37:17 <AnMaster> oh and: [251808.083600] gnome-settings-[1841]: segfault at 8 ip 00007fab28a5ad12 sp 00007fff642dfcb0 error 4 in libclipboard.so[7fab28a58000+5000]
11:37:31 <AnMaster> how? well I just closed kate....
11:37:38 <AnMaster> not sure what kate has to do with gnome at all
11:37:57 <AnMaster> also second time it happened, last time the computer was idle when it happened
11:38:14 <AnMaster> and it is not easy to reproduce on demand
11:42:26 -!- gm|lap has joined.
11:43:00 <gm|lap> next time you see cpressey, can you tell him that it's possible to get >5 keys in bubble escape 2K, and that, if you do that, you can't win?
11:43:55 <AnMaster> gm|lap, haven't seen cpressy for a long time
11:44:22 <AnMaster> gm|lap, and what on earth is bubble escape 2K?
11:44:33 <gm|lap> AnMaster: a game he wrote for the C64
11:44:46 <gm|lap> of course, i don't have a C64 so i just use VICE
11:46:45 * AnMaster tries to imagine a language for which qwerty is actually an efficient layout
11:47:58 <AnMaster> meh, I don't know enough about how to make efficient layouts to figure that out
11:48:14 <ais523> http://esolangs.org/wiki/Topline IIRC
11:48:48 <ais523> even so, though, you'd probably need 6 fingers on each hand
11:48:51 <AnMaster> ais523, I meant natural language
11:49:03 <ais523> and it doesn't really care about the letter keys, just the punctuation marks
11:49:19 <ais523> AnMaster: then you probably asked in the wrong channel
11:49:56 -!- ais523 has quit (Read error: Connection reset by peer).
11:51:36 -!- ais523 has joined.
11:51:57 <ais523> [11:50] <ais523> hmm, a problem I'm working on at University is basically an esolang issue
11:51:58 <ais523> [11:50] [CTCP] Sending CTCP-PING request to ais523.
11:52:00 <ais523> [11:50] <ais523> OTOH, I'd probably get in trouble if I asked here for the answer
11:52:01 <ais523> [11:50] <ais523> so I'll try to solve it first and then tell everyone what it is
11:56:40 <ais523> meh, I'm bored and slightly mad, I'm going to update Ubuntu to a new distro version unless someone stops me to tell me it's a really bad idea
11:57:07 <ais523> let's see if it fixes that intermittent problem with sound that I could never reliably reproduce
12:03:45 <Phantom_Hoover> If you install it Satan will jump through your monitor and eat your soul!
12:03:52 <gm|lap> fredfredfredfredfesdfresdresdfredfredfredfredfredred
12:03:58 <gm|lap> and i'm going to bed, gnight
12:04:11 -!- gm|lap has quit (Quit: ilua).
12:08:12 <ais523> this computer's a laptop, though
12:08:16 <ais523> so technically doesn't have a monitor
12:08:30 <ais523> also, I don't believe you
12:11:26 <ais523> if your soul has been eaten, why are you still talking here?
12:23:19 <Phantom_Hoover> If it does anything less than teleportation I will be sorely disappointed.
12:38:01 <Phantom_Hoover> Can someone else zoom in on that and tell me if it looks like a glider?
12:56:22 <ais523> it's only three pixels with a bit of antialiasing, hard to tell /what/ it looks like
13:08:41 <AnMaster> <ais523> [11:50] <ais523> so I'll try to solve it first and then tell everyone what it is <-- figured it out yet?
13:08:55 <ais523> in the middle of a distro upgrade atm
13:11:02 <AnMaster> which esolang? or is it general computation theory stuff or such?
13:12:12 <AnMaster> <Phantom_Hoover> ¸ <-- isn't it a . ?
13:12:37 <AnMaster> <Phantom_Hoover> Bah, Raymond has commandeered it for himself. <-- ?
13:13:08 <AnMaster> Phantom_Hoover, eh, that is in unicode?
13:13:26 <AnMaster> also, it looks like two pixels with antialias here
13:13:40 <AnMaster> just look up the unicode codepoint from that char
13:24:08 <ais523> AnMaster: if you want a bit of fun, https://review.source.android.com/#patch,sidebyside,14699,1,libc/memset.c
13:24:14 <ais523> it's the sort of diff I thought you would like
13:24:17 <ais523> although it requires JS to read
13:24:31 <ais523> sort-of worrying that that wasn't caught earlier...
13:25:57 <ais523> also, says a lot about how memset's used in practice
13:26:15 <AnMaster> ais523, android has it's own libc?
13:26:27 <ais523> it has its own pretty much everything
13:26:39 <ais523> it's Linux, but it isn't any of the things that are commonly bundled with Linux
13:26:41 <AnMaster> ais523, also strange it doesn't use something more efficient, considering embedded system and so on
13:26:50 <ais523> it's not /that/ embedded
13:27:10 <ais523> <petdance> Why are they even doing that as a loop? They should just be calling memset().
13:27:50 <ais523> not clear from context
13:27:56 <ais523> although I'm guessing serious
13:29:42 <ais523> it's sort of like the source-code to libgcc
13:29:52 <ais523> which defines the implementations of, say, multiplication
13:29:55 <ais523> on systems that don't do it natively
13:30:15 <ais523> it must be a real pain to try to write an efficient multiplication routine in C without accidentally defining it in terms of itself
13:42:45 -!- Mathnerd314 has quit (Ping timeout: 264 seconds).
13:46:18 -!- Tritonio_GR has joined.
13:50:21 <ais523> well, fixed it, but it was broken for ages
13:50:25 <ais523> set to 0 rather than to the argument
13:50:34 -!- oerjan has joined.
13:50:36 <ais523> hmm, apparently IE4 could be scripted using Haskell
13:53:54 <ais523> yes, Internet Explorer
13:54:03 <Phantom_Hoover> Fun thing: When I typed "internet explorer has" into Google for that, the suggestions was "internet explorer has stopped working".
13:54:10 <ais523> IE[0-9]+ seems to be an abbreviation reserved for Internet Explorer nowadays
13:54:15 <ais523> try the same thing with Bing
13:55:06 <ais523> what does Bing say? the same?
13:55:15 <oerjan> <Phantom_Hoover> OK, so can a von Neumann CA be represented as a non-planar graph?
13:55:16 <ais523> Bing was infamous for its autocompletions on "linux"
13:55:22 <ais523> which included "linux microsoft" and "linux windows"
13:55:36 <oerjan> the neighborhood graph for von Neumann CAs is planar
13:56:19 <oerjan> take a cell, draw a line to all four neighboring cells
13:57:06 <oerjan> doing that for all cells gives no crossing lines
13:57:07 <Phantom_Hoover> ais523: Try putting "Windows has" into Bin for a laugh.
13:57:34 <ais523> I'd need to remove several layers of browser lockdown to use Bing correctly
13:58:29 <oerjan> Phantom_Hoover: in that case, i suspect you've switched the definition of moore and von neumann neighborhoods
13:59:19 <oerjan> right, planar means no crossed lines, essentially
13:59:29 <ais523> Phantom_Hoover: in Bing's case, mostly against non-static content
13:59:35 <ais523> I wouldn't get suggestions, just a textbox
13:59:44 <ais523> in fact, I'm not even convinced I get the background image, I haven't used it in a while
14:00:03 <ais523> yep, it's just really busy and grey
14:00:29 <ais523> and I get no suggestions
14:00:38 <ais523> which is my standard query for testing search engines
14:01:25 <ais523> hmm, results better than they used to be, about as good as Google's and worse than Wikia's used to be
14:02:53 <Phantom_Hoover> I thought they just hosted innumerable pop-culture wikis.
14:04:39 <ais523> due to lack of funding
14:04:46 <ais523> it was pretty good while it was there, though
14:05:14 <ais523> in the sense of "provided good results that were different from Google's"
14:05:38 <ais523> maybe slightly worse than Google's on average, but being different was good as it let you find different pages
14:06:49 <AnMaster> <ais523> it must be a real pain to try to write an efficient multiplication routine in C without accidentally defining it in terms of itself <-- that is why it iirc uses a lot of inline asm versions for various arches
14:07:06 <ais523> actually, it isn't even inline asm
14:07:13 <ais523> it goes in the .md file, in sections reserved for writing asm
14:07:26 <ais523> so it's asm, but not "inline"
14:07:34 <ais523> the .md file's /meant/ to contain the asm
14:07:44 <ais523> because otherwise, htf could it know what asm instructions to output?
14:07:50 <AnMaster> ais523, but isn't that for defining codegen?
14:07:56 <AnMaster> rather than for writing libgcc
14:08:15 <ais523> it would be inline asm in libgcc if you wanted it to be an actual function
14:08:18 <Phantom_Hoover> Idea for an esolang: you specify a fitness metric, then the interpreter/compiler evolves a neural network according to it.
14:08:24 <ais523> but for things like multiplication, if it's short enough you probably want to inline it
14:08:37 <AnMaster> ais523, yes and then libgcc isn't involved at all afaik
14:09:01 <AnMaster> ais523, well I thought libgcc was the context: "<ais523> it's sort of like the source-code to libgcc"
14:09:03 <ais523> nothing either of us said in the conversation leading from that conversation implied libgcc was involved
14:09:10 <ais523> AnMaster: that was the context of a /different/ conversation
14:09:16 <AnMaster> ais523, well I was log reading
14:09:23 <ais523> just quoting a line from an old conversation doesn't necessarily drag in all its context
14:09:25 <AnMaster> as you could see from me quoting part of it
14:09:41 <ais523> Phantom_Hoover: evfunge, but the internet's short on details concerning it
14:09:53 <ais523> it would be pretty interesting to find a copy of the code involved, or its output
14:12:11 <AnMaster> ais523, why would you use bing anyway?
14:12:31 <ais523> AnMaster: I'm not sure
14:12:44 <ais523> amusement value, possibly
14:12:52 <ais523> or because you like pretty backgrounds
14:12:58 <AnMaster> I wonder when they will rename it next
14:13:00 <ais523> or because it's your default search engine and you don't know how to change
14:13:11 <ais523> win7 actually has a bing search widget on the start toolbar
14:13:35 <ais523> probably not by default, but I've seen installs that do
14:13:49 <AnMaster> <ais523> or because it's your default search engine and you don't know how to change <-- hm but doesn't IE ask about that during first run?
14:13:59 <AnMaster> "want to get more search engines" or such
14:14:08 <ais523> possibly, but people are trained to click Cancel
14:14:15 <ais523> or really, any button at random
14:14:34 <ais523> that's one of the things that makes Windows tech support so hard
14:14:44 <ais523> people dismiss dialog boxes out of habit, rather than reading them
14:15:10 <ais523> also, it's why Firefox greys out the "OK" button for three seconds on the "do you want to run/install this" confirmation screens
14:15:46 <ais523> on Linux, it's pretty rare that a dialog box comes up that isn't a) important, or b) asking for information needed to do what the user just asked it to do
14:15:48 <Phantom_Hoover> Also, I *hate* it when people type "google.com" into their address bar.
14:15:51 <ais523> which is a blessing, really
14:16:03 <AnMaster> ais523, now, firefox's "sure you want to continue with invalid ssl cert" is certainly hard to just get through by random clicks
14:16:22 <ais523> I love the way it has two buttons which are both "cancel", and the get-past isn't a button at all
14:16:23 <AnMaster> Phantom_Hoover, do they? Depends on browser
14:16:43 <ais523> AnMaster: w3m doesn't have an address bar either, though
14:16:49 <ais523> you can type U for a prompt for address
14:16:51 <ais523> but that isn't quite the same
14:16:53 -!- oerjan has quit (Remote host closed the connection).
14:16:56 <AnMaster> ais523, but iirc w3m-mode does
14:17:21 <ais523> AnMaster: I think there's one button that takes you to homepage, and another that explains why the site was blocked
14:17:28 <ais523> hmm, I wonder what the best way to check is
14:17:34 <ais523> is there a "test malicious page" for Firefox?
14:18:10 <AnMaster> ais523, I know some urls at my university with broken and/or self signed ssl certs
14:18:10 <ais523> sort-of like that test file which is not malicious, but is flagged as a virus by all antivirus software
14:18:14 <AnMaster> but since you filter urls anwyay
14:18:16 <ais523> due to existing as a test for antivirus software
14:18:24 <ais523> yep, I'll try to find one myself
14:18:30 <ais523> self-signed certs can be easy enough to find
14:18:43 <ais523> heh, I even have some self-signed files on my local computer
14:18:51 <AnMaster> ais523, plus I don't know if they fixed it. They seem to break it in different ways every now and then
14:19:09 <ais523> strange, mozilla.com's taking ages to load
14:19:12 <AnMaster> and randomly fixing it for a short period
14:20:06 <ais523> well, it's doing something, it's redirected once already
14:20:09 <ais523> but it's taken over a minute
14:20:10 <AnMaster> but considering the length of the traceroute... I'm not surpirsed
14:20:42 <AnMaster> took about 20 seconds to load here
14:20:43 <ais523> but unless mozilla.com was on a massively slow nextwork, you can easily do 21 hops in a minute
14:21:06 <AnMaster> ais523, also slow in what sense
14:21:23 <ais523> not sure if it's latency or throughput
14:21:29 <ais523> the headers seem to have arrived, though
14:21:33 <ais523> so I'm guessing it's throughput
14:21:43 <ais523> as in, the <title> has shown up, but the rest of the page hasn't
14:21:51 <AnMaster> ais523, could be packet loss, try ping?
14:22:09 <ais523> can't, I'm on a network that filters ICMP
14:22:23 <ais523> traceroutes only go about 3 hops before hitting a firewall
14:22:29 <ais523> it's strange, as it doesn't seem to filter anything else
14:22:39 <ais523> apart from possibly port 25
14:22:59 <AnMaster> ais523, what about non-PING ICMPs?
14:23:10 <AnMaster> like, destination unreachable or whatever
14:23:12 <ais523> not sure even how to generate those
14:23:25 <ais523> presumably it doesn't filter destination unreachable /replies/
14:23:38 <ais523> but if I send a destination unreachable into the internet as a whole, the recipient will wonder wtf is going on
14:23:52 <ais523> assuming they don't just discard it
14:23:57 <AnMaster> ais523, "internet as a whole"? 255.255.255.255?!
14:24:11 <AnMaster> well that won't get far I can tell youy
14:24:15 <ais523> except that all sane routers refuse to respect the defined semantics of such an address
14:24:51 <ais523> I still like the concept of having an IP address that means "the entire internet", even if you can't actually use it in practice
14:25:16 <ais523> hmm, stopped loading and tried again, it loaded in less than three seconds
14:25:31 <AnMaster> ais523, few broadcasts tend to get past any routers iirc
14:25:45 <AnMaster> mostly it only works on the same network segment or whatever the name is
14:29:15 <ais523> IIRC, the default for stealing variables in CLC-INTERCAL is to broadcast to the whole Internet in the hope of finding some other running INTERCAL program to steal from
14:29:22 <ais523> which makes the whole thing rather like Network Headache
14:29:30 <ais523> although ofc you can specify which IP address to steal from
14:29:42 <ais523> running multiple INTERCAL programs on the same address is an exercise for the reader
14:29:46 <ais523> although not really an insurmountable one
14:29:56 <ais523> AnMaster: I can't remember the format
14:30:04 <ais523> probably as one twospot
14:30:06 <AnMaster> ais523, so you can't change the port btw?
14:30:13 <ais523> and the port's specified in a config file
14:30:29 <AnMaster> ais523, well then you can use that to run more than one, no?
14:30:47 <AnMaster> ais523, or can't it have different local and remote port?
14:30:56 <ais523> you could just have a variable with a program identifier in
14:31:00 <ais523> leave it permanently ignored
14:31:08 <ais523> and just smuggle it to see if you're stealing from the right program
14:31:23 <AnMaster> ais523, another alternative: 127.0.0.2 and so on
14:31:34 <AnMaster> ais523, I assume you can configure which IP to bind on
14:31:47 <ais523> that's specified in the program
14:32:00 <ais523> btw, I know the whole of 127/8 is supposed to have the same semantics
14:32:08 <ais523> but in practice, doesn't it often work only for 127.0.0.1?
14:32:44 <AnMaster> I think for windows it is so, and it used to work for other ones in older versions
14:33:18 <AnMaster> under linux it should work for the whole block
14:33:27 <AnMaster> unless your distro does it in a silly way
14:33:49 <AnMaster> inet addr:127.0.0.1 Mask:255.0.0.0
14:33:50 <AnMaster> inet6 addr: ::1/128 Scope:Host
14:34:15 <AnMaster> Phantom_Hoover, why not? If it can somehow grow it's state that is.
14:34:28 <ais523> Phantom_Hoover: only if infinitely large, or has infinite state some other wa
14:34:28 <AnMaster> not that I'm any expert on how neural networks work
14:34:40 <ais523> apart from that, I don't see why it would fail to meet any of the other neural network requirements
14:34:51 <AnMaster> ais523, so just add ability to grow at runtime
14:35:05 <ais523> AnMaster: typical neural networks in computing are incapable of growing
14:35:28 <AnMaster> ais523, sure, but why not add that feature? Then you could add computer cancers to computer viruses in the future ;P
14:35:35 <Phantom_Hoover> You can make a function that takes a neural net and evaluates its fitness, then evolve accordingly.
14:35:40 <ais523> fwiw, even human brains stop growing new neurons after a while, although they keep making new connections
14:36:15 <AnMaster> ais523, iirc new ones grow but very slowly? Pretty sure I saw some relatively recent research result mentioned about that somewhere
14:36:19 <ais523> hmm, I'm glad that ::1/128 is the only address that's loopback in ipv6
14:36:23 <ais523> AnMaster: hmm, interesting
14:37:16 <AnMaster> ais523, a bit hazy on the details. But iirc after some certain age it was very very slow, so for "practical purposes" it was "stopped", but not completely
14:37:25 <ais523> hmm, distro update's now at the stage where Firefox breaks at random
14:37:34 <ais523> out of all the programs I use, it's the one that least likes being used while being upgraded
14:37:36 <AnMaster> ais523, hm I seem to have a /64...
14:37:44 <AnMaster> that is larger than ipv4 right?
14:38:00 <AnMaster> that is assigned to my ipv6 tunnel
14:38:10 <AnMaster> which is when you think about it, quite absurd
14:40:37 <ais523> it's actually over 4 billion times as large
14:40:49 <AnMaster> ais523, twice as many bits though
14:41:27 <AnMaster> ais523, I can't imagine what I need this for. It isn't even as if I got ipv6 tunnel sharing to the rest of the LAN to work...
14:41:44 <Phantom_Hoover> They made a format, decided it was too small, then made one so ridiculously huge there's more space than you could possible need.
14:41:59 <ais523> AnMaster: the idea's that you have enough to do for anything you might want to do
14:42:08 <ais523> say, why not give every program running on your computer its own public IP?
14:42:23 <ais523> that doesn't strike me as an insane thing to want to do
14:42:37 <AnMaster> ais523, I would need 172 ips currently for that
14:42:45 <AnMaster> but that includes kernel tasks
14:42:59 <ais523> why use dbus when you can use tcp?
14:43:09 <ais523> well, tcp/ip as a whole
14:43:17 <AnMaster> can't imagine what use giving [md2_raid1] it's own ip would be
14:43:52 <ais523> why send signals when you can send ICMP packets?
14:44:53 <AnMaster> ais523, because sending custom ICMP needs root
14:45:03 <AnMaster> ais523, why else did you think ping was suid root?
14:45:08 <ais523> obviously this would be in a hypothetical future OS
14:45:27 <AnMaster> ais523, there is no clear mapping of SIGTERM and so on
14:45:33 <ais523> besides, even with something like Linux, you could just change the rules so that pinging things in the same process group didn't need root
14:45:47 <AnMaster> why would you want to expose it outwards
14:45:59 <ais523> you don't, you firewall it
14:46:05 <ais523> but you still have it there for uniqueness
14:46:11 <AnMaster> ais523, well what if you make a small error in the firewall config
14:46:25 <AnMaster> what with netfilter/iptables being basically mad
14:46:26 <ais523> maybe, if you're doing remote procedure calls or whatever, you could open up a gap in the firewall to send signals back and forth between apps
14:46:39 <AnMaster> ais523, dumbed down last I looked
14:47:18 <ais523> the advanced syntax thing works well enough for all the firewalling I need
14:47:28 <ais523> although IIRC I don't actually have any ports open at all for any reason atm
14:48:51 <ais523> Status: active \ Logging: on (low) \ Default: deny (incoming), allow (outgoing)
14:49:45 <AnMaster> ais523, maybe it got better since I last looked at it
14:50:18 <AnMaster> ais523, I wanted to do rate limiting on new connections to port 22 and couldn't figure out how to. With iptables it was messy but possible
14:50:45 <ais523> ok, your firewall desires are a lot weirder than mine
14:51:06 <AnMaster> ais523, what did you use "the advanced syntax thing" for?
14:51:24 <ais523> specifying which IPs were allowed to connect
14:51:32 <ais523> when I open a port, I normally open it only to one other computer
14:51:56 <AnMaster> ais523, oh and of course I want to log strange packets, like those produced by nmap xmas scan
14:52:04 <ais523> AnMaster: ufw limit 22/tcp
14:52:13 <AnMaster> LOG tcp -- anywhere anywhere tcp flags:FIN,SYN,RST,PSH,ACK,URG/FIN,SYN,RST,PSH,ACK,URG limit: avg 5/min burst 5 LOG level alert prefix `XMAS:'
14:52:36 <AnMaster> (it also drops it after it logs it)
14:52:37 <ais523> on the other hand, ufw doesn't seem to have options to specify what sort of rate-limiting you want
14:52:39 <ais523> just whether you want it or not
14:52:49 <AnMaster> and yes that is rate limited to not flood the log
14:53:31 <ais523> hmm, are those the packets with every single option turned on simultaneously?
14:53:38 <ais523> I always wondered why you'd want to do that
14:53:41 <AnMaster> ais523, yes that is what the xmas scan is
14:53:52 <AnMaster> there are some other ones to log as well
14:54:25 <ais523> hmm, so why /do/ people send such packets?
14:54:54 <AnMaster> ais523, man nmap? Because for some OSes it fucks up system/reveals which OS it is/something else
14:55:00 <AnMaster> iirc that is the common reason
14:55:20 <ais523> I don't have nmap installed, and I'd get in trouble if I tried to
14:56:03 <AnMaster> ais523, huh? is this not your netbook?
14:56:29 -!- Phantom_Hoover has quit (Quit: Leaving).
14:56:32 <ais523> yes, but it's connected to a university network atm
14:56:52 <AnMaster> "The key advantage to these scan types is that they can sneak through certain non-stateful firewalls and packet filtering routers."
14:57:12 <AnMaster> that is from the section: "-sN; -sF; -sX (TCP NULL, FIN, and Xmas scans)"
14:57:17 <ais523> computer security is so weird
14:59:08 <AnMaster> what on earth is that front loader out there doing
15:01:14 <AnMaster> hm no idea. just driving back and forth a few meters, not doing anything in specific
15:19:26 <Ilari> Heh... Years ago I had firewall rules that dropped all packet flag combos that "should not happen"...
15:23:41 -!- Phantom_Hoover has joined.
15:25:03 <AnMaster> Ilari, mine functions more as a simple IDS in that it looks at some specific well known weird ones.
15:25:39 <AnMaster> anyway my point was that ufw is very limited
15:27:19 <Ilari> I think the valid combos are SYN, SYN+ACK, SYN+ACK+PSH, ACK, ACK+PSH, ACK+URG, ACK+URG+PSH, RST, FIN and FIN+ACK (and maybe some others).
15:28:00 <ais523> what does SYN+ACK mean?
15:28:12 <ais523> wouldn't the SYN and ACK need different sequence numbers unless you were very lucky?
15:28:31 -!- marchdown has joined.
15:28:32 <Ilari> IIRC, it is used to signal connection accepted.
15:29:16 <AnMaster> it is one of the first ones exchanged during a new connection
15:29:39 <AnMaster> forgot if it was SYN → SYN,ACK → ACK or SYN → ACK → SYN,ACK
15:30:06 -!- marchdown has quit (Client Quit).
15:30:21 <Ilari> Ah, SYN+ACK+PSH isn't valid (because second packet is not allowed to have payload and PSH impiles payload).
15:31:05 <AnMaster> Ilari, what about SYN ACK URG?
15:31:21 <Ilari> So SYN+ACK?, ACK+PSH?+URG?, RST and FIN+ACK?
15:31:43 <AnMaster> Ilari, can't RST be combined with anything?
15:31:58 <Ilari> I don't think it can...
15:33:06 <AnMaster> Ilari, according to wikipedia there are more flags
15:33:25 <Ilari> Yes, I ignore them here. They are used by congestion control.
15:34:43 * AnMaster forgot how much changed between ipv4 and ipv6 TCP
15:35:15 <AnMaster> ah yes, quite a different header format it seems like
15:37:23 <Ilari> Eh, do tcp4 and tcp6 have very different header formats?
15:40:05 <Ilari> TCP(v4?) header includes: Source port (16 bits), Destination port (16 bits), Sequence number (32 bits), Acknowledgement number (32 bits), data offset (4 bits), flags (8 bits), window (16 bits), checksum (16 bits), urgent pointer (16 bits), options (variable) and padding (variable).
15:41:03 <AnMaster> Ilari, didn't you miss packet length there?
15:41:20 * Phantom_Hoover notes that there seem to be no good neural network libraries for Common Lisp
15:41:22 <AnMaster> because that is what is wider. Also ipv6 seems to have a "next header" field
15:41:38 <Ilari> TCP header does not have packet length. IP header does.
15:41:57 <AnMaster> Ilari, ah that explains this. It is a combined view of both ip and tcp headers
15:43:03 <Ilari> IPv6 next header is equivalent to IPv4 protocol field.
15:46:30 <Ilari> Ah, its also used for options.
15:48:16 <Phantom_Hoover> I suppose I'll have to bolt a Lisp interface onto a C library, then...
15:48:59 -!- Tritonio_GR has quit (Quit: Leaving.).
15:52:46 <Phantom_Hoover> Yeah, so I decided the easiest way to do it would just be to make the source a CL function which takes a neural net and evaluates its fitness.
16:00:54 -!- augur_ has quit (Ping timeout: 245 seconds).
16:01:19 -!- Tritonio_GR has joined.
16:05:54 <pineapple> i think it's supposed to encourage good programming practices
16:05:58 <ais523> if you mark something private, you can then safely change it from then on
16:06:01 <ais523> without breaking existing code
16:09:45 <Phantom_Hoover> So it's purely to make sure you don't do something stupid?
16:11:30 -!- ais523 has quit (Ping timeout: 248 seconds).
16:11:34 -!- ais523 has joined.
16:11:35 <Deewiant> And in libraries, that users don't do anything stupid
16:12:52 -!- ais523_ has joined.
16:13:57 <Phantom_Hoover> Although I went off it for a while when I couldn't get any interesting libraries to compile.
16:16:41 -!- ais523 has quit (Ping timeout: 258 seconds).
16:16:44 <AnMaster> anyone know a command line tool to prepend a line number to every line
16:16:53 <AnMaster> as in foo | add_line_numbers | bar
16:18:03 <ais523_> hmm, upgrade mostly seems to have worked
16:18:04 -!- ais523_ has changed nick to ais523.
16:18:23 <ais523> although Konversation is currently interpreting "you have been made channel admin" as "you have been banned"
16:18:30 <ais523> which is a rather great misinterpretation
16:32:47 <AnMaster> also I have a cat here that doesn't know cat -n
16:33:10 <AnMaster> Deewiant, freebsd cat seems to know cat -n
16:33:47 <AnMaster> Deewiant, the one that didn't was from an old openbsd install
16:55:55 <ais523> heh, Reddit noticed Ursala
16:56:30 <ais523> Phantom_Hoover: I didn't notice it was a pun
16:56:42 <ais523> and still don't get it even after you've pointed it out
16:57:53 <Phantom_Hoover> It's a play on "cat", the animal, the collective noun for which is "clowder", and "cat", the UNIX utility, of which there are many variants, but for which there is no commonly-accepted collective noun.
16:59:56 <ais523> I've never heard that collective noun used for the animals before
17:00:11 <ais523> "herd" would make a good one, though, after the well-known simile
17:00:41 <Phantom_Hoover> So perhaps the simile should be "like clowdering cats".
17:03:38 * ais523 puts the window max/min/close buttons back where they were before, then puts the window menu back
17:04:07 <ais523> a) the placement's arbitrary anyway so I may as well have them where I'm used to, b) I'm used to being able to close a window from /both/ top corners
17:04:10 -!- tombom has joined.
17:04:24 <ais523> and you can't with Ubuntu's new layout
17:06:02 <AnMaster> <ais523> heh, Reddit noticed Ursala <--- ?
17:06:19 <ais523> no, the programming language
17:06:29 <ais523> which should be considered ontopic here despite not being deliberately an esolang
17:06:54 <ais523> http://www.basis.uklinux.net/ursala/
17:07:51 <ais523> <Deestan> Another important design goal of Ursala was to discourage "code obfuscation" techniques. This can be done elegantly by making sure that all syntactically valid programs are no more readable than their obfuscated counterparts.
17:08:02 <ais523> reddit criticism can be so scathing sometimes
17:12:22 <AnMaster> ais523, it isn't much worse than J
17:12:40 <ais523> there are some interesting ideas in there
17:12:53 <ais523> but they're buried under the mess of &pseudopointers~
17:14:06 <ais523> have you seen the virtual machine it compiles to?
17:14:41 <AnMaster> only looked at some of the examples on there
17:14:58 <ais523> beh, reading about the language itself is much more fun than the examples, which don't make much sense in the abstract
17:16:22 <AnMaster> ais523, "Programs that read this format should be as lenient as possible, accepting anything that looks remotely like a PGM. "
17:16:43 <AnMaster> ais523, it is discussing the ASCII variant of the file format
17:16:48 <ais523> what's that format for?
17:16:55 <AnMaster> ais523, PGM is an image format
17:17:08 <AnMaster> "pgm - Netpbm grayscale image format "
17:17:24 * AnMaster was messing with this to generate a map of bad pixels for his camera
17:17:44 <AnMaster> the ASCII format is rather easy to parse
17:18:49 <AnMaster> and gimps saves it as one pixel per line. So I can just use curves to hide everything but the hot pixels, save as ASCII pgm, then do: nl curve_hide.pgm | grep -Ev $'\t0$' > foo.txt
17:19:03 <AnMaster> and then remove some of the meta data at the start
17:19:44 <AnMaster> and then the x values can be fetched as (linenumber - 5) % image_with (or integer division for y coord)
17:19:55 <AnMaster> -5 is to compensate for a few lines of meta data
17:20:53 -!- BeholdMyGlory has joined.
17:21:46 <AnMaster> ais523, is that ursala open source?
17:21:53 <ais523> well, the main interp of it is
17:22:01 <ais523> I don't know if a programming language can be inherently open source
17:22:13 <AnMaster> and what did you say was the issue with it?
17:22:33 <AnMaster> Phantom_Hoover, I was discussing the implementation there
17:22:40 <ais523> AnMaster: read the docs, the concept is too complicated to easily describe in less than a few pages, but I'll try
17:22:58 <ais523> in Ursala's VM, everything is written as just lists
17:23:05 <ais523> which can only contain other lists (and can be empty)
17:23:09 <ais523> so all data structures work like that
17:23:20 <ais523> now, you can create a sort of "pointer", like "head of tail of tail of head"
17:23:27 <ais523> to access a particular location in a data structure
17:23:53 <ais523> imagine that you have more complicated pointers like "both the first and second element of a pair"; that contradicts what I just said, but it's possible in Ursala too
17:24:05 <ais523> and it has a crazy syntax for stringing these things together to make complicated accessors
17:24:17 <ais523> that looks like a load of letters between & and ~, normally
17:24:29 <ais523> now, imagine you have something that is completely different from everything I just said
17:24:33 <ais523> but it's shoehorned into the same syntax
17:24:36 <ais523> that's a pseudopointer
17:24:40 <AnMaster> <ais523> which can only contain other lists (and can be empty) <-- hm reminds me of lambda calculus in a remote kind of way.
17:24:51 <AnMaster> (not lists there, but everything is functions kind of)
17:25:06 <ais523> yes, the basic concept is not incredibly insane
17:25:09 -!- FireFly has joined.
17:25:14 <ais523> there are quite a few interesting thoughts behind the language
17:25:16 <AnMaster> ais523, for being efficient it probably is
17:25:19 <ais523> just enough insane ones that the result is mad
17:25:32 <AnMaster> ais523, also to quote the feature page:
17:25:35 <AnMaster> "Exact calculations involving integers and rationals are implemented by purpose written libraries packaged with the compiler. Standard floating point and complex numbers are manipulated natively using the host system's C library routines. Arbitrary precision floating point numbers are handled by the mpfr library, but without need of explicit storage allocation or reclamation at the source level."
17:25:44 <AnMaster> then how does it do the floating point
17:26:03 -!- Tritonio_GR has quit (Read error: Connection reset by peer).
17:26:06 <AnMaster> I mean, does it store it as some combination of lists or such
17:26:15 <ais523> if you try to look at a floating-point number's representation on the fly, it'll decompose it into lists for you
17:26:28 <ais523> and presumably vice versa, if you try to use a list as a floating-point number
17:26:31 <AnMaster> ais523, but what is it actually represented as in the VM?
17:26:45 <ais523> I don't know the list representation of a floating-point number off by heart
17:26:50 <ais523> I am not an expert in Ursala
17:26:56 <ais523> just someone who managed to get almost halfway through the docs
17:27:04 <ais523> which probably means I know more about it than almost anyone else
17:27:07 <AnMaster> <ais523> now, you can create a sort of "pointer", like "head of tail of tail of head" <-- caddar ?
17:27:23 <ais523> yes, actually the idea behind the syntax is the same
17:27:47 <ais523> but Lisp caddadadadar sorts of things don't have things like operator precedence and binary operators
17:28:08 <AnMaster> ais523, on the other hand caddar kind of stuff reads backwards (IMO)
17:28:22 <AnMaster> pointers having operator precedence?
17:28:55 <AnMaster> however, such a concept does sound interesting
17:29:22 <ais523> AnMaster: ok, more fun: instead of (abc), it uses the notation abc3
17:29:34 <AnMaster> ais523, is this the VM or the language?
17:29:35 <ais523> you don't use parens, you use numbers saying how many things to parenthesise
17:29:46 <ais523> oh, and you use P rather than 2
17:29:49 <AnMaster> ais523, and postfix notation for it?
17:29:59 <ais523> AnMaster: it's an effectively-postfix notation
17:30:05 <ais523> this is all just the pseudopointers
17:30:12 <ais523> the language has another entirely separate notation for other things
17:30:27 <AnMaster> ais523, what happens if you use a 2 instead of a P?
17:30:49 <ais523> you think I'm mad enough to /write/ in this language?
17:30:56 <ais523> I'll stick to INTERCAL, thank you very much
17:31:25 <AnMaster> ais523, you mean it is more insane than INTERCAL?
17:31:36 <ais523> INTERCAL is not insane
17:31:39 <ais523> it's deliberately strange
17:31:42 <ais523> but it has a kind-of logic to it
17:31:48 <ais523> it's hard to write, but I wouldn't call it insane
17:31:56 <AnMaster> ais523, hm BANKCAL, needs to be done
17:32:21 <AnMaster> (combining bankstar with INTERCAL)
17:32:30 <ais523> no, there'd be no point
17:32:46 <ais523> and I thought it was capitalised BancSTAR
17:32:46 <AnMaster> Phantom_Hoover, implement INTERCAL in BANCSTAR then ;P
17:32:50 <ais523> but I might be wrong on that
17:32:56 <ais523> also, there are no docs for BancSTAR anywhere
17:32:59 <ais523> just a bit of example code
17:33:07 <ais523> with no explanation, so that the person who wrote it didn't get sued
17:33:35 <ais523> because BancSTAR is secret and proprietary
17:33:43 <ais523> the person hosting the article about it actually got takedown notices
17:33:50 <AnMaster> ais523, the implementation would be. But hardly the language itself?
17:34:18 <ais523> and I guess the language was an in-house trade secret
17:34:41 <AnMaster> wasn't it a third party product they used?
17:34:57 <ais523> actually, IIRC the language was originally the VM for something else
17:35:06 <ais523> but the something else was too limited so they started using the lang directly
17:37:45 <AnMaster> ais523, a GUI generation tool iirc
17:39:33 <AnMaster> Phantom_Hoover, not enough info about it
17:40:12 <AnMaster> Phantom_Hoover, err, see esolang wiki
17:40:38 <AnMaster> I'm quite fond of /// (article on esowiki is "slashes" iirc)
17:50:52 -!- Sgeo has joined.
18:00:17 -!- atrapado has quit (Quit: Saliendo).
18:03:22 -!- KingOfKarlsruhe has joined.
18:08:43 <ais523> hmm, I thought we got http://esolangs.org/wiki//// to work?
18:15:44 <ais523> there's an environment variable you can set, IIRC
18:15:49 <ais523> but I've forgotten what it's called
18:15:51 <ais523> AnMaster probably knows
18:17:23 <ais523> does that replace or add to the standard locations?
18:18:24 <AnMaster> Phantom_Hoover, what Deewiant said or rpath
18:18:32 <AnMaster> rpath would be at link time of the application
18:18:40 <AnMaster> that way no need to use LD_LIBRARY_PATH
18:19:02 <AnMaster> of course you can also just make the path a standard path if you wish. /etc/ld.so.conf
18:19:13 <AnMaster> might or might not be a good idea
18:19:34 <AnMaster> Phantom_Hoover, iirc LD_LIBRARY_PATH prepends to the standard path btw
18:28:06 -!- pikhq_ has changed nick to pikhq.
18:59:09 <Phantom_Hoover> Googling "why language x is awful" gives some... interesting results.
19:02:28 <pikhq> "Why Befungebrainfuck is awful" :P
19:03:11 <pikhq> "OH GOD WITH THE TWO DIMENSIONAL NESS AND THE SELFMODIFICATION AND THE TAPE"
19:16:53 <Phantom_Hoover> What is it with people complaining about C's type system.
19:17:13 <pikhq> Phantom_Hoover: There are many legitimate complaints about it.
19:17:18 -!- tombom_ has joined.
19:17:22 <pikhq> For instance, it doesn't go far enough.
19:17:43 -!- tombom has quit (Ping timeout: 240 seconds).
19:17:47 <Phantom_Hoover> Well, obviously you don't expect 1+"hello" to even compile, but it still makes sense.
19:18:07 <pikhq> 1+"hello" certainly compiles.
19:18:27 <Phantom_Hoover> pikhq: Yes, but if you were learning C, you wouldn't expect it to.
19:18:54 <pikhq> Phantom_Hoover: And then they find out that C strings are syntactic sugar around array literals and it's all peachy.
19:19:33 -!- ais523 has quit (Remote host closed the connection).
19:47:05 <AnMaster> <Phantom_Hoover> And array literals are syntactic sugar for pointers. <-- not exactly
19:47:40 <AnMaster> quite different results in that case
19:48:03 <AnMaster> the former will point to a read only copy of the string literal "bar"
19:48:19 <AnMaster> the latter will contain a read write copy
19:49:00 <pikhq> AnMaster: The former is not const-correct.
20:01:16 -!- kar8nga has joined.
20:17:34 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
20:19:15 -!- Gracenotes has quit (Read error: Connection reset by peer).
20:39:22 -!- ais523 has joined.
20:41:56 -!- KingOfKarlsruhe has quit (Remote host closed the connection).
20:46:51 -!- tombom__ has joined.
20:48:28 -!- tombom_ has quit (Ping timeout: 258 seconds).
20:53:02 -!- tombom has joined.
20:55:23 -!- tombom__ has quit (Ping timeout: 240 seconds).
21:15:54 -!- Zuu has quit (Ping timeout: 276 seconds).
21:21:55 -!- Zuu has joined.
21:21:55 -!- Zuu has quit (Changing host).
21:21:56 -!- Zuu has joined.
21:25:15 -!- Phantom_Hoover has joined.
21:25:42 <Phantom_Hoover> What sort of insane client doesn't tell you when you get disconnected due to ping timeouts?
21:26:38 -!- kar8nga has quit (Remote host closed the connection).
21:26:58 <Phantom_Hoover> And neither comments nor variable names are in English.
21:28:27 -!- Sgeo_ has joined.
21:31:49 -!- Sgeo has quit (Ping timeout: 260 seconds).
22:16:44 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
22:17:26 -!- tombom has quit (Quit: Leaving).
22:17:54 -!- oerjan has joined.
22:26:08 <oerjan> Leonidas: did you find out what permute does?
22:28:57 -!- Tritonio_GR has joined.
22:46:05 <AnMaster> <pikhq> AnMaster: The former is not const-correct. <-- true, but it will by default compile without a warning
22:46:29 <AnMaster> needs more than -Wall -Wextra iirc
22:47:02 <pikhq> AnMaster: No, it just needs -std=gnu99 or -std=c99. Which makes it so that const exists. :)
22:59:53 -!- Gracenotes has joined.
23:04:58 -!- pikhq_ has joined.
23:06:28 -!- pikhq has quit (Ping timeout: 240 seconds).
23:08:18 -!- BeholdMyGlory has quit (Remote host closed the connection).
23:09:22 -!- Vegabondmx has joined.
23:11:08 -!- coppro has joined.
23:11:08 -!- coppro has quit (Changing host).
23:11:08 -!- coppro has joined.
23:13:57 -!- pikhq_ has quit (Ping timeout: 264 seconds).
23:18:42 -!- pikhq has joined.
23:27:31 -!- FireFly has quit (Quit: swatted to death).
23:29:51 -!- pikhq_ has joined.
23:31:00 -!- pikhq has quit (Ping timeout: 264 seconds).
23:31:35 <pikhq_> So much trouble staying connected.
23:33:48 -!- pikhq_ has changed nick to pikhq.
23:57:35 -!- poiuy_qwert has quit (Quit: This computer has gone to sleep).